Special Issue Reprint

Orthogonal Polynomials
and Special Functions
Recent Trends and Their Applications

Edited by
Yamilet Quintana

mdpi.com/journal/mathematics
Orthogonal Polynomials and Special
Functions: Recent Trends and Their
Applications
Orthogonal Polynomials and Special
Functions: Recent Trends and Their
Applications

Editor
Yamilet Quintana

Basel • Beijing • Wuhan • Barcelona • Belgrade • Novi Sad • Cluj • Manchester

Editor
Yamilet Quintana
Universidad Carlos III
de Madrid
Madrid
Spain

Editorial Office
MDPI AG
Grosspeteranlage 5
4052 Basel, Switzerland

This is a reprint of articles from the Special Issue published online in the open access
journal Mathematics (ISSN 2227-7390) (available at: https://www.mdpi.com/si/mathematics/
58IV0VBX80).

For citation purposes, cite each article independently as indicated on the article page online and as
indicated below:

Lastname, A.A.; Lastname, B.B. Article Title. Journal Name Year, Volume Number, Page Range.

ISBN 978-3-7258-1853-2 (Hbk)

ISBN 978-3-7258-1854-9 (PDF)
doi.org/10.3390/books978-3-7258-1854-9

© 2024 by the authors. Articles in this book are Open Access and distributed under the Creative
Commons Attribution (CC BY) license. The book as a whole is distributed by MDPI under the terms
and conditions of the Creative Commons Attribution-NonCommercial-NoDerivs (CC BY-NC-ND)
license.
Contents

Noor Alam, Shahid Ahmad Wani, Waseem Ahmad Khan and Hasan Nihal Zaidi
Investigating the Properties and Dynamic Applications of Δh Legendre–Appell Polynomials
Reprinted from: Mathematics 2024, 12, 1973, doi:10.3390/math12131973 . . . . . . . . . . . . . . . 1

Carlos Féliz-Sánchez, Héctor Pijeira-Cabrera and Javier Quintero-Roba

Asymptotic for Orthogonal Polynomials with Respect to a Rational Modification of a Measure
Supported on the Semi-Axis
Reprinted from: Mathematics 2024, 12, 1082, doi:10.3390/math12071082 . . . . . . . . . . . . . . . 15

Maryam Salem Alatawi, Waseem Ahmad Khan, Can Kızılateş and Cheon Seoung Ryoo
Some Properties of Generalized Apostol-Type Frobenius–Euler–Fibonacci Polynomials
Reprinted from: Mathematics 2024, 12, 800, doi:10.3390/math12060800 . . . . . . . . . . . . . . . . 31

Dionisio Peralta, Yamilet Quintana and Shahid Ahmad Wani

Mixed-Type Hypergeometric Bernoulli–Gegenbauer Polynomials
Reprinted from: Mathematics 2023, 11, 3920, doi:10.3390/math11183920 . . . . . . . . . . . . . . . 47

Chao Min and Pixin Fang

The Recurrence Coefficients of Orthogonal Polynomials with a Weight Interpolating between
the Laguerre Weight and the Exponential Cubic Weight
Reprinted from: Mathematics 2023, 11, 3842, doi:10.3390/math11183842 . . . . . . . . . . . . . . . 63

Mohra Zayed, Shahid Ahmad Wani and Yamilet Quintana

Properties of Multivariate Hermite Polynomials in Correlation with Frobenius–Euler
Polynomials
Reprinted from: Mathematics 2023, 11, 3439, doi:10.3390/math11163439 . . . . . . . . . . . . . . . 74

Héctor Pijeira-Cabrera, Javier Quintero-Roba and Juan Toribio-Milane

Differential Properties of Jacobi-Sobolev Polynomials and Electrostatic Interpretation
Reprinted from: Mathematics 2023, 11, 3420, doi:10.3390/math11153420 . . . . . . . . . . . . . . . 91

Juan Hernández, Dionisio Peralta and Yamilet Quintana

A Look at Generalized Degenerate Bernoulli and Euler Matrices
Reprinted from: Mathematics 2023, 11, 2731, doi:10.3390/math11122731 . . . . . . . . . . . . . . . 111

Abel Dı́az-González, Juan Hernández and Héctor Pijeira-Cabrera

Sequentially Ordered Sobolev Inner Product and Laguerre–Sobolev Polynomials
Reprinted from: Mathematics 2023, 11, 1956, doi:10.3390/math11081956 . . . . . . . . . . . . . . . 126

Clemente Cesarano, William Ramı́rez, Stiven Dı́az, Adnan Shamaoon and

Waseem Ahmad Khan
On Apostol-Type Hermite Degenerated Polynomials
Reprinted from: Mathematics 2023, 11, 1914, doi:10.3390/math11081914 . . . . . . . . . . . . . . . 141

Francisco Marcellán and Ridha Sfaxi

The Cauchy Exponential of Linear Functionals on the Linear Space of Polynomials
Reprinted from: Mathematics 2023, 11, 1895, doi:10.3390/math11081895 . . . . . . . . . . . . . . . 154

Mohamed Jleli and Bessem Samet

Integral Inequalities Involving Strictly Monotone Functions
Reprinted from: Mathematics 2023, 11, 1873, doi:10.3390/math11081873 . . . . . . . . . . . . . . . 172

v
Hari M. Srivastava, Azhar Y. Tantary and Firdous A. Shah
A New Discretization Scheme for the Non-Isotropic Stockwell Transform
Reprinted from: Mathematics 2023, 11, 1839, doi:10.3390/math11081839 . . . . . . . . . . . . . . . 186

Reem Alzahrani and Saiful R. Mondal

Redheffer-Type Bounds of Special Functions
Reprinted from: Mathematics 2023, 11, 379, doi:10.3390/math11020379 . . . . . . . . . . . . . . . . 195

vi
 mathematics

Article
Investigating the Properties and Dynamic Applications of Δh
Legendre–Appell Polynomials
Noor Alam 1 , Shahid Ahmad Wani 2, *, Waseem Ahmad Khan 3 and Hasan Nihal Zaidi 1

1 Department of Mathematics, College of Science, University of Ha’il, Ha’il 2440, Saudi Arabia;
[email protected] (N.A.); [email protected] (H.N.Z.)
2 Symbiosis Institute of Technology, Pune Campus, Symbiosis International (Deemed University),
Pune 412115, Maharashtra, India
3 Department of Mathematics and Natural Sciences, Prince Mohammad Bin Fahd University,
P.O. Box 1664, Al Khobar 31952, Saudi Arabia; [email protected]
* Correspondence: [email protected] or [email protected]

Abstract: This research aims to introduce and examine a new type of polynomial called the Δh
Legendre–Appell polynomials. We use the monomiality principle and operational rules to define the
Δh Legendre–Appell polynomials and explore their properties. We derive the generating function
and recurrence relations for these polynomials and their explicit formulas, recurrence relations, and
summation formulas. We also verify the monomiality principle for these polynomials and express
them in determinant form. Additionally, we establish similar results for the Δh Legendre–Bernoulli,
Euler, and Genocchi polynomials.

Keywords: Δh sequences; monomiality principle; Legendre–Appell polynomials; explicit forms;

determinant form

MSC: 33E20; 33B10; 33E30; 11T23

1. Introduction and Preliminaries

Complex system behavior has been modeled and described by special polynomials
Citation: Alam, N.; Wani, S.A.; Khan,
W.A.; Zaidi, H.N. Investigating the
in a variety of domains, including quantum mechanics and statistical mechanics. These
Properties and Dynamic Applications
unique polynomials have also been used to describe and analyze complex systems in a
of Δh Legendre–Appell Polynomials. number of other domains, such as quantum mechanics and statistics. Polynomial sequences
Mathematics 2024, 12, 1973. https:// are indispensable in several branches of mathematics, such as algebraic combinatorics,
doi.org/10.3390/math12131973 entropy, and combinatorics. The Legendre, Chebyshev, Laguerre, and Jacobi polynomials
are a few examples of polynomial sequences that are solutions to particular ordinary
Academic Editor: Valery Karachik
differential equations in approximation theory and physics. Legendre polynomials are a
Received: 15 May 2024 class of orthogonal polynomials with important applications in physics and mathematics.
Revised: 11 June 2024 The French mathematician Edmond Legendre, who first introduced them in the 19th
Accepted: 15 June 2024 century, is the reason behind their name. The Legendre differential equation, a second-
Published: 26 June 2024 order linear differential equation, has solutions that lead to the Legendre polynomials.
They are often represented as Sn (u) [1], where n is a non-negative integer that denotes the
degree of the polynomial. They are defined on the interval [0, +∞). There are numerous
noteworthy characteristics of Legendre polynomials: On the interval [0, +∞), the Legendre
polynomials form an orthogonal set with regard to the weight function e−u . This indicates
Copyright: © 2024 by the authors.
Licensee MDPI, Basel, Switzerland.
that, with the exception of situations in which the polynomials have the same degree,
This article is an open access article
the integral of the sum of two distinct Laguerre polynomials with the weight function
distributed under the terms and
equals zero. Moreover, the Legendre polynomials satisfy a recurrence relation, enabling
conditions of the Creative Commons
Attribution (CC BY) license (https://
the computation of higher-degree polynomials from lower-degree ones. This characteristic
creativecommons.org/licenses/by/
helps with efficient polynomial generation and numerical computations. Furthermore,
4.0/). the generating function of these polynomials permits the expansion of some functions

Mathematics 2024, 12, 1973. https://doi.org/10.3390/math12131973 1 https://www.mdpi.com/journal/mathematics

Mathematics 2024, 12, 1973

into a sequence of Legendre polynomials. This characteristic helps in differential equation

solving and yields closed-form solutions. Application areas for the Legendre polynomials
include the solutions of the Schrodinger equation for the hydrogen atom and other quantum
systems with spherical symmetry in mathematics, physics, and engineering. Furthermore,
issues involving diffusion equations, wave propagation, and heat conduction give rise to
these polynomials.
Mathematical physics two-variable special polynomials have been the subject of much
recent research. A class of polynomials known as two-variable special polynomials has
certain attributes, for example, [2,3]. They have numerous uses in mathematics and other
fields and are frequently researched in the area of algebraic geometry. Bivariate Chebyshev,
Hermite, Laguerre, and Laguerre polynomials are a few notable examples of two-variable
special polynomials. They are widely used in signal processing, numerical analysis, and
approximation theory. Bivariate Chebyshev polynomials are symmetric polynomials with
applications in least squares fitting and interpolation. Hermite polynomials of two variables
have applications in quantum mechanics, statistical mechanics, and waveguide theory.
Bivariate Hermite polynomials are often used in the study of harmonic oscillators in two
dimensions. Bivariate Legendre polynomials are a two-variable extension of the Legendre
polynomials. They satisfy a bivariate analogue of the Legendre differential equation and
have applications in quantum mechanics, potential theory, and random matrix theory.
Bivariate Legendre polynomials are particularly useful in studying the behavior of systems
with two degrees of freedom. These polynomials satisfy a certain orthogonality condition
with respect to a weight function and are thus extensively studied in mathematical physics,
probability theory, and approximation theory. The significance of these two-variable
special polynomials lies in their usefulness in solving problems in various mathematical
and scientific domains. They provide a rich framework for expressing and analyzing
multivariate functions and have specific properties that make them suitable for specific
applications. It is well known that huge classes of partial differential equations, which
are frequently encountered in physical issues, can be solved analytically by innovative
methods made possible by the special polynomials of two variables. The two-variable
Legendre polynomials Sω (u, v) [4] are of enormous mathematical significance and have
applications in physics, which makes their introduction intriguing.
The two-variable Legendre polynomials (2VLeP) Sω (u, v) are specified by means of
the following generating equation:

√ ∞
ξω
evξ J0 (2ξ −u) = ∑ Sω (u, v)
ω!
, (1)
ω =0

where J0 (uξ ) is the 0th order ordinary Bessel function of first kind [5] defined by

∞ √ ω +2ν
√ (−1)ν ( u)
Jω (2 u) = ∑ ν! (ω + ν)!
. (2)
ν =0

also note that

√ uω
exp(−γDu−1 ) = J0 (2 γu), Du−ω {1} := (3)
ω!
is the inverse derivative operator.
Or, alternatively, by
∞
ξω
evξ C0 (−uξ 2 ) = ∑ Sω (u, v)
ω!
, (4)
ω =0

where C0 (uξ ) is the 0th order Tricommi function of the first kind [5] with
−1 ξ 2
C0 (−uξ 2 ) = e Du . (5)

2
Mathematics 2024, 12, 1973

Thus, in view of Equation (3) or (5), the generating expression for Legendre polynomials
can be cast as: ∞
−1 2 ξω
evξ e Du ξ = ∑ Sω (u, v) . (6)
ω =0
ω!
Very recently, a large interest has been shown by mathematicians to introduce Δh forms
of special polynomials. Some extensions of the special polynomials were studied in [1,5–10].
After that, by using the classical finite difference operator Δh , a new form of the special poly-
nomials, known as the Δh special polynomials of different polynomials, were introduced
in [11,12]. These Δh special polynomials have been studied because of their remarkable
applications in different branches of mathematics, physics, and statistics.
These Δh Appell polynomials are represented as:
[h]
A ω ( u ) : = A ω ( u ) , ω ∈ N0 (7)

and defined by  
[h]
u Δh Aω (u) = ωh Aω −1 (u), ω ∈ N, (8)

where Δh is the finite difference operator:

[h]
u Δh H (u) = H(u + h) − H(u). (9)

The Δh Appell polynomials Aω (u) are specified by the following generating function [12]:
∞
u [h] ξω
γ(ξ )(1 + hξ ) h = ∑ Aω ( u )
ω!
, (10)
ω =0

where ∞
ξω
γ(ξ ) = ∑ γω,h
ω!
, γ0,h = 0. (11)
ω =0

Therefore, motivated by the results in [4,11–13], here we introduced the two-variable

Δh Legendre–Appell polynomials:
−1 ∞ ω
v Du [h] ξ
γ(ξ )(1 + hξ ) h (1 + hξ 2 ) h = ∑ S Aω ( u, v )
ω!
(12)
ω =0

through the generating function concept.

This article is designed as follows: Section 2 discusses how the Legendre–Appell poly-
nomials are generated and explores recurrence relations that govern their behavior. Section
3 presents formulas for summing or evaluating these Legendre–Appell polynomials over
certain ranges or with specific constraints. These formulas can be useful for calculating the
values of the polynomials efficiently. Section 4 discusses the monomiality principle, which
relates to how Legendre–Appell polynomials behave under certain operations. The deter-
minant form for these polynomials is also established. In Section 5, Symmetric identities
for these polynomials are derived. The conclusion section summarizes the findings of the
article and discusses implications, applications, and potential future research directions re-
lated to Legendre–Appell polynomials. Each of these sections likely delves deeper into the
mathematical properties and characteristics of Legendre–Appell polynomials, providing
insights into their behavior and utility in various mathematical contexts.

2. Two-Variable Δh Legendre–Appell Polynomials

The significance of this section lies in its exploration of a novel class of two-variable
Δh Legendre–Appell polynomials and its establishment of essential properties associated
with them. The research expands the existing knowledge base and opens doors to new
avenues of inquiry within polynomial theory and its applications.
The construction of the generating function for these Δh Legendre–Appell polynomi-
[h]
als, denoted as S Aω (u, v), marks a crucial step forward in understanding the behavior

3
Mathematics 2024, 12, 1973

and properties of these polynomials. Generating functions serve as powerful tools in

combinatorics, analysis, and mathematical physics, providing insights into the structure
and properties of sequences and functions. By proving the existence and constructing the
generating function for Δh Legendre–Appell polynomials, this section lays the foundation
for further exploration of their properties, such as orthogonality, recurrence relations, and
special function identities.
Moreover, by establishing a connection between the Δh Legendre–Appell polynomials
and their generating function, this research contributes to the broader mathematical com-
munity’s understanding of polynomial families and their applications. The traits listed in
this section provide valuable insights into the unique characteristics and behaviors of these
polynomials, paving the way for their utilization in various mathematical and scientific
domains. Overall, this section represents a significant advancement in polynomial theory,
offering fresh perspectives and potential applications that warrant further investigation and
exploration. First, we prove the following conclusion to construct the generating function
[h]
for these Δh Legendre–Appell polynomials S Aω (u, v) by proving the following result:

[h]
Theorem 1. For the two-variable Δh Legendre–Appell polynomials S Aω (u, v), the succeeding
generating relation holds true:
−1 ∞ ω
v Du [h] ξ
γ(ξ )(1 + hξ ) h (1 + hξ 2 ) h = ∑ S Aω ( u, v )
ω!
. (13)
ω =0

−1
Du
v
Proof. By expanding γ(ξ )(1 + hξ ) h (1 + hξ 2 ) h at u = v = 0 for finite differences by a
Newton series and the order of the product of the developments of the function γ(t)(1 +
−1
v Du [h]
hξ ) h (1 + hξ 2 ) h with respect to the powers of ξ, we observe the polynomials S Aω (u, v)
ω
expressed in Equation (13) as coefficients of ξω! as the generating function of two-variable
[h]
Δh Legendre–Appell polynomials S An (u, v).

[h]
Theorem 2. For the two-variable Δh Legendre–Appell polynomials S Aω (u, v), the succeeding
relations hold true:

v Δh [h] [h]
S Aω ( u, v ) = ω S Aω −1 (u, v) (14)
h

u Δh [h] [h]
S Aω ( u, v ) = ω (ω − 1) S Aω −2 (u, v), Du−1 → u. (15)
h

Proof. By differentiating (13) with respect to v by taking into consideration of expression (5),
we have
 v
−1 
Du v+h
−1
Du v
−1
Du
v Δh γ(t)(1 + hξ ) h (1 + hξ 2 ) h = γ(ξ )(1 + hξ ) h (1 + hξ 2 ) h − γ(ξ )(1 + hξ ) h (1 + hξ 2 ) h
−1
Du
= (1 + hξ − 1)γ(ξ )(1 + hξ ) h (1 + hξ 2 )
v
h
(16)
−1
Du
v
= hξ γ(ξ )(1 + hξ ) h (1 + hξ 2 ) h .
By substituting the righthand side of expression (13) in (16), we find
∞ ω ∞ ω +1
[h] ξ [h] ξ
v Δh ∑ S Aω ( u, v )
ω!
=h ∑ S Aω ( u, v )
ω!
. (17)
ω =0 ω =0

By replacing ω → ω − 1 in the righthand side of previous expression (16) and com-

paring the coefficients of the same exponents of t in the resultant expression, assertion (14)
is deduced.
Further, on similar grounds, expression (15) is established.

4
Mathematics 2024, 12, 1973

Next, we deduce the explicit form satisfied by these two-variable Δh Legendre–Appell

[h]
polynomials S Aω (u, v) by demonstrating the result:

[h]
Theorem 3. For the two-variable Δh Legendre–Appell polynomials S Aω (u, v), the explicit relation
holds true:
v   
v
[h]
h
ω [h]
S A ω ( u, v ) = ∑ d dh hd Aω−d (u). (18)
d =0

Proof. Expanding generating relation (13) in the given manner:

−1
v v ∞ ω
v Du h
(hξ )d [h] ξ
γ(ξ )(1 + hξ ) h (1 + hξ 2 ) h = ∑ h
d d! ∑ S Aω ( u, 0)
ω!
(19)
d =0 ω =0

which can further be written as

∞ ∞ [ vh ] v
[h] ξ ω
[h] ξ ω +d
∑ S Aω ( u, v )
ω!
= ∑ ∑ h
d
h d Aω ( u )
ω! d!
. (20)
ω =0 ω =0 d =0

By replacing ω → ω − d in the righthand side of the previous expression, it follows

that
∞ ω ∞ [ vh ] v
[h] ξ [h] ξω
∑ S Aω ( u, v )
ω!
= ∑ ∑ h
d
h d Aω ( u )
(ω − d)! d!
. (21)
ω =0 ω =0 d =0

On multiplying and dividing by ω! on the righthand side of previous expression (21)

and comparing the coefficients of the same exponents of ξ on both sides, assertion (18) is
deduced.
[h]
Theorem 4. Further, for the two-variable Δh Legendre–Appell polynomials S Aω (u, v), the explicit
relation holds true:
ω  
[h] ω [h]
S Aω ( u, v ) = ∑ γν,h Sω −ν (u, v). (22)
ν =0
ν

Proof. Expanding generating relation (13) in view of expressions (8) and (13) with γ(ξ ) = 1
in the given manner:
∞ ∞
v
−1
Du ξν [h] ξω
γ(ξ )(1 + hξ ) h (1 + hξ 2 ) h = ∑ γν,h ν! ∑ Sω (u, v)
ω!
, (23)
ν =0 ω =0

which can further be written as

∞ ∞ ∞
[h] ξ ω
[h] ξ ω +ν
∑ S Aω ( u, v )
ω!
= ∑ ∑ γν,h Sω (u, v) ω!ν! . (24)
ω =0 ω =0 ν =0

By replacing ω → ω − ν in the righthand side of the previous expression, it follows that

∞ ω ∞ ω
[h] ξ [h] ξω
∑ S Aω ( u, v )
ω!
= ∑ ∑ γν,h Sω−ν (u, v) (ω − ν)! ν! . (25)
ω =0 ω =0 ν =0

On multiplying and dividing by ω! on the righthand side of previous expression (25)

and comparing the coefficients of the same exponents of ξ on both sides, assertion (22)
is deduced.

3. Summation Formulae
This section establishes the summation formulae, or sigma notation, essential in
mathematical analysis. These formulae provide systematic methods for computing sums

5
Mathematics 2024, 12, 1973

involving special polynomials, facilitating the evaluation of complex expressions encoun-

tered in various mathematical contexts. By leveraging these formulae, mathematicians
can identify patterns and uncover hidden symmetries within polynomial structures, en-
hancing understanding and fostering innovative applications in combinatorics, probability
theory, and mathematical physics. Additionally, the study of summation formulae aids in
developing efficient computational techniques, enabling researchers to address challenging
problems precisely. These expressions concisely represent the sum of a sequence of terms,
providing a convenient way to compute the total of a series of numbers or expressions.
Thus, we demonstrate the summation formulae by proving the following results:

Theorem 5. For ω ≥ 0, we have

ω   
[h] ω 1 [h]
S Aω ( u, v + 1) = ∑ ν − (−h)ν S Aω −ν (u, v). (26)
ν =0
h ν

Proof. By (13), we have

∞ ∞ −1  
[h] ξω [h] ξω v Du 1
∑ S Aω ( u, v + 1) ω! − ∑ S Aω ( u, v ) ω!
= γ(ξ )(1 + hξ ) h (1 + hξ 2 ) h (1 + hξ ) h − 1
ω =0 ω =0    
∞ [h] ω ∞ ν
= ∑ S Aω (u, v) ξω! ∑ − 1h (−h)ν ξν! − 1 (27)
 ω =0 ν =0 ν
∞ ω   ω ∞ ω
[h] [h]
= ∑ ∑ ( ν ) − h (−h) S Aω −ν (u, v) ξω! − ∑ S Aω (v, u) ξω! .
ω 1 ν
ω =0 ν =0 ν ω =0

Comparing the coefficients of ξ, we obtain (26).

Theorem 6. For ω ≥ 0, we have

ω −ν
ω [ 2 ]
[h] v  u ω!
S Aω ( u, v ) = ∑ ∑ − (−h)ω − j−ν − (−1) j Aν,h . (28)
ν =0 j =0
h ω −2j−ν h j (ω − 2j − ν)!( j!)2 ν!

Proof. Using (13), we have

∞ ω −1
[h] ξ v Du
∑ S Aω ( u, v )
ω!
= γ(ξ )(1 + hξ ) h (1 + hξ 2 ) h
ω =0
 v
∞
ξω ∞  u ξ 2j
= γ(ξ ) ∑ − (−h)ω ∑ − (−1) j (−h) j
ω =0
h ω ω! j =0
h j j!j!
ω
∞ [ 2 ]
∞
ξν v  u ξω
= ∑ Aν,h ν! ∑ ∑ − (−h)ω − j − (−1) j
ν =0 ω =0 j =0
h ω −2j h j (ω − 2j)!( j!)2
ω −ν
ω [ 2 ]
∞
v  u ξω
= ∑ ∑ ∑ − (−h)ω − j−ν − (−1) j Aν,h . (29)
ω =0 ν =0 j =0
h ω −2j−ν h j (ω − 2j − ν)!( j!)2 ν!

Equating the coefficients of ξ, we obtain (28).

Now, we investigate the connection between the Stirling numbers of the first kind and
two-variable Δh Legendre polynomials.
∞
[log(1 + ξ )]ν ξi
= ∑ S1 (i, ν) , | ξ |< 1. (30)
ν! i=ν
i!
From the above definition, we have
i
( v )i = ∑ (−1)i−ν S1 (i, ν)vν . (31)
ν =0

6
Mathematics 2024, 12, 1973

Theorem 7. For ω ≥ 0, we have

ω   ν
[h] ω [h]
S Aω ( u, v ) = ∑ ν S Aω−ν (u, 0) ∑ v j S1 (ν, j)hν− j . (32)
ν =0 j =0

Proof. From (13), we have

∞ ω −1
[h] ξ Du
= e h log(1+hξ ) γ(ξ )(1 + hξ 2 )
v
∑ S Aω ( u, v )
ω!
h
ω =0
−1
Du ∞  v  j [log(1 + hξ )] j
= γ(ξ )(1 + hξ 2 ) h
∑ h j!
j =0
∞
ξω ∞ ν  v j ξν
[h]
= ∑ S Aω ( u, 0) ∑∑ S1 (ν, j)hν
ω =0
ω! ν=0 j=0 h ν!
  
∞ ω ν  j
ω [h] v ξω
= ∑ ∑ ν S Aω−ν (u, 0) ∑ h S1 (ν, j)hν ω!
. (33)
ω =0 ν =0 j =0

Comparing the coefficients of ξ, we obtain the result.

Theorem 8. For ω ≥ 0, we have

ω ω −l
[h] ω! [h]
S Aω ( u, v ) = ∑∑ hν A (u, 0)S1 (ν + l, l )vl . (34)
l =0 ν =0
( ω − ν − l ) ! ( ν + l ) ! S ω −ν−l

Proof. From (13), we have

∞ ω −1
[h] ξ v Du
∑ S Aω ( u, v )
ω!
= γ(ξ )(1 + hξ ) h (1 + hξ 2 ) h
ω =0
∞
ξ ω ∞ v ξν
[h]
= ∑ S Aω ( u, 0) ∑ − (−h)ν
ω =0 ω! ν=0 h ν ν!
  
∞ ω
ω v [h] ξω
= ∑ ∑ − (−h)ν S Aω −ν (u, 0) . (35)
ω =0 ν =0
ν h ν ω!

Comparing the coefficients of ξ, we obtain

 
[h]
ω
ω v [h]
S Aω ( u, v ) = ∑ ν − h ν (−h)ν S Aω−ν (u, 0). (36)
ν =0

Using, equality (31) in previous expression, we obtain

   
ω ν
[h] ω [h]
S Aω ( u, v ) = ∑ ν (−h) S Aω−ν (u, 0) ∑ (−1)ν−l S1 (ν, l )(−h)−l vl
ν
ν =0 l =0
ω ω
ω! [h]
= ∑ ∑ (ω − ν)!ν! (−h)ν−l S Aω−ν (u, 0)(−1)ν−l S1 (ν, l )vl
l =0 ν = l
ω ω −l
ω! [h]
= ∑∑ (−h)ν S Aω −ν−l (u, 0)(−1)ν S1 (ν + l, l )vl . (37)
l =0 ν =0
(ω − ν − l )!(ν + l )!

This completes the proof of the theorem.

7
Mathematics 2024, 12, 1973

Theorem 9. For ω ≥ 0, we have

ω ω −l
[h] ω! [h]
S Aω ( u, v + s ) = ∑∑ hν A (u, v)S1 (ν + l, l )sl . (38)
l =0 ν =0
(ω − ν − l )! (ν + l )! S ω −ν−l

Proof. Taking v + s instead of v in (13), we have

∞ ω −1
[h] ξ v+s Du
∑ S Aω ( u, v + s )
ω!
= γ(ξ )(1 + hξ ) h (1 + hξ 2 ) h
ω =0
 
∞
ξ ω ∞  s ξν
[h]
= ∑ S Aω ( u, v ) ∑ − (−h)ν . (39)
ω =0
ω! ν =0
h ν ν!

Using the Cauchy rule and after comparing the coefficients of ξ on both sides of the
resulting equation, we have
 
[h]
ω
ω s [h]
S Aω ( u, v + s ) = ∑ ν − h ν (−h)ν S Aω−ν (u, v). (40)
ν =0

Then, using (31) for − hs ν

, we obtain (38).

4. Monomiality Principle and Determinant Form

The monomiality principle is a fundamental concept in polynomial theory. It states
that any polynomial can be expressed uniquely as a combination of simple algebraic
terms called monomials. This representation simplifies the polynomial structure and
facilitates their analysis in various mathematical contexts. The principle plays a crucial
role in practical applications across scientific and engineering fields, such as computational
mathematics, signal processing, and physics, where polynomials are used to model complex
systems and phenomena. This highlights the broad applicability and significance of the
monomiality principle in advancing both theoretical understanding and practical problem-
solving capabilities. The exploration and utilization of the monomiality principle, along
with operational guidelines and other properties of hybrid special polynomials, have
been the focus of extensive study. Originating from Steffenson’s concept of poweroids
in 1941 [14], the notion of monomiality was further elaborated upon by Dattoli [15,16].
Central to this framework are the Jˆ and K̂ operators, which serve as multiplicative and
derivative operators, respectively, for a polynomial set gk (u1 )k∈N .
These operators adhere to the following expressions:

gk+1 (u1 ) = Jˆ { gk (u1 )} (41)

and
k gk−1 (u1 ) = K̂{ gk (u1 )}. (42)
Consequently, when these multiplicative and derivative operations are applied to the
polynomial set gk (u1 )m∈N , they yield a quasi-monomial domain. Of particular importance
is the following formula:
[K̂, Jˆ ] = K̂ Jˆ − Jˆ K̂ = 1̂, (43)
which exhibits a Weyl group structure.
Assuming the set { gk (u1 )}k∈N is quasi-monomial, the operators Jˆ and K̂ can be
leveraged to derive the significance of this set. Thus, the following axioms hold true:
For Jˆ and K̂ to exhibit differential traits, gk (u1 ) satisfies the differential equation:

Jˆ K̂{ gk (u1 )} = k gk (u1 ). (44)

8
Mathematics 2024, 12, 1973

The expression
gk (u1 ) = Jˆ k {1} (45)
represents the explicit form, with g0 (u1 ) = 1 and the expression
∞
wk
e w J {1} =
ˆ
∑ gk (u1 ) k! , |w| < ∞ , (46)
k =0

demonstrates generating expression behavior and is obtained by applying identity (45).

In this section, we will discuss the results of our validation efforts. These results aim
to strengthen the reliability and usefulness of the Δh Legendre–Appell polynomials as
important mathematical tools. As a result, we will be verifying the monomiality principle
[h]
for the Δh Legendre–Appell polynomials S Aω (u, v) by presenting the following results:

[h]
Theorem 10. The Δh Legendre–Appell polynomials S Aω (u, v) satisfy the succeeding multiplica-
tive and derivative operators:
 
v 2 Du−1 v Δh γ ( v Δh h )
MˆS A = + + (47)
1 + v Δh h + v Δh 2
γ( v Δh h )

and
v Δh
DSˆ A = . (48)
h

Proof. In consideration of expression (5), taking derivatives with respect to v of expres-

sion (13), we have
 v D −1  v+h
−1
Du v
−1
Du
2 u
v Δ h γ ( ξ )(1 + hξ ) h (1 + hξ ) h = γ(ξ )(1 + hξ ) h (1 + hξ 2 ) h − γ(ξ )(1 + hξ ) h (1 + hξ 2 ) h
−1
Du
v
= (1 + hξ − 1)γ(ξ )(1 + hξ ) h (1 + hξ 2 ) h
(49)
−1
Du
v
= hξ γ(ξ )(1 + hξ ) h (1 + hξ 2 ) h ,
thus, we have
−1 −1
v Δh v Du v Du
γ(ξ )(1 + hξ ) h (1 + hξ 2 ) h = ξ γ(ξ )(1 + hξ ) h (1 + hξ 2 ) h , (50)
h
which gives the identity

v Δh [h] [h]
S Aω ( u, v ) =ξ S Aω ( u, v ) . (51)
h
Now, differentiating expression (13) with respect to ξ, we have
 
∂  v
−1 
Du ∂ ∞
[h] ξ ω

∂ξ
γ(ξ )(1 + hξ ) h (1 + hξ 2 ) h =
∂ξ ∑ S Aω ( u, v )
ω!
, (52)
ω =0

  
∞ ∞
v D −1 ξ γ (ξ ) [h] ξ ω
[h] ξ ω −1
1 + hξ
+2 u 2 +
1 + hξ γ(ξ ) ∑ S Aω ( u, v )
ω!
= ∑ ω S Aω (u, v)
ω!
. (53)
ω =0 ω =0

On the usage of identity expression (51) and replacing ω → ω + 1 in the righthand

side of previous expression (53), assertion (47) is established.
Further, in view of identity expression (51), we have

v Δh [h] [h]
S Aω ( u, v ) = ω S Aω −1 (u, v) , (54)
h

9
Mathematics 2024, 12, 1973

which gives an expression for the derivative operator (48).

Next, we deduce the differential equation for the Δh Legendre–Appell polynomials

[h]
S Aω ( u, v ) by demonstrating the succeeding result:

[h]
Theorem 11. The Δh Legendre–Appell polynomials S Aω (u, v) satisfy the differential equation:
 
v 2 Du−1 v Δh γ ( v Δh h ) ωh [h]
+ + − S Aω ( u, v ) = 0. (55)
1 + v Δh h + v Δh 2
γ( v Δh h ) v Δh

Proof. Inserting expression (47) and (48) in the expression (44), the assertion (55) is proved.
[h]
Next, we give the determinant form of Δh Legendre–Appell polynomials S Aω (u, v)
[h]
in terms of Δh Legendre polynomials Sω (u, v) by proving the result listed below:

[h]
Theorem 12. The Δh Legendre–Appell polynomials S Aω ( u, v ) give rise to the determinant
represented by:
 
 [h] [h] [h] [h] 
 1 S1 (u, v) S2 (u, v) · · · Sω −1 (u, v) Sω (u, v) 
 
 
 
 γ0,h γ1,h γ2,h ··· γω −1,h γω,h 
 
 
 
(−1)ω  0 (21)γ1,h · · · (ω − 1
( 1 )γω −1,h ,
ω
γ0,h 1 ) γω −2,h
[h] 
S Aω ( u, v ) = ω +1   (56)
(γ0,h )  
 
 0 0 γ0,h ··· (ω − 1
2 ) γω −3,h
ω
( 2 )γω −2,h 
 
 . . . ··· . . 
 
 . . . ··· . . 
 ω 
 0 0 0 ··· γ0,h (ω −1)γ1,h 

where
1
γω,h , ω = 0, 1, · · · are the coefficients of Maclaurin series of .
γ(ξ )

ξ ω
Proof. Multiplying both sides of Equation (13) by 1
γ(ξ )
= ∑∞
ω =0 γω,h ω! , we find

∞ ∞ ∞
[h] ξω ξν [h] ξω
∑ Sω (u, v)
ω!
= ∑ ∑ γν,h
ν! S
Aω (u, v) ,
ω!
(57)
ω =0 ω =0 ν =0

which, on using the Cauchy product rule, becomes

ω  
[h] ω [h]
Sω (u, v) = ∑ γν,h S Aω −ν (u, v). (58)
ν =0
ν

[h]
This equality results in a set of ν equations with variables Sω (u, v), where ω = 0,
1, 2, · · · . Solving this set using Cramer’s rule, and exploiting the denominator as the
determinant of a lower triangular matrix with a determinant of (γ0,h )ω +1 , while transposing
the numerator and subsequently substituting the i-th row with the (i + 1)-th position for
i = 1, 2, · · · , n − 1 produces the desired outcome.

5. Examples
The Appell polynomial family is diverse, spanning various members derived by se-
lecting an appropriate function γ(ξ ). Each member boasts unique characteristics, including
distinct names, generating functions, and associated numerical properties. These polyno-
mials find applications across numerous mathematical domains due to their versatility and
rich properties. The selection of γ(ξ ) plays a crucial role in defining the specific polynomial

10
Mathematics 2024, 12, 1973

within the family, allowing for tailored solutions to various problems in mathematics and
physics. Understanding the generating functions associated with these polynomials is
essential for their practical utilization, enabling efficient computation and analysis. In the
following sections, we delve into the intricacies of the generating functions that underpin
the diverse set of Appell polynomials, shedding light on their mathematical elegance and
practical significance in a wide array of applications. The generating function for the Δh
[h]
Bernoulli polynomials β ω (v) is given by
1
∞
log(1 + hξ ) h v [h] ξω
1
(1 + hξ ) h = ∑ β ω (v)
ω!
, | ξ |< 2π. (59)
(1 + hξ ) − 1 h ω =0

[h]
The generating expression for Δh Euler polynomials Eω (v) is given by
∞
2 v [h] ξω
1
(1 + hξ ) h = ∑ Eω (v)
ω!
, | ξ |< π. (60)
(1 + hξ ) + 1 h ω =0

[h]
The generating expression for Δh Genocchi polynomials Gω (v) is given by
1
∞
2 log(1 + hξ ) h v [h] ξω
1
(1 + hξ ) h = ∑ Gω (v)
ω!
, | ξ |< π. (61)
(1 + hξ ) + 1 h ω =0

For h → 0, these polynomials reduce to the Bω (v), Eω (v) and Gω (v) polynomials [17].
The Bernoulli, Euler, and Genocchi numbers have found numerous applications in
various areas of mathematics, including number theory, combinatorics, and numerical
analysis. These applications extend to practical mathematics, where these polynomials and
numbers are utilized to solve problems and derive mathematical formulas.
For instance, the Bernoulli numbers appear in various mathematical formulas, such as
the Taylor expansion, trigonometric and hyperbolic tangent and cotangent functions, and
sums of powers of natural numbers. These numbers play a crucial role in number theory,
providing insights into patterns and relationships among integers.
Similarly, the Euler numbers arise in the Taylor expansion and have close connec-
tions to trigonometric and hyperbolic secant functions. They have applications in graph
theory, automata theory, and calculating the number of up/down ascending sequences,
contributing to the analysis of structures and patterns in discrete mathematics.
Moreover, the Genocchi numbers find utility in graph theory and automata theory.
They are particularly valuable in counting the number of up/down ascending sequences,
which involves studying the order and arrangement of elements in a sequence. Therefore,
these Δh polynomials and numbers of Bernoulli, Euler, and Genocchi play a significant role
in various mathematical domains, allowing for the exploration of mathematical relation-
ships, the derivation of formulas, and the analysis of patterns and structures.
By appropriately choosing the function γ(ξ ) in Equation (13), we can establish the
[h]
following generating functions for the Δh Legendre-based Bernoulli S Bω (u, v), Euler
[h] [h]
S Eω ( u, v ), and Genocchi S Gω ( u, v ) polynomials:
−1 ∞ ω
log(1 + hξ ) v Du [h] ξ
1
(1 + hξ ) h (1 + hξ 2 ) h = ∑ S Bω ( u, v )
ω!
, (62)
h(1 + hξ ) − h
h ω =0

−1 ∞ ω
2 v Du [h] ξ
1
(1 + hξ ) h (1 + hξ 2 ) h = ∑ S Eω ( u, v )
ω!
, (63)
(1 + hξ ) + 1
h ω =0

and
−1 ∞ ω
2 log(1 + hξ ) v Du [h] ξ
1
(1 + hξ ) h (1 + hξ 2 ) h = ∑ S Gω ( u, v )
ω!
. (64)
h(1 + hξ ) + h
h ω =0

11
Mathematics 2024, 12, 1973

[h] [h]
Further, in view of expression (22) and Table 1, the polynomials S Bω (u, v), S Eω (u, v)
[h]
and S Gω ( u, v ) satisfy the following explicit form:
ω  
[h] ω [h]
S Bω ( u, v ) = ∑ Bν,h S Aω −ν (u, v), (65)
ν =0 ν

n  
[h] ω [h]
S Eω ( u, v ) = ∑ ν Eν,h S Aω−ν (u, v) (66)
ν =0

and  
ω
[h] ω [h]
S Gω ( u, v ) = ∑ ν
Gν,h S Aω −ν (u, v). (67)
ν =0
[h] [h]
Furthermore, in view of expressions (56), the polynomials S Bω (u, v), S Eω (u, v) and
[h]
S Gω ( u, v ) satisfy the following determinant representations:
 
 [h] [h] [h] [h] 
 1 B1 (u, v) B2 (u, v) · · · Bω −1 (u, v) Bω (u, v) 
 
 
 
 γ0,h γ1,h γ2,h ··· γω −1,h γω,h 
 
 
 
(−1)ω  0 (21)γ1,h · · · (ω − 1
( 1 )γω −1,h ,
ω
γ0,h 1 ) γω −2,h
[h] 
S Bω ( u, v ) = ω +1   (68)
(γ0,h )  
 
 0 0 γ0,h · · · (ω − 1
2 ) γω −3,h
ω
( 2 )γω −2,h 
 
 . . . ··· . . 
 
 . . . ··· . . 
 ω 
 0 0 0 ··· γ0,h (ω −1)γ1,h 

 
 [h] [h] [h] [h] 
 1 E1 (u, v) E2 (u, v) · · · Eω −1 (u, v) Eω (u, v) 
 
 
 
 γ0,h γ1,h γ2,h ··· γω −1,h γω,h 
 
 
 
(−1)ω  0 (21)γ1,h · · · (ω − 1
( 1 )γω −1,h ,
ω
γ0,h 1 ) γω −2,h
[h] 
S Eω ( u, v ) = ω +1   (69)
(γ0,h )  
 
 0 0 γ0,h · · · (ω − 1
2 ) γω −3,h
ω
( 2 )γω −2,h 
 
 . . . ··· . . 
 
 . . . ··· . . 
 ω 
 0 0 0 ··· γ0,h (ω −1)γ1,h 

and
 
 [h] [h] [h] [h] 
 1 G1 (u, v) G2 (u, v) · · · Gω −1 (u, v) Gω (u, v) 
 
 
 
 γ0,h γ1,h γ2,h ··· γω −1,h γω,h 
 
 
 
(−1)ω  0 (21)γ1,h · · · (ω − 1
( 1 )γω −1,h .
ω
γ0,h 1 ) γω −2,h
[h] 
S Gω ( u, v ) = ω +1   (70)
(γ0,h )  
 
 0 0 γ0,h · · · (ω − 1
2 ) γω −3,h (ω2 )γω −2,h 
 
 . . . ··· . . 
 
 . . . ··· . . 
 ω 
 0 0 0 ··· γ0,h (ω −1)γ1,h 

12
Mathematics 2024, 12, 1973

Table 1. Several members of the Appell polynomials family.

S. No. Appell Polynomials Generating Function A(ξ )

ξ ξω ξ
I. The Bernoulli polynomials [11] euξ = ∑∞
ω =0 Bω ( u ) ω! A(ξ ) = e ξ −1
eξ − 1
2 ξω
II. The Euler polynomials [11] euξ = ∑∞
ω =0 Eω ( u ) ω! A(ξ ) = 2
e ξ +1
eξ + 1
2ξ uξ ξω
III. The Genocchi polynomials [11] e = ∑∞ ω =0 Gω ( u ) ω! A(ξ ) = 2ξ
e ξ +1
eξ + 1

6. Conclusions
The introduction and exploration of Δh Legendre–Appell polynomials mark a signifi-
cant advancement in polynomial theory, particularly in quantum mechanics and entropy
modeling. Integrating the monomiality principle and operational rules, these polynomials
offer fresh insights into uncharted mathematical territory. This research provides explicit
formulas and elucidates fundamental properties, deepening our understanding of Leg-
endre polynomials and linking them to established polynomial categories, enriching the
mathematical landscape.
Future research could delve into structural properties and algebraic aspects, uncover-
ing deeper insights and potential applications. Exploring their applicability in quantum
mechanics and mathematical physics may reveal new research directions and practical
implications. Additionally, bridging the gap between mathematical theory and real-world
applications could maximize their potential, especially in statistical mechanics, information
theory, and computational science. Collaborative interdisciplinary efforts could unlock the
full potential of Δh hybrid polynomials across diverse domains.
Therefore, introducing and investigating hybrid Δh polynomials represent a significant
milestone, fostering new research avenues and applications in various mathematical and
scientific fields. Continued exploration and collaboration are essential for realizing their
full potential and understanding their broader implications.

Author Contributions: Methodology, N.A., S.A.W. and W.A.K.; Validation, N.A. and H.N.Z.; Formal
analysis, W.A.K.; Investigation, S.A.W. and W.A.K.; Resources, N.A. and H.N.Z.; Writing—original
draft, S.A.W. All authors have read and agreed to the published version of the manuscript.
Funding: This research was funded by the Research Deanship at the University of Ha’il, Saudi Arabia,
through Project No. RG-23 206.
Data Availability Statement: No new data were created or analyzed in this study. Data sharing is
not applicable to this article.
Acknowledgments: The authors acknowledge the support received from the Research Deanship at
the University of Ha’il, Saudi Arabia, through Project No. RG-23 206.
Conflicts of Interest: The authors declare no conflicts of interest.

References
1. Khan, S.; Raza, N. Family of Legendre-Sheffer polynomials. Math. Compt. Mod. 2012, 55, 969–982. [CrossRef]
2. Dattoli, G.; Ricci, P.E.; Cesarano, C.; Vázquez, L. Special polynomials and fractional calculas. Math. Comput. Model. 2003, 37,
729–733. [CrossRef]
3. Dattoli, G.; Lorenzutta, S.; Mancho, A.M.; Torre, A. Generalized polynomials and associated operational identities. J. Comput.
Appl. Math. 1999, 108, 209–218. [CrossRef]
4. Dattoli, G.; Ricci, P.E. A note on Legendre polynomials. Int. J. Nonlinear Sci. Numer. Simul. 2001, 2, 365–370. [CrossRef]
5. Andrews, L.C. Special Functions for Engineers and Applied Mathematicians; Macmillan Publishing Company: New York, NY, USA,
1985.
6. Ramírez, W.; Cesarano, C. Some new classes of degenerated generalized Apostol-Bernoulli, Apostol-Euler and Apostol-Genocchi
polynomials. Carpathian Math. Publ. 2022, 14, 354–363. [CrossRef]
7. Roshan, S.; Jafari, H.; Baleanu, D. Solving FDEs with Caputo-Fabrizio derivative by operational matrix based on Genocchi
polynomials. Math. Methods Appl. Sci. 2018, 41, 9134–9141. [CrossRef]

13
Mathematics 2024, 12, 1973

8. Khan, W.A.; Alatawi, M.S. Analytical properties of degenerate Genocchi polynomials the second kind and some of their
applications. Symmetry 2022, 14, 1500. [CrossRef]
9. Hernandez, J.; Peralta, D.; Quintana, Y. A look at generalized Bernoulli and Euler matrices. Mathematics 2023, 11, 2731. [CrossRef]
10. Quintana, Y.; Ramirez, J.L.; Sirvent, V.F. On generalized Bernoulli-Barnes polynomials. Math. Rep. 2022, 24, 617–636.
11. Alyusof, R.; Wani, S.A. Certain Properties and Applications of Δh Hybrid Special Polynomials Associated with Appell Sequences.
Fractal Fract. 2023, 7, 233. [CrossRef]
12. Costabile, F.A.; Longo, E. Δh Appell sequences and related interpolation problem. Numer. Algor. 2013, 63, 165–186. [CrossRef]
13. Almusawa, M.Y. Exploring the Characteristics of Δh Bivariate Appell Polynomials: An In-Depth Investigation and Extension
through Fractional Operators. Fractal Fract. 2024, 8, 67. [CrossRef]
14. Steffensen, J.F. The poweriod, an extension of the mathematical notion of power. Acta Math. 1941, 73, 333–366. [CrossRef]
15. Dattoli, G. Hermite-Bessel and Laguerre-Bessel functions: A by-product of the monomiality principle. Adv. Spec. Funct. Appl.
1999, 1, 147–164.
16. Dattoli, G. Generalized polynomials operational identities and their applications. J. Comput. Appl. Math. 2000, 118, 111–123.
[CrossRef]
17. Carlitz, L. Eulerian numbers and polynomials. Math. Mag. 1959, 32, 247–260. [CrossRef]

Disclaimer/Publisher’s Note: The statements, opinions and data contained in all publications are solely those of the individual
author(s) and contributor(s) and not of MDPI and/or the editor(s). MDPI and/or the editor(s) disclaim responsibility for any injury to
people or property resulting from any ideas, methods, instructions or products referred to in the content.

14
 mathematics

Article
Asymptotic for Orthogonal Polynomials with Respect to a
Rational Modification of a Measure Supported on the Semi-Axis
Carlos Féliz-Sánchez 1 , Héctor Pijeira-Cabrera 2, * and Javier Quintero-Roba 3

1 Instituto de Matemáticas, Facultad de Ciencias, Universidad Autónoma de Santo Domingo, Av. Alma Mater,
Santo Domingo 10105, Dominican Republic; [email protected]
2 Departamento de Matemáticas, Universidad Carlos III de Madrid, Av. de la Universidad, 30,
28911 Leganés, Spain
3 Departamento de Teoría de la Señal y Comunicaciones y Sistemas Telemáticos y Computación,
Universidad Rey Juan Carlos, 28942 Fuenlabrada, Spain; [email protected]
* Corresponding: [email protected]

Abstract: Given a sequence of orthogonal polynomials { Ln }∞ n=0 , orthogonal with respect to a positive
Borel ν measure supported on R+ , let { Qn }∞ n=0 be the the sequence of orthogonal polynomials with
respect to the modified measure r ( x )dν( x ), where r is certain rational function. This work is devoted
(d)  √  Ak  √ √  Bj
Q (z) a +i z+ bj
⇒n ∏kN=1 1 √z+k√a ∏ j=21 √
N
to the proof of the relative asymptotic formula (nd) , on
Ln (z) k b j +i
compact subsets of C \ R+ , where ak and b j are the zeros and poles of r, and the Ak , Bj are their
respective multiplicities.

Keywords: orthogonal polynomials; asymptotic behavior; rational modifications

MSC: 41A60; 42C05; 41A20

1. Introduction
Citation: Féliz-Sánchez, C.; Let μ be a positive, finite, Borel measure on R+ = [0, +∞), such that for all n ∈ Z+
Pijeira-Cabrera, H.; Quintero-Roba, J. (the set of all non-negative integers)
Asymptotic for Orthogonal
 ∞
Polynomials with Respect to a
ηn = x n dμ( x ) < ∞. (1)
Rational Modification of a Measure 0
Supported on the Semi-Axis.  ∞
Mathematics 2024, 12, 1082. https://
If there is no other measure μ0 , such that ηn = x n dμ0 ( x ) for all n ∈ Z+ , it is said
doi.org/10.3390/math12071082 0
that the moment problem associated with { η n } n ∈Z+
is determined (see ([1] Ch. 4)). By a
Academic Editor: Manuel Manas classical result of T. Carleman (see ([1] Th. 4.3)), a sufficient condition in order to the
Received: 12 March 2024 moment problem associated with the sequence {ηn }n∈Z+ in (1) to be determined is
Revised: 1 April 2024
∞
1
Accepted: 2 April 2024
∑ √ = +∞.
2nηn
(2)
Published: 3 April 2024 n =1

We say that the measure μ belongs to the class M [R+ ] if {ηn }n∈Z+ satisfies (2) and
μ > 0 a.e. on R+ with respect to Lebesgue measure.
Copyright: © 2024 by the authors. α(z)
Let r (z) = be a rational function, where α and β are coprime polynomials with
Licensee MDPI, Basel, Switzerland. β(z)
This article is an open access article respective degrees A and B. We say that dμr ( x ) = r (z)dμ(z) is a rational modification (for
distributed under the terms and brevity, modification) of the measure μ. Write
conditions of the Creative Commons
Attribution (CC BY) license (https:// N1 N2
creativecommons.org/licenses/by/ α(z) = ∏ ( z − ai ) Ai , β(z) = ∏ ( z − b j ) Bj ,
4.0/). i =1 j =1

Mathematics 2024, 12, 1082. https://doi.org/10.3390/math12071082 15 https://www.mdpi.com/journal/mathematics

Mathematics 2024, 12, 1082

where ai , b j ∈ C \ R+ , Ai , Bj ∈ N. A = A1 + · · · + A N1 and B = B1 + · · · + BN2 .

We denote by { Ln }∞ n=0 the sequence of monic orthogonal polynomials with respect
to dμ. Assume that { Qn }∞ n=0 is the sequence of monic polynomials of least degree, not
identically equal to zero, such that
 ∞
x k Qn ( x ) r ( x )dμ( x ) = 0, for all k = 0, 1, 2, . . . , n − 1. (3)
0

The existence of Qn is an immediate consequence of (3). Indeed, it is deduced solving an

homogeneous linear system with n equations and n + 1 unknowns. Uniqueness follows
from the minimality of the degree of the polynomial. We call Qn the nth monic modified
orthogonal polynomial. In ([2] Th.1), explicit formulas are provided in order to compute
Qn when the poles and zeros of the rational modification have a multiplicity of one.
N
Suppose that { ai }i=11 , {b j } N
j=1 ⊂ C \ [−1, 1]. If μ is a positive (finite Borel) measure
2

on [−1, 1], such that μ is on the Nevai class M(0, 1), in ([3] Th. 1) the authors prove the
following asymptotic formula

(d) N1   Ai N2 Bj
Qn (z) ϕ ( z ) − ϕ ( ai ) 1
(d)
⇒ ∏ 2( z − a i ) ∏ 1−
ϕ(z) ϕ(b j )
, (4)
Ln (z) n i =1 j =1

on K ⊂ C \ [−1, 1]. The notation f n ⇒n f , K ⊂ U means that the sequence of functions

f n converges to f uniformly on a compact subset K of the region U, f (d) denotes the dth
derivative of f , d ∈ Z+ is fixed and
    
 
ϕ ( z ) = z + z2 − 1 z + z2 − 1 > 1, z ∈ C \ [−1, 1] .

In [3], the asymptotic formula (4) is pivotal in examining the asymptotic properties of
orthogonal polynomials across a broad range of inner products, encompassing Sobolev-type
inner products
 m dj
f, g S = f g dμ + ∑ ∑ λ j,i f (i ) ( ζ j ) g (i ) ( ζ j ),
j =1 i =0

where λ j,i ≥ 0, m, d j > 0, μ is certain kind of complex measure with compact support
is defined on the real line, and ζ j represents complex numbers outside the support of
μ. The authors compare the Sobolev-type orthogonal polynomials associated with this
measure to the orthogonal polynomials with respect to μ. These asymptotic results are of
interest for the electrostatic interpretation of zeros of Jacobi–Sobolev polynomials (cf. [4]).
On the other hand, the use of modified measures provides a stable way of computing
the coefficients of the recurrence relation associated to a family of orthogonal polynomials
(see ([5] Ch. 2)) and in [6,7] the interest of the modified orthogonal polynomials for the
study of the multipoint Padé approximation is shown.
For measures supported on [0, +∞) (or (−∞, +∞)) that satisfy the Carleman condition,
G. López in ([8] Th. 4) (or ([8] Th. 3) for (−∞, +∞)) proves a quite general version of
the relative asymptotic formula (4). In this case, if the modification function, ρ, is a non-
negative function on [0, +∞) in L1 (μ), such that there exists an algebraic polynomial G and
k ∈ N for which | G |ρ/(1 + x )k and | G |ρ−1 /(1 + x )k belong to L∞ (μ), then

Qn (z) S(ρ, C \ [0, +∞), z)

⇒ , K ⊂ C \ [0, +∞); (5)
Ln (z) n S(ρ, C \ [0, +∞), ∞)

16
Mathematics 2024, 12, 1082

where S(ρ, C \ [0, +∞), z) is the Szegő’s function for ρ with respect to C \ [0, +∞), i.e.,
 ∞ √ 
1 −z dx
S(ρ, C \ [0, +∞), z) =es(z) , s(z) = log ρ( x ) √ ;
2π 0 z−x x
S(ρ, C \ [0, +∞), ∞) = lim S(ρ, C \ [0, +∞), −r );
r →+∞
√
where the roots are selected from the condition 1 = 1. Additionally, it is requested that
f (z) = ρ(−((z + 1)/(z − 1))2 ) satisfies the Lipschitz condition in z = 1 and f (1) = 0.
Asymptotic results, analogous to those obtained in [3], are obtained in [9] for the
particular case of (5), when dμ( x ) = x a e− x dx with a > −1 (the Laguerre measure).
The aim of this paper is to obtain an analog of (4) for measures supported on R+ . We
prove the following theorem.

Theorem 1. Given a measure ν ∈ M [R+ ], it holds in compact subsets of C \ R

⎛√  ⎞ Bj
(d) N1  √  z + bj
Qn (z) ai + i Ai N2
(d)
⇒ ∏ √ √
z + ai
∏⎝  ⎠ , (6)
Ln (z) n i =1 j =1 bj + i

for d ∈ Z+ .

This situation is not a particular case of (5), because we consider ρ as a rational function
with complex coefficients and no necessarily ρ( x ) ≥ 0 on R+ .
The structure of the paper is as follows: Sections 2 and 3 are devoted to prove some
preliminary results on varying measures. On the other hand, in Section 4 we obtain an
essential theorem that allows us to finally prove Theorem 1 in Section 5.

2. Varying Measures and Carleman’s Condition

In this section, we introduce auxiliary results on varying measures and prove some
useful lemmas that allow us to extend results that hold for measures with bounded support
to the unbounded case. The following notations will be used throughout the paper:

1+z
Ψ(z) = for z ∈ C \ [−1, 1].
1−z
z−1
Ψ −1 ( z ) = for z ∈ C \ R+ . (7)
z+1
√
z+i
Φ(z) = √ where Φ(−1) = ∞ and z ∈ C \ [|z| ≤ 1].
z−i

If σ is a finite positive Borel measure on [−1, 1], we denote

 1
dσ(t) dσ(t)
dσn (t) = and ςn = . (8)
(1 − t)2n −1 (1 − t ) n
√ √ θ
In this paper, we consider the principal branch of the square root, i.e., reiθ = rei 2 , where
r > 0 and 0 ≤ θ < 2π.

Lemma 1. Let μ be a positive Borel measure supported on R+ and suppose that dσ(t) = (1 −
t) dμ(Ψ(t)). Then,
(a) μ > 0 a.e. on R+ implies that σ > 0 a.e. on [−1, 1],
∞ ∞
1 1
(b) if ∑ 2n √ = +∞, then ∑ 2n √ = +∞,
n =1
η n n =1
ςn
where, as in (1), ηn denotes the nth moment of the measure dμ.

17
Mathematics 2024, 12, 1082

2
Proof. To prove the first assertion note that if dσ(t) = μ (Ψ(t))dt, then
1−t
dσ dμ(Ψ(t)) 2
= (1 − t ) = μ (Ψ(t)) > 0 a.e. on [−1, 1].
dt dt 1−t
The second part is derived using the change of variable t = Ψ−1 ( x ) in the integral
 1  ∞ 
(1 − t ) x + 1 n −1
ςn = dμ(Ψ(t)) = dμ( x )
−1 (1 − t ) n 0 2
 1  n −1  ∞ 
x+1 x + 1 n −1
= dμ( x ) + dμ( x )
0 2 1 2
 ∞
≤ η0 + x n−1 dμ( x ) ≤ η0 + ηn . (9)
1

∞ ∞ ∞
As ∑ (ηn )−1/2n = +∞, from (9) we have ∑ (η0 + ηn )−1/2n = +∞, then ∑ (ς n )−1/2n =
n =0 n =0 n =0
+∞.
 
x+1 k
Lemma 2. Assume that dν ∈ M [R+ ], rk ( x ) = and consider the modification
2

dνrk ( x ) = rk ( x )dν( x ). Then dνrk ( x ) ∈ M [R+ ] for all k ∈ Z.

Proof. We now proceed by induction. Obviously, the initial case k = 0 is given by hypothesis.
• Case k > 0. Assume that dνr j ( x ) ∈ M [R+ ] for all j ≤ k − 1. Since dνrk ( x ) =
 
x+1 dνrk ( x )
dνrk−1 ( x ), it is immediate that dνrk ( x ) is positive and dx > 0 a.e. on R+ .
2
Let mn,k be the nth moment of the measure dνrk ( x ), then
 ∞  1    ∞  
x+1 x+1
mn,k = x n dνrk ( x ) = xn dνrk−1 ( x ) + xn dνrk−1 ( x ),
0 0 2 1 2
 1  ∞
≤ dνrk−1 ( x ) + x n+1 dνrk−1 ( x ) ≤ m0,k−1 + mn+1,k−1 ,
0 1

  
x+1 x+1
where we use that x n ≤ 1 for x ∈ [0, 1] and ≤ x, for x ∈ [1, +∞). Then,
2 2
using induction hypothesis, we obtain that mn,k < ∞ and the sequence of moments for
dνrk ( x ) satisfies Carleman’s condition.
• Case k < 0. Repeating the previous arguments, we obtain that if dνr j ( x ) ∈ M [R+ ] for
dνrk ( x )
all 0 < j ≤ k + 1 then dνrk ( x ) is positive and dx > 0 a.e. on R+ .
For the nth moment of the measure dνrk ( x ), we have
 ∞  1    ∞  
2 2
mn,k = x n dνrk ( x ) = xn dνrk+1 ( x ) + xn dνrk+1 ( x )
0 0 x+1 1 x+1
≤ 2 m0,k+1 + mn,k+1 ,
  
2 2
where we use that x n ≤ 2 for x ∈ [0, 1] and ≤ 1, for x ∈ [1, +∞). Then,
x+1 x+1
using induction hypothesis, we obtain that mn,k < ∞ and the sequence of moments for
dνrk ( x ) satisfies Carleman’s condition.

18
Mathematics 2024, 12, 1082

Lemma 3. [7], Th. 4, Cor. 1. Let Pn,k be the kth monic orthogonal polynomial with respect to dσn .
∞
1
If σ > 0 a.e. on [−1, 1] and ∑ 2n √ = +∞, then, for each integer k
n =1
ςn

Pn,n−k+1 ϕ(z)
(z) ⇒ ; K ⊂ C \ [−1, 1],
Pn,n−k n 2
√  √  
 
where ϕ(z) = z + z2 − 1 z + z2 − 1 > 1 z ∈ C \ [−1, 1] .

 2m
2
Lemma 4. Assume μ ∈ M [R+ ] and dμm ( x ) = dμ( x ), with m ∈ Z+ .
x+1
(a) Let m,n be the nth orthogonal polynomial with respect to μm , normalized by the condition
m,n (−1) = (−1)n , then for d ∈ Z+ , on K ⊂ C \ R+ it holds
(d)  m−k √  2( m − k )
m,n+m (z) z+1 z+i
(d)
⇒ Φm−k (z ) = . (10)
k,n+k (z) n 4 2

(b) Let Lm,n be the nth monic orthogonal polynomial with respect to μm , then on K ⊂ C \ R+
it holds
(d)
Lm,n+m (z) √ 2( m − k )
(d)
⇒ ( z + 1) m − k Φ m − k ( z ) = z+i . (11)
Lk,n+k (z) n

Proof. (Proof of a). Taking dσn (t) = (1 − t)1−2n dμ(Ψ(t)), from the assumptions and
Lemma 1, we obtain that dσn is a finite positive Borel measure on [−1, 1], σn > 0 a.e. on
∞
−1/(2n)
[−1, 1] and ∑ ςn = +∞, where ς n is as in (8).
n =1
Let Pn,k be the kth monic orthogonal polynomial with respect to dσn and denote
   
z + 1 n+m
∗m,n+m (z)= Pn,n+m Ψ−1 (z) . After a change of variable x = Ψ(t) in the next
2
integral, we obtain
 ∞   1
x+1 k 1
∗m,n+m ( x )dμm ( x ) = Pn,n+m (t)(1 − t)2m dμ(Ψ(t))
0 2 −1 (1 − t ) n + m + k
 1
dμ(Ψ(t))
= (1 − t)n+m−1−k Pn,n+m (t)
−1 (1 − t)2n−1
 1
= (1 − t)n+m−1−k Pn,n+m (t) dσn (t) = 0, (12)
−1
for k = 0, 1, · · · , n + m − 1.
   
z + 1 n+m
∗m,n+m (−1) = lim Pn,n+m Ψ−1 (z) = (−1)n+m . (13)
z→−1 2

From (12) and (13), we have m,n+m = ∗m,n+m . Therefore,

 n+m  
z+1
m,n+m (z) = Pn,n+m Ψ−1 (z) , (14)
2
m,n+m (z) Pn,n+m Ψ−1 (z)
=
(1 + z)m−k k,n+k (z) 2m−k Pn,n+k (Ψ−1 (z))
1 m −1 Pn,n+ j+1 Ψ−1 (z)
=
2m − k
∏ Pn,n+ j (Ψ−1 (z))
.
j=k

19
Mathematics 2024, 12, 1082

From Lemma 3, for j = k, . . . , m − 1;

Pn,n+ j+1 Ψ−1 (z) ϕ Ψ −1 ( z )

⇒ ; K ⊂ C \ R+ .
Pn,n+ j (Ψ−1 (z)) n 2

Thus,
 m−k  
m,n+m (z) z+1
⇒ ϕ m − k Ψ −1 ( z ) ; K ⊂ C \ R+ ,
k,n+k (z) n 4
which establishes (10) for d = 0. In order to proof (10) for d > 0, we proceed by induction
on d. ⎛ ⎞
( d +1) (d) (d) (d)
m,n+m (z) m,n+m (z) k,n+k (z) m,n+m (z)
( d +1)
= (d) + ( d +1) · ⎝ (d) ⎠.
k,n+k (z) k,n+k (z) k,n+k (z) k,n+k (z)
 
(d) (d) 
Assume that formula (10) holds for d ∈ Z+ , then m,n+m /0,n is uniformly bounded on
(d) ( d +1)
compact subsets K ⊂ C \ R+ . Note that 0,n /0,n ⇒ 0 on K ⊂ C \ R+ . This is proved
n
using an analogous of ([3] (2.9)), and the Bell’s polynomials version of the Faa Di Bruno
formula, see ([10] pp. 218, 219). The assertion (a) is proved.
(d)
m,n+m (z)
(Proof of b). Write f d,m,n (z) = (d)
and let κm,n+m be the leading coefficient of
zm 0,n (z)
m,n+m . Hence, for d > 1

(n + m) · · · (n + m − d + 1)κm,n+m
f d,m,,k,n (∞) =
(n + k) · · · (n + k − d + 1)κk,n+k
κm,n+m
f 0,m,k,n (∞) = .
κk,n+k

From (10),
 m−k
z+1
f d,m,k,n (z) ⇒ Φ m − k ( z ); K ⊂ C \ R+ , l ∈ Z+ . (15)
n 4z
  2( m − k )
κm,n+m 1
lim f d,m,k,n (∞) = lim = . (16)
n→∞ n→∞ κ k,n+k 2
(d)
(d) m,n+m (z)
As Lm,n+m (z) = κm,n+m for d ≥ 1, from (15) and (16), we get (11).

Denote by M[−1, 1] the class of admissible measures in [−1, 1] defined in ([11] Sec. 5).
Let σn a positive varying Borel measure supported on [−1, 1] and

pn,m (w) = τn,m wm + · · · , τn,m > 0

be the mth orthonormal polynomial with respect to σn , then ([11] Th. 7)

τn,n+k+1
lim = 2, k ∈ Z. (17)
n→∞ τn,n+k

Lemma 5. Let σn be an admissible measure, then for all v ∈ Z,

 1
pn,n+v (t) pn,n (t) 1
dσn (t) ⇒ √ ; K ⊂ C \ [−1, 1]. (18)
−1 w−t n ϕ | v | ( w ) w2 − 1

20
Mathematics 2024, 12, 1082

Proof. This proof is based on the proof of ([3] Lemma 2). Without loss of generality, let
us consider v ∈ Z+ . Applying the Cauchy–Schwarz inequality we have, for z ∈ K ⊂
C \ [−1, 1]  1 
 pn,n+v (t) pn,n (t)  1
 dσ ( t ) ≤ < ∞,
 −1 w−t
n  d(K, [−1, 1])
where d(K, [−1, 1]) denotes the Euclidian distance between the two sets. Thus, for (fixed)
values of v ∈ Z+ , the sequence of functions in the left hand side of (18) is normal. Thus,
we deduce uniform convergence from pointwise convergence. The pointwise limit follows
from ([11] Th. 9)
 1  1
pn,n+v (t) pn,n (t) 1 Tv (t) dt
lim dσn (t) = √ ,
n → ∞ −1 w−t π −1 w−t 1 − t2
here, Tv is the vth Chebyshev orthonormal polynomial of the first kind. Therefore, (18)
holds if we prove that
 1
1 Tv (t) dt 1
√ = √ . (19)
π −1 w−t 1 − t2 ϕv (w) w2 − 1

Note that T0 (t) = 1, T1 (t) = x, and, for v ≤ 1,

2tTv (t) = Tv+1 (t) + Tv−1 (t),

or equivalently
Tv+1 = 2tTv − Tv−1 . (20)
Next, proceed by induction. Start at v = 0, expression (18), is obtained from the residue
theorem and Cauchy’s integral formula. Then, for v = 1 we have
 1  1 
1 T1 (t) dt 1w dt 1 1 dt
√ = √ − √
π −1 w−t 1−t 2 − 1 w −πt 1−t 2 π − 1 1 − t2
w 1
= 2 −1 = √ .
w −1 ϕ ( w ) w2 − 1

Now, assume (19) holds for v = 0, 1, . . . , k; k ≥ 1, we will prove that it also holds for
v = k + 1. Combining (20) and the hypothesis of induction, we obtain

 1  
1 Tk+1 (t) dt 1 1 2tTk (t) dt 1 1 Tk−1 (t) dt
√ = √ − √
π −1 w−t 1 − t2 π −1 w − t 1 − t2 π −1 w − t 1 − t2
 
2z 1 Tk (t) dt 1 1 Tk−1 (t) dt
= √ − √
π −1 w − t 1 − t2 π −1 w − t 1 − t2
 
1 2w
= √ −1
ϕ k −1 ( w ) w2 − 1 ϕ ( z )
1
= √ ,
ϕ k +1 ( w ) w2 − 1

which we wanted to prove.

  A− B
x +1
Lemma 6. Let dμ( x ) = 2 dν( x ), where A, B ∈ Z+ , and dν ∈ M [R+ ]. We have on
compact subsets of C \ R+

21
Mathematics 2024, 12, 1082

 ∞ 
x + 1 k  A−k,n+ A−k ( x )− B,n− B ( x )
(v − 1)!τn,n
2
−B dν( x )
0 2 ( x − z)v
⎛ ⎞ ( v −1)
−1
⇒⎝  ⎠ .
(1 + z)(2Φ(z)) A+ B−k
2
n
(Ψ−1 (z)) − 1

where n,n+m is defined as in Lemma 4.

Proof. First, the sequence {n,n+m }n≥0 is well defined because the measure dν ∈ M [R+ ],
implies dμ ∈ M [R+ ] (see Lemma 2).
Let us use the connection formula (14) and the change of variable (7) to obtain
 ∞ 
x + 1 k  A−k,n+ A−k ( x )− B,n− B ( x )
(v − 1)!τn,n
2
−B dν( x ),
0 2 ( x − z)v
 1
Pn,n+ A−k (t) Pn,n− B (t) dσ(t)
= (v − 1)!τn,n
2
−B ,
−1 (Ψ(t) − z)v (1 − t)2n+ A− B
 1
( v −1) (v − 1)!τn,n− B 1 pn,n+ A−k (t) pn,n− B (t)
fn (z) = dσn (t).
τn,n+ A−k −1 1−t (Ψ(t) − z)v

where we use
dμ(Ψ(t)) (1 − t) B− A dν(Ψ(t))
dσn (t) = −
=
(1 − t ) 2n 1 (1 − t)2n−1
Take the (v − 1) primitive with respect to z of the previous expression
 1
τn,n− B 1 pn,n+ A−k (t) pn,n− B (t)
f n (z) = dσn (t). (21)
τn,n+ A−k −1 1−t Ψ(t) − z

Since we know that

 
(1 − t)(Ψ(t) − z) = (1 + z) t − Ψ−1 (z) ,

we rewrite (21) as

τn,n
2 
n,n+ A−k ( t ) Pn,n− B ( t )
1 P
−B
dσn (t),
1 + z −1 t − Ψ −1 ( z )
 1
τn,n− B pn,n+ A−k (t) pn,n− B (t)
= dσn (t).
(1 + z)τn,n+ A−k −1 t − Ψ −1 ( z )

Then, we use Lemma 5 and (17) to obtain on compact subsets of C \ R+ ,

 1
τn,n− B pn,n+ A−k (t) pn,n− B (t)
dσn (t)
(1 + z)τn,n+ A−k −1 t − Ψ −1 ( z )
⎛ ⎞
−1
⇒⎝  ⎠ = f ( z ).
A+ B−k 2
n
(1 + z)(2ϕ(Ψ−1 (z))) (Ψ−1 (z)) − 1

22
Mathematics 2024, 12, 1082

Note that by the Cauchy–Schwarz inequality we have for z ∈ C \ R+

    
 (v−1)   (v − 1)!τn,n− B 1 1 pn,n+ B−k (t) pn,n− A (t) 

 n
f ( z ) =
  dσn ( t ) 
τn,n+ A−k −1 1 − t (ψ(t) − z)v
B
≤ .
d(K, R+ )
 
( v −1)
Then, for each v, the family f n is uniformly bounded in each K ⊂ C \ R+ ,
n  
( v −1)
which means by Montel’s theorem (c.f. [12], §5.4, Th. 15) that f n is normal (see
n ≥0
([12] §5.1 Def. 2)), i.e., we have that from each sequence N ⊂ N we can take a subsequence
N1 ⊂ N such that
(v)
f n ⇒ g( v ) ; n ∈ N 1 , K ⊂ C \ R+ .
n

Now, taking the (v − 1) derivative and using the uniqueness of the limit we obtain
 1
(v − 1)!τn,n− B 1 pn,n+ A−k (t) pn,n− B (t)
dσn (t)
τn,n+ A−k −1 1−t (Ψ(t) − z)v
⎛ ⎞ ( v −1)
− 1
⇒⎝  ⎠ = f ( v −1) ( z ),
n A+ B−k − 2
(1 + z)(2Φ(z)) (Ψ (z)) − 1
1

on compact subsets K ⊂ C \ R+ , which establishes the formula.

3. Relative Asymptotic within Certain Class of Varying Measures

In this section, we obtain the asymptotic relation
  between orthogonal polynomials
m
x +1
with respect to different measures of the class 2 dμ( x ), where μ is any measure of
M [R+ ] and m ∈ Z. Note that, because of Lemma 2, the elements of this class belong to
M [ R+ ] .
To maintain a general tone in the expositions in this section we use μ and ν as two
measures in M [R+ ] having no relation with the previous use of the notation.
 Consider
 m ∈ Z+ and let hm,n (z) be the nth orthogonal polynomial with respect to
m
x +1
2 dν( x ), normalized as hm,n (−1) = (−1)n . Consider the following relations

 ∞ 
x+1 k
h0,n ( x )dν( x ) = 0,
0 2

for k = 0, . . . , n − 1. Apply the change of variable Ψ(t) = z given in (7) to obtain

 1 k
1
0= h0,n (Ψ(t))dν(Ψ(t))
−1 1−t
 1
(1 − t)dν(Ψ(t))
= (1 − t)n−k−1 (1 − t)n h0,n (Ψ(t)) .
−1 (1 − t)2n

Note that the polynomial Hn,n (t) = (1 − t)n h0,n (Ψ(t)) is the nth monic orthogonal
polynomial with respect to the varying measure modified by a polynomial term

(1 − t)dν(Ψ(t))
(1 − t)dσn∗ (t) = .
(1 − t)2n

Following the same reasoning, we obtain that

∗
Hn,n (t) = (1 − t)n h1,n (Ψ(t)),

23
Mathematics 2024, 12, 1082

is the nth monic orthogonal polynomial with respect to dσn∗ (t). It is not hard to prove that
the system {σ, {(1 − t)2n }, 0} is an admissible system, see ([11] Def. p 213). Therefore,
by ([11] Th. 10), we have

Hn,n (t) ϕ ( t ) − ϕ (1)

∗ (t) n
⇒ ; K ⊂ C \ [−1, 1]. (22)
Hn,n t−1

Theorem 2. Under the previous hypothesis we have on compact subsets of C \ R+

 
h0,n (z) z+1
⇒ (1 − Φ(z)), (23)
h1,n (z) n 4
 w−v
hv,n (z) z+1
⇒ (1 − Φ(z))w−v , (24)
hw,n (z) n 4

where v, w ∈ Z.

Proof. From (22) and taking the change of variable (7) we have

h0,n (Ψ(t)) (1 − t)n h0,n (Ψ(t))

=
h1,n (Ψ(t)) (1 − t)n h1,n (Ψ(t))
Hn,n (t) ϕ ( t ) − ϕ (1) Φ −1 ( z ) − 1
= ∗
⇒ = .
Hn,n (t) n t−1 Ψ(z) − 1

To prove (24), note that from Lemma 2.

 k
x+1
dμk = dμ ∈ M [R+ ] if μ ∈ M [R+ ].
2

The only hypothesis needed to obtain (23) is dν ∈ M [R+ ]. Thus if we let now
 k     k +1
dν = x+ 2
1
dμ = dμk , then x+
2
1
dν = x+
2
1
dμ = dμk+1 , where dν ∈ M [R+ ].
Therefore, h0,n = hk,n and h1,n = hk+1,n , where hk,n and hk+1,n are the orthogonal
polynomials with respect to the measures dμk and dμk+1 , respectively, normalized by
having the value (−1)k at −1. Therefore, we have
 
hk,n (z) z+1
⇒ − ( Φ ( z ) − 1). (25)
hk+1,n (z) n 4

Note that, without loss of generality, we can asume w > v, otherwise the relation between
the measures can be reverted, and they still belong to M [R+ ]. Stack formula (25) as

hv1 ,n (z) hv ,n (z) hv1 +1,n (z) hw −1,n (z)

= 1 · · ··· · 1 ,
hw1 ,n hv1 +1,n hv1 +2,n hw1 ,n

where v1 = v + k and w1 = w + k. Since the measure μ ∈ M [R+ ], (24) holds.

4. Asymptotic for Orthogonal Polynomials with Respect to a Measure Modified by a

Rational Factor
Let r = α/β, after canceling out common factors, where

N1 N2
α(z) = ∏ ( z − a i ) Ai , β(z) = ∏ ( z − b j ) Bj ,
i =1 j =1

ai ∈ C \ (R+ ∪ {−1}), b j ∈ C \ R+ , Ai , Bj ∈ N, (26)

N1 N2
A= ∑ Ai , B= ∑ Bj .
i =1 j =1

24
Mathematics 2024, 12, 1082

  A− B
x +1
Given a measure ν ∈ M [R+ ], denote by dμ( x ) = 2 dν( x ) a modified mea-
sure, note that according to Lemma 2 it holds ν ∈ M [R+ ].
Assume Sn is the polynomial of least degree not identically equal to zero, such that
 ∞
0= p( x )Sn ( x )r ( x ) dν( x ), p ∈ Pn −1 , (27)
0

normalized such that Sn (−1) = (−1)n , and Ln is the nth orthogonal polynomial with
respect to dν, normalized such that Ln (−1) = (−1)n . We are interested in the asymptotic
behavior of Sn /Ln , n ∈ Z+ in compact subsets of C \ R+ .

Theorem 3. Let μ ∈ M [R+ ] and α and β defined as before. Then for all sufficiently large n, for all
fixed d ∈ Z+ , in compact subsets of C \ R+ , it holds

N1   Ai N2 Bj
Sn ( z ) (−1) A α(−1) Φ ( z ) − Φ ( ai ) 1
⇒ A
0,n (z) n 4 (z + 1)− A ∏ z − ai ∏ 1−
Φ(z)Φ(b j )
. (28)
i =1 j =1

 k
x+1
Proof. First we focus on (27) for α( x ) = β( x ) where k = 0, . . . , n − B − 1, we have
2
 ∞ 
x+1 k
0= Sn ( x )α( x )dν( x ),
0 2

now, using the change of variables (7) and considering the expression dμ(Ψ(t)) = (1 − t) B− A dν(Ψ(t)),
the previous integral becomes
 1
dμ(Ψ(t))
0= (1 − t)n− B−k−1 (1 − t)n+ A Sn (Ψ(t)) α(Ψ(t)) . (29)
−1 (1 − t)2n−1

for k = 0, . . . , n − B − 1. Define the (n + A)-degree polynomial Rn+ A as

Rn+ A (t) := (1 − t)n+ A Sn (Ψ(t)) α(Ψ(t)).

dσ (t)
Thus, we can consider dσn (t) = (1−t)2n−1 with dσ (t) = dν(Ψ(t)). The measure dσn (t)
defines a varying orthogonal polynomial system, satisfying Lemma 3. We denote by
Pn,n+ A−k the (n + A − k )th monic orthogonal polynomial with respect to dσn (t). According
to (29), we have the following quasi-orthogonality of order n − A

A+ B
Rn+ A (t) := (1 − t)n+ A Sn (Ψ(t)) α(Ψ(t)) = ∑ λn,k Pn,n+ A−k (t). (30)
k =0

Back to (30), we use the connection formula (14) and the change of variables (7)
to obtain
 n+ A A+ B  
2
Sn (z)α(z) = ∑ λn,k Pn,n+ A−k Ψ−1 (z)
z+1 k =0
A+ B  n+ A−k
2
= ∑ λn,k  A−k,n+ A−k (z),
k =0
z+1
A+ B  
z+1 k
Sn (z)α(z) = ∑ λn,k  A−k,n+ A−k (z). (31)
k =0
2

25
Mathematics 2024, 12, 1082

Note that λn,0 = λ0 = (−1) A α(−1) or Sn has deg Sn < n. Dividing this relation by − B,n− B
we get
A+ B  
Sn ( z ) α ( z ) z + 1 k  A−k,n+ A−k (z)
= ∑ λn,k . (32)
− B,n− B (z) k =0
2 − B,n− B (z)
 −1
A+ B
Set λ∗∗ ∗
n,k = λn,k /λ0 , λn = ∑ |λ∗∗
n,k | < ∞ and introduce the polynomials
k =0

A+ B
pn (z) = ∑ λ∗∗
n,k z
A+ B−k
, p∗n = λ∗n pn (z).
k =0

We will prove that


N1   N2
Φ ( ai ) 1
pn (z) ⇒ p̂(z) =
n
∏ z−
2 ∏ z−
2Φ(b j )
; K ⊂ C.
i =1 j =1

To this end, it suffices to show that

 
p∗n (z) ⇒ c p̂(z) = c z A+ B + λ1∗∗ z A+ B−1 + · · · + λ∗∗
A+ B , (33)
n

where 
A+ B −1
c= lim λ∗ = ∑ |λk | . (34)
n→∞ n
k =0

Now, note that { p∗n }, for n ∈ Z+ is contained in P A+ B and the sum of the coefficients of
p∗n for each n ∈ Z+ , is equal to one. Therefore, this family of polynomials is normal. This
means that (33) can be prove if we check that, for all Λ ⊂ Z+ such that

lim p∗n (z) = pΛ , (35)

n→∞
n∈Λ

pΛ (z) = c p̂(z), where p̂(z) and c are defined as above. Since pΛ ∈ P A+ B and pΛ ≡ 0,
we can uniquely determine pΛ if we find its zeros and leading coefficient. Note that the
leading coefficient of pΛ is positive and the sum of the absolute value of its coefficients
is one. Therefore, we conclude that the leading coefficient is uniquely determined by the
zeros. This automatically implies that pΛ (z) = c p̂(z) if and only if it is divisible by p̂(z).
Note that the factor β is in (32) and all the zeros of − B,n− B concentrate on R+ . Thus,
we immediately obtain the following A equations, for n ≥ n0 :
 k   (v)
A+ B  A−k,n+ A−k
z+1
0= ∑ λ∗n λ∗∗ ( a i ),
k =0
n,k
2 − B,n− B

for i = 1, . . . , N1 and v = 0, . . . , A j − 1.
From Lemma 4 it follows that, for compact subsets K ⊂ C \ R+ , it holds
 k   (v)   A+ B   A+ B−k (v)
z+1 n+ A,n+ A−k (z) z+1 Φ(z)
⇒ . (36)
2 − B,n− B (z) n 2 2

Relations (35) and (36), together with the fact that Φ is holomorphic with Φ = 0 in C \ R+ ,
imply, using induction on v, that
 
(v) Φ ( ai )
pΛ = 0, i = 1, . . . , N1 , v = 0, . . . , Ai − 1; (37)
2

26
Mathematics 2024, 12, 1082

  A+ B A+ B   A+ B−k
z+1 Φ(z)
pΛ (z) = c ∑ λ∗∗
k .
2 k =0
2

On the other hand, take p(z) = β(z)− B,n− B (z)/(z − b j )v in (27), j = 1, . . . , N2 ;

λ∗
v = 1, . . . , Bj . Using (31) and multiplying by (v − 1)! λn0 τn,n
2
− B we have the additional rela-
tions
 ∞
λ∗n 2 ( v − 1) !
0= τ − B,n− B ( x )Sn ( x )α( x )dν( x ),
λ0 n,n− B 0 ( x − b j )v
 ∞
( v − 1) !
=τn,n
2
 (x)
−B
( x − b j )v − B,n− B
0
A+ B  
x+1 k
∑ λ∗n λ∗∗
n,k
2
 A−k,n+ A−k ( x )dν( x ),
k =0
A+ B
0= ∑ λ∗n λ∗∗
n,k ( v − 1) !τn,n− B
2
k =0
 ∞ 
x + 1 k  A−k,n+ A−k ( x )− B,n− B ( x )
dν( x ), (38)
0 2 ( x − b j )v

for each b j .
Relations (33),
(38) and Lemma 6 together with the fact that 1/Φ is holomorphic with
(1/Φ) = 0 and 1/
2
(ψ−1 (z)) − 1 = 0 in C \ R+ , give by induction

(v) 1
pΛ = 0, j = 1, . . . , N2 , v = 0, . . . , Bj − 1.
2Φ(b j )

From the previous expression and (37) it follows that pΛ is divisible by p0 (z). Therefore
(33) and (34) hold and
p n ( z ) ⇒ p0 ( z ), K ⊂ C.
n

From the previous expression, the definition of pn , (32), (36) with v = 0, we obtain
   
Sn ( z ) α ( z ) z + 1 A+ B Φ(z)
⇒ (−1) A α(−1) p̂ .
− B,n− B (z) n 2 2

Use the asymptotic formula (10) in the previous expression and group conveniently to obtain

 A
Sn ( z )  (z) z+1 (−1) A α(−1)Φ(z)− B
· − B,n− B ⇒
− B,n− B (z) 0,n (z) n 2 α(z)
   Bj
N1
Φ(z) − Φ( ai ) Ai N2 Φ(z) 1
∏ 2 ∏ 2 − 2Φ(bj )
i =1 i =1

and (28) follows for v = 0. To prove the formula for d ∈ Z+ , apply the same technique of
the proof of Lemma 4.

Remark 1.
1. The proof depends on the assumption of α(−1) = 0, we will remove this restriction in
Section 5.
2. We suppose that α, β are monic. We can remove that restriction without loss of generality due
to the fact that orthogonal polynomial systems are invariant under the constant modification
of measures.

27
Mathematics 2024, 12, 1082

Theorem 3 gives the ratio asymptotic between the orthogonal polynomials with respect
to a rational modification of kind r ( x )dν( x ) (a general rational modification with no zeros
at −1) denoted as Sn and those orthogonal with respect to a modified measure of type
 
x + 1 A− B
, denoted as 0,n .
2
To obtain the general formula we must find the following limit

0,n (z)
lim ,
n→∞ Ln (z)

on compact subsets of C \ R+ , where Ln (z) is the nth orthogonal polynomial with respect
to dν ∈ M [R+ ] normalized such that Ln (−1) = (−1)n .

5. Proof of Theorem 1
Next, we obtain an analogous of (4) for measures with support on R+ . Define α̂ as
 C
z+1
α̂(z) = α(z)
2

wherein α is defined in (26) and C ∈ Z+ is the multiplicity of the zero −1 in α̂/β. With-
out loss of generality we can assume that there are more zeros than poles on −1, if not C = 0.
Also, let Ln be the nth orthogonal polynomial with respect to dν̂ ∈ M [R+ ], normalized by
the condition Ln (−1) = (−1)n . Denote by Qn the nth orthogonal polynomial with respect
to r̂dν̂, where r = α̂/β, normalized as usual, Qn (−1) = (−1)n .
Note that if C = 0, r̂ = r and Qn = Sn , as defined in Section 4. Under this notation,
(6) is written as
⎛√  ⎞ Bj
(d)  C N1  √  z + bj
Qn (z) 2i ai + i Ai N2
(d)
⇒ √
z+i
∏ √ √
z + ai
∏⎝  ⎠ ,
Ln (z) n i =1 j =1 bj + i

in compact subsets of C \ R, for d ∈ Z+ .

 C
x +1 α
Proof of Theorem 1. Let us first observe that Qn is orthogonal with respect to 2 β d ν̂.
Then if we set  
x + 1 −C
dν̂ = dν, (39)
2
we obtain that Qn is orthogonal with respect to αβ dν, and satisfies the hypotheses of
Theorem 3, thus we have on compact subsets of C \ R+

Qn (z)
⇒ F( z ),
0,n (z) n

where F(z) is given in (28).

  A− B
On the other hand, 0,n is orthogonal with respect to x+ 2
1
dν. This means by (39)
  A− B+C
that 0,n is orthogonal with respect to x+2
1
dν̂. Thus, taking into account Theorem 2,
we have  
0,n (z) z + 1 B− A−C
⇒ (1 − Φ(z)) B− A−C .
Ln (z) n 4
Multiply the expressions corresponding to
  B− A−C
z+1
(1 − Φ(z)) B− A−C · F(z), (40)
4

28
Mathematics 2024, 12, 1082

Let us break down this expression into the following terms

   Bj
(−1) A α(−1) N1 Φ(z) − Φ( ai ) Ai N2 1
F( z ) = ∏
4 A ( z + 1) − A i =1 z − ai ∏ 1 − Φ(z)Φ(b j ) .
i =1
N1
(−1) A α(−1) = ∏(1 + ai ) Ai
i =1
2i
(1 − φ(z)) = − √
z−i
Φ ( z ) − Φ ( ai ) −2i
= √ √ √ √
z − ai ( z − i )( ai − i )( ai + z)
 √ 
1 2i bj + z
1− = √  .
Φ(z)Φ(b j ) ( z + i) bj + i

On the other hand

N1   Ai  A N1   Ai
Φ ( z ) − Φ ( ai ) −2i 1
∏ z − ai
= √
z−i (
∏
a − i
√
)(
√ √
ai + z )
i =1 i =1 i
 ⎛ √ ⎞ Bj
N2 Bj   B N2 bj + z
1 2i
∏ 1−
Φ(z)Φ(b j )
= √
z+i
∏⎝  ⎠
j =1 j =1 bj + i

Combining these terms in (40) we obtain

  B− A−C
z+1
(1 − Φ(z)) B− A−C · F(z)
4
  B−C − A 
  A B
1 N1
−2i z + 1 B−C −2i 2i
=
4A
∏ (1 + a i ) Ai √
z−i 4
√
z−i
√
z+i
i =1
⎛ √ ⎞ Bj
N1   Ai N2 bj + z
1
·∏ √
( a − i )(
√
a +
√
z )
∏⎝  ⎠ .
i =1 i i j =1 bj + i

Finally, taking into account

N1  √  N1   Ai
ai + i Ai N1
1
∏ √ √
ai + z
= ∏ (1 + a i ) Ai · ∏ √ √ √
i =1 i =1 i =1 ( ai − i )( ai + z )
 C      A B
2i 1 −2i B−C− A z + 1 B−C −2i 2i
√ = A √ √ √
z+i 4 z−i 4 z−i z+i

we obtain (6) for d = 0. To prove (6) for d ≥ 1, use induction in d and the method from the
proof of Lemma 4. The proof is complete.

Author Contributions: Conceptualization, H.P.-C. and J.Q.-R.; methodology, H.P.-C. and J.Q.-R.;
validation, C.F.-S., H.P.-C. and J.Q.-R.; formal analysis, C.F.-S. and J.Q.-R.; investigation, C.F.-S.,
H.P.-C. and J.Q.-R.; writing—original draft preparation, J.Q.-R.; writing—review and editing, C.F.-S.,
H.P.-C. and J.Q.-R.; supervision, H.P.-C.; project administration, C.F.-S.; funding acquisition, C.F.-S.
All authors have read and agreed to the published version of the manuscript.
Funding: The research of C. Féliz-Sánchez was partially supported by Fondo Nacional de Innovación y
Desarrollo Científico y Tecnológico (FONDOCYT), Dominican Republic, under grant 2020-2021-1D1-136.

29
Mathematics 2024, 12, 1082

Data Availability Statement: No new data were created or analyzed in this study. Data sharing is
not applicable to this article.
Conflicts of Interest: The authors declare no conflicts of interest.

References
1. Schmüdgen, K. The Moment Problem; Graduate Texts in Mathematics; Springer: Cham, Switzerland, 2017; Volume 27.
2. Uvarov, V.B. The connection between systems of polynomials orthogonal with respect to different distribution functions. USSR
Comput. Math. Math. Phys. 1969, 9, 25–36. [CrossRef]
3. Lagomasino, G.L.; Marcellán, F.; Assche, W.V. Relative asymptotics for orthogonal polynomials with respect to a discrete Sobolev
inner product. Constr. Approx. 1995, 11, 107–137.
4. Pijeira-Cabrera, H.; Quintero-Roba, J.; Toribio-Milane, J. Differential Properties of Jacobi-Sobolev Polynomials and Electrostatic
Interpretation. Mathematics 2023, 11, 3420. [CrossRef]
5. Gautschi, W. Orthogonal Polynomials: Computation and Approximation; Numerical Mathematics and Scientific Computation Series;
Oxford University Press: New York, NY, USA, 2004.
6. Lagomasino, G.L. Survey on multipoint Padé approximation to Markov-type meromorphic functions and asymptotic properties
of the orthogonal polynomials generated by them. In Polynômes Orthogonaux et Applications; Lecture Notes in Mathematics;
Springer, Berlin/Heidelberg, Germany, 1985; Volume 1171, pp. 309–316.
7. Lagomasino, G.L. Convergence of Padé approximants of Stieltjes type meromorphic functions and comparative asymptotics for
orthogonal polynomials. Mat. Sb. 1988, 136, 46–66; English transl. in Math. USSR Sb. 1989, 64, 207–227.
8. Lagomasino, G.L. Relative asymptotics for orthogonal polynomials on the real axis. Mat. Sb. 1988, 137, 500–525; English transl. in
Math. USSR Sb. 1990, 65, 505–529.
9. Díaz-González, A.; Hernández, J.; Pijeira-Cabrera, H. Sequentially Ordered Sobolev Inner Product and Laguerre–Sobolev
Polynomials. Mathematics 2023, 11, 1956. [CrossRef]
10. Johnson. W., The curious history of Faa di Bruno’s formula. Am. Math. Mon. 2003, 4, 358.
11. Lagomasino, G.L. Asymptotics of polynomials orthogonal with respect to varying measures. Constr. Approx. 1989, 5, 199–219.
[CrossRef]
12. Ahlfors, L.V. Complex Analysis; McGraw-Hill, Inc.: New York, NY, USA, 1979.

Disclaimer/Publisher’s Note: The statements, opinions and data contained in all publications are solely those of the individual
author(s) and contributor(s) and not of MDPI and/or the editor(s). MDPI and/or the editor(s) disclaim responsibility for any injury to
people or property resulting from any ideas, methods, instructions or products referred to in the content.

30
 mathematics

Article
Some Properties of Generalized Apostol-Type
Frobenius–Euler–Fibonacci Polynomials
Maryam Salem Alatawi 1 , Waseem Ahmad Khan 2, *, Can Kızılateş 3 and Cheon Seoung Ryoo 4

1 Department of Mathematics, Faculty of Science, University of Tabuk, Tabuk 71491, Saudi Arabia;
[email protected]
2 Department of Mathematics and Natural Sciences, Prince Mohammad Bin Fahd University,
P.O. Box 1664, Al Khobar 31952, Saudi Arabia
3 Department of Mathematics, Zonguldak Bülent Ecevit University, Zonguldak 67100, Turkey;
[email protected]
4 Department of Mathematics, Hannam University, Daejeon 34430, Republic of Korea; [email protected]
* Correspondence: [email protected]

Abstract: In this paper, by using the Golden Calculus, we introduce the generalized Apostol-type
Frobenius–Euler–Fibonacci polynomials and numbers; additionally, we obtain various fundamental
identities and properties associated with these polynomials and numbers, such as summation theo-
rems, difference equations, derivative properties, recurrence relations, and more. Subsequently, we
present summation formulas, Stirling–Fibonacci numbers of the second kind, and relationships for
these polynomials and numbers. Finally, we define the new family of the generalized Apostol-type
Frobenius–Euler–Fibonacci matrix and obtain some factorizations of this newly established matrix.
Using Mathematica, the computational formulae and graphical representation for the mentioned
polynomials are obtained.

Keywords: Golden Calculus; Apostol-type Frobenius–Euler polynomials; Apostol-type Frobenius–

Euler–Fibonacci polynomials; Stirling–Fibonacci numbers

MSC: 11B68; 11B83; 05A15; 05A19

Citation: Alatawi, M.S.; Khan, W.A.;
Kızılateş, C.; Ryoo, C.S. Some
Properties of Generalized Apostol-
Type Frobenius–Euler–Fibonacci 1. Introduction
Polynomials. Mathematics 2024, 12,
Recently, numerous scholars [1–3] have defined and developed methods of generating
800. https://doi.org/10.3390/
math12060800
functions for new families of special polynomials, including Bernoulli, Euler, and Genocchi
polynomials. These authors have established the basic properties of these polynomials
Academic Editor: Ioannis K. Argyros and have derived a variety of identities using the generating function. Furthermore,
by using the partial derivative operator to these generating functions, some derivative
Received: 10 February 2024 formulae and finite combinatorial sums involving the above-mentioned polynomials and
Revised: 6 March 2024 numbers have been obtained. These special polynomials also provide the straightforward
Accepted: 6 March 2024 derivation of various important identities. As a result, numerous experts in number theory
Published: 8 March 2024 and combinatorics have exhaustively studied their properties and obtained a series of
interesting results.
For any u ∈ C, u = 1 and ζ ∈ R, the Apostol-type Frobenius–Euler polynomials
(α)
Hw (ζ; u; λ) of order α ∈ C are introduced (see [4–7]).
Copyright: © 2024 by the authors.
    
Licensee MDPI, Basel, Switzerland.
1 − u α ζd ∞
(α) dw  
ln λ .
This article is an open access article
λed − u
e = ∑ w H ( ζ; u; λ )
w!
, | d | <  u  (1)
distributed under the terms and w =0
conditions of the Creative Commons
Attribution (CC BY) license (https://
creativecommons.org/licenses/by/
4.0/).

Mathematics 2024, 12, 800. https://doi.org/10.3390/math12060800 31 https://www.mdpi.com/journal/mathematics

Mathematics 2024, 12, 800

(α) (α)
For ζ = 0, Hw (u; λ) = Hw (0; u; λ) are called the Apostol-type Frobenius–Euler
numbers of order α. From (1), we known that
w  
(α) w (α)
Hw (ζ; u; λ) = ∑ Hs (u; λ)ζ w−s , (2)
s =0
s
and
(α) (α)
Hw (ζ; −1; λ) = Ew (ζ; λ), (3)
(α)
where Ew (ζ; λ) are the wth Apostol–Euler polynomials of order α.
The generalized λ-Stirling numbers of the second kind S(w, s; λ) are given by (see [8])
∞
(λed − 1)s dw
= ∑ S(w, s; λ) , (4)
s! w =0
w!

for λ ∈ C and s ∈ N = {0, 1, 2, · · · , }, where λ = 1 gives the well-known Stirling numbers

of the second kind; these are defined as follows (see [9,10]).
∞
( e d − 1) s dw
= ∑ S(w, s) . (5)
s! w =0 w!

By referring to (4), the λ-array type polynomials Ssw (ζ, λ) are defined by (see [11])
∞
(λed − 1)s ζd dw
e = ∑ S(w, s; ζ; λ) . (6)
s! w =0
w!

(α)
The Apostol-type Bernoulli polynomials Bw (ζ; λ) of order α, the Apostol-type Euler
(α) (α)
polynomials Ew (ζ; λ) of order α, and the Apostol-type Genocchi polynomials Gw (ζ; λ)
of order α are defined by (see [8,12]):
 α ∞
d (α) dw
eζd = ∑ Bw (ζ; λ) (| d + log λ |< 2π ), (7)
λed−1 w =0 w!
 α ∞
2 (α) dw
eζd = ∑ Ew (ζ; λ) (| d + log λ |< π ) (8)
λed + 1 w =0
w!
and  α ∞
2d (α) dw
eζd = ∑ Gw (ζ; λ) , (| d + log λ |< π ), (9)
λed+1 w =0
w!
respectively.
Clearly, we have

(α) (α) (α) (α) (α) (α)

Bw (λ) = Bw (0; λ), Ew (λ) = Ew (0; λ), Gw (λ) = Gw (0; λ).

The subject of Golden Calculus (or F-calculus) emerged in the nineteenth century due
to its wide-ranging applications in fields such as mathematics, physics, and engineering.
The ψ-extended finite operator calculus of Rota was studied by A.K. Kwaśniewski [13].
Krot [14] defined and studied F-calculus and gave some properties of these calculus types.
Pashaev and Nalci [15] dealt extensively with the Golden Calculus and obtained many
properties and used these concepts especially in the field of mathematical physics. The
definitions and notation of Golden Calculus (or F-calculus) are taken from [15–18].
The Fibonacci sequence is defined by the following recurrence relation:

Fw = Fw−1 + Fw−2 , w ≥ 2

32
Mathematics 2024, 12, 800

where F0 = 0, F1 = 1. Fibonacci numbers can be expressed explicitly as

φw − ψw
Fw = ,
φ−ψ
√ √
where φ = 1+2 5 and ψ = 1−2 5 . φ ≈ 16180339 · · · is called Golden ratio. The Golden ratio
is a frequently occurring number in many branches of science and mathematics. Pashaev
and Nalci [15] have thoroughly studied the miscellaneous properties of Golden Calculus.
Additional references include Pashaev [18], Krot [14], and Pashaev and Ozvatan [19].
The F-factorial was defined as follows:

F1 F2 F3 · · · Fw = Fw !, (10)

where F0 ! = 1. The binomial theorem for the F-analogues (or the Golden binomial theorem)
are given by
w  
w l
F = ∑
(ζ + η )w := (ζ + η )w (−1)(2) ζ w−l η l , (11)
l =0
l F

in terms of the Golden binomial coefficients, referred to as Fibonomials

 
w Fw !
= ,
l F Fw−l! Fl !

with w and l being non-negative integers, w ≥ l. The Fibonomial coefficients have follow-
ing identity:        
w l w w−m
= . (12)
l F m F m F l−m F
The F-derivative is introduced as follows:
∂F f (φζ ) − f (ψζ )
( f (ζ )) = . (13)
∂F ζ (φ − ψ)ζ

respectively. The first and second types of Golden exponential functions are defined as
∞ (ζ )w
e F (ζ ) = ∑ F !
F
, (14)
w =0 w

∞ w (ζ )w
EF ( ζ ) = ∑ (−1)( 2 ) F
Fw !
. (15)
w =0

Briefly, we use the following notations throughout the paper

∞
ζw
e F (ζ ) = ∑ Fw !
w =0

and
∞ w ζw
EF ( ζ ) = ∑ (−1)( 2 ) Fw ! .
w =0

e F (ζ ) and EF (ζ ) satisfy the following identity (see [17]).

ζ η (ζ +η ) F
e F EF = e F . (16)

33
Mathematics 2024, 12, 800

(α)
The Apostol-type Bernoulli–Fibonacci polynomials Bw,F (ζ; λ) of order α, the Apostol-
(α)
type Euler–Fibonacci polynomials Ew,F (ζ; λ) of order α and the Apostol-type Genocchi–
(α)
Fibonacci polynomials Gw,F (ζ; λ) of order α are defined by (see [20–22]):
 α ∞
d ζd (α) dw
λedF − 1
eF = ∑ Bw,F (ζ; λ)
Fw !
, (17)
w =0
 α ∞
2 ζd (α) dw
λedF +1
eF = ∑ Ew,F (ζ; λ)
Fw !
(18)
w =0

and  α ∞
2d ζd (α) dw
λedF + 1
eF = ∑ Gw,F (ζ; λ)
Fw !
, (19)
w =0

respectively.
Clearly, we have

(α) (α) (α) (α) (α) (α)

Bw,F (λ) = Bw,F (0; λ), Ew,F (λ) = Ew,F (0; λ), Gw,F (λ) = Gw,F (0; λ).

In light of the above studies, we define a new family of two-variable polynomials,

including the polynomials defined by Equation (1) with the help of the Golden Calcu-
lus. Namely, we introduce the concept of the generalized Apostol-type Frobenius–Euler–
Fibonacci polynomials and numbers. Thus, we give some properties of this polynomial
family, such as recurrence relations, sums formulae, and derivative relations, by using
their generating function and functional equations. Additionally, we establish relationships
between Apostol-type Frobenius–Euler–Fibonacci polynomials of order α and various
other polynomial sequences, including Apostol-type Bernoulli–Fibonacci polynomials,
Euler–Fibonacci polynomials, Genocchi–Fibonacci polynomials, and the Stirling–Fibonacci
numbers of the second kind. We also introduce the new family of the generalized Apostol-
type Frobenius–Euler–Fibonacci matrix and derive some factorizations of this newly estab-
lished matrix. Finally, we provide zeroes and graphical illustrations for the generalized
Apostol-type Frobenius–Euler–Fibonacci polynomials.

(α)
2. Generalized Apostol-Type Frobenius–Euler–Fibonacci Polynomials Hw,F (ζ, η; u; λ)
In this part, we introduce Apostol-type Frobenius–Euler–Fibonacci polynomials by
means of the Golden Calculus. Some relations for these polynomials are also obtained by
using various identities. At this point, we begin with the following definition.

Definition 1. Let λ ∈ C, α ∈ N, the generalized Apostol-type Frobenius–Euler polynomials

(α)
Hw,F (ζ, η; u; λ) of order α are defined by means of the following generating function:
 α ∞
1−u ζd ηd (α) dw
λedF − u
e F EF = ∑ Hw,F (ζ, η; u; λ)
Fw !
. (20)
w =0

(α) (α)
When ζ = η = 0 in (20), Hw,F (u; λ) = Hw,F (0, 0; u; λ) are called the wth Apostol-type
Frobenius–Euler–Fibonacci numbers of order α.

Theorem 1. The following summation formulas for the generalized Apostol-type Frobenius–Euler–
(α)
Fibonacci polynomials Hw,F (ζ, η; u; λ) of order α holds true:

w  
(α) w (α) −s
Hw,F (ζ, η; u; λ) = ∑ H (0, 0; u; λ)(ζ + η )w
s F s,F F , (21)
s =0

34
Mathematics 2024, 12, 800

w  
(α) w (α)
Hw,F (ζ, η; u; λ) = ∑ H (0, η; u; λ)ζ w−s (22)
s =0
s F s,F
and  
w
(α) s ( s −1) w (α)
Hw,F (ζ, η; u; λ) = ∑ (−1) 2 H
s F w−s,F
(ζ, 0; u; λ)η s . (23)
s =0

Proof. By virtue of (14)–(16) and (20), we obtain the desired results.

Theorem 2. The following recursive formulas for the generalized Apostol-type Frobenius–Euler–
(α)
Fibonacci polynomials Hw,F (ζ, η; u; λ) of order α hold true:

∂ F  (α) 
(α)
Hw,F (ζ, η; u; λ) = Fw Hw−1,F (ζ, η; u; λ), (24)
∂F ζ

∂ F  (α) 
(α)
Hw,F (ζ, η; u; λ) = Fw Hw−1,F (ζ, −η; u; λ). (25)
∂F η

Proof. Differentiating both sides of (20) with respect to ζ and η through Equation (13), we
obtain (24) and (25), respectively.

Theorem 3. The following difference formulas for the generalized Apostol-type Frobenius–Euler–
(α)
Fibonacci polynomials Hw,F (ζ, η; u; λ) of order α holds true:

(α) (α) ( α −1)

λHw,F (1, η; u; λ) − uHw,F (0, η; u; λ) = (1 − u)Hw,F (0, η; u; λ) (26)

and
(α) (α) ( α −1)
λHw,F (1, 0; u; λ) − uHw,F (1, −1; u; λ) = (1 − u)Hw,F (1, −1; u; λ). (27)

Proof. By virtue of (20), we can easily proof of Equations (26) and (27). We omit the proof.

Theorem 4. Let α, β ∈ N, the generalized Apostol-type Frobenius–Euler–Fibonacci polynomials

(α)
Hw,F (ζ, η; u; λ) of order α hold true:
w  
(α+ β) w (α) ( β)
Hw,F (ζ, η; u; λ) = ∑ s
Hw−s,F (0, 0; u; λ)Hs,F (ζ, η; u; λ), (28)
s =0 F

and  
w
(α− β) w (α) (− β)
Hw,F (ζ, η; u; λ) = ∑ Hw−s,F (0, 0; u; λ)Hs,F (ζ, η; u; λ). (29)
s =0 s F

Proof. Using generating function (20), we obtain Equations (28) and (29). We omit
the proof.

In the following theorems, we establish some results on the generalized Apostol-type

(α)
Frobenius–Euler–Fibonacci polynomials Hw,F (ζ, η; u; λ) of order α and some relationships
for Apostol-type Frobenius–Euler–Fibonacci polynomials of order α related to Apostol-
type Bernoulli–Fibonacci polynomials, Apostol-type Euler–Fibonacci polynomials, and
Apostol-type Genocchi–Fibonacci polynomials. We now begin with the following theorem.

Theorem 5. For the generalized Apostol-type Frobenius–Euler–Fibonacci polynomials Hw,F (ζ, η; u; λ),
one has
w  
w
(2u − 1) ∑ H (0, η; u; λ)Hw−l,F (ζ, 0; 1 − u; λ)
l =0
l F l,F

35
Mathematics 2024, 12, 800

= uHw,F (ζ, η; u; λ) − (1 − u)Hw,F (ζ, η; 1 − u; λ). (30)

Proof. We set
(2u − 1) 1 1
= d − .
(λedF − u)(λedF − (1 − u)) λe F − u λedF − (1 − u)

From the above equation, we see that

ηd
(1 − u)eζd
F (1 − (1 − u )) EF
(2u − 1)
(λedF − u)(λedF − (1 − u))
ηd ηd
(1 − u)eζd
F uEF (1 − u)eζd
F EF (1 − (1 − u ))
= − ,
λedF − u λedF − (1 − u)
through which, in using Equations (16) and (20) in both sides, we have
 
∞ ∞
dl dw
(2u − 1) ∑ Hl,F (0, η; u; λ) ∑ Hw,F (ζ, 0; 1 − u; λ)
l =0
Fl ! w =0
Fw!

∞ ∞
dw dw
=u ∑ Hw,F (ζ, η; u; λ)
Fw !
− (1 − u) ∑ Hw,F (ζ, η; 1 − u; λ)
Fw !
.
w =0 w =0

By applying the Cauchy product rule in the aforementioned equation and subsequently
comparing the coefficients of dw in both sides of the resulting equation, it can be deduced
that assertion (30) holds true.

Theorem 6. For the generalized Apostol-type Frobenius–Euler–Fibonacci polynomials Hw,F (ζ, η; u; λ),
one has
w  
w
uHw,F (ζ, η; u; λ) = λ ∑ H (ζ, η; u; λ) − (1 − u)(ζ + η )w F. (31)
l =0
l F l,F

Proof. Using the following identity

u 1 1
= − ,
λ(λedF − u)edF (λedF − u) λedF

we find that
ζd ηd ηd ηd
u (1 − u ) e F EF (1 − u)eζd
F EF (1 − u)eζd
F EF
= −
λ(λedF − u)edF λedF − u λedF
∞
dw
u ∑ Hw,F (ζ, η; u; λ)
Fw !
w =0
∞ ∞ ∞
dw dl dw
=λ ∑ Hw,F (ζ, η; u; λ)
Fw ! ∑ Fl !
− (1 − u ) ∑ ( ζ + η ) w
F
Fw !
.
w =0 l =0 w =0

By applying the Cauchy product rule in the aforementioned equation and subsequently
comparing the coefficients of dw in both sides of the resulting equation, it can be deduced
that assertion (31) holds true.

Theorem 7. For the generalized Apostol-type Frobenius–Euler–Fibonacci polynomials Hw,F (ζ, η; u; λ)

of order α, we obtain
w  
(α) 1 w (α)
Hw,F (ζ, η; u; λ) =
1−u ∑ l F
λHw−l,F (1, η; u; λ)Hl,F (ζ, 0; u; λ)
l =0

36
Mathematics 2024, 12, 800

(α)
−uHw−l,F (0, η; u; λ)Hl,F (ζ, 0; u; λ) . (32)

Proof. Using (20), we find that

   α
∞ λedF − u
(α) dw 1−u 1−u ζd ηd
∑ Hw,F (ζ, η; u; λ)
Fw !
=
λedF − u 1−u λedF − u
e F EF
w =0
  α
λ 1−u 1−u ζd ηd
= e d
e F EF
1 − u λedF − u F
λedF − u
  α
u 1−u 1−u ζd ηd
− e F EF .
1 − u λedF − u λedF − u
Simplifying the above equation and using Equation (20), we obtain
∞ ∞ ∞
(α) dw λ dw (α) dl
∑ Hw,F (ζ, η; u; λ)
Fw !
=
1−u ∑ Hw,F (1, η; u; λ)
Fw ! ∑ Hl,F (ζ, 0; u; λ) Fl ! −
w =0 w =0 l =0

∞ ∞
u dw (α) dl
1−u ∑ Hw,F (0, η; u; λ)
Fw ! ∑ l,F
H (ζ, 0; u; λ) .
Fl !
w =0 l =0

By applying the Cauchy product rule in the aforementioned equation and subsequently
comparing the coefficients of dw in both sides of the resulting equation, it can be deduced
that assertion (32) holds true.

Theorem 8. The following relation between the generalized Apostol-type Frobenius–Euler–Fibonacci

(α)
polynomials Hw,F (ζ, η; u; λ) and Apostol-type Bernoulli–Fibonacci polynomials Bw,F (ζ; λ) holds true:

w +1     
(α) w+1 l
l
Hw,F (ζ, η; u; λ) = ∑ l
λ ∑ r Bl−r,F (ζ; λ) − Bl,F (ζ; λ)
l =0 F r =0 F

(α)
×Hw−l +1,F (0, η; u; λ). (33)

Proof. Consider generating function (20), we have

∞
(α) dw
∑ Hw,F (ζ, η; u; λ)
Fw !
w =0
 α  
1−u ζd ηd d λedF − 1
= e F EF
λedF − u λedF − 1 d

∞ ∞ ∞
1 (α) dw dl dr
= λ ∑ Hw,F (0, η; u; λ) ∑ Bl,F (ζ; λ) Fl ! ∑
d w =0
Fw ! l =0 r =0
Fr !
∞ ∞
(α) dw dl
− ∑ Hw,F (0, η; u; λ)
Fw ! ∑ Bl,F (ζ; λ) Fl ! . (34)
w =0 l =0

Using the Cauchy product rule in (34), the assertion (33) is obtained.

Theorem 9. The following relation between the generalized Apostol-type Frobenius–Euler–Fibonacci

(α)
polynomials Hw,F (ζ, η; u; λ) and generalized Apostol-type Euler–Fibonacci polynomials Ew,F (ζ; λ)
holds true:

37
Mathematics 2024, 12, 800

   l  
(α) 1 w w l (α)
2 l∑
Hw,F (ζ, η; u; λ) = λ∑ El −r,F (ζ; λ) + El,F (ζ; λ) Hw−l,F (0, η; u; λ). (35)
=0
l F r = 0
r F

Proof. By virtue of (20), we have

∞
(α) dw
∑ Hw,F (ζ, η; u; λ)
Fw !
w =0
 α  
1−u ζd ηd 2 λedF + 1
= e F EF
λedF − u λedF + 1 2

∞ ∞ ∞
1 (α) dw dl dr
= λ ∑ Hw,F (0, η; u; λ) ∑ El,F (ζ; λ) Fl ! ∑
2 w =0 Fw ! l =0 r =0 Fr !
∞ ∞
(α) dw dl
+ ∑ Hw,F (0, η; u; λ)
Fw ! ∑ El,F (ζ; u; λ) Fl ! . (36)
w =0 l =0

Using the Cauchy product rule in (36), the assertion (35) is obtained.

Theorem 10. The following relation between the generalized Apostol-type Frobenius–Euler–Fibonacci
(α)
polynomials Hw,F (ζ, η; u; λ) and Apostol-type Genocchi–Fibonacci polynomials Gw,F (ζ; λ) holds true:
   l  
(α) 1 w +1 w + 1 l
Hw,F (ζ, η; u; λ) = ∑ λ∑ G (ζ; λ) + Gl,F (ζ; λ)
2 l =0 l F r =0
r F l −r,F

(α)
×Hw−l +1,F (0, η; u; λ). (37)

Proof. Using (20), we obtain

∞
(α) dw
∑ Hw,F (ζ, η; u; λ)
Fw !
w =0
 α  
1−u ζd ηd 2d λedF + 1
= e F EF
λedF − u λedF+1 2d

∞ ∞ ∞
1 (α) dw dl dr
= λ ∑ Hw,F (0, η; u; λ) ∑ Gl,F (ζ; λ) Fl ! ∑
2d w =0 Fw! l =0 r =0 Fr !
∞ ∞
(α) dw dl
+ ∑ Hw,F (0, η; u; λ)
Fw ! ∑ Gl,F (ζ; λ) Fl ! . (38)
w =0 l =0

Using the Cauchy product rule in (38), the assertion (37) is obtained.

(α)
Theorem 11. For the generalized Apostol-type Frobenius–Euler–Fibonacci polynomials Hw,F (ζ, η; u; λ)
of order α, we obtain
w  
( α +1) w (α)
Hw,F (ζ, η; u; λ) = ∑ s
Hw−s,F (u; λ)Hs,F (ζ, η; u; λ). (39)
s =0 F

Proof. From (20), we obtain

 α ∞
1−u 1−u ζd ηd 1−u (α) ds
λedF − u λedF − u
e F EF =
λedF − u
∑ Hs,F (ζ, η; u; λ) Fs !
s =0

38
Mathematics 2024, 12, 800

∞ ∞
dw (α) ds
= ∑ Hw,F (u; λ)
Fw ! ∑ Hs,F (ζ, η; u; λ) Fs ! .
w =0 s =0

Now, replacing w with w − s and equating the coefficients of dw leads to Formula (39).

(α)
Theorem 12. For the generalized Apostol-type Frobenius–Euler–Fibonacci polynomials Hw,F (ζ, η; u; λ)
of order α, we have
w w−l
 
∑ (−1)( 2 )
(α) w (α)
Hw,F (ζ + 1, η; u; λ) = H (ζ, η; u; λ). (40)
l =0
l F l,F

Proof. Using (20), we have

∞ ∞
(α) dw (α) dw
∑ Hw,F (ζ + 1, η; u; λ) − ∑ Hw,F (ζ, η; u; λ)
Fw ! w=0 Fw !
w =0
 α
1−u ζd ηd
= e F EF ( EFd − 1)
λedF − u
 
∞
(α) dl ∞ (w2 ) dw ∞
(α) dw
= ∑ Hl,F (ζ, η; u; λ) Fl ! ∑ (−1) Fw !
− ∑ Hw,F (ζ, η; u; λ)
Fw !
l =0 w =0 w =0
  
∞ w w−l ∞
dw dw
∑ (−1)( 2 ) l Hl,F (ζ, η; u; λ) Fw ! − ∑ Hw,F (ζ, η; u; λ) Fw ! .
w (α) (α)
= ∑
w =0 l =0 F w =0

Finally, equating the coefficients of the like powers of dw , we obtain (40).

Theorem 13. Let α and γ be non-negative integers. There is the following relationship between
the numbers S F (w, l; λ) and the generalized Apostol-type Frobenius–Euler–Fibonacci polynomials
(α)
Hw,F (ζ, η; u; λ) of order α, which holds true:
     
w
w (α) λ 1−u α
α! ∑ Hw−l,F (ζ, η; u; λ)SF l, α; = (ζ + η )w
F (41)
l =0
l F u u

and
 γ w    
(α−γ) u w (α) λ
Hw,F (ζ, η; u; λ) = γ!
1−u ∑ l H w−l,F ( ζ, η; u; λ ) S F l, γ;
u
, (42)
l =0 F

where S F (w, l; λ) is the Stirling–Fibonacci numbers of the second kind are defined by

(λedF − 1)l ∞
dw
= ∑ SF (w, l; λ) . (43)
l! w =0
Fw !

Proof. By virtue of (20), we find that

 α
∞
(α) dw 1−u ζd ηd
∑ Hw,F (ζ, η; u; λ)
Fw !
=
λedF − u
e F EF
w =0
 α ∞
dw ∞
dw
(α)
λedF − u ∑ Hw,F (ζ, η; u; λ) = (1 − u ) α ∑ ( ζ + η ) w
F
w =0
Fw ! w =0
Fw!
 α
λ d
−1 ∞  α ∞
u eF (α) dw 1−u dw
α!
α! ∑ Hw,F (ζ, η; u; λ)
Fw !
=
u ∑ (ζ + η )wF Fw !
w =0 w =0

39
Mathematics 2024, 12, 800

∞ ∞  
(α) dw λ dl
α! ∑ Hw,F (ζ, η; u; λ)
Fw ! ∑ SF l, α;
u Fl !
w =0 l =0
 α ∞
1−u dw
=
u ∑ (ζ + η )wF Fw ! ,
w =0

which, on rearranging the summation and then simplifying the resultant equation, yields
the relation (41).
Once more, we examine the following arrangement of generating function (20) as:
  γ
α  γ λ d
u eF − 1
∞
(α−γ) dw 1−u ζd ηd u
∑ Hw,F (ζ, η; u; λ) Fw ! = λed − u eF EF 1 − u γ! γ!
, (44)
w =0 F

on use of Equations (44) and (20). After evaluation, the desired result is obtained (42).

Now, we define the new family of generalized Apostol-type Frobenius–Euler–Fibonacci

matrices. By using this definition, we obtain the factorizations of this newly established
matrix in the following theorems.

(α)
Definition 2. Let Hw,F (ζ, η; u; λ) be the generalized Apostol-type Frobenius–Euler–Fibonacci
polynomials. The (n + 1) × (n + 1) generalized Apostol-type Frobenius–Euler-Fibonacci matrix,
(α) (α) n
Hn,F (ζ, η; u; λ) = hij (ζ, η; u; λ) is defined by
i,j=0
⎧ 
⎨ i H(α) (ζ, η; u; λ) i≥j
(α)
hij (ζ, η; u; λ)= j F i− j,F . (45)
⎩
0 i<j

(α)
Theorem 14. For the generalized Apostol-type Frobenius–Euler–Fibonacci matrix Hn,F (ζ, η; u; λ),
we have
(α+ β) (α) ( β)
Hn,F (ζ + ψ, η; u; λ) = Hn,F (ζ, η; u; λ)Hn,F (0, ψ; u; λ).

Proof. By virtue of (12), (16), (20), and (45), we find that


(α+ β) i (α+ β)
Hn,F (ζ + ψ, η; u; λ) = H (ζ + ψ, η; u; λ)
j F i− j,F
  i− j  
i i−j (α) ( β)
= ∑
j F k =0 k
Hi− j−k,F (ζ, η; u; λ)Hk,F (0, ψ; u; λ)
F
i    
i i−j (α) ( β)
= ∑ Hi−k,F (ζ, η; u; λ)Hk− j,F (0, ψ; u; λ)
k= j
j F k − j F
i    
i (α) k ( β)
= ∑ Hi−k,F (ζ, η; u; λ) H (0, ψ; u; λ)
k= j
k F j F k− j,F
(α) ( β)
= Hn,F (ζ, η; u; λ)Hn,F (0, ψ; u; λ).

Theorem 15. For the generalized Apostol-type Frobenius–Euler–Fibonacci matrix Hn,F (ζ, η; u; λ),
we have
Hn,F (ζ + η, 0; u; λ) = Pn,F (ζ )Hn,F (0, η; u; λ),

40
Mathematics 2024, 12, 800

 n
where Pn,F (ζ ) = pij (ζ ) i,j=0 is the generalized Pascal matrix [23] via Fibonomial coefficients of
the first kind is defined by ⎧ 
⎨ i ζ i− j i ≥ j
Pn,F (ζ ) = j F .
⎩
0 i<j

Proof. Using (45) and (12), we obtain

  i− j  
i i−j
Hn,F (ζ + η, 0; u; λ) = ∑ ζ i− j−k Hk,F (0, η; u; λ)
j F k =0 k F
i    
i i−j
= ∑ j ζ i−k k − j Hk− j,F (0, η; u; λ)
k= j F F
i    
i k
= ∑ k ζ i−k j Hk− j,F (0, η; u; λ)
k= j F F

= Pn,F (ζ )Hn,F (0, η; u; λ).

3. Some Values with Graphical Representations and Zeros of the Generalized

Apostol-Type Frobenius–Euler–Fibonacci Polynomials
In this section, evidence of the zeros of the generalized Apostol-type Frobenius–Euler–
Fibonacci polynomials is displayed, along with visually appealing graphical representa-
tions. A few of them are presented here:

41
Mathematics 2024, 12, 800

 
(α) −1 + u α
H0,F (ζ, η; u; λ) = ,
−λ + u
 α  α  α  α  α
(α)
uζ −u1−+λu uη −1+ u
u−λ αλ −1+ u
u−λ ζλ −1+ u
u−λ ηλ −1+ u
u−λ
H1,F (ζ, η; u; λ) =− − − + + ,
−u + λ −u + λ −u + λ −u + λ −u + λ
 α  α  α  α
(α)
u2 ζ 2 −u1−+λu u2 ζη −u1−+λu u2 η 2 −u1−+λu uα −u1−+λu λ
H2,F (ζ, η; u; λ) = + − +
(−u + λ)2 (−u + λ)2 (−u + λ)2 (−u + λ)2
 α  α  α  α
uαζ −u1−+λu λ 2uζ 2 −u1−+λu λ uαη −u1−+λu λ 2uζη −u1−+λu λ
+ − + −
(−u + λ)2 (−u + λ)2 (−u + λ)2 (−u + λ)2
 α  α  α  α
2
2uη u−λ − 1+ u
λ α u−λ − 1+ u
λ 2 α u−λ
2 − 1+ u
λ 2 αζ u1−+λu λ2
−
+ − + −
(−u + λ)2 2(−u + λ)2 2(−u + λ)2 (−u + λ)2
 α  α  α  α
ζ 2 −u1−+λu λ2 αη −u1−+λu λ2 ζη −u1−+λu λ2 η 2 −u1−+λu λ2
+ − + − ,
(−u + λ)2 (−u + λ)2 (−u + λ)2 (−u + λ)2
 α  α  α  
(α) −1 + u −1 + u −1 + u −1 + u α
H3,F (ζ, η; u; λ) = ζ 3 + 2ζ 2 η − 2ζη 2 − η3
u−λ u−λ u−λ u−λ
 α  α  α  α
u α u−λ
2 − 1+ u
λ 2uα u−λ 2 − 1+ u
λ 2 −
α u−λ 1+ u
λ 3 α u−λ
2 − 1+ u
λ3
− − + +
(−u + λ)3 (−u + λ)3 3(−u + λ)3 (−u + λ)3
 α  α  α  α
α3 −u1−+λu λ3 2uαζ −u1−+λu λ 2uαη −u1−+λu λ αζ −u1−+λu λ2
− + + −
3(−u + λ)3 (−u + λ)2 (−u + λ)2 (−u + λ)2
 α  α  α  α
α ζ u−λ
2 − 1+ u
λ 2 αη u−λ− 1+ u
λ 2 α η u−λ
2 − 1+ u
λ 2 2αζ 2 −u1−+λu λ
+ − + −
(−u + λ)2 (−u + λ)2 (−u + λ)2 −u + λ
 α  α
2αζη u−λ − 1+ u
λ 2αη u−λ 2 − 1+ u
λ
− + .
−u + λ −u + λ
We investigate the beautiful zeros of the generalized Apostol-type Frobenius–Euler
(α)
polynomials Hw,F (ζ, η; u; λ) = 0 of order α by using a computer. We plot the zeros of
(α)
generalized Apostol-type Frobenius–Euler polynomials Hw,F (ζ, η; u; λ) = 0 of order α for
w = 30 (Figure 1).

42
Mathematics 2024, 12, 800

1.0 1.0

0.5 0.5

Im(
) 0.0 Im(
) 0.0

-0.5 -0.5

-1.0 -1.0

-10 000 0 10 000 20 000 -20 000 -10 000 0 10 000 20 000 30 000

Re(
) Re(
)

1.0

0.5

0.5

Im(
) 0.0 Im(
) 0.0

-0.5

-0.5

-1.0

-50 000 0 50 000 100 000 -50 000 0 50 000 100 000 150 000

Re(
) Re(
)

(α)
Figure 1. Zeros of Hw,F (ζ, η; u; λ) = 0.

In Figure 1 (top left), we choose u = −2, λ = 4, α = 3 and η = 3. In Figure 1 (top

right), we choose u = −2, λ = 4, α = 3 and η = −3. In Figure 1 (bottom left), we choose
u = 2, λ = 6, α = 5 and η = 3. In Figure 1 (bottom right), we choose u = 2, λ = 6, α = 5
and η = −3.
Stacks of zeros of the generalized Apostol-type Frobenius–Euler polynomials
(α)
Hw,F (ζ, η; u; λ) = 0 of order α for 1 ≤ w ≤ 30, forming a 3D structure, are presented
(Figure 2).
In Figure 2 (top left), we choose u = −2, λ = 4, α = 3 and η = 3. In Figure 2 (top
right), we choose u = −2, λ = 4, α = 3 and η = −3. In Figure 2 (bottom left), we choose
u = 2, λ = 6, α = 5 and η = 3. In Figure 2 (bottom right),we choose u = 2, λ = 6, α = 5
and η = −3.
Plots of real zeros of the generalized Apostol-type Frobenius–Euler polynomials
(α)
Hw,F (ζ, η; u; λ) = 0 of order α for 1 ≤ w ≤ 30 are presented (Figure 3).
In Figure 3 (top left), we choose u = −2, λ = 4, α = 3 and η = 3. In Figure 3 (top
right), we choose u = −2, λ = 4, α = 3 and η = −3. In Figure 3 (bottom left), we choose
u = 2, λ = 6, α = 5 and η = 3. In Figure 3 (bottom right),we choose u = 2, λ = 6, α = 5
and η = −3.
Next, we calculated an approximate solution satisfying the generalized Apostol-type
(α)
Frobenius–Euler polynomials Hw,F (ζ, η; u; λ) = 0 of order α. The results are given in
Table 1. We choose u = 2, λ = 6, α = 5 and η = 3.

43
Mathematics 2024, 12, 800

(α)
Figure 2. Zeros of Hw,F (ζ, η; u; λ) = 0.

(α)
Figure 3. Real zeros of Hw,F (ζ, η; u; λ) = 0.

44
Mathematics 2024, 12, 800

(α)
Table 1. Approximate solutions of Hw,F (ζ, η; u; λ) = 0.

Degree w ζ
1 4.5000
2 −0.96131, 5.4613
3 −3.9141, 5.1740, 7.7401
−6.6036, 3.1453 − 2.1145i,
4
3.1453 + 2.1145i, 13.813
−10.775, 1.5396 − 0.9397i, 1.5396 + 0.9397i,
5
8.4352, 21.761
−17.428, −0.72148, 2.5863,
6
5.7586, 10.197, 35.608
−28.214, −1.9256, 2.6753 − 1.4884i, 2.6753 + 1.4884i,
7
6.7608, 19.152, 57.377
−45.645, −3.2315, 1.4614 − 1.2976i, 1.4614 + 1.2976i,
8
5.7479, 12.138, 29.581, 92.986
−73.860, −5.2463, 0.39703, 1.3178,
9 5.2440, 7.1909, 18.825, 48.769,
150.36
−119.51, −8.4883, −0.86402, 2.7850 − 0.2438i,
10 2.7850 + 0.2438i, 4.6030, 13.531, 30.944,
78.360, 243.35

4. Conclusions
In this article, our objective was to introduce the F-analogues of the Apostol-type
Frobenius–Euler polynomials, which we have denoted as generalized Apostol-type Frobenius–
Euler–Fibonacci polynomials. We have employed the Golden Calculus to introduce these
polynomials and subsequently explored their properties. Our work represents a generalization
of the previously published articles [24]. In our future research studies, we intend to utilize
the Golden Calculus to introduce the parametric types of certain special polynomials and to
derive a plethora of combinatorial identities through their generating functions.

Author Contributions: Conceptualization, M.S.A., W.A.K., C.K. and C.S.R.; formal analysis, M.S.A.,
W.A.K., C.K. and C.S.R.; funding acquisition, M.S.A. and W.A.K.; investigation, W.A.K. and C.K.;
methodology, M.S.A., W.A.K., C.K. and C.S.R.; project administration, M.S.A., W.A.K., C.K. and C.S.R.;
software, M.S.A., W.A.K., C.K. and C.S.R.; writing—original draft, W.A.K. and C.K.; writing—review
and editing, M.S.A., W.A.K., C.K. and C.S.R. All authors have read and agreed to the published
version of the manuscript.
Funding: This research received no external funding.
Data Availability Statement: No data were used to support this study.
Conflicts of Interest: The authors declare no conflicts of interest.

References
1. Alam, N.; Khan, W.A.; Ryoo, C.S. A note on Bell-based Apostol-type Frobenius–Euler polynomials of complex variable with its
certain applications. Mathematics 2022, 10, 2109. [CrossRef]
2. Khan, W.A.; Haroon, H. Some symmetric identities for the generalized Bernoulli, Euler and Genocchi polynomials associated
with Hermite polynomials. Springer Plus 2016, 5, 1920. [CrossRef] [PubMed]
3. Pathan, M.A.; Khan, W.A. A new class of generalized Apostol-type Frobenius–Euler–Hermite polynomials. Honam Math. J. 2020,
42, 477–499.
4. Carlitz, L. Eulerian numbers and polynomials. Mat. Mag. 1959, 32, 164–171. [CrossRef]
5. Kurt, B.; Simsek, Y. On the generalized Apostol-type Frobenius–Euler polynomials. Adv. Differ. Equ. 2013, 2013, 1. [CrossRef]
6. Kim, D.S.; Kim, T. Some new identities of Frobenius–Euler numbers and polynomials. J. Inequal. Appl. 2012, 2012, 307. [CrossRef]
7. Ryoo, C.S. A note on the Frobenius Euler polynomials. Proc. Jangjeon Math. Soc. 2011, 14, 495–501.

45
Mathematics 2024, 12, 800

8. Luo Q.M.; Srivastava, H.M. Some generalizations of the Apostol-Bernoulli and Apostol–Euler polynomials. J. Math. Anal. Appl.
2005, 308, 290–302. [CrossRef]
9. Milovanović, C. On generalized Stirling number and polynomials. Math. Balk. New Ser. 2004, 18, 241–248.
10. Jamei, M.J.; Milovanović, G., Dagli, M.C. A generalization of the array type polynomials. Math. Morav. 2022, 26, 37–46. [CrossRef]
11. Simsek Y. Generating functions for generalized Stirling type numbers, Array type polynomials, Eulerian type polynomials and
their applications. J. Fixed Point Theory Appl. 2013, 2013, 87. [CrossRef]
12. Luo, Q.M.; Srivastava, H.M. Some generalization of the Apostol-Genocchi polynomials and Stirling numbers of the second kind.
Appl. Math. Comput. 2011, 217, 5702–5728. [CrossRef]
13. Kwaśniewski, A.K. Towards ψ-extension of Rota’s finite operator calculus. Rep. Math. Phys. 2001, 47, 305–342. [CrossRef]
14. Krot, E. An introduction to finite fibonomial calculus. Cent. Eur. J. Math. 2004, 2, 754–766. [CrossRef]
15. Pashaev, O.K.; Nalci, S. Golden quantum oscillator and Binet–Fibonacci calculus. J. Phys. A Math. Theor. 2012, 45, 015303.
[CrossRef]
16. Kus, S.; Tuglu, N.; Kim, T. Bernoulli F-polynomials and Fibo-Bernoulli matrices. Adv. Differ. Equ. 2019, 2019, 145. [CrossRef]
17. Özvatan, M. Generalized Golden-Fibonacci Calculus and Applications. Master’s Thesis, Izmir Institute of Technology, Urla,
Türkiye, 2018.
18. Pashaev, O.K. Quantum calculus of Fibonacci divisors and infinite hierarchy of bosonic-fermionic golden quantum oscillators.
Int. J. Geom. Methods Mod. Phys. 2021, 18, 2150075. [CrossRef]
19. Pashaev, O.K.; Ozvatan, M. Bernoulli–Fibonacci Polynomials. arXiv 2020, arXiv:2010.15080.
20. Gulal, E.; Tuglu, N. Apostol-Bernoulli–Fibonacci polynomials, Apostol–Euler–Fibonacci polynomials and their generating
functions. Turk. J. Math. Comput. Sci. 2023, 15, 202–210. [CrossRef]
21. Kızılateş, C.; Öztürk, H. On parametric types of Apostol Bernoulli–Fibonacci Apostol Euler–Fibonacci and Apostol Genocchi–
Fibonacci polynomials via Golden Calculus. AIMS Math. 2023, 8, 8386–8402. [CrossRef]
22. Tuğlu, N.; Ercan, E. Some properties of Apostol Bernoulli Fibonacci and Apostol Euler Fibonacci Polynomials. In Proceedings of
the International Conference on Mathematics and Mathematics Education, Ankara, Turkey, 16–18 September 2021; pp. 32–34.
23. Tuglu, N.; Yesil, F.; Gokcen Kocer, E.; Dziemiańczuk, M. The F -Analogue of Riordan Representation of Pascal Matrices via
Fibonomial Coefficients. J. Appl. Math. 2014, 2014, 841826. [CrossRef]
24. Urielesa, A.; Ramırezb, W.; Ha, L.C.P.; Ortegac, M.J.; Arenas-Penalozac, J. On F-Frobenius–Euler polynomials and their matrix
approach. J. Math. Comput. Sci. 2024, 32, 377–386. [CrossRef]

Disclaimer/Publisher’s Note: The statements, opinions and data contained in all publications are solely those of the individual
author(s) and contributor(s) and not of MDPI and/or the editor(s). MDPI and/or the editor(s) disclaim responsibility for any injury to
people or property resulting from any ideas, methods, instructions or products referred to in the content.

46
 mathematics

Article
Mixed-Type Hypergeometric Bernoulli–Gegenbauer Polynomials
Dionisio Peralta 1, *, Yamilet Quintana 2,3 and Shahid Ahmad Wani 4

1 Instituto de Matemática, Facultad de Ciencias, Universidad Autónoma de Santo Domingo,

Santo Domingo 10105, Dominican Republic
2 Departamento de Matemáticas, Universidad Carlos III de Madrid, Avenida de la Universidad 30,
Leganés, 28911 Madrid, Spain; [email protected]
3 Instituto de Ciencias Matemáticas (ICMAT), Campus de Cantoblanco UAM, 28049 Madrid, Spain
4 Department of Applied Sciences, Symbiosis Institute of Technology, Symbiosis International
(Deemed University) (SIU), Lavale, Pune 412115, Maharashtra, India; [email protected]
* Correspondence: [email protected]

Abstract: In this paper, we consider a novel family of the mixed-type hypergeometric Bernoulli–
Gegenbauer polynomials. This family represents a fascinating fusion between two distinct categories
of special functions: hypergeometric Bernoulli polynomials and Gegenbauer polynomials. We focus
our attention on some algebraic and differential properties of this class of polynomials, including
its explicit expressions, derivative formulas, matrix representations, matrix-inversion formulas, and
other relations connecting it with the hypergeometric Bernoulli polynomials. Furthermore, we show
that unlike the hypergeometric Bernoulli polynomials and Gegenbauer polynomials, the mixed-type
hypergeometric Bernoulli–Gegenbauer polynomials do not fulfill either Hanh or Appell conditions.

Keywords: Gegenbauer polynomials; generalized Bernoulli polynomials; hypergeometric Bernoulli

polynomials

MSC: 33E20; 32A05; 11B83; 33C45

Citation: Peralta, D.; Quintana, Y.; 1. Introduction

Wani, S.A. Mixed-Type For a fixed integer m ∈ N, the mixed-type hypergeometric Bernoulli–Gegenbauer
[m−1,α]
Hypergeometric Bernoulli–
polynomials Vn ( x ) of order α ∈ (−1/2, ∞), where n ≥ 0, are defined through
Gegenbauer Polynomials.
generating the functions and series expansions as follows:
Mathematics 2023, 11, 3920. https://
  −α
doi.org/10.3390/math11183920 ∞
zm e xz xz z2 [m−1,α] zn
1 − + = ∑ Vn (x) , (1)
Academic Editor: Francesco Aldo
e −∑
z m −1 z l π 4π 2
n =0
n!
l =0 l!
Costabile

Received: 16 August 2023

where |z| < 2π, | x | ≤ 1, and α ∈ (−1/2, ∞) \ {0}.
Revised: 12 September 2023
Accepted: 12 September 2023
 
∞
zm e xz 2π − xz [m−1,0] zn
Published: 15 September 2023
−1 z l z2
= ∑ Vn (x) , |z| < 2π, | x | ≤ 1. (2)
ez − ∑m
l =0 l!
1− xz
π + 4π 2 n =0
n!
 
[m−1,α]
The polynomials Vn (x) represent a fascinating fusion between two classes
Copyright: © 2023 by the authors. n ≥0
Licensee MDPI, Basel, Switzerland.
of special functions: hypergeometric Bernoulli polynomials and Gegenbauer polynomials.
This article is an open access article
A significant amount of research has been conducted on various generalizations and
distributed under the terms and analogs of the Bernoulli polynomials and the Bernoulli numbers. For a comprehensive
conditions of the Creative Commons treatment of the diverse aspects, including summation formulas and applications, inter-
Attribution (CC BY) license (https:// ested readers can refer to recent works [1,2]. Inspired by recent articles [3–7] where au-
creativecommons.org/licenses/by/ thors explore analytic and numerical aspects of hypergeometric Bernoulli polynomials,
4.0/).

Mathematics 2023, 11, 3920. https://doi.org/10.3390/math11183920 47 https://www.mdpi.com/journal/mathematics

Mathematics 2023, 11, 3920

hypergeometric Euler polynomials, generalized mixed-type Bernoulli–Gegenbauer poly-

nomials, and Lagrange-based hypergeometric Bernoulli polynomials,
 this article
 focuses
[m−1,α]
on the algebraic and differential properties of the polynomials Vn (x) . These
n ≥0
properties include their explicit expressions, derivative formulas, matrix representations,
matrix-inversion formulas, and other relationships connecting them with hypergeometric
Bernoulli polynomials.
The paper is organized as follows. Section 2 provides relevant information about
hypergeometric Bernoulli polynomials and Gegenbauer polynomials. Section 3 is dedicated
to the study of the main algebraic and analytic properties of the HBG polynomials (1)
and (2), which are summarized in Theorems 1–4, and Proposition 6.

2. Background and Previous Results

Throughout this paper, let N, N0 , Z, R, and C denote, respectively, the sets of natural
numbers, non-negative integers, integers, real numbers, and complex numbers. As usual,
we always use the principal branch for complex powers, in particular, 1α = 1 for α ∈ C.
Furthermore, the convention 00 = 1 is adopted.
For λ ∈ C and k ∈ Z, we use the notations λ(k) and (λ)k for the rising and falling
factorials, respectively, i.e.,
⎧
⎨ 1, if k = 0,
λ(k) = ∏ik=1 (λ + i − 1), if k ≥ 1,
⎩
0, if k < 0,

and ⎧
⎨ 1, if k = 0,
(λ)k = ∏ik=1 (λ − i + 1), if k ≥ 1,
⎩
0, if k < 0.
From now on, we denote by Pn the linear space of polynomials with real coefficients
and a degree less than or equal to n. Moreover, to present some of our results, we require
the use of the generalized multinomial theorem (cf. [8,9] and the references therein).

2.1. Hypergeometric Bernoulli Polynomials

For a fixed m ∈ N, the hypergeometric Bernoulli polynomials are defined by means of
the following generating function [5,10–14]:
∞
zm e xz [ m −1] zn
= ∑ Bn (x) , |z| < 2π, (3)
− ∑lm=−01 zl!
l
ez n =0 n!

[ m −1] [ m −1]
and the hypergeometric Bernoulli numbers are defined by Bn := Bn (0) for all
n ≥ 0. The hypergeometric Bernoulli polynomials also are called generalized Bernoulli
polynomials of level m [5,6]. It is clear that if m = 1 in (3), then we obtain the definition of
the classical Bernoulli polynomials Bn ( x ) and classical Bernoulli numbers, respectively, i.e.,
[0] [0]
Bn ( x ) = Bn ( x ) and Bn = Bn , respectively, for all n ≥ 0.
The first four hypergeometric Bernoulli polynomials are as follows:

[ m −1]
B0 ( x ) = m!, 
[ m −1]
B1 ( x ) = m! x − m+
1
,
 1 
[ m −1]
B2 ( x ) = m! x − m+1 x + (m+1)22 (m+2) ,
2 2
 
[ m −1] 6( m −1)
B3 ( x ) = m! x3 − m+3
1 x + ( m +1)2 ( m +2) x +
2 6
(m+1)3 (m+2)(m+3)
.

The following results summarize some properties of the hypergeometric Bernoulli

polynomials (cf. [5,6,11,12,15]).

48
Mathematics 2023, 11, 3920

 
[ m −1]
Proposition 1 ([5], Proposition 1). For a fixed m ∈ N, let Bn (x) be the sequence of
n ≥0
hypergeometric Bernoulli polynomials. Then the following statements hold:
(a) Summation formula. For every n ≥ 0,
n  
[ m −1] n [ m −1] n − k
Bn (x) = ∑ k k
B x . (4)
k =0

(b) Differential relations (Appell polynomial sequences). For n, j ≥ 0 with 0 ≤ j ≤ n, we have

[ m −1] n! [ m −1]
[ Bn ( x )]( j) = B ( x ). (5)
(n − j )! n− j

(c) Inversion formula. ([12], Equation (2.6)) For every n ≥ 0,

n  
n k! [ m −1]
xn = ∑ k (m + k)! Bn−k (x). (6)
k =0

(d) Recurrence relation. ([12], Lemma 3.2) For every n ≥ 1,

  n −2  
[ m −1] 1 [ m −1] 1 n [ m −1] [ m −1]
Bn (x) = x−
m+1
Bn−1 ( x ) −
n ( m − 1) ! ∑ B
k n−k k
B ( x ).
k =0

(e) Integral formulas.

 x
1 [ m −1] 1 [ m −1] [ m −1]
Bn ( x )dx = B ( x1 ) − Bn+1 ( x0 )
x0 n + 1 n +1
n  
1 n [ m −1]
= ∑ B (( x1 )n−k+1 − ( x0 )n−k+1 ).
k =0
n − k + 1 k k

 x
[ m −1] [ m −1] [ m −1]
Bn (x) = n Bn−1 (t)dt + Bn .
0

[ m −1]
(f) ([12], Theorem 3.1) Differential equation. For every n ≥ 1, the polynomial Bn ( x ) satisfies
the following differential equation
[ m −1] [ m −1] [ m −1]  
Bn Bn−1 B 1
y(n) + y ( n −1) + · · · + 2 y + (m − 1)! − x y + n(m − 1)!y = 0.
n! ( n − 1) ! 2! m+1

As a straightforward consequence of the inversion Formula (6), the following expected

algebraic property is obtained.

Proposition
 2 ([5], Proposition 2). For  a fixed m ∈ N and each n ≥ 0, the set
[ m −1] [ m −1] [ m −1]
B0 ( x ), B1 ( x ), . . . , Bn ( x ) is a basis for Pn , i.e.,
 
[ m −1] [ m −1] [ m −1]
Pn = span B0 ( x ), B1 ( x ), . . . , Bn (x) .

Let ζ (s) be the Riemann zeta function defined by

∞
1
ζ (s) = ∑ ns
, (s) > 1.
n =1

The following result provides a formula for evaluating ζ (2r ) in terms of the hypergeo-
metric Bernoulli numbers.

49
Mathematics 2023, 11, 3920

Proposition 3 ([6], Theorem 3.3). For a fixed m ∈ N and any r ∈ N, the following identity holds.

[ m −1]
(−1)r−1 22r−1 π 2r B2r [ m −1]
ζ (2r ) = + Δr ,
m!(2r )!

where
⎡   ⎤
[ m −1] [ m −1] [ m −1] [ m −1]
2r −2j+1 (1) − B2r −2j+1
[ m −1] [ m −1] r −1 B
[ m −1] (−1)r−1 22r−1 π 2r ⎣ B2r (1) − B2r B2r+1 (1) − B2r+1 B2j
⎦.
Δr = − −∑
m! 2(2r )! (2r + 1)! j =1
(2r − 2j + 1)! (2j)!

2.2. Gegenbauer Polynomials

(α)
For α > − 12 , we denote by {Ĉn ( x )}n≥0 the sequence of Gegenbauer polynomials,
orthogonal on [−1, 1] with respect to the measure dμ( x ) = (1 − x2 )α− 2 dx (cf. [16], Chap-
1

ter IV), normalized by

(α) Γ(n + 2α)
Ĉn (1) = .
n!Γ(2α)
More precisely,
 1  1
(α) (α) (α) (α)
Ĉn ( x )Ĉm ( x )(1 − x2 )α− 2 dx = Mnα δn,m ,
1
Ĉn ( x )Ĉm ( x ) dμ( x ) = n, m ≥ 0,
−1 −1

where the constant Mnα is positive. It is clear that the normalization above does not allow α
to be zero or a negative integer. Nevertheless, the following limits exist for every x ∈ [−1, 1]
(see [16], (4.7.8))
(α)
(α) Ĉ ( x ) 2
lim Ĉ0 ( x ) = T0 ( x ), lim n = Tn ( x ),
α →0 α →0 α n
where Tn ( x ) is the nth Chebyshev polynomial of the first kind. In order to avoid confusing
(0)
notation, we define the sequence {Ĉn ( x )}n≥0 as follows:

(0) (0) 2 (0) 2

Ĉ0 (1) = 1, Ĉn (1) = , Ĉn ( x ) = Tn ( x ), n ≥ 1.
n n
We denote the nth monic Gegenbauer orthogonal polynomial by

(α) (α)
Cn ( x ) = (kαn )−1 Ĉn ( x ),

where the constant kαn (cf. [16], Formula (4.7.31)) is given by

2n Γ ( n + α )
kαn = , α = 0,
n!Γ(α)
k α 2 n
k0n = lim n = , n ≥ 1.
α →0 α n
Then for n ≥ 1, we have

(0) (α) 1
Cn ( x ) = lim (kαn )−1 Ĉn ( x ) = Tn ( x ). (7)
α →0 2n −1
Gegenbauer polynomials are closely connected with axially symmetric potentials
in n dimensions (cf. [4] and the references cited therein), and contain the Legendre and
Chebyshev polynomials as special cases. Furthermore, they inherit practically all the
formulas known in the classical theory of Legendre polynomials.

50
Mathematics 2023, 11, 3920

(α)
Proposition 4 ([17], cf. Proposition 2.1). Let {Cn }n≥0 be the sequence of monic Gegenbauer
orthogonal polynomials. Then the following statements hold.
(a) Three-term recurrence relation.

(α) (α) (α) (α) 1

xCn ( x ) = Cn+1 ( x ) + γn Cn−1 ( x ), α > − , α = 0, (8)
2
(α) (α) (α)
with initial conditions C−1 ( x ) = 0, C0 ( x ) = 1 and recurrence coefficients γ0 ∈ R,
(α) n(n+2α−1)
γn = 4(n+α)(n+α−1)
, n ∈ N.
(b) For every n ∈ N (see [16], (4.7.15))
 1
(α) (α) n!Γ(n + 2α)
hαn := Cn 2μ = [Cn ( x )]2 dμ( x ) = π21−2α−2n . (9)
−1 Γ ( n + α + 1) Γ ( n + α )

(c) Rodrigues formula.

(α) (−1)n Γ(n + 2α) dn

(1 − x2 )α− 2 Cn ( x ) = (1 − x 2 ) n + α − 2 ,
1 1
x ∈ (−1, 1).
Γ(2n + 2α) dx n

(d) Structure relation (see [16], (4.7.29)). For every n ≥ 2

( α −1) (α) (α) (α)

Cn ( x ) = Cn ( x ) + ξ n−2 Cn−2 ( x ),

where
(α) (n + 2)(n + 1)
ξn = , n ≥ 0.
4(n + α + 1)(n + α)
(e) For every n ∈ N (see [16], Formula (4.7.14))

d (α) ( α +1)
Cn ( x ) = nCn−1 ( x ).
dx

(f) For every n ∈ N (see [18], Proposition 2.1)

d (0) n (1)
Cn ( x ) = Cn−1 ( x ).
dx 2

As is well known, the monic Gegenbauer orthogonal polynomials admit other dif-
ferent definitions [16,19–21]. In order to deal with the definitions (1) and (2) of the HBG
polynomials, we also are interested in the definition of the monic Gegenbauer orthogonal
polynomials by means of the following generating functions:
 −α ∞
xz z2 Γ ( n + α ) (α) zn
1−
π
+ 2
4π
= ∑ π
C (x) ,
n Γ(α) n n!
|z| < 2π, | x | ≤ 1, α ∈ (−1/2, ∞) \ {0}, (10)
n =0

and
∞ ∞
2π − xz 1 (0) Γ ( n + 1 ) (0) zn
= ∑ C ( x )zn =
n −1 n ∑ n −1
Cn ( x ) , |z| < 2π, | x | ≤ 1. (11)
n =0 π π
z2
1− xz
π + 4π 2 n =0 n!

51
Mathematics 2023, 11, 3920

Remark 1. Note that (10) and (11) are suitable modifications of the generating functions for the
(α)
Gegenbauer polynomials Ĉn ( x ):
 −α ∞
(α)
1 − 2xz + z2 = ∑ Ĉn ( x )zn , |z| < 1, | x | ≤ 1, α ∈ (−1/2, ∞) \ {0},
n =0
∞
1 − xz n (0)
= 1 + ∑ Ĉn ( x )zn , |z| < 1, | x | ≤ 1.
1 − xz + z 2
n =1
2

[m−1,α]
3. The Polynomials Vn ( x) and Their Properties
Now, we can proceed to investigate some relevant properties of the HBG polynomials.
 
[m−1,α]
Proposition 5. For α ∈ (−1/2, ∞), let Vn (x) be the sequence of HBG polynomials
n ≥0
of order α. Then the following explicit formulas hold.
n  
[m−1,α] n Γ ( k + α ) (α) [ m −1]
Vn (x) = ∑ C ( x ) Bn−k ( x ), n ≥ 0, α = 0, (12)
k =0
k π k Γ(α) k
n  
[m−1,0] n k! (0) [ m −1]
Vn (x) = ∑ C ( x ) Bn−k ( x ), n ≥ 0. (13)
k =0
k π k −1 k

Proof. On account of the generating functions (1) and (10), it suffices to make a suitable
use of Cauchy product of series in order to deduce the expression (12).
Similarly, taking into account the generating functions (2) and (11), we can use an
analogous reasoning to the previous one to obtain expression (13).

Thus, the suitable use of (8) and (12) allow us to check that for α ∈ (−1/2, ∞) \ {0},
the first five HBG polynomials are:
[m−1,α]
V0 ( x ) =m! v0 (α),

# $
[m−1,α] 1
V1 ( x ) =m! v1 (α) x − ,
m+1

# $
[m−1,α] 2( π + α ) 4π 2 (α + 1) + α(m + 1)2 (m + 2)
V2 ( x ) =m! v2 (α) x2 − x+ ,
π ( m + 1) 2π 2 (m + 1)2 (m + 2)(1 + α)

#    
[m−1,α] 3 2 α α (1 + α )
V3 ( x ) =m! v3 (α) x3 − v2 ( α ) x 2 + 3 1 + − 1 + x
m+1 ( m + 1)2 ( m + 2) π 2π 2 π
 $
2( m − 1) α
+3 − ,
(m + 1)3 (m + 2)(m + 3) 2π 2 (m + 1)

# 
[m−1,α] 4 m−2 8α α 2(1 + α ) α
V4 ( x ) =m! v4 (α) x4 − v3 ( α ) x 3 + 3 + − −
m+1 (m + 1)(m + 2) π (m + 1)2 (m + 2) π 2 π3
 
(2 + α)(1 + α)α 2 5−m 4( m − 1) α α
− x +6 + +
π4 (m + 1)2 (m + 2)(m + 3) π (m + 1)3 (m + 2)(m + 3) π 2 (m + 1)
 $
(1 + α ) α 6(m3 − 3m2 − 6m + 36) 6(1 + 2α)α 3(1 + α ) α
+ 3 x2 − + + ,
π ( m + 1) (m + 1)2 (m + 2)2 (m + 3)(m + 4) π 2 (m + 1)2 (m + 2) 4π 4
n   (k)
n α
where vn (α) = ∑ k πk
, 0 ≤ n ≤ 4.
k =0

52
Mathematics 2023, 11, 3920

In contrast to the hypergeometric Bernoulli polynomials and Gegenbauer polynomials,

the HBG polynomials neither satisfy a Hanh condition nor an Appell condition. More
precisely, we have the following result.
 
[m−1,α]
Theorem 1. For α ∈ (−1/2, ∞), let Vn (x) be the sequence of HBG polynomials of
n ≥0
order α. Then we have
d [m−1,α] α [m−1,α+1] [m−1,α]
V ( x ) = ( n + 1 ) Vn ( x ) + Vn (x) , α = 0, (14)
dx n+1 π

   
d [m−1,0] [m−1,0] 1 n n ( k + 1 ) ! (1) [ m −1]
V ( x ) = ( n + 1 ) Vn (x) + ∑ Ck ( x ) Bn−k ( x ) , α = 0.
dx n+1 2 k =0 k πk
(15)

Proof. From (12), we have

n +1  
[m−1,α] n + 1 Γ ( k + α ) (α) [ m −1]
Vn+1 (x) = ∑ k π k Γ(α) k
C ( x ) Bn+1−k ( x ),
k =0

differentiating this last equation, and using part (e) of Proposition 4, (14) follows.

Furthermore, it is possible to establish an integral formula connecting the HBG po-

lynomials with the monic Gegenbauer polynomials. This integral formula allows us to
deduce a concise expression for the Fourier coefficients of the HBG polynomials in terms of
the basis of monic Gegenbauer polynomials.
 
[m−1,α]
Lemma 1. For α ∈ (−1/2, ∞), let Vn (x) be the sequence of HBG polynomials of
n ≥0
order α. Then, the following formula holds.
⎧ m!n!Γ(n+2α) Γ(k+α)
 1 ⎪
⎨
n
∑nk=0 ( k ) π k−1 Γ(α) , α = 0,
π 2α+2n Γ(n+α+1)Γ(n+α)
[m−1,α] (α)
Vn ( x )Cn ( x )dμ( x ) = (16)
−1 ⎪
⎩ n
m!π
2n ∑nk=0 ( k ) π k!
k −1 , α = 0,

whenever n ≥ 0.

Proof. In order to obtain (16), it suffices to use the orthogonality properties of the monic
Gegenbauer polynomials (4), (7), (9), (12) and (13).

Regarding the zero distribution of these polynomials, the numerical evidence indicates
that this distribution does not align with the behavior of either Bernoulli hypergeometric
polynomials or Gegenbauer polynomials. For instance, in Figure 1, the plots for the zeros
[m−1,α] [m−1,α]
of V28 ( x ) and V30 ( x ) are shown for m = 2 and α = − 14 .

53
Mathematics 2023, 11, 3920

[1,− 1 ] [1,− 1 ]
(a) Zeros of V28 4 ( x ). (b) Zeros of V30 4 ( x ).
[1,− 1 ] [1,− 1 ]
Figure 1. Zeros of V28 4 ( x ) and V30 4 ( x ).

As expected, the symmetry property of Gegenbauer polynomials is not inherited by

[m−1,α]
the HBG polynomials. For instance, Figure 2 displays the induced mesh of V j (x)
[m−1,α]
for m = 2, α = 1, and j = 1, . . . , 21. Each point on this mesh takes the form ( x j , j ),
(α)
j = 1, . . . , 21. In contrast, Figure 3 displays the induced mesh of Cj ( x ) for α = 1, and
j = 1, . . . , 19.

[m−1,α]
Figure 2. Induced mesh of V j ( x ) for m = 2, α = 1, and j = 1, . . . , 21.

(α)
Figure 3. Induced mesh of Cj ( x ) for α = 1, and j = 1, . . . , 19.

54
Mathematics 2023, 11, 3920

For any α ∈ (−1/2, ∞), it is possible to deduce interesting relations connecting the
[m−1,α] [ m −1]
HBG polynomials Vn ( x ) and the hypergeometric Bernoulli polynomials Bn ( x ).
The following two results concern these relations.

[m−1,α]
Proposition 6. For a fixed m ∈ N, let Vn ( x ) be the nth HBG polynomial of order α ∈
(−1/2, ∞) \ {0}. Then, the following relation is satisfied:
∞ ∞  
[ m −1] zn (−1) j α j [m−1,α] zn+2k+ j
∑ Bn (x)
n!
= ∑ ∑ 22k π 2k+ j j, k
x V n ( x )
n!
. (17)
n =0 n =0 0≤ j,k ≤|α|

Proof. On the account of generalized multinomial theorem, we deduce that

 α  
xz z2 (−1) j α
1−
π
+ 2
4π
= ∑ 22k π2k+ j j, k x j z2k+ j . (18)
0≤ j,k ≤|α|

Next, (1), (3) and (18) imply that

∞  α ∞
[ m −1] zn xz z2 [m−1,α] z
n
∑ Bn (x)
n!
= 1−
π
+ 2
4π ∑ Vn n!
n =0 n =0
⎛ ⎞
  ∞
(−1) j α [m−1,α] z
n
= ⎝ ∑ 2k π 2k+ j j, k
x j z2k+ j ⎠ ∑ Vn . (19)
0≤ j,k≤|α|
2 n =0 n!

Since the sum on the right-hand side of (18) is finite, (17) follows directly from (19).

[m−1,0]
Theorem 2. For a fixed m ∈ N, the HBG polynomials Vn ( x ) are related with the hypergeo-
[ m −1]
metric Bernoulli polynomials Bn ( x ) by means of the following identities.

[ m −1] [m−1,0]
2πB0 ( x ) = V0 ( x ),
[ m −1] [ m −1] [m−1,0] [m−1,0]
2πB1 ( x ) − xB0 ( x ) = V1 ( x ) − πx V0 ( x ), (20)
[ m −1] [ m −1] [m−1,0] nx [m−1,0] n−1) [m−1,0]
2πBn ( x ) − nxBn−1 ( x ) = Vn ( x ) − π Vn−1 ( x ) + n(4π 2 Vn −2 ( x ), n ≥ 2.

Proof. From the identities (2) and (3), we have

∞   ∞
[ m −1] zn xz z2 [m−1,0] zn
(2π − xz) ∑ Bn (x)
n!
= 1−
π
+ 2
4π ∑ Vn (x)
n!
.
n =0 n =0

Multiplying, respectively,
 the left-hand
 side of the above expression by (2π − xz) and
z2
the right-hand side by 1 − xzπ + 4π 2
, we obtain the following equivalent expression:

∞   zn
[ m −1] [ m −1] [ m −1] [ m −1] [ m −1]
2πB0 ( x ) + 2πB1 ( x )z − xB0 ( x )z + ∑ 2πBn ( x ) − nxBn−1 ( x )
n!
n =2
[m−1,0] [m−1,0] x [m−1,0]
= V0 ( x ) + V1 ( x )z − V ( x )z
π 0
∞   n
[m−1,0] nx [m−1,0] n(n − 1) [m−1,0] z
+ ∑ Vn (x) − Vn −1 (x) + V n −2 ( x ) . (21)
n =2
π 4π 2 n!

Therefore, by comparing the coefficients on both sides of (21), we obtain the identi-
ties (20).

55
Mathematics 2023, 11, 3920

Remark 2. When α = r ∈ N, Equation (18) becomes

 r  
xz z2 (−1) j r
1− + 2 = ∑ 2k π 2k + j j, k
x j z2k+ j .
π 4π j + k =r 2

Thus, for r = 1 we can combine the above identity with (17), and obtain the following connecting
relations:
[ m −1] [m−1,1]
B0 ( x ) = V0 (x)
[ m −1] [m−1,1] x [m−1,1]
B1 (x) = V1 (x) −
V (x) (22)
π 0
[ m −1] [m−1,1] nx [m−1,1] n(n − 1) [m−1,1]
Bn ( x ) = Vn (x) − V (x) + Vn −2 ( x ), n ≥ 2,
π n −1 4π 2
[m−1,1]
Hence, as a straightforward consequence of (17) and (20), the HBG polynomials Vn (x)
[m−1,0]
and Vn (x) are related by means of the following identities:
[m−1,1] [m−1,0]
2πV0 ( x ) = V0 (x)
[m−1,1] [m−1,1] x [m−1,0] [m−1,0]
2πV1 ( x ) − 3xV0 (x) V= V1 (x) (x) −
π 0
 
[m−1,1] [m−1,1] n ( n − 1) n ( n − 1) x 2 [m−1,1] n(n − 1)(n − 2) x [m−1,1] (23)
2πVn ( x ) − 3nxVn−1 ( x ) + + Vn −2 (x) − Vn −3 (x)
2π π 4π 2
[m−1,0] nx [m−1,0] n(n − 1) [m−1,0]
= Vn (x) − V (x) + Vn −2 ( x ), n ≥ 2.
π n −1 4π 2

Using (12), (13), and employing a matrix approach, we can obtain a matrix representa-
[m−1,α]
tion for Vn ( x ), n ≥ 0. In order to implement that, we follow some ideas from [4,5].
First of all, we must point out that for r = 0, 1, . . . , n, Equations (12) and (13) allow us
[m−1,α]
to deduce the following matrix form of Vr ( x ):
[m−1,α] (α)
Vr ( x ) = C r ( x ) B [ m −1] ( x ), r = 0, 1, . . . , n, (24)

where
⎧ (α) (α) (α)
⎪
⎪ (r ) Γ (r + α ) C ( x )
⎨ r π r Γ(α) r
(r−r 1) Γπ(rr−−11Γ+(αα)) Cr−1 ( x ) · · · C0 ( x ) 0 · · · 0 , if α = 0,
(α)
Cr ( x ) =
⎪
⎪
⎩ (r) r! C (0) ( x ) (0) (0)
(r−r q) (πr−r−12)! Cr−1 ( x ) · · · C0 ( x ) 0 · · · 0 , if α = 0,
r π r −1 r

(α)
the null entries of the matrix Cr ( x ) appear (n − r )-times, and the matrix B[m−1] ( x ) is
T
given by B[m−1] ( x ) = B0[m−1] ( x ) [ m −1]
B1 (x) · · ·
[ m −1]
Bn (x) .

Now, for α ∈ (−1/2, ∞), let C(α) ( x ) be the (n + 1) × (n + 1) whose rows are precisely
(α)
the matrices Cr ( x ) for r = 0, 1, . . . , n. That is,

56
Mathematics 2023, 11, 3920

⎡ ⎤
(α)
C0 ( x ) 0 ··· 0
⎢ ⎥
⎢ ⎥
⎢ 1 Γ (1+ α ) ( α ) (α) ⎥
⎢ ( ) ⎥
⎢ 1 πΓ(α) C1 ( x ) C0 ( x ) ··· 0 ⎥
⎢ ⎥
⎢ ⎥
⎢ 2 Γ (2+ α ) ( α ) (α) ⎥
⎢ (1+ α )
(21) ΓπΓ ⎥ 1
C(α) ( x ) = ⎢ (2) π2 Γ(α) C2 ( x ) (α) 1
C (x) ··· 0 ⎥, α > − , α = 0,
⎢ ⎥ 2
⎢ ⎥
⎢ ⎥
⎢ .. .. .. .. ⎥
⎢ . . . . ⎥
⎢ ⎥
⎢ ⎥
⎣ ⎦
(α) (α) (α)
(nn) Γπ(nnΓ+(αα)) Cn ( x ) n Γ ( n −1+ α )
(n− 1) π n−1 Γ(α) Cn−1 ( x ) · · · C0 ( x )

and from (7):

⎡ ⎤
1 0 ··· 0
⎢ ⎥
⎢ ⎥
⎢ (11)πT1 ( x ) ··· 0 ⎥
⎢ 1 ⎥
⎢ ⎥
⎢ ⎥
⎢ ⎥
(0) ⎢ 2 1
(2) π T2 ( x ) 2
(1) T1 ( x ) · · · 0 ⎥.
C (x) = ⎢ ⎥
⎢ ⎥
⎢ ⎥
⎢ .. .. .. . ⎥
⎢ . . . . ⎥
.
⎢ ⎥
⎢ ⎥
⎣ ⎦
n n ( n −1) !
(n) (2π )n−1 Tn ( x ) (n−1) (2π )n−2 Tn−1 ( x ) · · · 1
n!

(α)
 the matrix C ( x ) is a lower triangular matrix for each x ∈ R, so
It is clear that
that det C ( x ) = 1. Therefore, C(α) ( x ) is a nonsingular matrix for each x ∈ R and
( α )

α ∈ (−1/2, ∞).
 
[m−1,α]
Theorem 3. For a fixed m ∈ N and any α ∈ (−1/2, ∞), let Vn (x) be the sequence
n ≥0
of HBG polynomials. Then, the following matrix representation holds.

V[m−1,α] ( x ) = C ( α ) ( x ) B [ m −1] ( x ), (25)

T
where V[m−1,α] ( x ) = V0[m−1,α] ( x ) [m−1,α]
V1 ( x ) · · · Vn
[m−1,α]
(x) .

[m−1,α]
Proof. For each r = 0, 1, . . . , n, consider the matrix form (24) of Vr ( x ). Then, it is not
difficult to see that the matrix V[m−1,α] ( x ) becomes
T
V[m−1,α] ( x ) = V0[m−1,α] ( x ) V1
[m−1,α]
( x ) · · · Vn
[m−1,α]
(x) = C ( α ) ( x ) B [ m −1] ( x ),

and (25) follows.

The following examples show how Theorem 3 can be used.

57
Mathematics 2023, 11, 3920

Example 1. Let us consider m = 1, n = 3, and α = 1, then,

⎡ ⎤ −1
1 0 0 0
⎢ ⎥
⎢ ⎥
⎢ x
1 0 0 ⎥
  −1 ⎢ π ⎥
⎢ ⎥
(1)
B( x ) = C ( x ) V [0,1]
(x) = ⎢
⎢
⎥
⎥ V[0,1] ( x ), (26)
⎢ 4x2 −1
⎢ 2π 2
2x
π 1 0 ⎥
⎥
⎢ ⎥
⎣ ⎦
6x3 −3x 3(4x2 −1) 3x
π3 2π 2 π 1

where
⎡ ⎤
1
⎢ ⎥
⎢   ⎥
⎢ ⎥
⎢ 1+ 1
π x− 1
⎥
⎢ 2 ⎥
[0,1] ⎢ ⎥
V (x) = ⎢     ⎥.
⎢ ⎥
⎢ 1+ 2
π + 1
x2 − 1 + 1
π x+ 1
− 1
⎥
⎢ π2 6 2π 2 ⎥
⎢ ⎥
⎣       ⎦
1+ 3
π + 6
π2
+ 6
π3
x3 − 3
2 1+ 2
π + 2
π2
x2 + 1
2 1+ 1
π − 3
π2
− 6
π3
x+ 3
4π 2

Since
⎡ ⎤ −1 ⎡ ⎤
1 0 0 0 1 0 0 0
⎢ ⎥ ⎢ ⎥
⎢ ⎥ ⎢ ⎥
⎢ 0 ⎥ ⎢ −x 0 ⎥
x
⎢ 1 0 ⎥ 1 0
  −1 ⎢ π
⎥ ⎢ π ⎥
⎢ ⎥
C (1) ( x ) =⎢
⎢
⎥
⎥ =⎢ ⎥,
⎢ 4x2 −1 2x
0 ⎥ ⎢ 1 ⎥
⎢ 2π 2 π 1 ⎥ ⎢ 2π2 − 2x
π 1 0 ⎥
⎢ ⎥ ⎢ ⎥
⎣ ⎦ ⎣ ⎦
6x3 −3x 3(4x2 −1) 3x 0 3
− 3x 1
π3 2π 2 π 1 2π 2 π

then (26) becomes ⎡ ⎤

1
⎢ ⎥
⎢ ⎥
⎢ x− 1 ⎥
⎢ 2 ⎥
⎢ ⎥
B( x ) = ⎢ ⎥.
⎢ ⎥
⎢ x2 − x + 1 ⎥
⎢ 6 ⎥
⎣ ⎦
x3 − 32 x2 + 12 x
That is, the entries of the matrix B( x ) are the first four classical Bernoulli polynomials.
[0,1]
It is worth noting that for α = m = 1, the HBG polynomials Vn ( x ) coincide with the GBG
(1)
polynomials Vn ( x ), for all n ≥ 0 (cf. [4]).

Example 2. Let m = n = 3 and α = − 14 . From (25), we obtain

58
Mathematics 2023, 11, 3920

⎡ ⎤⎡ ⎤
1 0 0 0 6
⎢ ⎥⎢ ⎥
⎢ ⎥⎢ ⎥
⎢ − 4π
x
1 0 0⎥ ⎢ 6x − 32 ⎥
⎢ ⎥⎢ ⎥
( −4)
1
[2] ⎢ ⎥⎢ ⎥
C ( x )B ( x ) = ⎢ ⎥⎢ ⎥
⎢ ⎥⎢ ⎥
⎢ − 16π
3
2 x − 3
2 2
− 2π
x
1 0⎥⎢ 6x2 − 3x + 20 3
⎥
⎢ ⎥⎢ ⎥
⎣ ⎦⎣ ⎦
2
− 64π
21
3 x − 7
3 6x
− 16π
9
2 x −
2 2
3 − 4π
3x
1 6x3 − 9x2 + 9x
20 +
3
80
⎡ ⎤
6
⎢ ⎥
⎢ ⎥
⎢ − 2π
3x
+ 6x − 3 ⎥
⎢ 2 ⎥
⎢ ⎥
=⎢
⎢
⎥.
⎥
⎢ −45x2 +6π 2 (40x2 −20x +1)+30π (1−4x ) x +30 ⎥
⎢ 40π 2 ⎥
⎢ ⎥
⎣ ⎦
3(−6π 2 x (40x2 −20x +1)+15x (6−7x2 )−15π (12x3 −3x2 −8x +2)+π 3 (320x3 −240x2 +24x +2))
160π 3
Straightforward calculations show that this last matrix coincides with
⎡ ⎤
6
⎢ ⎥
⎢   ⎥
⎢ ⎥
⎢ 6 1− 1
x− 3
⎥
⎢ 4π 2 ⎥
[ 2,− 4 ]
1 ⎢ ⎥
V (x) = ⎢     ⎥.
⎢ ⎥
⎢ 6 1− 1
− 3
x2 − 3 1 − 1
x+ 3
+ 3
⎥
⎢ 2π 16π 2 4π 20 4π 2 ⎥
⎢ ⎥
⎣       ⎦
6 1− 3
4π − 9
16π 2
− 21
64π 2
x3 − 9
2 1− 1
2π − 3
16π 2
x2 + 9
4
1
5 − 1
20π + 1
π2
+ 3
4π 3
x+ 3
80 − 9
16π 2

Hence, C(− 4 ) ( x )B[2] ( x ) = V[2,− 4 ] ( x ).

1 1

We can now proceed as outlined in [5]. From the summation Formula (4) it follows

[ m −1] [ m −1]
Br ( x ) = Mr T ( x ), r = 0, 1, . . . , n,

where
[ m −1]
Mr = (rr) Br[m−1] (r−r 1) Br−1
[ m −1]
· · · (0r ) B0
[ m −1]
0 ··· 0 , (27)
[ m −1]  T
the null entries of the matrix Mr appear (n − r )-times, and T( x ) = 1 x ··· xn .
Analogously, by (27) the matrix B[m−1] ( x ), can be expressed as follows:

B [ m −1] ( x ) = M [ m −1] T ( x )
⎡ [ m −1] ⎤
B0 0 ··· 0
⎢ ⎥
⎢ ⎥
⎢ 1 [ m −1] [ m −1] ⎥
⎢ (1) B1 (10) B0 ··· 0 ⎥
⎢ ⎥
⎢ ⎥
⎢ ⎥
⎢ 2 [ m −1] (21) B1
[ m −1]
··· ⎥
= ⎢ (2) B2 0 ⎥ T ( x ). (28)
⎢ ⎥
⎢ ⎥
⎢ ⎥
⎢ .. .. .. .. ⎥
⎢ . . . . ⎥
⎢ ⎥
⎣ ⎦
[ m −1] [ m −1] [ m −1]
(nn) Bn n
(n− 1) Bn−1 · · · (n0 ) B0

59
Mathematics 2023, 11, 3920

Notice that according to (27) the rows of the matrix M[m−1] are precisely the matrices
[ m −1]
Mr for r = 0, . . . , n. Furthermore, the matrix M[m−1] is a lower triangular matrix, so

that det M[m−1] = (m!)n+1 . Therefore, M[m−1] is a nonsingular matrix.
Another interesting algebraic property of the HBG polynomials is related with the
following matrix-inversion formula.
 
[m−1,α]
Theorem 4. For a fixed m ∈ N and any α ∈ (−1/2, ∞), let Vn (x) be the sequence
n ≥0
of HBG polynomials. Then, the following formula holds.
  −1
T( x ) = Q[m−1,α] ( x ) V[m−1,α] ( x ), (29)

where Q[m−1,α] ( x ) = C(α) ( x )M[m−1] .

Proof. Using the inversion Formulas (6), (25) and (28), and the nonsingularity of the
matrices C(α) ( x ) and M[m−1] , it is possible to deduce that
  −1   −1
T ( x ) = M [ m −1] C(α) ( x ) V[m−1,α] ( x ),

and (29) follows.

A simple and important consequence of Theorem 4 is:

 
[m−1,α] [m−1,α]
Corollary 1. For a fixed m ∈ N and any α ∈ (−1/2, ∞) the set V0 ( x ), . . . , Vn (x)
is a basis for Pn , n ≥ 0, i.e.,
 
[m−1,α] [m−1,α] [m−1,α]
Pn = span V0 ( x ), V1 ( x ), . . . , Vn (x) .

4. Conclusions
In the present paper, we introduced the mixed-type hypergeometric Bernoulli–Gegenbauer
polynomials and analyzed some algebraic and differential properties of these polynomials,
including their explicit expressions, derivative formulas, matrix representations, matrix-
inversion formulas, and other relations connecting them with the hypergeometric Bernoulli
polynomials. Furthermore, we demonstrated that unlike the hypergeometric Bernoulli
polynomials and Gegenbauer polynomials, the HBG polynomials do not fulfill either Hanh
or Appell conditions.
It is worth noting that the utilization of a matrix approach, specifically employing the
operational matrix method based on hypergeometric Bernoulli polynomials, underpins
several of our formulations. The matrix approaches using operational matrix methods
associated with special polynomials and their practical applications constitute a relatively
recent area of interest, as evidenced by the substantial body of literature (see, for instance,
refs. [22–28] and the references therein). However, within the context of mixed special
polynomials, to the best of our knowledge, there are no other published works that have
adopted a similar approach, with the possible exception of a recent investigation [4].
Furthermore, we provided some examples to illustrate that the class of HBG polyno-
mials does not generalize to the classical Bernoulli polynomials, although the latter can be
recovered using Theorem 3. Unfortunately, the numerical evidence suggests that the zero
distribution of the HBG polynomials does not align with the behavior of either Bernoulli
hypergeometric polynomials or Gegenbauer polynomials.
Finally, by employing the determinantal approach introduced by Costabile and
Longo [29], which implies that hypergeometric Bernoulli polynomials have a correspon-
ding determinant form, and considering Theorem 3, it becomes feasible to investigate
the determinantal forms associated with the HBG polynomials. Furthermore, Theorem 4

60
Mathematics 2023, 11, 3920

and the differential equation presented in part (f) of Proposition 1 (cf. [12], Theorem 3.1)
suggest that the HBG polynomials satisfy a differential equation of order n. These two
properties, along with their implications and potential applications, will be the focus of our
future work.

Author Contributions: Conceptualization, D.P. and Y.Q.; methodology, D.P. and Y.Q.; formal analysis,
D.P., Y.Q. and S.A.W.; investigation, D.P., Y.Q. and S.A.W.; writing—original draft preparation, Y.Q.;
writing—review and editing, D.P., Y.Q. and S.A.W.; supervision, Y.Q.; project administration, Y.Q.
and S.A.W.; funding acquisition, Y.Q. All authors have read and agreed to the published version of
the manuscript.
Funding: The research of Y. Quintana has been partially supported by the grant CEX2019-000904-
S funded by MCIN/AEI/10.13039/501100011033, and by the Madrid Government (Comunidad
de Madrid-Spain) under the Multiannual Agreement with UC3M in the line of Excellence of Uni-
versity Professors (EPUC3M23), in the context of the Fifth Regional Programme of Research and
Technological Innovation (PRICIT).
Data Availability Statement: Data sharing is not applicable to this article.
Acknowledgments: The authors express their profound gratitude to the referees and the academic
editor for their meticulous review of our manuscript and their invaluable comments and suggestions,
which significantly contributed to the enhancement of this paper.
Conflicts of Interest: The authors declare no conflicts of interest.

References
1. Leinartas, E.K.; Shishkina, O.A. The discrete analog of the Newton-Leibniz formula in the problem of summation over simplex
lattice points. J. Sib. Fed. Univ.-Math. Phys. 2019, 12, 503–508. [CrossRef]
2. Cuchta, T.; Luketic, R. Discrete hypergeometric Legendre polynomials. Mathematics 2021, 9, 2546. [CrossRef]
3. Albosaily, S.; Quintana, Y.; Iqbal, A.; Khan, W. Lagrange-based hypergeometric Bernoulli polynomials. Symmetry 2022, 14, 1125.
[CrossRef]
4. Quintana, Y. Generalized mixed type Bernoulli-Gegenbauer polynomial. Kragujev. J. Math. 2023, 47, 245–257. [CrossRef]
5. Quintana, Y.; Ramírez, W.; Urieles, A. On an operational matrix method based on generalized Bernoulli polynomials of level m.
Calcolo 2018, 55, 30. [CrossRef]
6. Quintana, Y.; Torres-Guzmán, H. Some relations between the Riemann zeta function and the generalized Bernoulli polynomials
of level m. Univers. J. Math. Appl. 2019, 2, 188–201. [CrossRef]
7. Quintana, Y.; Urieles, A. Quadrature formulae of Euler-Maclaurin type based on generalized Euler polynomials of level m. Bull.
Comput. Appl. Math. 2018, 6, 43–64.
8. Comtet, L. Advanced Combinatorics: The Art of Finite and Infinite Expansions, 2nd ed.; D. Reidel Publishing Company, Inc.: Boston,
MA, USA, 1974.
9. Kargin, L.; Kurt, V. On the generalization of the Euler polynomials with the real parameters. Appl. Math. Comput. 2011, 218,
856–859. [CrossRef]
10. Hassen, A.; Nguyen, H.D. Hypergeometric Bernoulli polynomials and Appell sequences. Int. J. Number Theory 2008, 4, 767–774.
[CrossRef]
11. Howard, F.T. Some sequences of rational numbers related to the exponential function. Duke Math. J. 1967, 34, 701–716. [CrossRef]
12. Natalini, P.; Bernardini, A. A generalization of the Bernoulli polynomials. J. Appl. Math. 2003, 2003, 155–163. [CrossRef]
13. Srivastava, H.M.; Choi, J. Zeta and q-Zeta Functions and Associated Series and Integrals, 1st ed.; Elsevier: London, UK, 2012.
14. Srivastava, H.M.; Manocha, H.L. A Treatise on Generating Functions, 1st ed.; Ellis Horwood Ltd.: West Sussex, UK, 1984.
15. Hernández-Llanos, P.; Quintana, Y.; Urieles, A. About extensions of generalized Apostol-type polynomials. Results Math. 2015, 68,
203–225. [CrossRef]
16. Szegő, G. Orthogonal Polynomials, 4th ed.; American Mathematical Society: Providence, RI, USA, 1975.
17. Paschoa, V.G.; Pérez, D.; Quintana, Y. On a theorem by Bojanov and Naidenov applied to families of Gegenbauer-Sobolev
polynomials. Commun. Math. Anal. 2014, 16, 9–18.
18. Pijeira, H.; Quintana, Y.; Urbina, W. Zero location and asymptotic behavior of orthogonal polynomials of Jacobi-Sobolev. Rev. Col.
Mat. 2001, 35, 77–97.
19. Askey, R. Orthogonal Polynomials and Special Functions, 1st ed.; SIAM: Philadelphia, PA, USA, 1975.
20. Temme, N.M. Special Functions. An Introduction to the Classical Functions of Mathematical Physics, 1st ed.; John Wiley & Sons Inc.:
New York, NY, USA, 1996.
21. Andrews, L.C. Special Functions for Engineers and Applied Mathematicians, 1st ed.; Macmillan Publishing Company: New York, NY,
USA, 1985.

61
Mathematics 2023, 11, 3920

22. Costabile, F.A.; Gualtieri, M.I.; Napoli, A. Matrix calculus-based approach to orthogonal polynomial sequences. Mediterr. J. Math.
2020, 17, 118. [CrossRef]
23. Ricci, P.E.; Tavkhelidze, I. An introduction to operational techniques and special polynomials. J. Math. Sci. 2009, 157, 161–189.
[CrossRef]
24. Dattoli, G.; Khomasuridze, I.; Cesarano, C.; Ricci, P.E. Bilateral generating functions of Laguerre polynomials and operational
methods. South East Asian J. Math. Math. Sci. 2006, 4, 1–6.
25. Balaji, S. Legendre wavelet operational matrix method for solution of fractional order Riccati differential equation. J. Egypt. Math.
Soc. 2015, 23, 263–270. [CrossRef]
26. Golbabai, A.; Ali Beik, S.P. An efficient method based on operational matrices of Bernoulli polynomials for solving matrix
differential equations. Comput. Appl. Math. 2015, 34, 159–175. [CrossRef]
27. Yousefi, S.A.; Behroozifar, M. Operational matrices of Bernstein polynomials and their applications. Internat. J. Syst. Sci. 2010,
41, 709–716. [CrossRef]
28. Yousefi, S.A.; Behroozifar, M.; Dehghan, M. The operational matrices of Bernstein polynomials for solving the parabolic equation
subject to specification of the mass. J. Comput. Appl. Math. 2010, 41, 709–716. [CrossRef]
29. Costabile, F.; Longo, E. A determinantal approach to Appell polynomials. J. Comput. Appl. 2010, 234, 1528–1542. [CrossRef]

Disclaimer/Publisher’s Note: The statements, opinions and data contained in all publications are solely those of the individual
author(s) and contributor(s) and not of MDPI and/or the editor(s). MDPI and/or the editor(s) disclaim responsibility for any injury to
people or property resulting from any ideas, methods, instructions or products referred to in the content.

62
 mathematics

Article
The Recurrence Coefficients of Orthogonal Polynomials with
a Weight Interpolating between the Laguerre Weight and the
Exponential Cubic Weight
Chao Min * and Pixin Fang

School of Mathematical Sciences, Huaqiao University, Quanzhou 362021, China; [email protected]

* Correspondence: [email protected]

Abstract: In this paper, we consider the orthogonal polynomials with respect to the weight w( x ) =
w( x; s) := x λ e− N [ x+s( x − x)] , x ∈ R+ , where λ > 0, N > 0 and 0 ≤ s ≤ 1. By using the ladder
3

operator approach, we obtain a pair of second-order nonlinear difference equations and a pair of
differential–difference equations satisfied by the recurrence coefficients αn (s) and β n (s). We also
establish the relation between the associated Hankel determinant and the recurrence coefficients.
From Dyson’s Coulomb fluid approach, we prove that the recurrence coefficients converge and the
limits are derived explicitly when q := n/N is fixed as n → ∞.

Keywords: orthogonal polynomials; Laguerre weight; exponential cubic weight; ladder operators;
difference equations; Coulomb fluid

MSC: 33C45; 42C05

1. Introduction
In this paper, we are concerned with the coefficients in the three-term recurrence
Citation: Min, C.; Fang, P. The
relation for the orthogonal polynomials with respect to the weight
Recurrence Coefficients of
w( x ) = w( x; s) := x λ e− N [ x+s( x
3 − x )]
Orthogonal Polynomials with a , x ∈ R+ , (1)
Weight Interpolating between the
Laguerre Weight and the Exponential
with parameters λ > 0, N > 0 and 0 ≤ s ≤ 1.
Cubic Weight. Mathematics 2023, 11,
If s = 0, the weight (1) is the classical (scaled with N) Laguerre weight. If s = 1, it
3842. https://doi.org/10.3390/ is an exponential cubic weight. Orthogonal polynomials associated with the exponential
math11183842 cubic weight have been well studied (see e.g., [1–4]), and have important applications in
numerical analysis [5] and random matrix theory [6–8]. Furthermore, orthogonal poly-
Academic Editor: Yamilet Quintana
nomials and the Hankel determinant for the so-called semi-classical Laguerre weight
w̃( x ) = x λ e− N [ x+s( x − x)] , x ∈ R+ have been studied in [9,10], which is also the motivation
2
Received: 15 August 2023
Revised: 2 September 2023 of the present paper.
Accepted: 6 September 2023 Let { Pn ( x; s)}∞n=0 be a sequence of monic polynomials, Pn ( x ) of degree n, orthogonal
Published: 7 September 2023 with respect to the weight (1); that is,
 ∞
Pm ( x; s) Pn ( x; s)w( x; s)dx = hn (s)δmn , m, n = 0, 1, 2, . . . , (2)
0
Copyright: © 2023 by the authors.
Licensee MDPI, Basel, Switzerland. where hn (s) > 0 and Pn ( x; s) has the expansion
This article is an open access article
distributed under the terms and Pn ( x; s) = x n + p(n, s) x n−1 + · · · + Pn (0; s),
conditions of the Creative Commons
Attribution (CC BY) license (https://
where p(n, s), the sub-leading coefficient of Pn ( x; s), will play a significant role in the
creativecommons.org/licenses/by/ following discussions. Note that Pn ( x; s) and p(n, s) also depend on the parameters λ
4.0/). and N.

Mathematics 2023, 11, 3842. https://doi.org/10.3390/math11183842 63 https://www.mdpi.com/journal/mathematics

Mathematics 2023, 11, 3842

One of the most important properties of the orthogonal polynomials is that they satisfy
the three-term recurrence relation of the form

xPn ( x; s) = Pn+1 ( x; s) + αn (s) Pn ( x; s) + β n (s) Pn−1 ( x; s), (3)

with the initial conditions

P0 ( x; s) := 1, β 0 (s) P−1 ( x; s) := 0.

As an easy consequence, we have

αn (s) = p(n, s) − p(n + 1, s), (4)

hn (s)
β n (s) = > 0. (5)
h n −1 ( s )
Taking a telescopic sum of (4) and noting that p(0, s) := 0, we obtain an important identity

n −1
∑ α j (s) = −p(n, s). (6)
j =0

It is known that (see, e.g., [11] (p. 17)) Pn ( x; s) can be expressed as the determinant
 
 μ0 ( s ) μ1 ( s ) ··· μn (s) 

 μ1 ( s ) μ2 ( s ) ··· μn+1 (s) 
1  .. .. .. 
Pn ( x; s) =  . . . 
Dn ( s )  
μ μ2n−1 (s)
 n −1 ( s ) μn (s) ···
 1 x ··· xn 

and
Dn + 1 ( s )
hn (s) = , (7)
Dn ( s )
where Dn (s) is the Hankel determinant for the weight (1) defined by
 
 μ0 ( s ) μ1 ( s ) ··· μn−1 (s) 

 μ1 ( s ) μ2 ( s ) ··· μn (s) 
n −1 
Dn (s) := det(μi+ j (s))i,j =0 = .. .. .. ,
 . . . 
 
μ (s) μn (s) ··· μ (s)
n −1 2n−2

and μ j (s) is the jth moment given by the integral

 ∞
μ j (s) := x j w( x; s)dx.
0

We mention that the moment μ j (s) can be expressed in terms of the generalized
hypergeometric functions after some calculations.
Furthermore, it is easy to see from (7) that the Hankel determinant Dn (s) can be
expressed as the product of h j (s) in the form

n −1
Dn ( s ) = ∏ h j ( s ). (8)
j =0

Obviously, the recurrence coefficients αn (s), β n (s) and the Hankel determinant Dn (s)
are all dependent on the parameters λ and N in our problem. For more information about
orthogonal polynomials, see [11–13].

64
Mathematics 2023, 11, 3842

The remainder of the paper is organized as follows. In Section 2, by using the ladder
operator approach, we derive the discrete system for the recurrence coefficients αn (s)
and β n (s). We also obtain an important identity in the representation of the sub-leading
coefficient p(n, s) in terms of the recurrence coefficients. In Section 3, we derive the
differential–difference equations satisfied by the recurrence coefficients. We establish the
relation between the Hankel determinant Dn (s) and the recurrence coefficients, and also
obtain the differential–difference equations satisfied by Dn (s). In Section 4, by making
use of Dyson’s Coulomb fluid approach, we find that the large n limits of the recurrence
coefficients exist in the sense that n/N is fixed as n → ∞. The expressions of the limits are
also given explicitly. Finally, the conclusions and some remarks are outlined in Section 5.

2. Ladder Operators and Second-Order Difference Equations

The ladder operator approach has been applied to solve a series of problems about
semi-classical orthogonal polynomials and the related Hankel determinants, especially the
relationship to Painlevé equations; see, e.g., [14–16] and the references therein. Note that,
in order to simplify the notations, the s-dependence of many quantities such as Pn ( x ), w( x ),
hn , αn and β n will not be displayed unless it is needed. Following the general set-up of Chen
and Ismail [17,18], the lowering and raising operators for our orthogonal polynomials are
 
d
+ Bn ( x ) Pn ( x ) = β n An ( x ) Pn−1 ( x ),
dx
 
d
− Bn ( x ) − v ( x ) Pn−1 ( x ) = − An−1 ( x ) Pn ( x ),
dx
where the functions An ( x ) and Bn ( x ) are defined by
 ∞ 
1 v ( x ) − v ( y ) 2
An ( x ) := Pn (y)w(y)dy, (9)
hn 0 x−y
 ∞ 
1 v ( x ) − v ( y )
Bn ( x ) := Pn (y) Pn−1 (y)w(y)dy, (10)
h n −1 0 x−y
and v( x ) = − ln w( x ).
The associated compatibility conditions for the functions An ( x ) and Bn ( x ) are

Bn+1 ( x ) + Bn ( x ) = ( x − αn ) An ( x ) − v ( x ), (11)

1 + ( x − αn )( Bn+1 ( x ) − Bn ( x )) = β n+1 An+1 ( x ) − β n An−1 ( x ), (12)

n −1
Bn2 ( x ) + v ( x ) Bn ( x ) + ∑ A j ( x ) = β n A n ( x ) A n −1 ( x ). (13)
j =0

Here, (13) is obtained by the combination of (11) and (12), and is usually more useful
compared to (12).
For our problem with the weight (1), we have

v( x ) = − ln w( x ) = N [ x + s( x3 − x )] − λ ln x,

and
v ( x ) − v ( y ) λ
= 3Ns( x + y) + . (14)
x−y xy
Using (14), we compute the functions An ( x ) and Bn ( x ) in the following lemma.

65
Mathematics 2023, 11, 3842

Lemma 1. For our problem, the expressions of An ( x ) and Bn ( x ) are given by

Rn (s)
An ( x ) = 3Ns( x + αn ) + , (15)
x

rn (s)
Bn ( x ) = 3Nsβ n + , (16)
x
where Rn (s) and rn (s) are the auxiliary quantities
 ∞
λ 1 2
Rn (s) := Pn (y)w(y)dy,
hn 0 y
 ∞
λ 1
rn (s) := Pn (y) Pn−1 (y)w(y)dy.
h n −1 0 y

Proof. Substituting (14) into the definitions of An ( x ) and Bn ( x ) in (9) and (10), we obtain
the desired results by using the orthogonality condition (2) and the three-term recurrence
relation (3).

From the compatibility conditions (11) and (13), we have the following results.

Proposition 1. The recurrence coefficients αn , β n and the auxiliary quantities Rn (s), rn (s)
satisfy the relations as follows:

3Ns( β n+1 + β n ) = Rn (s) − N (1 − s) − 3Nsα2n , (17)

r n +1 ( s ) + r n ( s ) = λ − α n R n ( s ), (18)

rn (s) + n = 3Nsβ n (αn + αn−1 ), (19)

n −1
3Nsβ2n + N (1 − s) β n + ∑ αj = βn Rn (s) + Rn−1 (s) + 3Nsαn αn−1 , (20)
j =0

n −1
Nrn (s)(6sβ n + 1 − s) + ∑ R j (s) = 3Nsβ n α n R n −1 ( s ) + α n −1 R n ( s ) + λ , (21)
j =0

rn2 (s) − λrn (s) = β n Rn (s) Rn−1 (s). (22)

Proof. Substituting (15) and (16) into (11), and comparing the coefficients of z0 and
z−1 on both sides, we obtain (17) and (18), respectively. Similarly, substituting (15)
and (16) into (13), and comparing the coefficients of z1 , z0 , z−1 and z−2 on both sides,
we obtain (19), (20), (21) and (22), respectively.

Now we are ready to derive the main result of this section on the discrete system for
the recurrence coefficients.

Theorem 1. The recurrence coefficients αn and β n satisfy a pair of second-order nonlinear difference
equations:

2n + λ + 1
3s α3n + β n (2αn + αn−1 ) + β n+1 (2αn + αn+1 ) + (1 − s)αn = , (23a)
N

66
Mathematics 2023, 11, 3842

n n+λ  
3sβ n (αn + αn−1 ) − 3sβ n (αn + αn−1 ) − = β n 3s(α2n + β n + β n+1 ) + 1 − s
N N
 
× 3s(α2n−1 + β n−1 + β n ) + 1 − s . (23b)

Proof. From (17) and (19), we can express Rn (s) and rn (s) in terms of the recurrence coefficients:

Rn (s) = 3Ns(α2n + β n + β n+1 ) + N (1 − s), (24)

rn (s) = 3Nsβ n (αn + αn−1 ) − n. (25)

Substituting (24) and (25) into (18) and (22), we obtain (23a) and (23b), respectively.

Remark 1. When s = 0, the results in the above theorem are reduced to

Nαn (0) = 2n + λ + 1, N 2 β n (0) = n ( n + λ ), (26)

which are consistent with the recurrence coefficients of the classical monic Laguerre polynomials.

At the end of this section, we give an expression of the sub-leading coefficient p(n, s),
which will be very useful in the analysis of the next section.

Corollary 1. The sub-leading coefficient p(n, s) can be expressed in terms of the recurrence coeffi-
cients as follows:
 
p(n, s) = − Nβ n 3s α2n−1 + αn−1 αn + α2n + β n−1 + β n + β n+1 + 1 − s . (27)

Proof. Substituting (6) into (20), we have

p(n, s) = 3Nsβ2n + N (1 − s) β n − β n Rn (s) + Rn−1 (s) + 3Nsαn αn−1 .

Eliminating Rn (s) and Rn−1 (s) by (24), we obtain (27).

3. S Evolution and Differential-Difference Equations

Note that all the quantities discussed in this paper, such as the recurrence coefficients
αn and β n , depend on the parameter s. We consider the s evolution in this section.
We start from taking a derivative with respect to s in the equation
 ∞
Pn2 ( x; s) x λ e− N [ x+s( x
3 − x )]
hn (s) = dx,
0

which gives

d 3Ns ∞
3s ln hn (s) = ( x − x3 ) Pn2 ( x )w( x )dx
ds hn 0
 
3Ns ∞ 2 3Ns ∞ 3 2
= xPn ( x )w( x )dx − x Pn ( x )w( x )dx. (28)
hn 0 hn 0

By the three-term recurrence relation (3), we obtain the first term

 ∞
3Ns
xPn2 ( x )w( x )dx = 3Nsαn (29)
hn 0

and the second term

 ∞
3Ns
x3 Pn2 ( x )w( x )dx = 3Ns α3n + β n (2αn + αn−1 ) + β n+1 (2αn + αn+1 )
hn 0
= 2n + λ + 1 − N (1 − s)αn , (30)

67
Mathematics 2023, 11, 3842

where we have used (23a) in the second step to simplify the result.
From (28)–(30), it follows that

d
3s ln hn (s) = N (1 + 2s)αn − (2n + λ + 1). (31)
ds
Using (5), we have

d d d
3s ln β n (s) = 3s ln hn (s) − 3s ln hn−1 (s) = N (1 + 2s)(αn − αn−1 ) − 2;
ds ds ds
that is,
3sβn (s) = β n [ N (1 + 2s)(αn − αn−1 ) − 2].
On the other hand, differentiating with respect to s in the equation
 ∞
Pn ( x; s) Pn−1 ( x; s) x λ e− N [ x+s( x
3 − x )]
dx = 0
0

produces
 ∞  ∞
d 3Ns 3Ns
3s p(n, s) = x3 Pn ( x ) Pn−1 ( x )w( x )dx − xPn ( x ) Pn−1 ( x )w( x )dx. (32)
ds h n −1 0 h n −1 0

The first term is

  
3Ns ∞ 3
x Pn ( x ) Pn−1 ( x )w( x )dx = 3Nsβ n α2n + αn αn−1 + α2n−1 + β n+1 + β n + β n−1
h n −1 0
= −p(n, s) − N (1 − s) β n , (33)

where we have used (27) to simplify the result in the second equality. The second term reads
 ∞
3Ns
xPn ( x ) Pn−1 ( x )w( x )dx = 3Nsβ n . (34)
h n −1 0

Substituting (33) and (34) into (32), we find

d
3s p(n, s) = −p(n, s) − N (1 + 2s) β n . (35)
ds
Taking account of (4), it follows that

3sαn (s) = −αn + N (1 + 2s)( β n+1 − β n ).

To sum up, we have the following theorem.

Theorem 2. The recurrence coefficients αn and β n satisfy the coupled differential–difference equations:

3sαn (s) = −αn + N (1 + 2s)( β n+1 − β n ),

3sβn (s) = β n [ N (1 + 2s)(αn − αn−1 ) − 2].

We also derive some results about the Hankel determinant Dn (s) as follows.

Theorem 3. The logarithmic derivative of the Hankel determinant is expressed in terms of the
recurrence coefficients as follows:

d  
3s ln Dn (s) = N 2 (1 + 2s) β n 3s α2n−1 + αn−1 αn + α2n + β n−1 + β n + β n+1 + 1 − s − n(n + λ).
ds

68
Mathematics 2023, 11, 3842

Proof. From (8) and (31), we have

n −1
d d
3s
ds
ln Dn (s) = ∑ 3s ds ln h j (s)
j =0
n −1 
= ∑ N (1 + 2s)α j − (2j + λ + 1) .
j =0

Taking account of (6) and using (27), we find

d
3s ln Dn (s) = − N (1 + 2s)p(n, s) − n(n + λ) (36)
ds  
= N 2 (1 + 2s) β n 3s α2n−1 + αn−1 αn + α2n + β n−1 + β n + β n+1 + 1 − s
− n ( n + λ ).

The proof is complete.

Corollary 2. The Hankel determinant Dn (s) satisfies the differential–difference equation

d2 d D (s) D (s)
9s2 (1 + 2s) ln Dn (s) + 6s(2 + s) ln Dn (s) + n(n + λ)(1 − 4s) = N 2 (1 + 2s)3 n+1 2 n−1 .
ds2 ds Dn ( s )

Proof. From (36), we have

d
3s ds ln Dn (s) + n(n + λ)
p(n, s) = − . (37)
N (1 + 2s)

A combination of (5) and (7) gives

Dn + 1 ( s ) Dn − 1 ( s )
β n (s) = . (38)
Dn2 (s)

Substituting (37) and (38) into (35), we obtain the desired result.

4. Asymptotics of the Recurrence Coefficients

Recall that, for our problem, the weight function is

w ( x ) = x λ e− N [ x + s ( x
3 − x )]
, x ∈ R+ (39)

and the potential is

v( x ) = N [ x + s( x3 − x )] − λ ln x, x ∈ R+ , (40)

where λ > 0, N > 0 and 0 ≤ s ≤ 1.

In random matrix theory [19–21], it is known that our Hankel determinant Dn (s) is
equal to the partition function for the unitary random matrix ensemble associated with the
weight (39) [11] (Corollary 2.1.3), i.e.,
 n
1
( xi − x j )2 ∏ xkλ e− N [ xk +s( xk − xk )] dxk ,
3
Dn ( s ) =
n! (0,∞)n
∏
1≤ i < j ≤ n k =1

where { x j }nj=1 are the eigenvalues of n × n Hermitian matrices from the ensemble with the
joint probability density function
n
1
( xi − x j )2 ∏ xkλ e− N [ xk +s( xk − xk )] .
3

n! Dn (s) 1≤∏
p ( x1 , x2 , . . . , x n ) =
i< j≤n k =1

69
Mathematics 2023, 11, 3842

If we interpret { x j }nj=1 as the positions of n charged particles, then the collection of

particles can be approximated as a continuous fluid with an equilibrium density σ( x ) in the
limit of large n according to Dyson’s Coulomb fluid approach [22]. Since our potential v( x )
in (40) is convex for x ∈ R+ , the density σ( x ) is supported on an single interval denoted by
(0, b); see Chen and Ismail [23] and also [24] (p. 198).
Following [23], the equilibrium density σ ( x ) is obtained by minimizing the free en-
ergy functional
 b  b b
F [σ] := σ ( x )v( x )dx − σ ( x ) ln | x − y|σ (y)dxdy
0 0 0

subject to the normalization condition

 b
σ ( x )dx = n. (41)
0

Upon minimization, the density σ ( x ) satisfies the integral equation

 b
v( x ) − 2 ln | x − y|σ (y)dy = A, x ∈ (0, b),
0

where A is the Lagrange multiplier for the constraint (41). Taking a derivative with respect
to x in the above equation gives the singular integral equation
 b
σ(y)
v ( x ) − 2P dy = 0, x ∈ (0, b), (42)
0 x−y

where P denotes the Cauchy principal value. From the theory of singular integral equa-
tions [25], the solution of (42) is given by
(  b  (
1 b−x v (y) y
σ( x) = P dy. (43)
2π 2 x 0 y−x b−y

Substituting (40) into (43) and after some elaborate computations, we find
( #  $
N b−x 3bx 9b2
σ( x) = 1 + s 3x2 + + −1 .
2π x 2 8

The normalization condition (41) then becomes

1
Nb 15sb2 + 8(1 − s) = n. (44)
32
Motivated by the works [9,10], we consider the case that q := n/N is fixed when
n → ∞. Equation (44) is actually a cubic equation for b,

15sb3 + 8(1 − s)b − 32q = 0,

which has a unique real solution given by

24/3
b= ξ 1/3 − 101/3 s(1 − s)ξ −1/3 ,
3 × 52/3 s
where 
ξ = 45qs2 + s 5s[2 + 3(135q2 − 2)s + 6s2 − 2s3 ].

70
Mathematics 2023, 11, 3842

It was shown in Chen and Ismail [23] that, as n → ∞,

 2 
b ∂ A
αn (s) = + O ,
2 ∂s∂n
  3 
b2 ∂ A
β n (s) = 1+O .
16 ∂n3
Hence, we have the following theorem.

Theorem 4. Let q := n/N be fixed when n → ∞. Then, the limits of αn and β n as n → ∞ exist
and are given by
21/3
lim αn = ξ 1/3 − 101/3 s(1 − s)ξ −1/3 , (45)
n→∞ 3 × 52/3 s

22/3
lim β n = ξ 2/3 + 102/3 s2 (1 − s)2 ξ −2/3 − 2 × 101/3 s(1 − s) , (46)
n→∞ 180 × 51/3 s2
where 
ξ = 45qs2 + s 5s[2 + 3(135q2 − 2)s + 6s2 − 2s3 ].

Remark 2. It is an interesting phenomenon that the limits of the recurrence coefficients in (45) and
(46) are independent of the parameter λ.

Remark 3. When s → 0+ , we find from (45) and (46) that

lim αn = 2q, lim β n = q2 ,

n→∞ n→∞

which coincides with the classical results for the Laguerre polynomials; see (26).

Remark 4. We conjecture that αn and β n have the following large n asymptotic expansion
∞ aj ∞ bj
αn = ∑ nj , βn = ∑ nj ,
j =0 j =0

where a0 and b0 are given by the right hand sides of (45) and (46), respectively. Then, one can
determine the expansion coefficients a j and b j recursively by using the discrete system for the
recurrence coefficients in (23) following the procedure in [14–16]. However, the results are too
complicated to write down here.

5. Conclusions
In this paper, we studied the monic polynomials orthogonal with respect to a semi-
classical weight, which interpolates between the classical Laguerre weight and the exponen-
tial cubic weight. By making use of the ladder operator approach, we derived the discrete
system for the recurrence coefficients αn (s) and β n (s). Considering the s evolution, we
obtained the coupled differential–difference equations satisfied by αn (s) and β n (s). We
also studied the relations between the associated Hankel determinant, the sub-leading
coefficient of the monic orthogonal polynomials and the recurrence coefficients. Finally,
we proved that the large n limits of the recurrence coefficients exist and are given when
n/N is fixed as n → ∞. The large n asymptotic expansions of the recurrence coefficients,
the sub-leading coefficient p(n, s) and the Hankel determinant Dn (s) in the sense that n/N
is fixed as n → ∞ can be considered based on the results in this paper; however, we found
that the computations are very cumbersome.

71
Mathematics 2023, 11, 3842

Author Contributions: Methodology, C.M.; Software, P.F.; Validation, C.M.; Formal analysis, P.F.;
Investigation, C.M. and P.F.; Resources, C.M.; Writing—original draft, P.F.; Writing—review & editing,
C.M.; Supervision, C.M.; Funding acquisition, C.M. All authors have read and agreed to the published
version of the manuscript.
Funding: This work was partially supported by the National Natural Science Foundation of China
under grant number 12001212, by the Fundamental Research Funds for the Central Universities
under grant number ZQN-902 and by the Scientific Research Funds of Huaqiao University under
grant number 17BS402.
Data Availability Statement: Data sharing not applicable to this article as no datasets were generated
or analyzed during the current study.
Acknowledgments: The authors thank the reviewers for giving many useful comments, which
improve the presentation of this paper.
Conflicts of Interest: The authors have no competing interest to declare that are relevant to the
content of this article.

References
1. Clarkson, P.A.; Jordaan, K. Generalised Airy polynomials. J. Phys. A Math. Theor. 2021, 54, 185202 . [CrossRef]
2. Magnus, A.P. Painlevé-type differential equations for the recurrence coefficients of semi-classical orthogonal polynomials.
J. Comput. Appl. Math. 1995, 57, 215–237. [CrossRef]
3. Martínez-Finkelshtein, A.; Silva, G.L.F. Critical measures for vector energy: Asymptotics of non-diagonal multiple orthogonal
polynomials for a cubic weight. Adv. Math. 2019, 349, 246–315. [CrossRef]
4. Assche, W.V.; Filipuk, G.; Zhang, L. Multiple orthogonal polynomials associated with an exponential cubic weight. J. Approx.
Theory 2015, 190, 1–25. [CrossRef]
5. Deaño, A.; Huybrechs, D.; Kuijlaars, A.B.J. Asymptotic zero distribution of complex orthogonal polynomials associated with
Gaussian quadrature. J. Approx. Theory 2010, 162, 2202–2224. [CrossRef]
6. Bleher, P.; Deaño, A. Topological expansion in the cubic random matrix model. Int. Math. Res. Not. 2013, 2013, 2699–2755.
[CrossRef]
7. Bleher, P.; Deaño, A. Painlevé I double scaling limit in the cubic random matrix model. Random Matrices Theor. Appl. 2016,
5, 1650004. [CrossRef]
8. Bleher, P.; Deaño, A.; Yattselev, M. Topological expansion in the complex cubic log-gas model: One-cut case. J. Stat. Phys. 2017,
166, 784–827. [CrossRef]
9. Deaño, A.; Simm, N.J. On the probability of positive-definiteness in the gGUE via semi-classical Laguerre polynomials. J. Approx.
Theory 2017, 220, 44–59. [CrossRef]
10. Han, P.; Chen, Y. The recurrence coefficients of a semi-classical Laguerre polynomials and the large n asymptotics of the associated
Hankel determinant. Random Matrices Theor. Appl. 2017, 6, 1740002. [CrossRef]
11. Ismail, M.E.H. Classical and Quantum Orthogonal Polynomials in One Variable; Encyclopedia of Mathematics and Its Applications 98;
Cambridge University Press: Cambridge, UK, 2005.
12. Chihara, T.S. An Introduction to Orthogonal Polynomials; Dover: New York, NY, USA, 1978.
13. Szegö, G. Orthogonal Polynomials, 4th ed.; American Mathematical Society: Providence, RI, USA, 1975.
14. Min, C.; Chen, Y. Differential, difference, and asymptotic relations for Pollaczek-Jacobi type orthogonal polynomials and their
Hankel determinants. Stud. Appl. Math. 2021, 147, 390–416. [CrossRef]
15. Min, C.; Chen, Y. A note on the asymptotics of the Hankel determinant associated with time-dependent Jacobi polynomials. Proc.
Amer. Math. Soc. 2022, 150, 1719–1728. [CrossRef]
16. Min, C.; Chen, Y. Painlevé IV, Chazy II, and asymptotics for recurrence coefficients of semi-classical Laguerre polynomials and
their Hankel determinants. Math. Meth. Appl. Sci. 2023, 46, 15270–15284. [CrossRef]
17. Chen, Y.; Ismail, M.E.H. Ladder operators and differential equations for orthogonal polynomials. J. Phys. A Math. Gen. 1997, 30,
7817–7829. [CrossRef]
18. Chen, Y.; Ismail, M.E.H. Jacobi polynomials from compatibility conditions. Proc. Am. Math. Soc. 2005, 133, 465–472. [CrossRef]
19. Deift, P. Orthogonal Polynomials and Random Matrices: A Riemann-Hilbert Approach; American Mathematical Society: Providence, RI,
USA, 2000.
20. Forrester, P.J. Log-Gases and Random Matrices; Princeton University Press: Princeton, NJ, USA, 2010.
21. Mehta, M.L. Random Matrices, 3rd ed.; Elsevier: New York, NY, USA, 2004.
22. Dyson, F.J. Statistical theory of the energy levels of complex systems, I, II, III. J. Math. Phys. 1962, 3, 140–156. 157–165. 166–175.
[CrossRef]
23. Chen, Y.; Ismail, M.E.H. Thermodynamic relations of the Hermitian matrix ensembles. J. Phys. A Math. Gen. 1997, 30, 6633–6654.
[CrossRef]

72
Mathematics 2023, 11, 3842

24. Saff, E.B.; Totik, V. Logarithmic Potentials with External Fields; Springer: Berlin, Germany, 1997.
25. Mikhlin, S.G. Integral Equations and Their Applications to Certain Problems in Mechanics, Mathematical Physics and Technology, 2nd ed.;
Pergamon Press: New York, NY, USA, 1964.

Disclaimer/Publisher’s Note: The statements, opinions and data contained in all publications are solely those of the individual
author(s) and contributor(s) and not of MDPI and/or the editor(s). MDPI and/or the editor(s) disclaim responsibility for any injury to
people or property resulting from any ideas, methods, instructions or products referred to in the content.

73
 mathematics

Article
Properties of Multivariate Hermite Polynomials in Correlation
with Frobenius–Euler Polynomials
Mohra Zayed 1,† , Shahid Ahmad Wani 2, *,† and Yamilet Quintana 3,4, *,†

1 Mathematics Department, College of Science, King Khalid University, Abha 61413, Saudi Arabia;
[email protected]
2 Department of Applied Sciences, Symbiosis Institute of Technology, Symbiosis International (Deemed
University) (SIU), Lavale, Pune 412115, Maharashtra, India
3 Departamento de Matemáticas, Universidad Carlos III de Madrid, Avenida de la Universidad 30, Leganés,
28911 Madrid, Spain
4 Instituto de Ciencias Matemáticas (ICMAT), Campus de Cantoblanco UAM, 28049 Madrid, Spain
* Correspondence: [email protected] or [email protected] (S.A.W.);
[email protected] (Y.Q.)
† These authors contributed equally to this work.

Abstract: A comprehensive framework has been developed to apply the monomiality principle
from mathematical physics to various mathematical concepts from special functions. This paper
presents research on a novel family of multivariate Hermite polynomials associated with Apostol-
type Frobenius–Euler polynomials. The study derives the generating expression, operational rule,
differential equation, and other defining characteristics for these polynomials. Additionally, the
monomiality principle for these polynomials is verified. Moreover, the research establishes series
representations, summation formulae, and operational and symmetric identities, as well as recurrence
relations satisfied by these polynomials.

Keywords: multivariate special polynomials; monomiality principle; explicit form; operational

connection; symmetric identities; summation formulae

Citation: Zayed, M.; Wani, S.A.; MSC: 33E20; 33C45; 33B10; 33E30; 11T23
Quintana, Y. Properties of
Multivariate Hermite Polynomials in
Correlation with Frobenius–Euler
Polynomials. Mathematics 2023, 11, 1. Introduction and Preliminaries
3439. https://doi.org/10.3390/
A current field of study with practical applications involves investigating the con-
math11163439
volution of multiple polynomials as a method for introducing innovative multivariate
Academic Editor: Francesco Aldo generalized polynomials. These polynomials hold immense importance due to their useful
Costabile characteristics, which include recurring and explicit relations, functional and differential
equations, summation formulae, symmetric and convolution identities, determinant forms,
Received: 8 July 2023
Revised: 26 July 2023
and more.
Accepted: 1 August 2023
Multivariate hybrid special polynomials exhibit a wide range of features that show
Published: 8 August 2023 great promise for their utilization in various areas of pure and practical mathematics, such
as number theory, combinatorics, classical and numerical analysis, theoretical physics, and
approximation theory. The development of diverse new classes of hybrid polynomials is
motivated by the desire to harness their utility and potential for application.
Copyright: © 2023 by the authors. Sequences of polynomials hold significant relevance in various domains of applied
Licensee MDPI, Basel, Switzerland. mathematics, theoretical physics, approximation theory, and other branches of mathemat-
This article is an open access article ics. Particularly, the Bernstein polynomials of degree n serve as a foundational basis for
distributed under the terms and the space of polynomials with degrees less than or equal to n. Dattoli and collaborators
conditions of the Creative Commons utilized operational approaches to examine Bernstein polynomials [1], exploring the Ap-
Attribution (CC BY) license (https:// pell sequences—a broad class encompassing several well-known polynomial sequences,
creativecommons.org/licenses/by/
including the Miller–Lee, Bernoulli, and Euler polynomials, among others.
4.0/).

Mathematics 2023, 11, 3439. https://doi.org/10.3390/math11163439 74 https://www.mdpi.com/journal/mathematics

Mathematics 2023, 11, 3439

The introduction and study of classes of hybrid special polynomials connected to

the Appell sequences, as seen in references [2–7], play a significant role in engineering,
biological, medical, and physical sciences. These hybrid polynomials are of paramount
importance due to their key characteristics, such as differential equations, generating
functions, series definitions, integral representations, and more. In numerous scientific and
technical fields, problems are often expressed as differential equations, and their solutions
typically manifest as special functions. Consequently, the challenges encountered in the
development of scientific fields can be addressed by utilizing the differential equations
satisfied by these hybrid special polynomials.
The multivariate special polynomials are extremely important in many areas of mathe-
matics and have many uses. They are crucial in algebraic geometry, which examines the
geometric properties of algebraic varieties. They are used to define and study significant
geometric objects such as algebraic curves, surfaces, and higher-dimensional varieties.
These polynomials describe the intersection of curves and surfaces, the singularities of
algebraic varieties, and the properties of their coordinate rings. They may also be observed
in many areas of theoretical physics, including quantum mechanics and quantum field
theory. They show up as differential equation solutions in mathematical physics, especially
when eigenvalue issues, boundary value issues, and symmetry analysis are involved. These
polynomials have applications in quantum field theory, statistical mechanics, the study of
integrable systems, etc. Due to such significance, several authors introduced multivariate
Hermite and other special polynomials. Datolli et al. [8] introduced the generating function:
∞
tn
e u1 t + u2 t
2 +u 3
3t = ∑ Hn (u1 , u2 , u3 ) n! , (1)
n =0

representing three-variable Hermite polynomials (3VHPs) Hn (u1 , u2 , u3 ).

Further, by taking u3 = 0, 3VHPs reduce to the polynomials Hn (u1 , u2 ) widely
known as 2-v Hermite Kampé de Fériet polynomials (2VHKdFPs) [9] and on taking
u3 = 0, u1 = 2u1 and u2 = −1 3VHPs become the classical Hermite polynomials
Hn (u1 ) [10] (Equation (5.1), p. 167).
At this point, it is noteworthy to mention that many semi-classical orthogonal poly-
nomials, serving as generalizations of classical orthogonal polynomials such as Hermite,
Laguerre, and Jacobi polynomials, have been extensively studied in recent years. Enthusias-
tic readers are encouraged to explore the works of [11,12] (and the references cited therein),
along with the valuable insights presented in the book [13]. Furthermore, other interesting
results concerning recurrence relations for generalized Appell polynomials and summation
problems involving simplex lattice points or operators with a summing effect can be found
in [14–16].
[m]
Recently, the polynomials represented by Yn (u1 , u2 , . . . , um ), known as multivariate
Hermite polynomials (MHPs), were introduced in [17] and are given by generating relation:
∞
[m] ξn
exp(u1 ξ + u2 ξ 2 + · · · + um ξ m ) = ∑ Yn ( u1 , u2 , . . . , u m )
n!
, (2)
n =0

with the operational rule:

 
∂2 ∂3 ∂m [m]
exp u2 + u3 + · · · + um u1n = Yn (u1 , u2 , . . . , um ), (3)
∂u1 2 ∂u1 3 ∂u1 m

and series representation:

[n/m] [m]
[m] urm Yn−mr (u1 , u2 , . . . , um−1 )
Yn (u1 , u2 , . . . , um ) = n! ∑ r! (n − mr )!
. (4)
r =0

Several mathematicians are keen to introduce different forms of various special poly-
nomials. The unified forms of Apostol-type polynomials are introduced in the study of [18].

75
Mathematics 2023, 11, 3439

These polynomials are known as the Apostol-type Frobenius–Euler polynomials and they
are represented mathematically by the symbol Fn (u1 ; u) [19]. For λ = 1, these polynomials
reduce to the Frobenius–Euler polynomials [20]. We now recall the generating expression
of these Frobenius–Euler polynomials, which is as follows:
  ∞
1 − u u1 ξ ξn
eξ − u
e = ∑ Fn (u1 ; u) n! , (5)
n =0

where u ∈ C, u = 1.
Therefore, on taking u1 = 0, expression (5) gives the Frobenius–Euler numbers (FENs)
Fn (u), defined by
∞
1−u ξn
= ∑ Fn (u) . (6)
eξ − u n =0
n!

Further, on taking u = −1, the FEPs becomes Euler polynomials (EPs) An (u1 ) [21].
Extensive research has been dedicated to the advancement and integration of the
monomiality principle, operational rules, and other properties within the domain of hy-
brid special polynomials. This line of investigation traces its roots back to 1941 when
Steffenson initially proposed the concept of poweroids as a means to understanding
monomiality [22]. Building upon Steffenson’s work, Dattoli further refined the theory,
offering valuable insights and refinements [2]. Their contributions have paved the way
for a more comprehensive understanding of the monomiality principle and its appli-
cation within the context of the so-called hybrid special polynomials. Therefore, on a
[m]
combination of multivariate Hermite polynomials Yn (u1 , u2 , . . . , um ) given by (2) and
Frobenius–Euler polynomials [23,24] given by (5) by using the concept of the monomiality
principle and operational rules, the convoluted new polynomial, namely, multivariate
Hermite–Frobenius–Euler polynomials are given by the formal expression:
  ∞
1−u [m] ξn
eξ − u
exp(u1 ξ + u2 ξ 2 + · · · + um ξ m ) := ∑ Y Fn ( u1 , u2 , . . . , u m ; u )
n!
. (7)
n =0

The rest of the article is as follows: The multivariate Hermite–Frobenius–Euler poly-

nomials are introduced and studied in Section 2. Also, operational formulae for these
polynomials are derived. In Section 3, the monomiality principle is verified and the dif-
ferential equation is deduced. Further, several identities satisfied by these multivariate
Hermite–Frobenius–Euler polynomials are established by using operational formalism. In
Section 4, summation formulae and symmetric identities for these polynomials are estab-
lished. Further, several special cases of these polynomials are taken and the corresponding
results are deduced. Section 5 is devoted to some illustrative examples. Finally, Section 6
consists of concluding remarks.

2. Multivariate Hermite–Frobenius–Euler Polynomials

In this section, a novel and comprehensive method is introduced for determining
[m]
the multivariate Hermite–Frobenius–Euler polynomials (MHFEPs) Y Fn (u1 , u2 , . . . , um ; u).
The approach presents an alternative viewpoint and methodology when compared to
existing methods. By employing this innovative technique, our objective is to enrich the
comprehension and investigation of these polynomial sequences, offering a new outlook
on their properties and potential applications. As a result, we have introduced a fresh
perspective to advance the understanding and utilization of these polynomials.
Now, we will use two different approaches to show that the representation series (7) is
meaningful. Thus, MHFEPs are well-defined through the generating function method.

76
Mathematics 2023, 11, 3439

[m]
Theorem 1. The MHFEPs represented by Y Fn (u1 , u2 , · · · , um ; u) satisfy the generating expres-
sion:
  ∞
1−u [m] ξn
exp(u1 ξ + u2 ξ 2 + · · · + um ξ m ) = ∑ Y Fn (u1 , u2 , . . . , um ; u) . (8)
eξ − u n =0
n!

Proof. We prove the result in two alternative ways:

 
(i) Expanding the product of terms e1ξ−−uu and exp(u1 ξ + u2 ξ 2 + · · · + um ξ m ) by New-
 
ton series and ordering the product of the developments of functions e1ξ−−uu and
exp(u1 ξ + u2 ξ 2 + · · · + um ξ m ) w.r.t. the powers of ξ, we obtain the polynomials
[m]
expressed in (7) as coefficients of ξn! .
n
Y F n ( u1 , u2 , . . . , u m ; u )
(ii) Substituting the multiplicative operator M̂ = u1 + 2u2 ∂u1 + 3u3 ∂2u1 + · · · + mum ∂m −1
u1
of MHFEPs given in [17] in expression (5) in place of u1 on both sides, we find
  ∞
1 − u (u1 +2u2 ∂u1 +3u3 ∂2u +···+mum ∂um−1 )ξ ξn
eξ − u
e 1 1 = ∑ Fn (u1 + 2u2 ∂u1 + 3u3 ∂2u1 + · · · + mum ∂mu1−1 ; u) n! (9)
n =0

In view of the identity given in [5], (Equation (7)) gives the l.h.s. of (8) and, denoting
the r.h.s. Y Fn (u1 + 2u2 ∂u1 + 3u3 ∂2u1 + . . . + mum ∂m −1
u1 ; u ) by Y Fn ( u1 , u2 , . . . , um ; u ),
assertion (8) is deduced.

The following result shows that the MHFEPs behave component-wise as Appell-type
polynomial sequences.

[m]
Theorem 2. The multivariate Hermite–Frobenius–Euler polynomials Y F n ( u1 , u2 , . . . , u m ; u )
satisfy the following differential relations:

∂ [m] [m]
[ Fn (u1 , u2 , . . . , um ; u)] = (n) j Y Fn− j (u1 , u2 , . . . , um ; u), 1 ≤ j ≤ m ≤ n, (10)
∂u j Y

where (n) j denotes the falling factorial, given by

⎧
⎪
⎨ 1, if j = 0,
j
(n) j = ∏ i =1 ( n − i + 1), if j ≥ 1,
⎪
⎩
0, if j < 0.

Proof. By taking derivatives of expression (7) w.r.t. u1 , it follows that

     
∂ 1−u 1−u
exp(u1 ξ + u2 ξ + · · · + um ξ ) = ξ
2 m
exp(u1 ξ + u2 ξ + · · · + um ξ ) .
2 m
(11)
∂u1 eξ − u eξ − u

Substituting the r.h.s. of (7) into (11), we find

 
∞ ∞
∂ [m] ξn [m] ξ n +1
∂u1 ∑ Y F n ( u1 , u2 , . . . , u m ; u )
n!
= ∑ Y Fn ( u1 , u2 , . . . , u m ; u )
n!
, (12)
n =0 n =0

By replacing n → n − 1 on the r.h.s. of the previous expression and then equating

the coefficients of like exponents of ξ, the first expression of the system of expressions (10)
is deduced.
Next, on taking derivatives of expression (7) w.r.t. u2 , it follows that
     
∂ 1−u 1−u
exp ( u 1 ξ + u 2 ξ 2
+ · · · + u m ξ m
) = ξ 2
exp ( u 1 ξ + u 2 ξ 2
+ · · · + u m ξ m
) . (13)
∂u2 eξ − u eξ − u

77
Mathematics 2023, 11, 3439

Substituting the r.h.s. of expression (7) into (13), we find

 
∞ ∞
∂ [m] ξn [m] ξ n +2
∂u2 ∑ Y F n ( u1 , u2 , . . . , u m ; u )
n!
= ∑ Y Fn ( u1 , u2 , . . . , u m ; u )
n!
, (14)
n =0 n =0

by replacing n → n − 2 on the r.h.s. of the previous expression and then equating the
coefficients of like exponents of ξ, the second expression of the system of expressions (10)
is deduced.
Similarly, continuing in the same fashion, we deduce other expressions of system (10).

Concerning the operational formalism satisfied by the multivariate polynomials

Y F n ( u1 , u2 , . . . , u m ; u ),
we have the following:

Theorem 3. For MHFEPs Y Fn (u1 , u2 , . . . , um ; u), the operational rule:

 
∂2 ∂3 ∂m  
[m]
exp u2 + u3 + · · · + um F n ( u1 ; u ) = Y F n ( u1 , u2 , . . . , u m ; u ) (15)
∂u1 2 ∂u1 3 ∂u1 m

holds true.

Proof. To prove result (15), we proceed by taking derivatives of expression (7) as:

∂ [m] [m]
[ Fn (u1 , u2 , . . . , um ; u)] = n Y Fn−1 (u1 , u2 , . . . , um ; u),
∂u1 Y
∂2 [m] [m]
[Y Fn (u1 , u2 , . . . , um ; u)] = n(n − 1) Y Fn−2 (u1 , u2 , . . . , um ; u),
∂u1 2
∂3 [m] [m]
[Y Fn (u1 , u2 , . . . , um ; u)] = n(n − 1)(n − 2) Y Fn−3 (u1 , u2 , . . . , um ; u),
∂u1 3
.. ..
. .
∂m [m] [m]
[ Fn (u1 , u2 , . . . , um ; u)] = (n)m Y Fn−m (u1 , u2 , . . . , um ; u), (16)
∂u1 m Y

and

∂ [m] [m]
[ Fn (u1 , u2 , . . . , um ; u)] = n(n − 1) Y Fn−2 (u1 , u2 , . . . , um ; u),
∂u2 Y
∂ [m] [m]
[ Fn (u1 , u2 , . . . , um ; u)] = n(n − 1)(n − 2) Y Fn−3 (u1 , u2 , . . . , um ; u),
∂u3 Y
.. ..
. .
∂ [m] [m]
[ Fn (u1 , u2 , . . . , um ; u)] = (n)m Y Fn−m (u1 , u2 , . . . , um ; u). (17)
∂um Y

In consideration of the system of Equations (16) and (17), we find that the MHFEPs
are solutions of the equations:

78
Mathematics 2023, 11, 3439

∂ [m] ∂2 [m]
[Y Fn (u1 , u2 , . . . , um ; u)] = [Y Fn (u1 , u2 , . . . , um ; u)],
∂u2 ∂u1 2
∂ [m] ∂3 [m]
[Y Fn (u1 , u2 , . . . , um ; u)] = [Y Fn (u1 , u2 , . . . , um ; u)],
∂u3 ∂u1 3
... ..
.
∂ [m] ∂m [m]
[ Fn (u1 , u2 , . . . , um ; u)] = [ Fn (u1 , u2 , . . . , um ; u)], (18)
∂um Y ∂u1 m Y

under the initial conditions:

[m]
Y Fn ( u1 , 0, 0, . . . , 0; u ) = F n ( u1 ; u ). (19)

Therefore, in cognizance of previous expressions (18) and (19), assertion (15) is ob-
tained.

Next, we will obtain the series representation of MHFEPs Y Fn (u1 , u2 , . . . , um ; u) by

proving the succeeding results:

Theorem 4. For MHFEPs Y Fn (u1 , u2 , . . . , um ; u), the succeeding series representations are demon-
strated:
n  
[m] n [m]
Y F n ( u1 , u2 , . . . , u m ; u ) = ∑ Fs (u)Yn−s (u1 , u2 , . . . , um ) (20)
s =0 s

and  
n
[m] n [m]
Y F n ( u1 , u2 , . . . , u m ; u ) = ∑ s
Fs (u1 ; u)Yn−s (u2 , u3 , . . . , um ). (21)
s =0

Proof. Inserting expressions (6) and (2) on the l.h.s. of (7), we find
∞ ∞ ∞
ξs [m] ξn [m] ξn
∑ Fs (u) s! ∑ Yn ( u1 , u2 , . . . , u m )
n!
= ∑ Y Fn ( u1 , u2 , . . . , u m ; u )
n!
. (22)
s =0 n =0 n =0

Interchanging the expressions and replacing n → n − s in the resultant expression in

view of the Cauchy product rule, it follows that
∞ ∞ n
[m] ξn [m] ξn
∑ Y Fn ( u1 , u2 , . . . , u m ; u )
n!
= ∑ ∑ Fs (u)Yn ( u1 , u2 , . . . , u m )
(n − s)! s!
. (23)
n =0 n =0 s =0

Multiplying and dividing by n! on the r.h.s. of the previous expression and then
equating the coefficients of the same exponents of ξ on both sides, assertion (20) is deduced.
In a similar fashion, inserting expressions (5) and (2) (with u1 = 0) on the l.h.s. of (7),
we find
∞ ∞ ∞
ξs [m] ξn [m] ξn
∑ Fs (u1 ; u) s! ∑ Yn ( u2 , u3 , . . . , u m )
n!
= ∑ Y Fn ( u1 , u2 , . . . , u m ; u )
n!
. (24)
s =0 n =0 n =0

Interchanging the expressions and replacing n → n − s in the resultant expression in

view of the Cauchy product rule, it follows that
∞ ∞ n
[m] ξn [m] ξn
∑ Y Fn ( u1 , u2 , . . . , u m ; u )
n!
= ∑ ∑ Fs (u1 ; u)Yn ( u2 , u3 , . . . , u m )
(n − s)! s!
. (25)
n =0 n =0 s =0

Multiplying and dividing by n! on the r.h.s. of the previous expression and then equat-
ing the coefficients of the same exponents of ξ on both sides, assertion (21) is deduced.

79
Mathematics 2023, 11, 3439

3. Monomiality Principle
The development and incorporation of the monomiality principle, operational rules,
and other properties in hybrid special polynomials have been extensively studied. The
concept of monomiality was first introduced by Steffenson in 1941 through the notion
of poweroids [22] and was further refined by Dattoli [2]. In this context, the M̂ and D̂
operators play a crucial role as multiplicative and derivative operators for a polynomial set
bk (u1 )k∈N . These operators satisfy the following expressions:

bk+1 (u1 ) = M̂{bk (u1 )} (26)

and
k bk−1 (u1 ) = D̂{bk (u1 )}. (27)
Subsequently, the polynomial set bk (u1 )m∈N under the manipulation of multiplicative
and derivative operators is known as a quasi-monomial. It is essential for this quasi-
monomial to adhere to the following formula:

[D̂ , M̂] = D̂ M̂ − M̂D̂ = 1̂, (28)

and, as a result, it shows a Weyl group structure.

The significance and usage of the operators M̂ and D̂ can be exploited to extract the
significance of the set {bk (u1 )}k∈N , provided it is quasi-monomial. Hence, the succeeding
axioms hold:
(i) bk (u1 ) gives the differential equation

M̂D̂{bk (u1 )} = k bk (u1 ), (29)

provided M̂ and D̂ exhibit differential traits.

(ii) The expression
bk (u1 ) = M̂k {1}, (30)
gives the explicit form, with b0 (u1 ) = 1.
(iii) Further, the expression
∞
wk
ewM̂ {1} = ∑ bk (u1 ) k! , |w| < ∞ , (31)
k =0

behaves as a generating expression, which is derived by usage of identity (30).

Many branches of mathematical physics, quantum mechanics, and classical optics
still employ these methods today. As a result, these methods offer strong and efficient
research tools. We thus confirm the monomiality concept for MHFEPs by taking into
account the importance of this method. Thus we verify the monomiality principle for
[m]
MHFEPs Y Fn (u1 , u2 , . . . , um ; u) in this section by demonstrating the succeeding results:

[m]
Theorem 5. The MHFEPs Y F n ( u1 , u2 , . . . , u m ; u ) satisfy the succeeding multiplicative and
derivative operators:

−1 e ∂ u1
MˆY F = u1 + 2u2 ∂u1 + 3u3 ∂2u1 + · · · + mum ∂m
u1 − ∂ u1
(32)
e −u
and
DˆY F = ∂u1 , (33)
∂
where ∂u1 = ∂u1 .

Proof. By differentiating expression (7) w.r.t. ξ on both sides, we find

80
Mathematics 2023, 11, 3439

  
eξ 1−u
u1 + 2u2 ξ + 3u3 ξ 2 + · · · + mum ξ m−1 − ξ
exp(u1 ξ + u2 ξ 2 + · · · + um ξ m )
e −u eξ − u
∞
[m] ξ n −1
= ∑ n Y Fn ( u1 , u2 , . . . , u m ; u )
n!
. (34)
n =0

which further can be written as follows:

  ∞
m −1 eξ [m] ξ n −1
u1 + 2u2 ξ + 3u3 ξ + · · · + mum ξ
2
− ξ
e −u
∑ Y Fn ( u1 , u2 , . . . , u m ; u )
n!
n =0
∞
[m] ξ n −1
= ∑ n Y Fn ( u1 , u2 , . . . , u m ; u )
n!
. (35)
n =0

Also, by taking a derivative of (7) w.r.t. u1 , we find the identity

   
∂ 1−u 1−u
exp(u1 ξ + u2 ξ 2 + · · · + um ξ m ) = ξ exp ( u 1 ξ + u 2 ξ 2
+ · · · + u m ξ m
) ,
∂u1 eξ − u eξ − u
 
∞ ∞
∂ [m] ξ n −1 [m] ξ n −1
∑
∂u1 n=0 Y F n ( u1 , u2 , . . . , u m ; u )
n!
= ξ ∑ Y Fn ( u1 , u2 , . . . , u m ; u )
n!
. (36)
n =0

By replacing n → n + 1 on the r.h.s. of (35) and equating the coefficients of same

exponents of ξ in view of expressions (37) and (26) in the resultant expression, asser-
tion (32) is demonstrated.
Moreover, the second part of expression (36) can be written as:
 
∞ n −1 ∞
∂ [m] ξ [m] ξ n +1
∂u1 ∑ Y F n ( u1 , u2 , . . . , u m ; u )
n!
= ∑ Y Fn ( u1 , u2 , . . . , u m ; u )
n!
. (37)
n =0 n =0

By replacing n → n − 1 on the r.h.s. of (37) and equating the coefficients of the same
exponents of ξ in view of (27) in the resultant expression, assertion (33) is demonstrated.
[m]
Next, we deduce the differential equation for MHFEPs Y F n ( u1 , u2 , . . . , u m ; u ) by
demonstrating the succeeding result:

[m]
Theorem 6. The MHFEPs Y Fn (u1 , u2 , . . . , um ; u) satisfy the differential equation:

e ∂ u1 [m]
u1 ∂u1 + 2u2 ∂2u1 + 3u3 ∂3u1 + · · · + mum ∂m
u1 − ∂ u1
∂ u1 − n Y F n ( u1 , u2 , . . . , u m ; u ) = 0. (38)
e −u

Proof. Inserting expression (32) and (33) into the expression (29), assertion (38) is proved.

The operational formalism developed in Theorem 6 can be applied to numerous

identities related to the Frobenius–Euler polynomials, which are widely investigated to
[m]
produce MHFEPs Y Fn (u1 , u2 , . . . , um ; u). To do this, we carry out the subsequent action
2 3
of operator (O) given by exp u2 ∂u∂ + u3 ∂u∂ 3 + · · · + um ∂u∂ 1 m on the identities involving
m
2
1 1
Frobenius–Euler polynomials Fn (u1 ; u) [25]:
n  
n
u F n ( u1 ; u −1 ) + F n ( u1 ; u ) = (1 + u ) ∑ F ( u −1 ) F k ( u1 ; u ), (39)
k =0
k n−k

n −1 (nk) n
1
F ( u ; u ) + F n − k ( u1 ; u ) =
n+1 k 1 ∑ ∑
n − k + 1 l =k
((−u)Fl −k (u)Fn−l (u) + 2uFn−k (u))Fk (u1 ; u)Fn (u1 ; u), (40)
k =0

81
Mathematics 2023, 11, 3439

n  
n
F n ( u1 ; u ) = ∑ F
k n−k
( u ) F k ( u1 ; u ), n ∈ Z+ = N ∪ {0}. (41)
k =0
[m]
The MHFEPs Y Fn (u1 , u2 , . . . , um ; u) are obtained after operating (O) on both sides
of (39)–(41):
n  
[m] [m] n [m] [m]
u Y F n ( u1 , u2 , . . . , u m ; u −1 ; u ) + Y F n ( u1 , u2 , . . . , u m ; u ) = (1 + u ) ∑ F ( u −1 ) Y F k ( u1 , u2 , . . . , u m ; u ),
k =0
k Y n−k

1 [m] [m]
Y F k ( u1 , u2 , u3 , · · · , u m ; u ) + Y F n − k ( u1 , u2 , u3 , . . . , u m ; u )
n+1
n −1 (nk) n
[m] [m]
= ∑ ∑
n − k + 1 l =k
((−u)Fn−l (u)Fl −k (u) + 2uFn−k (u)) Y Fk (u1 , u2 , u3 , . . . , um ; u)Y Fn (u1 , u2 , u3 , . . . , um ; u),
k =0

n  
[m] n [m]
Y F n ( u1 , u2 , u3 , . . . , u m ; u ) = ∑ F
k n−k
( u ) Y F k ( u1 , u2 , u3 , . . . , u m ; u ), n ∈ Z+ = N ∪ {0}.
k =0

4. Summation Formulae and Symmetric Identities

[m]
To derive the summation formulae for the MHFEPs Y Fn (u1 , u2 , u3 , . . . , um ; u), the
succeeding results are demonstrated:

[m]
Theorem 7. For the MHFEPs Y Fn (u1 , u2 , u3 , . . . , um ; u), the succeeding implicit summation
formula holds true:
n  
[m] n [m]
Y Fn ( u1 + w, u2 , u3 , . . . , um ; u ) = ∑ k Y F k ( u1 , u2 , u3 , . . . , u m ; u ) w n − k . (42)
k =0

Proof. On taking u1 → u1 + w in expression (7), it follows that

  ∞
1−u [m] ξn
eξ − u
exp((u1 + w)ξ + u2 ξ 2 + · · · + um ξ m ) = ∑ Y Fn (u1 + w, u2 , . . . , um ; u)
n!
n =0

which further can be written as

  ∞
1−u [m] ξn
eξ − u
exp(u1 ξ + u2 ξ 2 + · · · + um ξ m ) exp(wξ ) = ∑ Y Fn (u1 + w, u2 , . . . , um ; u)
n!
,
n =0

By making use of the series expansion of exp(wξ ) on the l.h.s. of the previous expres-
sion, we have
∞ ∞
[m] ξ n+k [m] ξn
∑ Y Fn (u1 + w, u2 , . . . , um ; u)wn
n!k!
= ∑ Y Fn (u1 + w, u2 , . . . , um ; u)
n!
. (43)
k =0 n =0

This results in the deduction of assertion (42) by substituting n → n − k into the r.h.s.
of consequent expression and then equating the coefficients of the identical powers of ξ in
the resulting equation.

Corollary 1. For w = 1 in expression (42), we have

n  
[m] n [m]
Y F n ( u1 + 1, u2 , u3 , . . . , um ; u) = ∑ F ( u1 , u2 , u3 , . . . , u m ; u ).
k Y k
(44)
k =0

[m]
Theorem 8. For the MHFEPs Y Fn (u1 , u2 , u3 , . . . , um ; u), the succeeding implicit summation
formula holds true:

82
Mathematics 2023, 11, 3439

n  
[m] n [m]
Y F n ( u1 + x, u2 + y, u3 + z, . . . , um ; u) = ∑ F
k Y n−k 1
(u , u2 , u3 , . . . , um ; u)Yk ( x, y, z). (45)
k =0

Proof. On taking u1 → u1 + x, u2 → u1 + y and u3 → u3 + z in expression (7), it follows that

 
1−u
exp((u1 + x )ξ + (u2 + y)ξ 2 + (u3 + z)ξ 3 + · · · + um ξ m ) =
eξ − u
∞
[m] ξn
∑ Y Fn (u1 + x, u2 + y, u3 + z, . . . , um ; u)
n!
(46)
n =0

which further can be written as

 
1−u
exp(u1 ξ + u2 ξ 2 + · · · + um ξ m ) exp( xξ + yξ 2 + zξ 3 ) =
eξ − u
∞
[m] ξn
∑ Y Fn (u1 + x, u2 + y, u3 + z, . . . , um ; u)
n!
. (47)
n =0

By making use of the series expansion of exp( xξ + yξ 2 + zξ 3 ) on the l.h.s. of the

previous expression, we have

∞ ∞
[m] ξ n+k [m] ξn
∑ Y Fn (u1 , u2 , u3 , . . . , um ; u)Yk ( x, y, z)
n!k!
= ∑ Y Fn (u1 + x, u2 + y, u3 + z, . . . , um ; u)
n!
. (48)
n =0 n =0

This results in the deduction of assertion (45) by substituting n → n − k on the l.h.s. of

the consequent expression and then equating the coefficients of the identical powers of xi
in the resulting equation.

Corollary 2. For z = 0 in expression (45), we have

n  
[m] n [m]
Y F n ( u1 + x, u2 + y, u3 , . . . , um ; u) = ∑ F
k Y n−k 1
(u , u2 , u3 , . . . , um ; u)Yk ( x, y). (49)
k =0

[m]
Theorem 9. For the MHFEPs Y Fn (u1 , u2 , u3 , . . . , um ; u), the succeeding implicit summation
formula holds true:

n,s   
[m] n s [m]
Y Fn+s ( q, u2 , u3 , . . . , um ; u ) = ∑ ( q − u1 ) l + m Y F n + s − l − m ( u1 , u2 , u3 , . . . , u m ; u ). (50)
l,m=0
l m

Proof. By replacing ξ → ξ + η and in view of the expression:

∞ ∞
( u1 + u2 ) M ul um
∑ g( M)
M!
= ∑ g(l + m) 1 2
l! m!
(51)
M =0 l,m=0

in relation (7) and afterward simplifying the resultant expression, we have

∞  
[m] ξ n ηs 1−u
e − u1 ( ξ + η ) ∑ Y F n + s ( u 1 , u 2 , u 3 , . . . , u m ; u ) = ξ + η
exp(u2 (ξ + η )2 + · · · + um (ξ + η )m ). (52)
n,s=0 n! s! e −u
Substituting u1 → q into (52) and comparing the resultant expression to the previous
expression and further expanding the exponential function gives
∞ ∞ ∞
[m] ξn ηs (ξ + η ) M [m] ξ
n ηs
∑ Y F n + s ( u1 , u2 , u3 , . . . , u m ; u )
n! s!
= ∑ ( q − u1 ) M M!
× ∑ Y F n + s ( u1 , u2 , u3 , . . . , u m ; u )
n! s!
. (53)
n,s=0 M =0 n,s=0

83
Mathematics 2023, 11, 3439

Thus, in view of expression (51) in expression (53) and then replacing n → n − l and
s → s − m in the resultant expression, we find
∞ ∞
[m] ξn ηs ( q − u1 ) l + m
n,s
[m] ξ n ηs
∑ Y F n + s ( u1 , u2 , u3 , . . . , u m ; u )
n! s!
= ∑ ∑ l! m!
× Y F n + s − l − m ( u1 , u2 , u3 , . . . , u m ; u )
(n − l )! (s − m)!
. (54)
n,s=0 n,s=0 l,m=0

On comparison of the coefficients of the like exponents of ξ and η on both sides of the
previous expression, assertion (50) is established.

Corollary 3. For n = 0 in expression (50), we find

s  
[m] s [m]
Y Fs ( q, u2 , u3 , . . . , um ; u ) = ∑ m
( q − u1 ) m Y F s − m ( u1 , u2 , u3 , . . . , u m ; u )
m =0

Corollary 4. Substituting q → q + u1 and taking m = 2 in expression (50), we have

n,s   
[m] n s [m]
Y F n + s ( q + u1 , u2 , u3 , . . . , u m ; u ) = ∑ ( q ) l + m Y F n + s − l − m ( u1 , u2 , u3 , . . . , u m ; u ).
l,m=0
l m

Corollary 5. Substituting q → q + u1 and taking m = 1 in expression (50), we have

n,s   
[m] n s [m]
Y F n + s ( q + u1 ; u ) = ∑ ( q ) l + m Y F n + s − l − m ( u1 ; u ).
l,m=0
l m

Corollary 6. Substituting q = 0 in expression (50), we have

n,s   
[m] n s [m]
Y F n + s ( u2 , u3 , · · · , um ; u) = ∑ (−u1 )l +m Y Fn+s−l −m (u1 , u2 , u3 , . . . , um ; u).
l,m=0
l m

In physics and applied mathematics, it is common to encounter problems where

finding a solution requires evaluating infinite sums that involve special functions. The
applications of generalized special functions can be found in various fields, including
electromagnetics and combinatorics. Several authors [23–34] established and examined
different types of identities related to Apostol-type polynomials. These investigations serve
as a motivation to establish symmetry identities for the MHFEPs. Let us now review the
following definitions:

Definition 1. The generalized sum of integer powers Sk (n) is defined by the generating function
shown below for:
∞
ξj e ( n +1) ξ − 1
∑ Sj (n) j! = eξ − 1 . (55)
j =0

(l )
Definition 2. The multiple power sums Sk (m) are defined by the generating function shown below:
 
∞ n    l
n n−q (l ) ξn 1 − emξ
∑ ∑ q (−l ) Sk (m) n! = 1 − eξ . (56)
n =0 q =0

[m]
In order to derive the symmetry identities for the MHFEPs Y Fn (u1 , u2 , u3 , . . . , um ; u),
we prove the following results:

84
Mathematics 2023, 11, 3439

Theorem 10. The following symmetry connection between the MHFEPs and generalized integer
power sums is valid for any integers with μ, η > 0 and n ≥ 0, u ∈ C:

n   k  
n n−k [m] k 1
∑ k μ F
Y n−k ( ηu 1 , η 2
u 2 , η 3
u 3 , . . . , η m
u m ; u ) ∑ l η k Sl (μ − 1; u )
k =0 l =0
[m]
× Y Fk−1 (μU1 , μ2 U2 , μ3 U3 , . . . , μm Um ; u)
n   k  
n n−k [m] k 1
= ∑ η F
Y n−k ( μu 1 , μ 2
u 2 , μ 3
u 3 , . . . , μ m
u m ; u ) ∑ l μk Sl (η − 1; u )
k =0
k l =0
[m]
× Y Fk−1 ( ηU1 , η U2 , η U3 , . . . , η Um ; u ).
2 3 m
(57)

Proof. Consider

(1 − u) eμu1 ηξ +u2 (μηξ )

2 +u
3 ( μηξ )
(eμηξ − u) eμηξU1 +U2 (μηξ )
3 2 +U ( μηξ )3
3
G( ξ ) : = , (58)
(eμξ − u) (eηξ − u)

which in consideration of the Cauchy product rule becomes

∞  n   k  
n n−k [m] k k 1
G( ξ ) = ∑ ∑ k
μ Y Fn−k ( ηu1 , η u2 , η u3 , . . . , η um ; u ) ∑
2 3 m
l
η Sl (μ − 1; )
u
n =0 k =0 l =0
[m] ξn
× Y Fk−1 (μU1 , μ2 U2 , μ3 U3 , . . . , μm Um ; u) . (59)
n!
Continuing in a similar fashion, we find

∞  n   k  
n n−k [m] k 1
G( ξ ) = ∑ ∑ k η F
Y n−k ( μu 1 , μ 2
u 2 , μ 3
u 3 , . . . , μ m
u m ; u ) ∑ l μk Sl (η − 1; u )
n =0 k =0 l =0
[m] ξn
× Y Fk−1 (ηU1 , η 2 U2 , η 3 U3 , . . . , η m Um ; u) . (60)
n!
On comparison of the coefficients of like exponents of ξ in expressions (59) and (60),
assertion (57) is deduced.

Theorem 11. The following symmetry connection for the MHFEPs is valid for any integers with
μ, η > 0 and n ≥ 0, u ∈ C:
n   μ −1 η −1
n 1 i+ j [m] μ [m] η
∑ k ∑ ∑ uμ+η −2 u μn−k η k Y Fk (μU1 + η j, μ2 U2 , μ3 U3 , . . . , μm Um ; u) × Y Fn−k (ηu1 + μ i, η 2 u2 , η 3 u3 , . . . , η m um ; u)
k =0 i =0 j =0
n   η −1 μ −1
n 1 i+ j [m] η
= ∑ k ∑ ∑ uμ+η −2 u η n−k μk Y Fk (ηU1 + μ j, μ2 U2 , μ3 U3 , . . . , μm Um ; u)
k =0 i =0 j =0
[m] μ 2
× Y Fn−k (μu1 + i, η u2 , η 3 u3 , . . . , η m um ; u). (61)
η

Proof. Consider

(eμηξ − uμ ) (eμηξ − uη )eμηξU1 +U2 (μηξ ) +U3 (μηξ ) +···+Um (μηξ )m

2 3

(1 − u)2 eμηξu1 +u2 (μηξ ) +u3 (μηξ )3 +···+um (μηξ )m

2
H( ξ ) := × , (62)
(eμξ − u) (eηξ − u)

(eμηξ −uμ ) (eμηξ −uη )

which in consideration of series representations of (eηξ −u)
and (eμξ −u)
in final expression
gives

85
Mathematics 2023, 11, 3439

 1−u  μ −1   i
1 ηξi
eηu1 (μξ )+η u2 (μξ ) +η u3 (μξ ) +···+η um (μξ ) uμ−1 ∑
2 2 3 3 m m
H( ξ ) = μξ
e
e −u i =0
u
 1−u  η −1   j
1
eμU1 (ηξ )+μ U2 (ηξ ) +μ U3 (ηξ ) +···+μ Um (ηξ ) uη −1 ∑
2 2 3 3
eμξ j .
m m
× ηξ
(63)
e −u j =0
u

Thus, in view of (7) and the usage of the Cauchy product rule in the previous expres-
sion (63), we find
   μ −1 η −1
∞ n
n 1 i+ j n−k k [m] μ 2
H( ξ ) : = ∑ ∑ k i∑ ∑ u μ + η −2 u μ η Y Fk ( μU1 +
η
j, μ U2 , μ3 U3 , . . . , μm Um ; u)
n =0 k =0 =0 j =0

[m] η 2
× Y Fn−k (ηu1 + i, η u2 , η 3 u3 , . . . , η m um ; u) . (64)
μ

Continuing in a similar fashion, we find another identity

   η −1 μ −1
∞ n
n 1 i+ j n−k k [m] η
H( ξ ) : = ∑ ∑ k ∑ ∑ u μ + η −2 u η μ Y Fk ( ηU1 +
μ
j, μ2 U2 , μ3 U3 , · · · , μm Um ; u)
n =0 k =0 i =0 j =0

[m] μ 2
× Y Fn−k (μu1 + i, η u2 , η u3 , . . . , η um ; u) .
3 m
(65)
η

On comparison of the coefficients of like exponents of ξ in expressions (64) and (65),

assertion (61) is deduced.

Theorem 12. The following symmetry connection for the MHFEPs is valid for any integers with
μ, η > 0 and n ≥ 0, u ∈ C:

η −1 n  
1 n [m]
∑ u η −1 F (μu1 , μ2 u2 , μ3 u3 , . . . , μm um ; u)η n−i (μk)i
k
u ∑ i Y n −i
k =0 i =0
μ −1  
1 k n n [m]
= ∑ u μ −1 ∑ F (ηu1 , η 2 u2 , η 3 u3 , . . . , η m um ; u)μn−i (ηk)i . (66)
k =0
u i =0 i Y n − i

Proof. Consider

(1 − u) eμηξu1 +u2 (μηξ ) +u3 (μηξ ) +···+um (μηξ )

2 3 m

N( ξ ) : = .
(eμηξ − uη ) (eμξ − u) (eηξ − u)

By continuing in a similar fashion to that performed in Theorem 11, assertion (4) is

deduced.

Theorem 13. The following symmetry connection between the MHFEPs and multiple power sums
is valid for any integers with μ, η > 0 and n ≥ 0, u ∈ C:

86
Mathematics 2023, 11, 3439

n   k   l  
n [m] k l 1
∑ η
Y Fn−k ( ηu1 , η u2 , η u3 , . . . , η um ; u ) u ∑ (−1)l −r Sk (η; )
l r∑
2 3 m
k =0
k l =0 =0 r u
[ m +1]
× Y Fk−l (μU1 , μ2 U2 , μ3 U3 , . . . , μm Um ; u)μn−k+l η k−l
n   k   l  
n [m] k l 1
= ∑ μ
Y Fn−k ( μu1 , μ u2 , μ u3 , . . . , μ um ; u ) u ∑ (−1)l −r Sk (μ; )
l r∑
2 3 m
k =0
k l =0 =0 r u
[ m +1]
× Y Fk−l (ηU1 , η 2 U2 , η 3 U3 , . . . , η m Um ; u)η n−k+l μk−l . (67)

Proof. Consider

F(ξ ) := (1 − u)2 eμu1 (ηξ )+μ 2 ( ηξ )

2 + μ3 u
3 ( ηξ )
3 +···+ μm u
m ( ηξ )
2u m

(eμηξ − uη ) eμU1 (ηξ )+μ U2 (ηξ ) +μ U3 (ηξ )

2 2 3 3 +···+ μm U ( ηξ )m
m
× , (68)
(eηξ − u) (eμξ − u)

which on simplifying the exponents and usage of expressions (7) and (56) in the final
expression gives
∞ ∞ m  
[m] ξn m 1 ξm
F(ξ ) := ∑ Y Fn (ηu1 , η 2 u2 , η 3 u3 , lcdots, η m um ; u)μn uη ∑ ∑ (−1)m−r Sk (η; )μm
n =0
n! m=0 r=0 r u m!
[ m +1] ξl
× Y Fk−l (μU1 , μ2 U2 , μ3 U3 , . . . , μm Um ; u)η l . (69)
l!
Therefore, in view of the Cauchy product rule, we have
  
∞ n l   m  
n [m] n−l η l m 1
F( ξ ) : = ∑ ∑ l Y Fn−l ( ηu1 , η u2 , η u3 , . . . , η um ; u ) μ
2 3 m
u ∑
m ∑ r (−1)m−r Sk (η; u )
n =0 l =0 m =0 r =0

[ m +1] ξn
× Y Fl −m (μU1 , μ2 U2 , μ3 U3 , . . . , μm Um ; u)μm η l −m . (70)
n!

Continuing in a similar fashion, we have

  
∞ n l   m  
n [m] n−l μ l m 1
∑ ∑ u ∑ (−1)m−r Sk (μ; )
m r∑
F( ξ ) : = Y Fn−l ( μu1 , μ u2 , μ u3 , . . . , μ um ; u ) η
2 3 m
n =0 l =0
l m =0 =0 r u

[ m +1] ξn
× Y Fl −m (ηU1 , η 2 U2 , η 3 U3 , . . . , η m Um ; u)μm μl −m . (71)
n!

On comparison of the coefficients of like exponents of ξ in expressions (70) and (71),

assertion (67) is deduced.

Theorem 14. The following symmetry connection between the MHFEPs and generalized integer
power sums is valid for any integers with μ, η > 0 and n ≥ 0, u ∈ C:
n   m  
n [m] n−m μ m 1
∑ 2 3
Y Fn−m ( ηu1 , η u2 , η u3 , . . . , η um ; u ) μ
m
u ∑ (−1)m−r Sk (μ; )η m
m =0
m r =0
r u
n   m  
n [m] n−m η m 1
= ∑ 2 3
Y Fn−m ( μu1 , μ u2 , μ u3 , . . . , μ um ; u ) η
m
u ∑ (−1)m−r Sk (η; )μm . (72)
m =0
k r =0
r u

87
Mathematics 2023, 11, 3439

Proof. Consider

(1 − u) eηu1 (μξ )+η 2 ( μξ )

2 +η 3 u
3 ( μξ )
3 +···+ η m u
m ( μξ )
2u
(eμηξ − uμ )
m

M( ξ ) : = .
(eηξ − u) (eμξ − u)
(73)

By continuing in a similar fashion to that performed in the previous Theorem, assertion

(72) is deduced.

5. Some Illustrative Examples

Here, we give some specific examples of MHFEPs by taking their special cases:
[3]
For m = 3, the MHFEPs reduce to three-variable HFEPs Y Fn (u1 , u2 , u3 ; u) specified
by the generating expression:
  ∞
1−u [3] ξn
eξ − u
exp(u1 ξ + u2 ξ 2 + u3 ξ 3 ) = ∑ Y Fn ( u1 , u2 , u3 ; u )
n!
, (74)
n =0

operational rule:
 
∂2 ∂3  
[3]
exp u2 + u 3 F n ( u1 ; u ) = Y F n ( u1 , u2 , u3 ; u ), (75)
∂u1 2 ∂u1 3

series representations:
n  
[3] n [3]
Y F n ( u1 , u2 , u3 ; u ) = ∑ s Fs (u)Yn−s (u1 , u2 , u3 ) (76)
s =0

and  
n
[3] n [3]
Y F n ( u1 , u2 , u3 ; u ) = ∑ s
Fs (u1 ; u)Yn−s (u2 , u3 ). (77)
s =0

For m = 2, the MHFEPs reduce to two-variable HFEPs Y Fn (u1 , u2 , u3 ; u) specified by

the generating expression:
  ∞
1−u ξn
eξ − u
exp(u1 ξ + u2 ξ 2 ) = ∑ Y Fn (u1 , u2 ; u) n! , (78)
n =0

operational rule:
 
∂2  
exp u2 F n ( u 1 ; u ) = Y F n ( u1 , u2 ; u ), (79)
∂u1 2
series representations:
n  
n
Y F n ( u1 , u2 ; u ) = ∑ Fs (u)Yn−s (u1 , u2 ) (80)
s =0 s

and  
n
n
Y F n ( u1 , u2 ; u ) = ∑ Fs (u1 ; u)Yn−s (u2 ). (81)
s =0 s

For m = 1, they reduce to Frobenius–Euler polynomials.

6. Conclusions
We develop the generation function and recurrence rules for the multivariate Hermite-
type Frobenius–Euler polynomials in this context. We may investigate the polynomials’
characteristics and potential applications to physics and related fields using this approach.
The generating function is derived and gives a compact representation of the polynomials,

88
Mathematics 2023, 11, 3439

which makes it simpler to analyze their algebraic and analytical properties. The recurrence
relations also enable rapid computation and analysis of polynomial values through the use
of recursive computing.
The multivariate Hermite-type Frobenius–Euler polynomials offer a strong foundation
for further research. They provide opportunities to explore several algebraic and analytical
characteristics, including differential equations, orthogonality, and others. Quantum me-
chanics, statistical physics, mathematical physics, engineering, and other areas of physics
all make use of these polynomials. By developing the generating function and recurrence
relations of extended hybrid-type polynomials, this technique is reinforced. These discov-
eries not only add to our understanding of multivariate Hermite-type Frobenius–Euler
polynomials but also open up new avenues for investigation into their characteristics and
potential applications in physics and related fields.
Operational techniques are effective in constructing new families of special functions
and deriving features related to both common and generalized special functions. By em-
ploying these techniques, explicit solutions for families of partial differential equations,
including those of the Heat and D’Alembert type, can be obtained. The approach de-
scribed in this article, in conjunction with the monomiality principle, enables the analysis
of solutions for a wide range of physical problems involving various types of partial
differential equations.

Author Contributions: Conceptualization, M.Z., Y.Q., and S.A.W.; Data curation, Y.Q. and M.Z.;
Formal analysis, Y.Q.; Funding acquisition, M.Z. and S.A.W.; Investigation, M.Z., Y.Q., and S.A.W.;
Methodology, S.A.W.; Project administration, M.Z.; Resources, M.Z.; Software, S.A.W.; Supervision,
Y.Q. and S.A.W.; Validation, M.Z. and Y.Q.; Visualization, M.Z.; Writing—original draft, S.A.W., Y.Q.,
and M.Z.; Writing—review and editing, Y.Q. All authors have read and agreed to the published
version of the manuscript.
Funding: This research work was funded by the Deanship of Scientific Research at King Khalid
University through a large group Research Project under grant number RGP2/237/44.
Data Availability Statement: Data sharing is not applicable to this article.
Acknowledgments: The authors sincerely thank the reviewers for their careful review of our
manuscript and valuable comments and suggestions, which have improved the paper presentation.
M. Zayed extends her appreciation to the Deanship of Scientific Research at King Khalid University
for funding this work through a large group Research Project under grant number RGP2/237/44.
Conflicts of Interest: The authors declare no conflict of interest.

References
1. Dattoti, G.; Lorenzutta, S.; Cesarano, C. Bernstein polynomials and operational methods. J. Comput. Anal. Appl. 2006, 8, 369–377.
2. Dattoli, G. Hermite-Bessel and Laguerre-Bessel functions: A by-product of the monomiality principle, Advanced Special Functions
and Applications. Adv. Spec. Funct. Appl. 1999, 1, 147–164.
3. Nahid, T.; Choi, J. Certain hybrid matrix polynomials related to the Laguerre-Sheffer family. Fractal Fract. 2022, 6, 211. [CrossRef]
4. Wani, S.A.; Abuasbeh, K.; Oros, G.I.; Trabelsi, S. Studies on special polynomials involving degenerate Appell polynomials and
fractional derivative. Symmetry 2023, 15, 840. [CrossRef]
5. Alyusof, R.; Wani, S.A. Certain properties and applications of Δh hybrid special polynomials associated with Appell sequences.
Fractal Fract. 2023, 7, 233. [CrossRef]
6. Srivastava, H.M.; Yasmin, G.; Muhyi, A.; Araci, S. Certain results for the twice-iterated 2D q-Appell polynomials. Symmetry 2019,
11, 1307. [CrossRef]
7. Obad, A.M.; Khan, A.; Nisar, K.S.; Morsy, A. q-Binomial convolution and transformations of q-Appell polynomials. Axioms 2021,
10, 70. [CrossRef]
8. Dattoli, G. Generalized polynomials operational identities and their applications. J. Comput. Appl. Math. 2000, 118, 111–123.
[CrossRef]
9. Appell, P.; Kampé de Fériet, J. Fonctions Hypergéométriques et Hypersphériques: Polynômes d’ Hermite; Gauthier-Villars: Paris,
France, 1926.
10. Andrews, L.C. Special Functions for Engineers and Applied Mathematicians; Macmillan Publishing Company: New York, NY,
USA, 1985.
11. Clarkson P.A.; Jordaan, K. Properties of generalized Freud polynomials. J. Approx. Theory 2018, 225, 148–175. [CrossRef]

89
Mathematics 2023, 11, 3439

12. Min, C.; Chen, Y. Painlevé IV, Chazy II, and asymptotics for recurrence coefficients of semi-classical Laguerre polynomials and
their Hankel determinants. Math. Methods Appl. Sci. 2023. [CrossRef]
13. Van Assche, W. Orthogonal Polynomials and Painlevé Equations; Australian Mathematical Society Lecture Series 27; Cambridge
University Press: Cambridge, UK, 2018.
14. Luzón, A.; Morón, M.A. Recurrence relations for polynomial sequences via Riordan matrices. Linear Algebra Appl. 2010, 433,
1422–1446. [CrossRef].
15. Leinartas, E.K.; Shishkina, O.A. The discrete analog of the Newton-Leibniz formula in the problem of summation over simplex
lattice points. J. Sib. Fed. Univ. Math. Phys. 2019, 12, 503–508. [CrossRef]
16. Grigoriev, A.A.; Leinartas, E.K.; Lyapin, A.P. Summation of functions and polynomial solutions to a multidimensional difference
equation. J. Sib. Fed. Univ. Math. Phys. 2023, 16, 153–161.
17. Dattoli. G. Summation formulae of special functions and multivariable Hermite polynomials. Nuovo Cimento Soc. Ital. Fis. 2004,
119, 479–488. [CrossRef]
18. Özarslan, M.A. Unified Apostol-Bernoulli, Euler and Genocchi polynomials. Comput. Math. Appl. 2011, 62, 2452–2462. [CrossRef]
19. Luo, Q.M. Apostol-Euler polynomials of higher order and the Gaussian hypergeometric function. Taiwanese J. Math. 2006, 10,
917–925. [CrossRef]
20. Carlitz, L. Eulerian numbers and polynomials. Math. Mag. 1959, 32, 247–260. [CrossRef]
21. Erdélyi, A.; Magnus, W.; Oberhettinger, F.; Tricomi, F. Higher Transcendental Functions; McGraw Hill: New York, NY, USA, 1953;
Volume 1–3. [CrossRef]
22. Steffensen, J.F. The poweriod, an extension of the mathematical notion of power. Acta Math. 1941, 73, 333–366. [CrossRef]
23. Kurt, B.; Simsek, Y. Frobenius-Euler type polynomials related to Hermite-Bernoulli polyomials. In Proceedings of the Numer-
ical Analysis and Applied Mathematics ICNAAM 2011: International Conference, Halkidiki, Greece, 19–25 September 2011;
Volume 1389, pp. 385–388. [CrossRef]
24. Simsek, Y. Generating functions for q-Apostol type Frobenius-Euler numbers and polynomials. Axioms 2012, 1, 395–403. [CrossRef]
25. Kim, D.S.; Kim, T. Some new identities of Frobenius-Euler numbers and polynomials. J. Inequal. Appl. 2012, 2012, 307. [CrossRef]
26. Yang, S.L. An identity of symmetry for the Bernoulli polynomials. Discrete Math. 2008, 308, 550–554. [CrossRef]
27. Zhang, Z.; Yang, H. Several identities for the generalized Apostol-Bernoulli polynomials. Comput. Math. Appl. 2008, 56, 2993–2999.
[CrossRef].
28. Kurt, V. Some symmetry identities for the Apostol-type polynomials related to multiple alternating sums. Adv. Differ. Equ. 2013,
2013, 32. [CrossRef]
29. Kim, T. Identities involving Frobenius-Euler polynomials arising from non-linear differential equations. J. Number Theory 2012,
132, 2854–2865. [CrossRef]
30. Kim, T.; Lee, B. Some identities of the Frobenius-Euler polynomials. Abstr. Appl. Anal. 2009, 2009, 639439. [CrossRef]
31. Kim, T.; Seo, J.J. Some identities involving Frobenius-Euler polynomials and numbers. Proc. Jangjeon Math. Soc. 2016, 19, 39–46.
[CrossRef]
32. Bayad, A.; Kim, T. Identities for Apostol-type Frobenius-Euler polynomials resulting from the study of a nonlinear operator. Russ.
J. Math. Phys. 2016, 23, 164–171. [CrossRef]
33. Kim, T. An identity of the symmetry for the Frobenius-Euler polynomials associated with the fermionic p-adic invariant q-integrals
on Z p . Rocky Mountain J. Math. 2011, 41, 239–247. [CrossRef]
34. Kim, T.; Lee, B.-J.; Lee, S.H.; Rim, S.H. Some identities for the Frobenius-Euler numbers and polynomials. J. Comput. Anal.Appl.
2013, 15, 544–551. [CrossRef]

Disclaimer/Publisher’s Note: The statements, opinions and data contained in all publications are solely those of the individual
author(s) and contributor(s) and not of MDPI and/or the editor(s). MDPI and/or the editor(s) disclaim responsibility for any injury to
people or property resulting from any ideas, methods, instructions or products referred to in the content.

90
 mathematics

Article
Differential Properties of Jacobi-Sobolev Polynomials and
Electrostatic Interpretation
Héctor Pijeira-Cabrera 1, *, Javier Quintero-Roba 1 and Juan Toribio-Milane 2

1 Departamento de Matemáticas, Universidad Carlos III de Madrid, 28911 Madrid, Spain;

[email protected]
2 Instituto de Matemáticas, Facultad de Ciencias, Universidad Autónoma de Santo Domingo,
Santo Domingo 10105, Dominican Republic; [email protected]
* Correspondence: [email protected]

Abstract: We study the sequence of monic polynomials {Sn }n0 , orthogonal with respect to the Jacobi-
)1 dj (k) ( c ) g(k) ( c ), where
Sobolev inner product f , g s = −1 f ( x ) g( x ) dμα,β ( x ) + ∑ N j=1 ∑k =0 λ j,k f j j
N, d j ∈ Z+ , λ j,k  0, dμα,β ( x ) = (1 − x )α (1 + x ) β dx, α, β > −1, and c j ∈ R \ (−1, 1). A con-
nection formula that relates the Sobolev polynomials Sn with the Jacobi polynomials is provided, as
well as the ladder differential operators for the sequence {Sn }n0 and a second-order differential
equation with a polynomial coefficient that they satisfied. We give sufficient conditions under which
the zeros of a wide class of Jacobi-Sobolev polynomials can be interpreted as the solution of an
electrostatic equilibrium problem of n unit charges moving in the presence of a logarithmic potential.
Several examples are presented to illustrate this interpretation.

Keywords: Jacobi polynomials; Sobolev orthogonality; second-order differential equation; electro-

static model

MSC: 30C15; 42C05; 33C45; 33C47; 82B23

Citation: Pijeira-Cabrera, H.;

Quintero-Roba, J.; Toribio-Milane, J. 1. Introduction
Differential Properties of It is well known that the classical orthogonal polynomials (i.e., Jacobi, Laguerre, and
Jacobi-Sobolev Polynomials and Hermite) satisfy a second-order differential equation with polynomial coefficients, and its
Electrostatic Interpretation.
zeros are simple. Based on these facts, Stieltjes gave a very interesting interpretation of the
Mathematics 2023, 11, 3420. https://
zeros of the classical orthogonal polynomials as a solution of an electrostatic equilibrium
doi.org/10.3390/math11153420
problem of n movable unit charges in the presence of a logarithmic potential (see [1] Sec. 3).
Academic Editor: Carsten An excellent introduction to Stieltjes’ result on this subject and its consequences can be
Schneider found in ([1] Sec. 3) and ([2] Sec. 2). See also the survey [3] and the introduction of [4,5].
In order to make this paper self-contained, it is convenient to briefly recall the Jacobi,
Received: 6 July 2023
Laguerre, and Hermite cases. We begin with Jacobi. Let us consider n unit charges at
Revised: 29 July 2023
Accepted: 4 August 2023
the points x1 , x2 , . . . , xn distributed in [−1, 1] and add two positive fixed charges of mass
Published: 6 August 2023
(α + 1)/2 and ( β + 1)/2 at 1 and −1, respectively. If the charges repel each other according
to the logarithmic potential law (i.e., the force is inversely proportional to the relative
distance), then the total energy E(·) of this system is obtained by adding the energy of the
mutual interaction between the charges. This is
Copyright: © 2023 by the authors.
Licensee MDPI, Basel, Switzerland. 1
This article is an open access article E ( ω1 , ω2 , . . . , ω n ) = ∑ log  
ωi − ω j 
1i < jn
distributed under the terms and
conditions of the Creative Commons α+1 n
1 β+1 n
1
Attribution (CC BY) license (https://
+
2 ∑ log 1 − ω  + 2 ∑ log 1 + ω  . (1)
j =1 j j =1 j
creativecommons.org/licenses/by/
4.0/).

Mathematics 2023, 11, 3420. https://doi.org/10.3390/math11153420 91 https://www.mdpi.com/journal/mathematics

Mathematics 2023, 11, 3420

The minimum of (1) gives the electrostatic equilibrium. The points x1 , x2 , . . . , xn where
the minimum is obtained are the places where the charges will settle down. It is obvious
that, for the minimum, all the x j are distinct and different from ±1.
∂Et
For a minimum, it is necessary that ∂ω j = 0 (1  k  n), from which it follows that the
polynomial Pn ( x ) = ∏nj=1 ( x − x j ) satisfies the differential equation
 
1 − x2 Pn ( x ) + ( β − α − (α + β + 2) x ) Pn ( x ) = −n(n + α + β + 1) Pn ( x ), (2)

α,β
which is the differential equation for the monic Jacobi polynomial Pn ( x ) = Pn ( x ) (see [6]
(Theorems 4.2.2 and 4.21.6)). The proof of the uniqueness of the minimum, based on the
inequality between the arithmetic and geometric means, can be found in [6] (Section 6.7).
In conclusion, the global minimum of (1) is reached when each of the n charges is located
α,β
on a zero of the nth Jacobi polynomial Pn ( x ).
For the other two families of classical orthogonal polynomials on the real line (i.e.,
Laguerre and Hermite), Stieltjes also gave an electrostatic interpretation. Since, in this
situation, the free charges move in an unbounded set, they can escape to infinity. Stieltjes
avoided this situation by constraining the first (Laguerre) or second (Hermite) moment of
his zero-counting measures (see [6] (Theorems 6.7.2 and 6.7.3) and [1] (Section 3.2)).
The electrostatic interpretation of the zeros of the classical orthogonal polynomials,
in addition to Stieltjes, was also studied by Bôcher, Heine, and Van Vleck, among others.
These works were developed between the end of the 19th century and the beginning of
the 20th century. After that, the subject remained dormant for almost a century, until it
received new impulses from advances in logarithmic potential theory, the extensions of the
notion of orthogonality, and the study of new classes of special functions.
Let μ be a finite positive Borel measure with finite moments whose support supp(μ) ⊂ R
contains an infinite set of points. Assume that { Pn }n0 denotes the monic orthogonal
polynomial sequence with respect to the inner product

f, g μ = f ( x ) g( x )dμ( x ). (3)

In general, an inner product is referred to as “standard” when the multiplication operator

exhibits symmetry with respect to the inner product, i.e., x f , g μ = f , xg μ . As (3) is a stan-
dard inner product, we have that Pn has exactly n simple zeros on ( a, b) = Ch(supp(μ))◦ ⊂ R,
where Ch( A) denotes the convex hull of a real set A and A◦ denotes the interior set of A.
Furthermore, the sequence { Pn }n0 satisfies the three-term recurrence relation

xPn ( x ) = Pn+1 ( x ) + γ1,n Pn ( x ) + γ2,n Pn−1 ( x ); P0 ( x ) = 1, P−1 ( x ) = 0,


where γ2,n =  Pn μ 2 / Pn−1 μ 2 for n  1, γ1,n = Pn , xPn μ /  Pn μ ,
2 and  · μ = ·, · μ
denotes the norm induced by (3). See [6–8] for these and other properties of { Pn }n0 .
Let ( a, b) be as above, N, d j ∈ Z+ , λ j,k  0, for j = 1, . . . , N, k = 0, 1, . . . , d j ,
{c1 , c2 , . . . , c N } ⊂ R \( a, b), where ci = c j if i = j and I+ = {( j, k) : λ j,k > 0}. We
consider the following Sobolev-type inner product:

N dj
f, g s = f, g μ +∑ ∑ λ j,k f (k) (c j ) g(k) (c j )
j =1 k =0

= f ( x ) g( x )dμ( x ) + ∑ λ j,k f (k) (c j ) g(k) (c j ), (4)
( j,k)∈ I+

92
Mathematics 2023, 11, 3420

where f (k) denotes the kth derivative of the function f . We also assume, without restriction
of generality, that {( j, d j )} N
j=1 ⊂ I+ and d1  d2  · · ·  d N . Let us denote by Sn ( n ∈ Z+ )
the lowest degree monic polynomial that satisfies

x k , Sn s = 0, for k = 0, 1, . . . , n − 1. (5)

Henceforth, we refer to the sequence {Sn }n0 of monic polynomials as the system
of monic Sobolev-type orthogonal polynomials. It is not difficult to see that for all n  0,
there exists a unique polynomial Sn of the degree n. Note that the coefficients of Sn are
the solution of a homogeneous linear system (5) of n + 1 unknowns and n equations. The
uniqueness is a consequence of the required minimality on the degree. For more details on
this type of nonstandard orthogonality, we refer the reader to [9,10].
It is not difficult to see that, in general, (4) is nonstandard, i.e., xp, q s = p, xq s . The
properties of orthogonal polynomials concerning standard inner products are distinct from
those of Sobolev-type polynomials. For instance, the roots of Sobolev-type polynomials
either can be complex or, if real, might lie beyond the convex hull of the measure μ support,
as demonstrated in the following example:

Example 1. Let
 1
f, g s = f ( x ) g( x )dx + f  (−2) g (−2) + f  (2) g (2),
−1

then the corresponding third-degree

 monic Sobolev-type
 orthogonal polynomial is S3 (z) = z3 −
183
20 z, whose zeros are 0 and ± 183
20 . Note that ± 183
20 ≈ ±3 ∈ [−2, 2].

We will denote by P the linear space of all polynomials and by dgr( p) the degree of
p ∈ P. Let

d j +1 d j +1
ρ*( x ) = ∏ x − cj ∏ cj − x and dμρ*( x ) = ρ*( x )dμ( x ).
c j a c j b

Note that ρ*( x ) > 0 for all x ∈ ( a, b) and dgr(ρ*) = d = ∑ N

j=1 ( d j + 1). Additionally, for
n > d, from (5), we have that {Sn } satisfies the following quasi-orthogonality relations:

Sn , f μρ* = Sn , ρ* f μ = Sn ( x ) f ( x )ρ*( x )dμ( x ) = Sn , ρ* f s = 0,

for f ∈ Pn−d−1 , where Pn denotes the linear space of polynomials with real coefficients
and degree less than or equal to n ∈ Z+ . Thus, Sn is a quasi-orthogonal of order d with
respect to the modified measure μρ*. Therefore, Sn has at least (n − d) changes of sign in
( a, b).
Taking into account the known results for measures of bounded support (see [11]
(1.10)), the number of zeros located in the interior of the support of the measure is closely
related to d∗ = #( I+ ), where the symbol #( A) denotes the cardinality of a given set A. Note
that d∗ is the number of terms in the discrete part of ·, · s ( i.e., λ j,k > 0).
From Section 3 onward, we will restrict our attention to the case when in (4) the
measure dμ is the Jacobi measure dμα,β ( x ) = (1 − x )α (1 + x ) β dx (α, β > −1) on [−1, 1].
Some of the results we obtain are generalizations of previous work, with derivatives up to
order one. For more details, we refer the reader to [12,13] and the references therein.
The aim of this paper is to give an electrostatic interpretation for the distribution of
zeros of a wide class of Jacobi-Sobolev polynomials, following an approach based on the
works [4,14,15] and the original ideas of Stieltjes in [16,17].
In the next section, we obtain a formula that allows us to express the polynomial Sn
as a linear combination of Pn and Pn−1 , whose coefficients are rational functions. We refer

93
Mathematics 2023, 11, 3420

to this formula as “connection formula”. Sections 3 and 4 deal with the ladder (raising
and lowering) equations and operators of {Sn }n0 . We combine the ladder (raising and
lowering) operators to prove that the sequence of monic polynomials {S*n ( x )}n0 satisfies
the second-order linear differential Equation (35), with polynomial coefficients.
In the last section, we give a sufficient condition for an electrostatic interpretation of
the distribution of the zeros of {S*n ( x )}n0 as the logarithmic potential interaction of unit
positive charges in the presence of an external field. Several examples are given to illustrate
whether or not this condition is satisfied.

2. Connection Formula
Let μ be a finite positive Borel measure with finite moments, whose support supp(μ) ⊂ R
contains an infinite set of points. Assume that { Pn }n0 denotes the monic orthogonal
polynomial sequence with respect to the inner product (3). We first recall the well-known
Christoffel-Darboux formula for Kn ( x, y), the kernel polynomials associated with { Pn }n0 .
⎧
⎪
⎪ Pn ( x ) Pn−1 (y) − Pn (y) Pn−1 ( x )
n −1 ⎪
⎨ , if x = y,
Pk ( x ) Pk (y)  Pn−1 2μ ( x − y)
Kn−1 ( x, y) = ∑ = Pn ( x ) Pn−1 ( x ) − Pn ( x ) Pn −1 ( x ) (6)
 Pk 2μ ⎪
⎪
k =0 ⎪
⎩ , if x = y.
 Pn−1 2μ

∂ j+k Kn ( x, y)
( j,k)
We denote by Kn ( x, y) = the partial derivatives of the kernel (6).
∂x j ∂yk
Then, from the Christoffel-Darboux Formula (6) and the Leibniz rule, it is not difficult to
verify that

n −1 (k)
(0,k) Pi ( x ) Pi (y)
Kn−1 ( x, y) = ∑  Pi 2μ
i =0
k!( Qk ( x, y; Pn−1 ) Pn ( x ) − Qk ( x, y; Pn ) Pn−1 ( x ))
= , (7)
 Pn−1 2μ ( x − y)k+1

f (ν) (y)
where Qk ( x, y; f ) = ∑kν=0 ν! ( x − y)ν is the Taylor polynomial of the degree k of f
centered at y. Observe that (7) becomes the usual Christoffel-Darboux formula (6) if k = 0.
From (4), if i < n
(k) (k) (k) (k)
Sn , Pi μ = Sn , Pi s − ∑ λ j,k Sn (c j ) Pi (c j ) = − ∑ λ j,k Sn (c j ) Pi (c j ). (8)
( j,k)∈ I+ ( j,k)∈ I+

Therefore, from the Fourier expansion of Sn in terms of the basis { Pn }n0 and using (8),
we obtain
n −1 n −1 (k)
Pi ( x ) (k) Pi ( x ) Pi (c j )
Sn ( x ) = Pn ( x ) + ∑ Sn , Pi μ
 Pi 2μ
= Pn ( x ) − ∑ λ j,k Sn (c j ) ∑  Pi 2μ
i =0 ( j,k)∈ I+ i =0
(k) (0,k)
= Pn ( x ) − ∑ λ j,k Sn (c j )Kn−1 ( x, c j ). (9)
( j,k)∈ I+

94
Mathematics 2023, 11, 3420

Now, replacing (7) in (9), we have the connection formula

Sn ( x ) = F1,n ( x ) Pn ( x ) + G1,n ( x ) Pn−1 ( x ), (10)

(k)
λ j,k k! Sn (c j ) Qk ( x, c j ; Pn−1 )
where F1,n ( x ) = 1 − ∑  Pn−1 2μ ( x − c j ) k +1
( j,k)∈ I+
(k)
λ j,k k! Sn (c j ) Qk ( x, c j ; Pn )
and G1,n ( x ) = ∑  Pn−1 2μ ( x − c j ) k +1
.
( j,k)∈ I+

Deriving Equation (9) -times and evaluating then at x = ci for each ordered pair
(i, ) ∈ I+ , we obtain the following system of d∗ = #( I+ ) linear equations and d∗ unknowns
(k)
Sn ( c j ).
 
() (,) () (,k) (k)
Pn (ci ) = 1 + λi, Kn−1 (ci , ci ) Sn (ci ) + ∑ λ j,k Kn−1 (ci , c j )Sn (c j ). (11)
( j,k)∈ I+
( j,k)=(i,)

The remainder of this section is devoted to proving that system (11) has a unique
solution. The following lemma is essential to achieve this goal.

Lemma 1. Let I ⊂ R × Z+ be a (finite) set of d∗ pairs. Denote {c j } N j=1 = π1 ( I ) where π1 is the

projection function over the first coordinate, i.e., π1 ( x, y) = x, d j = max{νi : (c j , νi ) ∈ I } and
d = ∑N j=1 ( d j + 1). Let Pk be an arbitrary polynomial of the degree k for 0  k  n − 1. Then, for
all n  d, the d∗ × n matrix
 
(ν)
A∗ = Pk−1 (c)
(c,ν)∈ I,k=1,2,...,n

has a full rank d∗ .

Proof. First, note that, using elementary column transformations, we can reduce the proof
to the case when Pk ( x ) = x k , for k = 0, 1, . . . , n − 1. On the other hand, d∗j = #({νi :
(c j , νi ) ∈ I })  d j + 1 for j = 1, 2, . . . , N, so d∗ = ∑ N ∗
j=1 d j  d  n, and it is sufficient to
prove the case n = d. Consider the m × n matrix
⎛ ⎞
1 x x2 x3 ··· x n −1
⎜ 0 1 2x 3x2 ··· ( n − 1) x n −2 ⎟
⎜ ⎟
⎜ 0 0 2 6x ··· (n − 1)(n − 2) x n−3 ⎟
Am ( x ) = ⎜ ⎟,
⎜ .. .. .. .. .. .. ⎟
⎝ . . . . . . ⎠
0 0 0 0 · · · ( n − 1) · · · ( n − m + 1) x n − m

where m  n. Without loss of generality, we can rearrange the rows of A∗ such that
⎡ ⎤
A1∗
⎢ A2∗ ⎥  
⎢ ⎥ (ν)
A∗ = ⎢ .. ⎥, where A∗j = Pk−1 (c j ) .
⎣ . ⎦ (c j ,ν)∈ I,k=1,2,...,n
A∗N

Note that A∗j is obtained by taking some rows from Ad j +1 (c j ), the rows ν, such that
(c j , ν − 1) ∈ I. Consider the matrix

95
Mathematics 2023, 11, 3420

⎡ ⎤
A d1 +1 ( c 1 )
⎢ A d +1 ( c2 ) ⎥
⎢ 2 ⎥
A=⎢ .. ⎥.
⎣ . ⎦
A d N +1 ( c N )

From [18] (Theorem 20), we compute det(A) as

N dj
det(A) = det(A ) = ∏ ∏ i! ∏ (c j1 − c j2 )(d j1 +1)(d j2 +1) = 0.
j =1 i =1 1 j1 < j2  N

Then the n row vectors of A are linearly independent, and consequently, the d∗ rows
of A∗are also linearly independent.

Now we can rewrite (11) in the matrix form

Pn (C) = (Id∗ + Kn−1 (C , C)L)Sn (C), where (12)

Id∗ is the identity matrix of the order d∗ .

L is the d∗× d∗ -diagonal matrix with the diagonal entries λ j,k , ( j, k) ∈ I+ .
C is the column vector C = (c1 , . . . , c1 , c2 , . . . , c2 , . . . , c N , . . . , c N ) .
- ./ 0 - ./ 0 - ./ 0
d1∗ -times d2∗ -times d∗N -times
(k) (k)
Pn (C) and Sn (C) are column vectors with the entries Pn (c j ), and Sn (c j ), ( j, k) ∈ I+
respectively.
Kn−1 (C , C) is a d∗ × d∗ matrix whose entry associated to the (i, )th row and the ( j, k)th
n−1 P() ( c ) P(k ) ( c )
(,k) ν i ν j
column, (i, ), ( j, k) ∈ I+ , is Kn−1 (ci , c j ) = ∑ .
ν =0  Pν 2μ
 (k)

Pν−1 (c j )
Clearly, we can write Kn−1 (C , C) = F F  , where F =  Pν−1 μ
is a
( j,k)∈ I+ ,ν=1,...,n,
matrix of the order d∗× n and full rank for all n  d, according to Lemma 1.
Then the matrix Kn−1 (C , C) is a d∗× d∗ positive definite matrix for all n  d; see [19]
(Theorem 7.2.7(c)). Since L is a diagonal matrix with positives entries, it follows that
L−1 + Kn−1 (C , C) is also a positive definite matrix, and consequently, Id∗ + Kn−1 (C , C)L =
L−1 + Kn−1 (C , C) L is nonsingular. Then the linear system (12) has the unique solution

Sn (C) = (Id∗ + Kn−1 (C , C)L)−1 Pn (C). (13)

Using this notation, we can rewrite (9) in the compact form

Sn ( x ) = Pn ( x ) − Kn−1 ( x, C) L Sn (C), (14)

(0,k)
where Kn−1 ( x, C) is a row vector with the entries Kn−1 ( x, c j ), for ( j, k ) ∈ I+ . Now, replac-
ing (13) into (14), we obtain the matrix version of the connection Formula (10)

Sn ( x ) = Pn ( x ) − Kn−1 ( x, C)L(Id∗ + Kn−1 (C , C)L)−1 Pn (C).

3. Ladder Equations for Jacobi-Sobolev Polynomials

Henceforth, we will restrict our attention to the Jacobi-Sobolev case. Therefore, we
consider in the inner product (4) the measure dμ( x ) = dμα,β ( x ) = (1 − x )α (1 + x ) β dx,
where α, β > −1 and whose support is [−1, 1]. To simplify the notation, we will continue to
α,β
write Sn instead of Sn to denote the corresponding nth Jacobi-Sobolev monic polynomial.
In the following, we omit the parameters α and β when no confusion arises.

96
Mathematics 2023, 11, 3420

From [6] ((4.1.1), (4.3.2), (4.3.3), (4.5.1), and (4.21.6)), for the monic Jacobi polynomials,
we have
    
α,β 2n + α + β −1 n n + α n+β
Pn ( x ) = ∑ ( x − 1) ν ( x + 1) n − ν .
n ν =0 n − ν ν
1 1 Γ ( n + 1) Γ ( n + α + 1) Γ ( n + β + 1) Γ ( n + α + β + 1)
α,β 1 α,β 12
hn = 1Pn 1 α,β =22n+α+ β+1 .
μ Γ(2n + α + β + 2)Γ(2n + α + β + 1)
α,β 2n Γ ( n + α + 1 ) Γ ( n + α + β + 1)
Pn (1) = .
Γ(α + 1)Γ(2n + α + β + 1)
α,β α,β α,β α,β α,β α,β
xPn ( x ) = Pn+1 ( x ) + γ1,n Pn ( x ) + γ2,n Pn−1 ( x ); P0 ( x ) = 1, P−1 ( x ) = 0, (15)

where
α,β β2 − α2
γ1,n =γ1,n = ,
(2n + α + β)(2n + α + β + 2)
(16)
α,β 4n(n + α)(n + β)(n + α + β)
γ2,n =γ2,n = .
(2n + α + β)2 ((2n + α + β)2 − 1)
Let I be the identity operator. We define the two ladder Jacobi differential operators
on P as

an ( x )
* ↓n := − * 1 − x2 d
L I+ (lowering Jacobi differential operator),
*
bn *
bn dx
*
c
* ↑n := − n ( x ) 1 − x2 d
L I+ (raising Jacobi differential operator).
d*n d*n dx

where
n((2n + α + β) x + β − α) 4n(n + α)(n + β)(n + α + β)
an ( x ) = −
* , *bn = ,
2n + α + β (2n + α + β)2 (2n + α + β − 1)
(17)
(n + α + β)((2n + α + β) x + α − β)
c*n ( x ) = and d*n = −(2n + α + β − 1).
2n + α + β
 
α,β
From [6] (4.5.7 and 4.21.6), if n  1, the sequence Pn satisfies the relations
n 0

an ( x ) α,β
* ↓n Pnα,β ( x ) = − * 1 − x2  α,β  α,β
L Pn ( x ) + Pn ( x ) = Pn−1 ( x ),
*
bn *bn
2  (18)
* ↑n Pα,β ( x ) = − c*n ( x ) Pα,β ( x ) + 1 − x Pα,β ( x ) = Pnα,β ( x ).
L n −1 * n −1 * n −1
dn dn

In this case, the connection Formula (10) becomes

α,β α,β
Sn ( x ) = A1,n ( x ) Pn ( x ) + B1,n ( x ) Pn−1 ( x ), (19)
(k) α,β
α,β λ j,k k! Sn (c j ) Qk ( x, c j ; Pn−1 )
where A1,n ( x ) = A1,n ( x ) = 1 − ∑ α,β ( x − c j ) k +1
( j,k)∈ I+ h n −1
(k) α,β
α,β λ j,k k! Sn (c j ) Qk ( x, c j ; Pn )
and B1,n ( x ) = B1,n ( x ) = ∑ α,β ( x − c j ) k +1
.
( j,k)∈ I+ h n −1

97
Mathematics 2023, 11, 3420

N
d j +1
Let ρ( x ) = ∏ x − cj and define the (d − k − 1)th degree polynomial
j =1

ρ( x ) d j −k
N
ρ j,k ( x ) := = ( x − c j ) ∏ ( x − c i ) d i +1 , (20)
( x − c j ) k +1 i =1
i= j

for every ( j, k ) ∈ I+ . The following four lemmas are essential for defining ladder operators
(lowering and raising operators).

α,β
Lemma 2. For the sequences of polynomials {Sn }n0 and { Pn }n0 , we obtain
α,β α,β
ρ ( x ) Sn ( x ) = A2,n ( x ) Pn ( x ) + B2,n ( x ) Pn−1 ( x ), (21)
 
 α,β α,β
1−x 2
(ρ( x )Sn ( x )) = A3,n ( x ) Pn ( x ) + B3,n ( x ) Pn−1 ( x ), (22)

where
⎛ ⎞
(k)  
k!λ j,k Sn (c j ) α,β
A2,n ( x ) =ρ( x ) A1,n ( x ) = ρ( x ) − ∑ ⎝ Qk x, c j ; Pn−1 ⎠ρ j,k ( x ),
α,β
( j,k)∈ I+ h n −1
⎛ ⎞
(k)  
k!λ j,k Sn (c j ) α,β
B2,n ( x ) =ρ( x ) B1,n ( x ) = ∑ ⎝ Qk x, c j ; Pn ⎠ρ j,k ( x ),
α,β
( j,k)∈ I+ h n −1
 

A3,n ( x ) = A2,n ( x ) 1 − x2 + *an ( x ) A2,n ( x ) + d*n B2,n ( x ),
 

B3,n ( x ) = B2,n ( x ) 1 − x2 + *bn A2,n ( x ) + c*n ( x ) B2,n ( x ),

where A2,n , B2,n , A3,n , and B3,n are polynomials of degree at most d, d − 1, d + 1 and d, respectively,
and the coefficients *a n ( x ), *
bn , c*n ( x ), and d*n are given by (17).

Proof. From (19) and (20), Equation (21) is immediate. To prove (22), we can take deriva-
tives with respect to x in both hand sides of (21) and then multiply by 1 − x2
      
α,β
1 − x2 (ρ( x )Sn ( x )) = 1 − x2 A2,n

Pn ( x ) + A2,n 1 − x2 Pn ( x )
    
 α,β α,β
+ 1 − x2 B2,n Pn−1 ( x ) + B2,n 1 − x2 Pn−1 ( x ) .

Using (18) in the above expression, we obtain

   
α,β
1 − x2 (ρ( x )Sn ( x )) = A2n

( x ) 1 − x2 + * an ( x ) A2,n ( x ) + B2,n ( x )d*n Pn ( x )
 
α,β
+ B2,n
( x ) 1 − x2 + *bn A2,n ( x ) + B2,n ( x )c*n ( x ) Pn−1 ( x ),

which is (22).
 
α,β
Lemma 3. The sequences of the monic polynomials {Sn }n0 and Pn are also related by
n 0
the equations
α,β α,β
ρ( x )Sn−1 ( x ) = C2,n ( x ) Pn ( x ) + D2,n ( x ) Pn−1 ( x ), (23)
 
 α,β α,β
1−x 2
(ρ( x )Sn−1 ( x )) = C3,n ( x ) Pn ( x ) + D3,n ( x ) Pn−1 ( x ), (24)

98
Mathematics 2023, 11, 3420

where
 
B2,n−1 ( x ) x − γ1,n−1
C2,n ( x ) =− , D2,n ( x ) = A2,n−1 ( x ) + B2,n−1 ( x ) ,
γ2,n−1  γ2,n−1 
B (x) x − γ1,n−1
C3,n ( x ) = − 3,n−1 , D3,n ( x ) = A3,n−1 ( x ) + B3,n−1 ( x ) ,
γ2,n−1 γ2,n−1

where C2,n ( x ), D2,n ( x ), C3,n ( x ), and D3,n ( x ) are polynomials of degree at most d − 1, d, d and
d + 1, respectively.

Proof. The proof of (23) and (24) is a straightforward consequence of Lemma 2 and the
three-term recurrence relation (15), whose coefficients are given in (16).
 
α,β
Lemma 4. The monic orthogonal Jacobi polynomials Pn can be expressed in terms of the
n 0
monic Sobolev-type polynomials {Sn }n0 in the following way:

α,β ρ( x )
Pn ( x ) = ( D2,n ( x )Sn ( x ) − B2,n ( x )Sn−1 ( x )), (25)
Δn ( x )
α,β ρ( x )
Pn−1 ( x ) = ( A2,n ( x )Sn−1 ( x ) − C2,n ( x )Sn ( x )). (26)
Δn ( x )

where  
A2,n ( x ) B2,n ( x )
Δn ( x ) = det = A2,n ( x ) D2,n ( x ) − C2,n ( x ) B2,n ( x ) (27)
C2,n ( x ) D2,n ( x )
is a polynomial of the degree 2d.

Proof. Note that (21) and (23) form a system of two linear equations with the two un-
α,β α,β
knowns Pn ( x ) and Pn−1 ( x ). Therefore, from Cramer’s rule, we obtain (25) and (26).
A2,n ( x )
As dgr(C2,n B2,n )  2d − 2 and lim = 1, we obtain
x →∞ x2d
⎧
⎪
⎨ 1, if dgr( B2,n−1 ) < d − 1,
Δn ( x ) D2,n( x)
Λ n −1
lim = lim = 1 + , if dgr( B2,n−1 ) = d − 1, (28)
x →∞ x2d x →∞ xd ⎪
⎩ α,β
γ2,n−1 hn−2
 (k)
(k) α.β
where Λn−1 = ∑ λk,j Sn−1 (c j ) Pn−1 (c j ) = (Sn−1 (C)) T LPn−1 (C). From (12),
(i,j)∈ I+

Λn−1 = (Sn−1 (C)) T L(Id∗ + Kn−2 (C , C)L)Sn−1 (C).

Since the matrix L(Id∗ + Kn−2 (C , C)L) is positive definite, we conclude that

Λn−1 > 0, for all n ∈ N; (29)

i.e., Δn ( x ) is a polynomial of the degree 2d.

Remark 1. Obviously, from (25) (or (26)), we have that Δn ( x ) = ρ( x )δn ( x ), where δn is a
polynomial of the degree d. Hence, from (27),

δn ( x ) = A1,n ( x ) D2,n ( x ) − B1,n ( x )C2,n ( x ).

99
Mathematics 2023, 11, 3420

Theorem 1. Under the above assumptions, we have the following ladder equations:

A4,n ( x )Sn ( x ) + B4,n ( x )Sn ( x ) = Sn−1 ( x ), (30)

C4,n ( x )Sn−1 ( x ) + D4,n ( x )Sn −1 ( x ) = Sn ( x ), (31)

where
q2,n ( x ) q0,n ( x ) q3,n ( x ) q0,n ( x )
A4,n ( x ) = , B (x) = , C (x) = , D4,n ( x ) = .
q1,n ( x ) 4,n q1,n ( x ) 4,n q4,n ( x ) q4,n ( x )
 
q0,n ( x ) = 1 − x2 Δn ( x ), dgr(q0,n ) = 2d + 2.
q1,n ( x ) = B3,n ( x ) A2,n ( x ) − A3,n ( x ) B2,n ( x ), dgr(q1,n ) = 2d.
q2,n ( x ) = (1 − x2 )ρ ( x )δn ( x ) + B3,n ( x )C2,n ( x ) − A3,n ( x ) D2,n ( x ), dgr(q2,n ) = 2d + 1.

q3,n ( x ) = (1 − x )ρ ( x )δn ( x ) + C3,n ( x ) B2,n ( x ) − D3,n ( x ) A2,n ( x ),
2
dgr(q3,n ) = 2d + 1.
q4,n ( x ) = C3,n ( x ) D2,n ( x ) − D3,n ( x )C2,n ( x ), dgr(q4,n ) = 2d.

Proof. Replacing (25) and (26) in (22) and (24), the two ladder Equations (30) and (31)
follow.
1. ⎧
q1,n ( x ) ⎨ *
⎪ bn , if dgr( B2,n ) < d − 1,
lim = * Λn
x →∞ x 2d ⎪ b + (2n + α + β + 1) α,β , if dgr( B2,n ) = d − 1,
⎩ n h n −1

where, according to (29), Λn > 0, i.e., dgr(q1,n ) = 2d.

δn ( x ) D (x)
2. From (28), lim = lim 2,nd = κ2 > 0.
x →∞ x d x →∞ x
 
q2,n ( x ) (1 − x 2 ) ρ  ( x ) A3,n ( x )
lim 2d+1 = κ2 lim − lim = κ2 (−d + n + d)
x →∞ x x →∞ x d +1 x → ∞ x d +1
⎧
⎪
⎨n, if dgr( B2,n−1 ) < d − 1,
= n + nΛn−1 , if dgr( B
⎪
⎩ α,β 2,n−1 ) = d − 1,
γ2,n−1 hn−2

where, according to (29), Λn−1 > 0, i.e., dgr(q2,n ) = 2d + 1.

3.
⎧ *
bn − 1
⎪
q4,n ( x ) ⎨ − γ2,n−1 ,
⎪ if dgr( B2,n−1 ) < d − 1,
* Λ n −1
lim = b n −1 + ( 2n + α + β − 1 ) α,β
x →∞ x2d ⎪
⎪ h n −2
⎩ − , if dgr( B2,n−1 ) = d − 1.
γ2,n−1
Then, according to (29), dgr(q4,n ) = 2d.
4.

q3,n ( x ) D (x)
lim = −dκ2 − lim 3,n
x →∞ x2d+1 x → ∞ x d +1
⎧
⎨−(n + α + β),
⎪ if dgr( B2,n−1 ) < d − 1,
= Λ n −1
⎩−(n + α + β) 1 +
⎪ α,β
, if dgr( B2,n−1 ) = d − 1,
γ2,n−1 hn−2

where, according to (29), Λn−1 > 0, i.e., dgr(q3,n ) = 2d + 1.

In the previous theorem, the polynomials qk,n were defined. Note that these poly-
nomials are closely related to certain determinants. The following result summarizes

100
Mathematics 2023, 11, 3420

some of their properties that will be of interest later. For brevity, we introduce the follow-
ing notations:

Δ1,n ( x ) = B3,n ( x ) A2,n ( x ) − A3,n ( x ) B2,n ( x ).

Δ2,n ( x ) = B3,n ( x )C2,n ( x ) − A3,n ( x ) D2,n ( x ).
Δ3,n ( x ) = B2,n ( x )C3,n ( x ) − A2,n ( x ) D3,n ( x ).

N N
ρ( x )
∏ ∏
dj
Lemma 5. Let ρ N ( x ) = x − c j and ρd− N ( x ) = x − cj = . Then, the above
j =1 j =1
ρ N (x)
polynomial determinants admit the following decompositions:

Δ1,n ( x ) = ρd− N ( x ) ϕ1,n ( x ), where dgr( ϕ1,n ) = d + N.

Δ2,n ( x ) = ρd− N ( x ) ϕ2,n ( x ), where dgr( ϕ2,n ) = d + N + 1. (32)
Δ3,n ( x ) = ρd− N ( x ) ϕ3,n ( x ), where dgr( ϕ3,n ) = d + N + 1.

Proof. Multiplying (21) by B3,n and (22) by B2,n and taking their difference, we have
α,β
Δ1,n ( x ) Pn ( x ) = ρ( x ) B3,n ( x )Sn ( x ) − (1 − x2 ) B2,n ( x ) ρ ( x )Sn ( x ) + ρ( x )Sn ( x )

= ρd− N ( x ) ρ N ( x ) B3,n ( x )Sn ( x ) − (1 − x2 ) B2,n ( x )
 N 
∑ (d j + 1) ρ j,dj (x) Sn (x) + ρ N (x) Sn (x) .
j =1

α,β
As Pn (c j ) = 0 for j = 1, . . . , N and dgr(Δ1,n ) = dgr(q1,n ) = 2d (see the proof of
Theorem 1), then there exists a polynomial ϕ1,n of the degree d + N such that Δ1,n ( x ) =
ρd− N ( x ) ϕ1,n ( x ).
For the decomposition of Δ2,n (Δ3,n ) the procedure of the proof is analogous, using the
linear system of (22) and (23) ((21)–(24)).

4. Ladder Jacobi-Sobolev Differential Operators and Consequences

Definition 1 (Ladder Jacobi-Sobolev differential operators). Let I be the identity operator. We
define the two ladder differential operator on P as

d
L↓n := A4,n ( x )I + B4,n ( x ) (lowering Jacobi-Sobolev differential operator),
dx
d
L↑n := C4,n ( x )I + D4,n ( x ) (raising Jacobi-Sobolev differential operator).
dx

Remark 2. Assume in (4) that dμ( x ) = dμα,β ( x ) = (1 − x )α (1 + x ) β dx (α, β > −1), whose
support is [−1, 1] and λ j,k ≡ 0 for all pairs ( j, k ). Under these conditions, it is not difficult to verify
that L↓n ≡ L* ↓n and L↑n ≡ L
* ↑n .

Now, we can rewrite the ladder Equations (30) and (31) as

 
d
L↓n [Sn ( x )] = A4,n ( x )I + B4,n ( x ) Sn ( x ) = Sn −1 ( x ), (33)
dx
 
d
L↑n [Sn−1 ( x )] = C4,n ( x )I + D4,n ( x ) S ( x ) = Sn ( x ). (34)
dx n−1

In this section, we state several consequences of Equations (33) and (34), which gener-
alize known results for classical Jacobi polynomials to the Jacobi-Sobolev case.

101
Mathematics 2023, 11, 3420

First, we are going to obtain a second-order differential equation with polynomial

coefficients for Sn . The procedure is well known and consists in applying the raising
operator L↑n to both sides of the formula L↓n [Sn ] = Sn−1 . Thus, we have

0 =L↑n L↓n [Sn ( x )] − Sn ( x )

= B4,n ( x ) D4,n ( x )Sn ( x )

+ A4,n ( x ) D4,n ( x ) + B4,n ( x )C4,n ( x ) + D4,n ( x ) B4,n ( x ) Sn ( x )

+ A4,n ( x )C4,n ( x ) + D4,n ( x ) A4,n ( x ) − 1 Sn ( x )
q20,n ( x )
= S ( x )
q1,n ( x )q4,n ( x ) n
 
q0,n ( x ) q1,n ( x )q2,n ( x ) + q1,n ( x )q3,n ( x ) + q0,n  ( x )q ( x ) − q ( x )q ( x )
1,n 0,n 1,n
+ Sn ( x )
q4,n ( x )q21,n ( x )
⎛   ⎞
 ( x )q ( x ) − q ( x )q ( x )
q1,n ( x )q2,n ( x )q3,n ( x ) + q0,n ( x ) q2,n 1,n 2,n 1,n
+⎝ − 1⎠ Sn ( x ),
q4,n ( x )q21,n ( x )

from where we conclude the following result.

Theorem 2. The nth monic orthogonal polynomial with respect to the inner product (4) is a
polynomial solution of the second-order linear differential equation, with polynomial coefficients

P2,n ( x )Sn ( x ) + P1,n ( x )Sn ( x ) + P0,n ( x )Sn ( x ) = 0, (35)

where

P2,n ( x ) =q1,n ( x )q20,n ( x ),

 
P1,n ( x ) =q0,n ( x ) q1,n ( x )q2,n ( x ) + q1,n ( x )q3,n ( x ) + q0,n ( x )q1,n ( x ) − q0,n ( x )q1,n (x) ,
 
P0,n ( x ) =q1,n ( x )q2,n ( x )q3,n ( x ) + q0,n ( x ) q2,n ( x )q1,n ( x ) − q2,n ( x )q1,n (x) (36)
− q4,n ( x )q21,n ( x ),
dgr(P2,n ) = 6d + 4, dgr(P1,n )  6d + 3 , and dgr(P0,n )  6d + 2.

Remark 3 (The classical Jacobi differential equation). Under the conditions stated in Remark 2,
α,β
(4) becomes to the classical Jacobi inner product and Sn ( x ) = Pn (x).
Note that, here, A1,n ( x ) ≡ 1, B1,n ( x ) = 0 and ρ( x ) ≡ 1. For the rest of the expressions
involved in the coefficients of the differential Equation (35), we have

ρ( x ) ≡ 1, A1,n ( x ) ≡ A2,n ( x ) ≡ D2,n ( x ) = 1, B1,n ( x ) ≡ B2,n ( x ) ≡ C2,n ( x ) ≡ 0,

an ( x ), B3,n ( x ) = *
Δn ( x ) ≡ 1, A3,n ( x ) = * bn , C3,n ( x ) = −γ−1 *bn−1 and
2,n−1
−1 *
D3,n ( x ) = *
an−1 ( x ) + γ2,n −1 bn−1 ( x − γ1,n−1 ).

Thus,  
q0,n ( x ) = 1 − x2 , q1,n ( x ) = *
bn , q2,n ( x ) = −*
a n ( x ),
−1 *
q3,n ( x ) = −*
an−1 ( x ) − γ2,n −1 bn−1 ( x − γ1,n−1 )
(n + α + β) (α − β) (37)
= −(n + α + β) x + and
2n + β + α
−1 *
q4,n ( x ) = −γ2,n −1 bn−1 = −(2n + α + β − 1).

102
Mathematics 2023, 11, 3420

Substituting (37) in (36), the reader can verify that the differential Equation (35) becomes (2),
i.e.,
 
P2,n ( x ) = 1 − x2 , P1,n ( x ) = β − α − (α + β + 2) x and P0,n ( x ) = n(n + α + β + 1).

Second, we can obtain the polynomial nth degree of the sequence {Sn }n0 as the
repeated action (n times) of the raising differential operator on the first Sobolev-type
polynomial of the sequence (i.e., the polynomial of degree zero).

Theorem 3. The nth Jacobi-Sobolev polynomial Sn (n  0) can be given by

 
Sn ( x ) = L↑n L↑n−1 L↑n−2 · · · L1↑ S0 ( x ),

where S0 ( x ) = 1.

Proof. Using (34), the theorem follows for n = 1. Next, the expression for Sn is a straight-
forward consequence of the definition of the raising operator.

To conclude this section, we prove an interesting three-term recurrence relation with

rational coefficients, which satisfies the Jacobi-Sobolev monic polynomials. From the
explicit expression of the ladder operators, shifting n to n + 1 in (34), we obtain

d
C4,n ( x )Sn ( x ) + D4,n ( x ) Sn ( x ) = Sn −1 ( x ),
dx
d
A4,n ( x )Sn ( x ) + B4,n ( x ) Sn ( x ) = Sn+1 ( x ).
dx
Next, we multiply the first equation by − B4,n ( x ) and the second equation by D4,n ( x ),
and adding two resulting equations, we have the following three-term recurrence reaction
with rational coefficients for the Jacobi-Sobolev monic orthogonal polynomials.

Theorem 4. Under the assumptions of Theorem 2, we have the recurrence relation

q4,n+1 ( x )q0,n ( x )Sn+1 ( x ) =[q3,n+1 ( x )q0,n ( x ) − q2,n ( x )q0,n+1 ( x )]Sn ( x )

(38)
+ q1,n ( x )q0,n+1 ( x )Sn−1 ( x ),
where the explicit formula of the coefficient is given in Theorem 1.

Proof. From (30), and (31) for n + 1, we have

q2,n ( x )Sn ( x ) + q0,n ( x )( x )Sn ( x ) = q1,n ( x )Sn−1 ( x ).

q3,n+1 ( x )Sn ( x ) + q0,n+1 ( x )Sn ( x ) = q4,n+1 ( x )Sn ( x ).

Multiplying by q0,n+1 ( x ) and q0,n ( x ), respectively, we subtract both equations to

eliminate the derivative term obtaining

(q3,n+1 ( x )q0,n ( x ) − q2,n ( x )q0,n+1 ( x ))Sn ( x )

= q4,n+1 ( x )q0,n ( x )Sn+1 ( x ) − q1,n+1 ( x )q0,n+1 ( x )Sn−1 ( x ),

which is the required formula.

103
Mathematics 2023, 11, 3420

Remark 4 (The classical Jacobi three-term recurrence relation). Under the assumptions of
Remark 2, substituting (37) in (38), the reader can verify that the three-term recurrence relation (38)
becomes (35), i.e.,

q3,n+1 ( x )q0,n ( x ) − q2,n ( x )q0,n+1 ( x ) q1,n ( x )q0,n+1 ( x )

= x − γ1,n and = −γ2,n .
q4,n+1 ( x )q0,n ( x ) q4,n+1 ( x )q0,n ( x )

5. Electrostatic Interpretation
Let us begin by recalling the definition of a sequentially ordered Sobolev inner product,
which was stated in [20] (Definition 1) or [21] (Definition 1).

Definition 2. Let {(r j , νj )} jM=1 ⊂ R × Z+ be a finite sequence of M ordered pairs and A ⊂ R. We

say that {(r j , νj )} jM=1 is sequentially ordered with respect to A, if
1. 0  ν1  ν2  · · ·  νM .
2. / Ch( A ∪ {r1 , r2 , . . . , rk−1 })◦ for k = 1, 2, . . . , M, where Ch( B)◦ denotes the interior of
rk ∈
the convex hull of an arbitrary set B ⊂ C.
If A = ∅, we say that {(r j , νj )} jM=1 is sequentially ordered for brevity.
We say that the discrete Sobolev inner product (4) is sequentially ordered if the set of ordered
pairs {(c j , i ) : 1  j  N, 0  i  d j and η j,i > 0} may be arranged to form a finite sequence of
ordered pairs, which is sequentially ordered with respect to (−1, 1).

From the second condition of Definition 2, the coefficient λ j,d j is the only coefficient
λ j,i (i = 0, 1, . . . , d j ) different from zero, for each j = 1, 2, . . . , N. Hence, (4) takes the form

 1 N
f, g s =
−1
f ( x ) g( x ) dμα,β ( x ) + ∑ λ j,dj f (dj ) (c j ) g(dj ) (c j ), (39)
j =1

where dμα,β ( x ) = (1 − x )α (1 + x ) β dx, with α, β > −1.

Hereinafter, we will restrict our attention to sequentially ordered discrete Sobolev
inner products. The following two lemmas show our reasons for this restriction.

Lemma 6 ([20, Th. 1] and [21, Prop. 4]). If (39) is a sequentially ordered discrete Sobolev inner
product, then Sn has at least n − N changes of sign on (−1, 1).

Lemma 7 ([20, Lem. 3.4] and [21, Th. 7]). Let (39) be a sequentially ordered Sobolev inner
product. Then, for all n sufficiently large, each sufficiently small neighborhood of c j , j = 1, . . . , N,
contains exactly one zero of Sn , and the remaining n − N zeros lie on (−1, 1).

As the coefficient of Sn is real, under the same hypotheses of Lemma 7, for all n
sufficiently large, the zeros of Sn are real and simple.
In the rest of this section, we will assume that the zeros of Sn are simple. Note that
sequentially ordered Sobolev inner products provide us with a wide class of Sobolev inner
products such that the zeros of the corresponding orthogonal polynomials are simple.
Therefore, for all n sufficiently large, we have
n n n n n
Sn ( x ) = ∑ j=∏1, (x − xn,j ), Sn ( x ) = ∑ ∑ ∏ (x − xn,l ),
i =1 i =1 j=1, l =1,
j  =i j  =i i  = j  = l

n n n
Sn ( xn,k ) = ∏ (xn,k − xn,j ), Sn ( xn,k ) = 2 ∑ ∏ ( xn,k − xn,j ).
j=1, i =1, j=1,
j=k i=k i= j=k

104
Mathematics 2023, 11, 3420

2 3n we evaluate the polynomials P2,n ( x ), P1,n ( x ), and P0,n ( x ) in (35) at xn,k , where
Now
xn,k k=1 are the zeros of Sn ( x ) arranged in an increasing order. Then, for k = 1, 2, . . . , n,
we obtain

0 =P2,n ( xn,k )Sn ( xn,k ) + P1,n ( xn,k )Sn ( xn,k ) + P0,n ( xn,k )Sn ( xn,k )
=P2,n ( xn,k )Sn ( xn,k ) + P1,n ( xn,k )Sn ( xn,k ).
Sn ( xn,k ) P1,n ( xn,k ) n
1 P1,n ( xn,k )
0= 
+ =2∑ + . (40)
Sn ( xn,k ) P2,n ( xn,k ) i =1 x n,k − x n,i P2,n ( xn,k )
i=k

Let us recall that, from (32),

Δ1,n ( x )
ϕ1,n ( x ) = , dgr( ϕ1,n ) = d + N,
ρd− N ( x )
Δ2,n ( x )
ϕ2,n ( x ) = , dgr( ϕ2,n ) = d + N + 1,
ρd− N ( x )
Δ3,n ( x )
ϕ3,n ( x ) = , dgr( ϕ3,n ) = d + N + 1.
ρd− N ( x )

Hence, from Theorems 1 and 2 and Lemma 5,

 
P1,n ( x ) q1,n ( x )q2,n ( x ) + q1,n ( x )q3,n ( x ) + q0,n ( x )q1,n ( x ) − q0,n ( x )q1,n ( x )
=
P2,n ( x ) q1,n ( x )q0,n ( x )
 
q2,n ( x ) + q3,n ( x ) q0,n ( x ) q1,n ( x )
= + −
q0,n ( x ) q0,n ( x ) q1,n ( x )
ρ ( x ) Δ ( x ) + Δ3,n ( x ) Δ ( x ) 2x Δ ( x )
=2 + 2,n 2 + n + 2 − 1,n
ρ( x ) (1 − x )ρ( x )δn ( x ) Δn ( x ) x − 1 Δ1,n ( x )
ρ ( x ) ϕ ( x ) + ϕ3,n ( x ) δ ( x ) 1 1
=3 + 2,n 2 + n + +
ρ( x ) (1 − x )ρ N ( x )δn ( x ) δn ( x ) x − 1 x + 1
ϕ ( x ) ρd− N ( x )
− 1,n − . (41)
ϕ1,n ( x ) ρd− N ( x )

ρ ( x ) N ρd− N ( x )
dj + 1 N dj
Let us write
ρ( x )
= ∑ x − cj
.
ρd− N ( x )
=∑
x − cj
.
j =1 j =1
 
As ψ1 ( x ) = ϕ2,n ( x ) + ϕ3,n ( x ) and ψ2 ( x ) = 1 − x2 ρ N ( x )δn ( x ) are polynomials of
ψ1 ( x )
the degree d + N + 1 and d + N + 2, respectively, we have that is a rational proper
ψ2 ( x )
fraction. Therefore,
N r (c ) d r (u )
ψ1 ( x ) r (1) r (−1) j j ψ1 ( x )
=− + +∑ +∑ , where r ( x ) = .
ψ2 ( x ) x−1 x+1 j =1
x − c j j =1
x − uj ψ2 ( x )

Based on the results of our numerical experiments, in the remainder of the section, we
will assume certain restrictions with respect to some functions and parameters involved
in (41). In that sense, we suppose that
1. The zeros of δn are real, simple, and different from xn,k for all k = 1, . . . , n. Therefore,
d
δ ( x ) d
1
δn ( x ) = ∏ (x − u j ), where ui = u j if i = j, and δnn (x) = ∑ x − uj .
k =1 j =1

105
Mathematics 2023, 11, 3420

N1
2. Let ϕ1,n ( x ) = κ1 ∏ ( x − e j )5,j , where e j ∈ C \ Ch([−1, 1] ∪ {c1 , . . . , c N }) for all
j =1
 (x)
N1 ϕ1,n N1 5,j
j = 1, . . . , N − 1, and ∑ 5,j = d + N. Therefore, ϕ1,n ( x )
= ∑ x − ej .
j =1 j =1
3. Substituting into (41) the previous decompositions, we have

P1,n ( x )   N 3,j d 4,j N1 

5,j
= 1 + 2 +∑ +∑ −∑ ,
P2,n ( x ) x − 1 x + 1 j =1 x − c j j =1 x − u j j =1 x − e j

where 1 = 1 − r (1), 2 = 1 + r (−1), 3,j = 2d j + r (c j ) + 3, and 4,j = r (u j ) + 1. We

will assume that 1 , 2 , 3,j , 4,j  0.
From (40), for k = 1, . . . , n,
n
1 1 1  1
0= ∑ xn,k − xn,i + 2 xn,k − 1
+ 2
2 xn,k + 1
i =1
i=k

1 N 3,j 1 d 4,j 1 N1 5,j

+
2 ∑ xn,k − c j + 2 ∑ xn,k − u j + 2 ∑ e j − xn,k . (42)
j =1 j =1 j =1

Let ω = (ω1 , ω2 , · · · , ωn ), x n = ( xn,1 , xn,2 , · · · , xn,n ) and denote

1
E(ω ) := ∑ log
| ω j − ωk |
+ F ( ω ) + G ( ω ), (43)
1≤ k < j ≤ n

1 n 1 1 N
1
F (ω ) := ∑ log 
+ log 
+ ∑ log ,
2 k =1 |1 − ω k | 1 |1 + ω k | 2 j =1 |c j − ωk |3,j

N1
1 n d
1 1
G (ω ) := ∑ ∑ log 4,j
+ ∑ log .
2 k =1 j =1 | u j − ωk | j =1 |e j − ωk |5,j

Let us introduce the following electrostatic interpretation:

Consider the system of n movable positive unit charges at n distinct points of the
real line, {ω1 , ω2 , · · · , ωn }, where their interaction obeys the logarithmic potential
law (that is, the force is inversely proportional to the relative distance) in the presence
of the total external potential Vn (ω ) = F (ω ) + G (ω ). Then, E(ω ) is the total energy
of this system.

Following the notations introduced in [14] (Section 2), the Jacobi-Sobolev inner product
creates two external fields. One is a long-range field whose potential is F (ω ), and the other
is a short-range field whose potential is G (ω ). Therefore, the total external potential Vn (ω )
is the sum of the short- and long-range potentials, which is dependent on n (i.e., varying
external potential).
∂E
Therefore, for each k = 1, . . . , n, we have ( x n ) = 0; i.e., the zeros of Sn are the
∂ωk
zeros of the gradient of the total potential of energy E(ω ) (∇ E( x n ) = 0).

Theorem 5. The zeros of Sn ( x ) are a local minimum of E(ω ), if for all k = 1, . . . , n;

∂E
1. ( x n ) = 0.
∂ωk
∂ Vn
2 ∂2 F ∂2 G
2. ( xn ) = ( xn ) + ( x n ) > 0.
∂wk2 ∂wk2 ∂w2k

106
Mathematics 2023, 11, 3420

Proof. The Hessian matrix of E at x n is given by

⎧
⎪
⎪ ∂2 E
⎪
⎪ ( x ) = −( xk − x j )−2 , if k = j,
⎨ ∂wk ∂w j n
∇ ω ω E ( x n ) = ∂2 E
2 n
1 ∂ (Vn )
2 (44)
⎪
⎪ ( xn ) = ∑ + ( x n ), if k = j.
⎪
⎪
⎩ ∂w2k i =1 ( x n,k − x n,i ) ∂w2k
2
i=k

Note that (44) is a symmetric real matrix with negative values in the nondiagonal
entries. Additionally, note that
n
∂2 E ∂2 E ∂2 Vn
∑ ∂wk ∂w j (xn ) + ∂w2 (xn ) = ∂w2k
( x n ).
j =1 k
i=k

Since this is positive, we conclude according to Gershgorin’s theorem [19] (Theorem 6.1.1)
that the eigenvalues of the Hessian are positive, and therefore, (44) is positive definite.
Combining this with the fact that ∇ E( x n ) = 0, we conclude that x n is a local minimum
of (43).

The computations of the following examples have been performed using the symbolic
computer algebra system Maxima [22]. In all cases, we fixed n = 12 and considered sequen-
tially ordered Sobolev inner products (see Definition 2 and Lemmas 6 and 7). From (42), it is
obvious that ∇ E( x12 ) = 0, where x12 = ( x12,1 , x12,2 , · · · , x12,n ) and
S12 ( x12,k ) = 0 for k = 1, 2, . . . , 12. Under the above condition, x12 is a local minimum
(maximum) of E if the corresponding Hessian matrix at x12 is positive (negative) definite;
in any other case, x12 is said to be a saddle point. We recall that a square matrix is positive
(negative) definite if all its eigenvalues are positive (negative).

Example 2 (Case in which the conditions of Theorem 5 are satisfied).

 1
1. Jacobi-Sobolev inner product f , g s = f ( x ) g( x )(1 + x )100 dx + f  (2) g (2).
−1
2. Zeros of S12 ( x ).

x12 =(0.44845, 0.563364, 0.653317, 0.728094, 0.791318, 0.844674,

0.889402, 0.925746, 0.954364, 0.97639, 0.989824, 0.998408).

1
3. Total potential of energy E(ω ) = ∑ log
| ω j − ωk |
+ F (ω ) + G (ω ), where
1≤k < j≤12

1 12 1 1 1
2 k∑
F (ω ) = log + log + log ,
=1
| ω k − 1| |ωk + 1|101 | ω k − 2|3
1 12
2 k∑
G (ω ) = log|(ωk − 1.04563)τ (ωk )| and τ ( x ) = x2 − 3.8812x + 3.76606 > 0.
=1

∂E
4. From (42), ( x ) = 0, for j = 1, . . . , 12.
∂ω j 12
5. Computing the corresponding Hessian matrix at x12 , we have that the approximate values of
its eigenvalues are

{81.7737, 220.5813, 383.5185, 586.5056, 857.6819, 1248.8, 1857.7, 2927.5, 5039.9,

9986.6, 26185, 214620}.

Thus, Theorem 5 holds for this example, and we have the required local electrostatic equilibrium
distribution.

107
Mathematics 2023, 11, 3420

Example 3 (Case in which the conditions of Theorem 5 are satisfied).

1. Jacobi-Sobolev inner product
 1
f, g s = f ( x ) g( x )(1 + x )110 dx + f  (1) g (1) + f  (2) g (2).
−1

2. Zeros of S12 ( x ).

x12 =(0.482433, 0.590159, 0.674139, 0.74379, 0.802629, 0.852355,

0.894142, 0.928255, 0.955716, 0.976239, 0.990307, 0.998211).

1
3. Total potential of energy E(ω ) = ∑ log
| ω j − ωk |
+ F (ω ) + G (ω ), where
1≤k < j≤12

1 12 1 1 1
2 k∑
F (ω ) = log + log + log ,
=1 | ω k − 1|3 |ωk + 1|111 | ω k − 2|4
1 12
2 k∑
G (ω ) = log|(ωk − 1.22268)(ωk − 1.94089)τ (ωk )|
=1
and τ ( x ) = x2 − 3.8196x + 3.65881 > 0.

∂E
4. From (40), ( x ) = 0, for j = 1, . . . , 12.
∂ω j 12
5. Computing the corresponding Hessian matrix at x12 , we have that the approximate values of
its eigenvalues are

{102.3077, 265.8911, 459.368, 702.7009, 1030.2, 1504.8, 2247.1, 3563.2, 6146,

12806, 38783, 488410}.

Thus, Theorem 5 holds for this example, and we have the required local electrostatic equilibrium
distribution.

Example 4 (Case in which the conditions of Theorem 5 are not satisfied).

 1
1. Jacobi-Sobolev inner product f , g s = f ( x ) g( x )dx + f  (2) g (2).
−1
2. Zeros of S12 ( x ).

x12 =(−0.979635, −0.894154, −0.746211, −0.545446, −0.305098, −0.0412552,

0.227973, 0.483321, 0.705221, 0.87481, 0.975632, 2.1607).

1
3. Total potential of energy E(ω ) = ∑ log
| ω j − ωk |
+ F (ω ) + G (ω ), where
1≤k < j≤12

1 12 1 1 1
2 k∑
F (ω ) = log + log + log ,
=1
| ω k − 1 | | ω k + 1 | | ω k − 2|3
1 12
2 k∑
G (ω ) = log|(ωk − 2.12065)τ (ωk )| and τ ( x ) = x2 − 3.74216 x + 3.51112 > 0.
=1

∂E
4. From (42), ( x ) = 0, for j = 1, . . . , 12.
∂ω j 12

108
Mathematics 2023, 11, 3420

5. Computing the corresponding Hessian matrix at x12 , we have that the approximate values of
its eigenvalues are

{1388.3, 975.7989, 242.5338, 179.5748, 107.6368, 86.754, 70.7275, 62.6406, 50.3046,

34.4135, 14.0599, −258.3366}.

Then, x12 is a saddle point of E(ω ).

Remark 5. As can be noticed, in some cases, the configuration given by the external field includes
complex points; they correspond to e j . Specifically, in the examples, these points are given as the
zeros of τ ( x ). Since φ1,n ( x ) is a polynomial of real coefficients, the nonreal zeros arise as complex
conjugate pairs. Note that

a a 2x + 2z
+ =a 2
x−z x−z x + 2 z + | z |2

where z denotes the real part of z. The antiderivative of the previous expression is a ln( x2 +
2z + |z|2 ). This means in our current case that the presence of complex roots does not change the
formulation of the energy function.

What Happens If the Hessian Is Not Positive Definite? A Case Study

Theorem 5 gives us a general condition to determine whether the electrostatic inter-
pretation is a mere extension of the classical cases. However, in Example 4, the Hessian has
one negative eigenvalue of about −258 corresponding to the last variable ωn . Therefore,
we do not have the nice interpretation given in Theorem 5. However, note that the rest of
the eigenvalues are positive, which means that the number

∂2 (Vn )
( xn )
∂w2k

remains positive for k = 1, . . . , 11. In this case, the potential function exhibits a saddle point.
The presence of the saddle point is somehow justified by the attractor point a ≈ −2.121
having a zero ( x12,12 ≈ 2.161) in its neighborhood. In this case, we are able to give an
interpretation of the position of the zeros by considering a problem of conditional extremes.
Assume that, when checking the Hessian, we obtained that the eigenvalues λi , for
i ∈ E ⊂ {1, 2, . . . , n}, are negative or zero. Without loss of generality, assume that this
happens for the last mE = |E | variables. This is a saddle point. However, the rest of the
eigenvalues are positive, which means that the truncated Hessian ∇2ωm ωm E formed by
E E
taking the first n − mE rows and columns of ∇2ω ω ER is a positive definite matrix by the
same arguments used in the proof of Theorem 5.
Let us define the following problem of conditional extremum on ω = ω n ∈ Rn

min E(ω n )
ω n ∈ Rn
subject to ωk − xk = 0, for all k = n − mE + 1, . . . , n.
Note that this problem is equivalent to solve

min ER ( ω n − m E , x m E +1 , . . . , x n ).
ω n − m E ∈ Rn − m E

Let us prove that x n−mE is a minimum of this problem. Note that the gradient of this
function corresponds to the first n − mE conditions of (42), and the second-order condition
is given by the truncated Hessian ∇2ωm ωm E( x mE ), which is by hypothesis positive definite.
E E
Therefore, the configuration x n corresponds to the local equilibrium of the energy
function (43) once mE charges are fixed.

109
Mathematics 2023, 11, 3420

Author Contributions: Conceptualization, H.P.-C. and J.Q.-R.; methodology, H.P.-C.; software, J.Q.-R.
and J.T.-M.; validation, J.Q.-R. and J.T.-M.; formal analysis, H.P.-C. and J.Q.-R.; investigation, H.P.-C.,
J.Q.-R. and J.T.-M.; writing—original draft preparation, H.P.-C.; writing—review and editing, H.P.-C.,
J.Q.-R. and J.T.-M.; supervision, H.P.-C.; funding acquisition, J.T.-M. All authors have read and agreed
to the published version of the manuscript.
Funding: The research of J. Toribio-Milane was partially supported by Fondo Nacional de Innovación
y Desarrollo Científico y Tecnológico (FONDOCYT), Dominican Republic, under grant 2020-2021-
1D1-137.
Data Availability Statement: Not applicable.
Conflicts of Interest: The authors declare no conflict of interest.

References
1. Van Assche, W. The impact of Stieltjes work on continued fractions and orthogonal polynomials. In Thomas Jan Stieltjes Oeuvres
Complètes—Collected Papers; van Dijk, G., Ed.; Springer: Berlin/Heidelberg, Germany, 1993; pp. 5–37.
2. Valent, G.; Van Assche, W. The impact of Stieltjes’s work on continued fractions and orthogonal polynomials: Additional material.
J. Comput. Appl. Math. 1995, 65, 419–447. [CrossRef]
3. Marcellán, F.; Martínez-Finkelshtein, A.; Martínez, P. Electrostatic models for zeros of polynomials: Old, new, and some open
problems. J. Comput. Appl. Math. 2007, 207, 258–272. [CrossRef]
4. Huertas, E.J.; Marcellán, F.; Pijeira-Cabrera, H. An electrostatic model for zeros of perturbed Laguerre polynomials. Proc. Amer.
Math. Soc. 2014, 142, 1733–1747. [CrossRef]
5. Orive, R.; García, Z. On a class of equilibrium problems in the real axis. J. Comput. Appl. Math. 2020, 235, 1065–1076. [CrossRef]
6. Szegő, G. Orthogonal Polynomials, 4th ed.; American Mathematical Society Colloquium Publications; American Mathematical
Society: Providence, RI, USA, 1975; Volume 23.
7. Chihara, T.S. An Introduction to Orthogonal Polynomials; Gordon and Breach: New York, NY, USA, 1978.
8. Freud, G. Orthogonal Polynomials; Pergamon Press: Oxford, UK, 1971.
9. Marcellán, F.; Xu, Y. On Sobolev orthogonal polynomials. Expo. Math. 2015, 33, 308–352. [CrossRef]
10. Martinez-Finkelshtein, A. Analytic properties of Sobolev orthogonal polynomials revisited. J. Comput. Appl. Math. 2001,
127, 255–266. [CrossRef]
11. López Lagomasino, G.; Marcellán, F.; Van Assche, W. Relative asymptotics for orthogonal polynomials with respect to a discrete
Sobolev inner product. Constr. Approx. 1995, 11, 107–137. [CrossRef]
12. Arvesú, J.; Álvarez-Nodarse, R.; Marcellán, F.; Pan, K. Jacobi-Sobolev-type orthogonal polynomials: Second-order differential
equation and zeros. J. Comput. Appl. Math. 1998, 90, 135–156. [CrossRef]
13. Dueñas, H.A.; Garza, L.E. Jacobi-Sobolev-type orthogonal polynomials: Holonomic equation and electrostatic interpretation—A
non-diagonal case. Integral Transforms Spec. Funct. 2013, 24, 70–83. [CrossRef]
14. Ismail, M.E.H. An electrostatics model for zeros of general orthogonal polynomials. Pacific J. Math. 2000, 193, 355–369. [CrossRef]
15. Ismail, M.E.H. More on electrostatic models for zeros of orthogonal polynomials. Numer. Funct. Anal. Optimiz. 2000, 21, 191–204.
[CrossRef]
16. Stieltjes, T.J. Sur quelques théorèmes d’algèbre, Comptes Rendus de l’Academie des Sciences. Paris 1885, 100, 439–440.
17. Stieltjes, T.J. Sur les polynômes de Jacobi, Comptes Rendus de l’Academie des Sciences. Paris 1885, 100, 620–622.
18. Krattenthaler, C. Advanced Determinant Calculus, The Andrews Festschrift: Seventeen Papers on Classical Number Theory and Combina-
torics; Springer: Berlin/Heidelberg, Germany, 2001; pp. 349–426.
19. Horn, R.A.; Johnson, C.R. Matrix Analysis; Cambridge University Press: Cambridge, UK, 1990.
20. Díaz-González, A.; Pijeira-Cabrera, H.; Pérez-Yzquierdo, I. Rational approximation and Sobolev-type orthogonality. J. Approx.
Theory 2020, 260, 105481-1–105481-19. [CrossRef]
21. Díaz-González, A.; Pijeira-Cabrera, H.; Quintero-Roba, J. Polynomials of Least Deviation from Zero in Sobolev p-Norm. Bull.
Malays. Math. Sci. Soc. 2022, 45, 889–912. [CrossRef]
22. Öchsner, A.; Makvandi, R. Numerical Engineering Optimization. Application of the Computer Algebra System Maxima; Springer: Cham,
Switzerland, 2020.

Disclaimer/Publisher’s Note: The statements, opinions and data contained in all publications are solely those of the individual
author(s) and contributor(s) and not of MDPI and/or the editor(s). MDPI and/or the editor(s) disclaim responsibility for any injury to
people or property resulting from any ideas, methods, instructions or products referred to in the content.

110
 mathematics

Article
A Look at Generalized Degenerate Bernoulli and Euler Matrices
Juan Hernández 1, *, Dionisio Peralta 1 and Yamilet Quintana 2,3

1 Escuela de Matemáticas, Facultad de Ciencias, Universidad Autónoma de Santo Domingo,

Santo Domingo 10105, Dominican Republic; [email protected]
2 Departamento de Matemáticas, Universidad Carlos III de Madrid, Avenida de la Universidad 30,
Leganés, 28911 Madrid, Spain; [email protected]
3 Instituto de Ciencias Matemáticas (ICMAT), Campus de Cantoblanco UAM, 28049 Madrid, Spain
* Correspondence: [email protected]

Abstract: In this paper, we consider the generalized degenerate Bernoulli/Euler polynomial matrices
and study some algebraic properties for them. In particular, we focus our attention on some matrix-
inversion formulae involving these matrices. Furthermore, we provide analytic properties for the
so-called generalized degenerate Pascal matrix of the first kind, and some factorizations for the
generalized degenerate Euler polynomial matrix.

Keywords: generalized degenerate Bernoulli polynomials; generalized degenerate Euler polyno-

mials; generalized degenerate Bernoulli matrix; generalized degenerate Euler matrix; generalized
degenerate Pascal matrix

MSC: 33E20; 11B83; 11B68

1. Introduction
Matrices play an important role in all branches of science, engineering, social science,
and management. In many settings (see, e.g., [1–4] and the references therein), a number of
interesting and useful identities involving binomial (q-binomial or λ-binomial) coefficients
can be obtained from a matrix representation of a particular counting sequence. Such a
Citation: Hernández, J.; Peralta, D.; matrix representation provides a powerful computational tool for deriving identities and
Quintana, Y. A Look at Generalized an explicit formula related to the sequence.
Degenerate Bernoulli and Euler There are many special types of matrices such as Pascal, Vandermonde, Stirling,
Matrices. Mathematics 2023, 11, 2731. Riordan arrays, and others. These matrices are of specific importance in many scientific
https://doi.org/10.3390/ and engineering applications. For instance, Pascal matrices appear in combinatorics, image
math11122731 processing, signal processing, numerical analysis, probability, and surface reconstruction.
Academic Editor: Sitnik Sergey In the case of generalized Pascal matrices of the first kind, extensive research has
been devoted to them (cf., e.g., [3–10] and the references therein). Situations with a matrix
Received: 24 May 2023 representation—including analogs of generalized Pascal matrices of the first kind and
Revised: 14 June 2023 degenerate versions of special classes of polynomials (e.g., Bernstein, Bernoulli, and Euler
Accepted: 14 June 2023
polynomials, etc.)—are of particular interest.
Published: 16 June 2023
Motivated by recent articles [1–4,11–14] that consider degenerate Bernstein polynomi-
als, degenerate Euler polynomials, generalized degenerate Euler–Genocchi polynomials
of order α, and algebraic properties of the generalized Euler and generalized Apostol-
Copyright: © 2023 by the authors.
type polynomial matrices, in the present article, we consider the generalized degenerate
Licensee MDPI, Basel, Switzerland. Bernoulli/Euler polynomial matrix. In particular, we focus our attention on some inversion-
This article is an open access article type formulae from a matrix framework. Furthermore, we show some analytic properties
distributed under the terms and for the so-called generalized degenerate Pascal matrix of the first kind. Furthermore, some
conditions of the Creative Commons factorizations for the generalized degenerate Euler polynomial matrix in terms of such a
Attribution (CC BY) license (https:// matrix are given.
creativecommons.org/licenses/by/ The paper is organized as follows. Section 2 is a preliminary section containing the
4.0/). definitions, notations, and terminology needed. Section 3 contains the main results of this

Mathematics 2023, 11, 2731. https://doi.org/10.3390/math11122731 111 https://www.mdpi.com/journal/mathematics

Mathematics 2023, 11, 2731

paper. First, we provide the corresponding inversion-type formulae for the degenerate
Bernoulli and Euler polynomials, respectively (Theorems 1 and 2). Second, we show that the
generalized degenerate Pascal matrix of the first kind is a matrix exponential (Theorem 4),
and, as a consequence, we obtain an Appell-type property for this matrix (Corollary 5). In
addition, factorizations for the generalized degenerate Pascal matrix of the first kind in
terms of the degenerate Bernoulli/Euler matrices are given (Theorems 6 and 7, respectively).
The remainder of this section is devoted to establishing the corresponding product formulae
for generalized degenerate Euler polynomial matrices and their factorizations in terms of
generalized degenerate Pascal matrices of the first kind (Theorems 8 and 9).

2. Background and Previous Results

Throughout this paper, let N, N0 , Z, R, and C denote, respectively, the set of all natural
numbers, the set of all non-negative integers, the set of all integers, the set of all real
numbers, and the set of all complex numbers. As usual, we will always use the principal
branch for complex powers, in particular, 1α = 1 for α ∈ C. Furthermore, the convention
00 = 1 will be adopted.
For w ∈ C and k ∈ Z, we use the notations w(k) and (w)k for the rising and falling
factorials, respectively, i.e.,
⎧
⎪
⎨ 1, if k = 0,
w(k) = ∏ik=1 (w + i − 1), if k ≥ 1,
⎪
⎩
0, if k < 0,

and ⎧
⎪
⎨ 1, if k = 0,
(w)k = ∏ k ( w − i + 1), if k ≥ 1,
⎪ i =1
⎩
0, if k < 0.
Any matrix is assumed an element of Mn+1 (R), the set of all (n + 1)-square matri-
ces over the real field R. Moreover, for i, j, any nonnegative integers, and any matrix
A ∈ Mn+1 (R) we adopt, respectively, the following conventions

i
= 0, whenever j > i, and A0 = In+1 = diag(1, 1, . . . , 1),
j

where In+1 denotes the identity matrix of order n + 1.

For λ, x ∈ R and z ∈ C, the degenerate exponentials are defined as follows (cf., [15]):
⎧ x
⎨ (1 + λz) λ , if λ ∈ R \ {0},
eλx (z) = (1)
⎩
e xz , if λ = 0.

As usual, for x = 1, we use the notation eλ (z) = eλx (z).

It follows immediately from (1) that
⎧ ∞
⎪
⎪ zn
⎪
⎪
⎪ ∑ (x)n,λ n! , |λz| < 1, if λ ∈ R \ {0},
⎨ n =0
eλx (z) = (2)
⎪
⎪ ∞
⎪
⎪ nz
n
⎩ ∑ x n! , if λ = 0.
⎪
n =0

where the generalized falling factorials ( x )n,λ , are given by (cf., [1,2,12–15]):

112
Mathematics 2023, 11, 2731

⎧
⎪
⎨ 1, if n = 0,
( x )n,λ = ∏in=1 ( x − (i − 1)λ), if n ≥ 1,
⎪
⎩
0, if n < 0,
where x, λ ∈ R and n ∈ Z.
It is clear that lim eλx (z) = e0x (z) = e xz , and for n ∈ N0 , the polynomial in two variables
λ →0
Qn ( x, λ), given by

1, if n = 0,
Qn ( x, λ) =
∏in=1 ( x − ( i − 1) λ ), if n ≥ 1,

is a continuous function on R2 , and consequently, ( x )n,0 = x n .

In [16,17], Carlitz introduced the degenerate Bernoulli (Euler) and the generalized
degenerate Bernoulli (Euler) polynomials of order α ∈ C, respectively, by means of the
generating functions and series expansions:
∞
z zn
e x (z)
eλ ( z ) − 1 λ
= ∑ Bn,λ (x) n! , (3)
n =0
∞
2 zn
e x (z)
eλ ( z ) + 1 λ
= ∑ En,λ (x) n! , (4)
n =0
 α ∞
z (α) zn
eλ ( z ) − 1
eλx (z) = ∑ Bn,λ (x) n! , (5)
n =0
 α ∞
2 (α) zn
eλ ( z ) + 1
eλx (z) = ∑ En,λ (x) n! . (6)
n =0

These are valid in a suitable neighborhood of z = 0 and represent degenerate versions

of the classical Bernoulli and Euler polynomials, respectively. In [8], the notation β n (λ, x )
is used for the degenerate Bernoulli (3).
Since the degenerate exponentials (1) satisfy the same exponent product law as the
exponentials functions, i.e.,
x +y y
eλ (z) = eλx (z) eλ (z),
we can use the generating relations (2), (5) and (6) to deduce the following addition
formulas:
n  
n
( x + y)n,λ = ∑ ( x )k,λ (y)n−k,λ , n ≥ 0, (7)
k =0
k
n  
(α+ β) n (α) ( β)
Bn,λ ( x + y) = ∑ Bk,λ ( x )Bn−k,λ (y), n ≥ 0, (8)
k =0
k
n  
(α+ β) n (α) ( β)
En,λ ( x + y) = ∑ E ( x )En−k,λ (y), n ≥ 0. (9)
k =0
k k,λ

For a treatment of diverse aspects of some summation formulas and their applications,
the interested reader is referred to the relatively recent works [18–20].

113
Mathematics 2023, 11, 2731

For r ∈ N0 , λ ∈ R, and α ∈ C, definitions of generalized degenerate Euler–Genocchi

and generalized degenerate Euler–Genocchi polynomials of order α, respectively, have
recently been introduced in [14] (Section 2):
∞
2zr (r ) zn
e x (z)
eλ ( z ) + 1 λ
= ∑ An,λ (x) n! , (10)
n =0
 α ∞
2 (r,α) zn
zr
eλ ( z ) + 1
eλx (z) = ∑ An,λ (x)
n!
. (11)
n =0

Remark 1. Notice that:

(i) If r ∈ N, then it follows immediately from (2), (4) and (10), that
(r ) (r ) (r )
A0,λ ( x ) = A1,λ ( x ) = · · · = Ar−1,λ ( x ) = 0, and

(r ) n! (0)
An,λ ( x ) = ( x ) = n(r) En−r,λ ( x ), n ≥ r.
(n − r )! n,λ
(0)
Furthermore, An,λ ( x ) = En,λ ( x ), n ≥ 0.
The first above identities guarantee that, up to multiplicative constants, it suffices to take
generalized degenerate Euler polynomials of order 0 instead of the so-called generalized
degenerate Euler–Genocchi polynomials as the main family to study. Similarly, the last identity
tells us that the generalized degenerate Euler polynomials coincides with the generalized
degenerate Euler–Genocchi polynomials of order 0.
(ii) In [14], Theorem 4 proves the following reduction formula:

(r,α) (α)
An,λ ( x ) = n(r) En−r,λ ( x ), n ≥ r, n, r ∈ N0 .

In particular, we obtain that up to multiplicative constants, the generalized degenerate Euler–

Genocchi polynomials of order α = 1 can be reduced to the generalized degenerate Euler
polynomials (4).
Hence, in order to avoid essentially redundant definitions (cf., [21]), the families of polynomials
eqrefeul-gen1 and (11) will not be considered in this paper.

3. The Generalized Degenerate Bernoulli and Euler Matrices and Their Properties
In this section, we present some novel properties for the generalized degenerate
Bernoulli and Euler matrices. Before that, we show the corresponding inversion-type
formulae for the generalized degenerate Bernoulli and Euler polynomials, respectively.

Theorem 1. For every n ≥ 0 and λ ∈ R, the degenerate Bernoulli polynomials satisfy the following
inversion-type formula:
n  
1 n+1
n + 1 k∑
( x )n,λ = (1)k+1,λ Bn−k,λ ( x ) (12)
=0
k+1
n  
1 n+1
= ∑
n + 1 k =0 k + 1
(1 − λ)k,λ Bn,λ ( x ). (13)

Proof. Let λ ∈ R. In view of (2) and (3), and the identity

∞ ∞
zn z n +1
z ∑ (x)n,λ n! = ∑ (n + 1)(x)n,λ (n + 1)! ,
n =0 n =0

114
Mathematics 2023, 11, 2731

we have
  
∞ ∞ ∞
z n +1 zn zn
∑ (n + 1)(x)n,λ ( n + 1) !
= ∑ (1)n,λ n!
−1 ∑ Bn,λ ( x )
n!
n =0 n =0 n =0
  
∞ ∞
z n +1 zn
= ∑ (1)n+1,λ ( n + 1) ! ∑ Bn,λ ( x )
n!
. (14)
n =0 n =0

From the use of the Cauchy product rule on the right-hand side of (14), it follows that
 
∞ ∞ n  
z n +1 n+1 z n +1
∑ (n + 1)(x)n,λ (n + 1)! = ∑ ∑ k + 1 (1)k+1,λ Bn−k,λ (x) (n + 1)! . (15)
n =0 n =0 k =0

Hence, comparing the coefficients of zn+1 on both sides of (15), we obtain (12).
Finally, (13) is a simple consequence of the identity (1)k+1,λ = (1 − λ)k,λ , for all
k ∈ N0 .

Remark 2. Notice that the substitution of λ = 0 into (12) recovers the inversion formula for the
classical Bernoulli polynomials (cf., [22] (Equation (9))).

From a matrix framework, Theorem 1 has the following consequence.

T
Corollary 1. For n ∈ N0 and λ ∈ R, the matrix Tλ ( x ) = 1 ( x )1,λ · · · ( x )n,λ can be
expressed as follows:

Tλ ( x ) = Mλ Bλ ( x )
⎛ ⎞
(11)(1)1,λ 0 0 ··· 0
⎜ 1 2 1 2 ⎟
⎜ 2 (2)(1)2,λ 2 (1)(1)1,λ 0 ··· 0 ⎟
⎜ 1 3 1 3 1 3 ⎟
= ⎜
⎜ 3 (3)(1)3,λ 3 (2)(1)2,λ 3 (1)(1)1,λ ··· 0 ⎟Bλ ( x )
⎟
⎜ .. .. .. .. .. ⎟
⎝ . . . . . ⎠
1 n +1 1 n +1 1 n +1 1 n +1
n+1 (n+1)(1)n+1,λ n+1 ( n )(1)n,λ n+1 (n−1)(1)n−1,λ ··· n+1 ( 1 )(1)1,λ
⎛ ⎞
1 0 0 ··· 0
⎜ 2 (1)2,λ
1
1 0 · · · 0⎟
⎜ ⎟
⎜ · · · 0⎟
3 (1)3,λ (1)2,λ
1
= ⎜ 1 ⎟ B λ ( x ), (16)
⎜ .. .. .. .. .⎟
⎝ . . . . .. ⎠
n+1 )n+1,λ
( (1)n,λ 2 (1)n−1,λ ··· 1
1 1
1

T
where Bλ ( x ) = B0,λ ( x ) B1,λ ( x ) · · · Bn,λ ( x ) .

Theorem 2. For every n ≥ 0 and λ ∈ R. The degenerate Euler polynomials satisfy the following
inversion-type formula:
 
1 n n
2 k∑
( x )n,λ = (1 + ak (λ))(1)k,λ En−k,λ ( x ) (17)
=0
k

where 
1, if k = 0,
ak (λ) =
0, if 1 ≤ k ≤ n,

Proof. From (2) and (4) we have

115
Mathematics 2023, 11, 2731

  
∞ ∞ ∞
zn zn zn
2 ∑ (x)n,λ n!
= ∑ (1)n,λ n!
+1 ∑ En,λ ( x )
n!
n =0 n =0 n =0
  
∞ ∞
zn zn
= ∑ (1 + ak (λ))(1)n,λ n! n∑
E n,λ ( x )
n!
n =0 =0
   
∞ n
n zn
= ∑ ∑ ( 1 + a k ( λ ))
k
( 1 ) E
k,λ n−k,λ ( x )
n!
,
n =0 k =0

where 
1, if k = 0,
ak (λ) =
0, if 1 ≤ k ≤ n.
Therefore, by comparing the coefficients of zn on both sides, we obtain the identity.

Remark 3. Notice that if λ = 0 in (17), then we recover the inversion formula for the classical
Euler polynomials (cf., [22] (Equation (27))).

Theorem 2 has the following consequence.

T
Corollary 2. For n ∈ N0 and λ ∈ R, the matrix Tλ ( x ) = 1 ( x )1,λ · · · ( x )n,λ can be
expressed as follows:

1
Tλ ( x ) = N E (x)
2 ⎛λ λ ⎞
(00)(1 + a0 (λ))(1)0,λ 0 ··· 0
⎜ 1 ⎟
⎜ (1)(1 + a1 (λ))(1)1,λ (10)(1 + a0 (λ))(1)0,λ ··· 0 ⎟
1⎜⎜ (22)(1 + a2 (λ))(1)2,λ (21)(1 + a1 (λ))(1)1,λ ···
⎟
⎟Eλ ( x )
= 0
2⎜⎜ .. .. .. ..
⎟
⎟
⎝ . . . . ⎠
(nn)(1 + an (λ))(1)n,λ n
(n− 1)(1 + an−1 ( λ ))(1)n−1,λ · · · (n0 )(1 + a0 (λ))(1)0,λ
⎛ ⎞
2 0 0 0 ··· 0
⎜ (1) 2 0 0 ··· 0⎟
⎜ 1,λ ⎟
1⎜
⎜ (1)2,λ 2(1)1,λ 2 0 ··· 0⎟ ⎟ E λ ( x ),
= (18)
2⎜
⎜ .. .. .. .. .. .. ⎟
⎟
⎝ . . . . . .⎠
( n )2 ( n )3
(1)n,λ n(1)n−1,λ 2! (1)n−2,λ 3! (1)n−3,λ ··· 2

T 1, if k = 0,
where Eλ ( x ) = E0,λ ( x ) E1,λ ( x ) · · · En,λ ( x ) and ak (λ) =
0, if 1 ≤ k ≤ n.

Clearly, when λ ∈ R, the matrix Nλ is an invertible matrix.

Corollary 3. For n ∈ N0 and λ ∈ R, we have

E λ ( x ) = 2( N λ ) −1 M λ B λ ( x ).

The degenerate Pascal matrices corresponding to the generalized falling factorials can
be defined as follows:

116
Mathematics 2023, 11, 2731

Definition 1. Let x be any nonzero real number. For λ ∈ R, the generalized degenerate Pascal
matrix of the first kind Pλ [ x ], is an (n + 1) × (n + 1) matrix whose entries are given by
⎧
⎪ i
⎨ ( j)( x )i− j,λ , i ≥ j,
pi,j,λ ( x ) := (19)
⎪
⎩ 0, otherwise.

Remark 4.
(i) It is clear that the matrix Pλ [ x ] tends to the generalized Pascal matrix of the first kind P[ x ] as
λ → 0.
(ii) For n ∈ N0 , x ∈ R \ {0}, λ ∈ R, it is clear that P−λ [ x ] = Pn,λ [ x ], where Pn,λ [ x ] is the
Pascal functional matrix introduced in [5]. Hence, all results corresponding to P−λ [ x ] given
in [5] hold in this setting.
(iii) It is worth mentioning that the matrix entries (19) coincide with the entries of the variation
of Pascal functional matrix Pn [ x, λ] introduced by Can and Cihat-Dağli in [8]. Hence, all
results corresponding to factorizing the matrix Pn [ x, λ] by the summation matrices also hold
for Pλ [ x ], taking into account the suitable shift on the respective order for these matrices (cf.,
[8] (Lemma 1 and Theorem 2)).
(iv) If for x ∈ R \ {0}, λ ∈ R we consider the truncated exponential generating function for the
binomial-type polynomial sequence {( x )n,λ }n≥0 (cf., [9]):
n
tk
f (t; x ) = ∑ (x)k,λ k! ,
k =0

then, it is easy to see that

 
n
tk 

Pλ [ x ] = Pn [ f ( x, t)]|t=0 = Pn ∑ (x)k,λ k! 

,
k =0 t =0

where Pn [ f (t; x )] denotes the generalized Pascal functional matrix introduced by Yang and
Micek in [9].

From now on, we denote Pλ = Pλ [1]. The following theorem summarizes some
properties of Pλ [ x ].

Theorem 3. Let Pλ [ x ] ∈ Mn+1 (R) be the generalized degenerate Pascal matrix of the first kind.
Then, the following statements hold.
(a) Special value. If the convention (0)0,λ = 1 is adopted, then it is possible to define

Pλ [0] := In+1 .

(b) For x, y ∈ R, we have

Pλ [ x + y] = Pλ [ x ] Pλ [y]. (20)
(c) Pλ [ x ] is an invertible matrix and its inverse is given by

Pλ−1 [ x ] := ( Pλ [ x ])−1 = Pλ [− x ]. (21)

Proof. Since part (a) is a straightforward consequence of the extension of Definition 1 for
the case x = 0, we shall omit its proof. Thus, we focus our efforts on the proof of parts (b)
and (c).

117
Mathematics 2023, 11, 2731

Let Ai,j,λ ( x, y) be the (i, j)-th entry of the matrix product Pλ [ x ] Pλ [y]. Then, by (7),
we have
n    
i k
Ai,j,λ ( x, y) = ∑ ( x )i−k,λ (y)k− j,λ
k =0
k j
i    
i k
= ∑ ( x )i−k,λ (y)k− j,λ
k= j
k j
i   
i i−j
= ∑ ( x )i−k,λ (y)k− j,λ
k= j
j i−k
  i− j  
i i−j
j k∑
= ( x )i− j−k,λ (y)k,λ
=0
k

i
= ( x + y)i− j,λ ,
j

which implies (20).

The substitution y = − x into (20) yields

Pλ [0] = Pλ [ x ] Pλ [− x ] = Pλ [− x ] Pλ [ x ].

By part (a), we have Pλ [0] = In+1 , thus

Pλ [ x ] Pλ [− x ] = In+1 = Pλ [− x ] Pλ [ x ],

and (21) follows.

Corollary 4. For any λ ∈ R, r ∈ Z and s ∈ Z \ {0} we have

(a) Pλr = Pλ [r ].
  s
(b) Pλ rs = Pλr .

Proof. Making the corresponding modifications, we apply the same reasoning as in the
proof of [7] (Corollary 3). Since Pλ = Pλ [1], Pλ [0], and Pλ0 coincide with the identity matrix,
it follows from Theorem 3, by induction on r, that Pλ [r ] = Pλr , for all r ∈ N0 . Again, by
Theorem 3, we have that Pλ [−1] = Pλ−1 , and a similar induction on |r | shows Pλ [r ] = Pλr ,
for all r < 0.   s
Next, by Theorem 3 and part (a), we obtain Pλ rs = Pλ [r ] = Pλr .

Remark 5. Part (b) of Corollary 4 shows that for a fixed λ ∈ R and any rational number x, Pλ [ x ]
is the x-th power of Pλ . Indeed, this property could be expected in the sense that it is satisfied for the
generalized Pascal matrix of the first kind P[ x ] (cf., [7]).

From the addition Formula (20), we proceed according to [7] and conclude that the
degenerate Pascal matrix Pλ [ x ] has an exponential form as follows: Assume that for λ ∈ R,
there is a matrix Lλ , such that Pλ [ x ] = e xLλ . Then,

d
P [ x ] = Lλ e xLλ = Lλ Pλ [ x ],
dx λ
and 
d 
Pλ [ x ] = Lλ Pλ [0] = Lλ In+1 = Lλ .
dx x =0

118
Mathematics 2023, 11, 2731

Thus, there is at most one matrix Lλ such that Pλ [ x ] = e xLλ . For instance, in the case
n = 3, we can find the only possible value as follows:
⎡ ⎤
0 0 0 0
d ⎢ 1 0 0 0 ⎥
⎢
P [x] = ⎣ ⎥,
dx λ −λ + 2x 2 0 0 ⎦
x (−2λ + x ) + x (−λ + x ) + (−2λ + x )(−λ + x ) 3(−λ + 2x ) 3 0

and ⎡ ⎤
 0 0 0 0
d  ⎢ 1 0 0 0 ⎥
Lλ = Pλ [ x ] =⎢
⎣ −λ
⎥.
dx x =0 2 0 0 ⎦
2λ2 −3λ 3 0
While, in the case n = 7, we have
⎡ ⎤
0 0 0 0 0 0 0 0
⎢ 1 0 0 0 0 0 0 0 ⎥
⎢ ⎥
⎢ −λ ⎥
 ⎢ 2 0 0 0 0 0 0 ⎥
 ⎢ ⎥
d ⎢ 2λ2 −3λ 3 0 0 0 0 0 ⎥
Lλ = Pλ [ x ] =⎢ ⎥.
dx x =0 ⎢ − 6λ 3 8λ2 −6λ 4 0 0 0 0 ⎥
⎢ ⎥
⎢ 24λ 4 −30λ3 20λ2 −10λ 5 0 0 0 ⎥
⎢ ⎥
⎣ −120λ5 144λ4 −90λ3 40λ2 −15λ 6 0 0 ⎦
720λ6 −840λ5 504λ4 −210λ3 70λ2 −21λ2 7 0

This suggests a general way of choosing Lλ . More precisely, the entries of Lλ are
given by
⎧ i
⎨sλ (i − j, 1)( j), if i ≥ j + 1,
⎪
( Lλ )i,j =
⎪
⎩
0, otherwise,
where sλ (n, k ) denotes the degenerate Stirling number of the first kind, defined as follows
(cf., [17,23] or [24] (Ch. 5)):
n
∑ sλ (n, k)xk = (x)n,λ . (22)
k =0

Furthermore, the entries of the matrix Lkλ , for 1 ≤ k ≤ n and n ∈ N can be explicitly
represented as follows.

Lemma 1. For every n ∈ N and 1 ≤ k ≤ n, the entries of Lkλ are given by the formula
⎧ i
  ⎨k!sλ (i − j, k)( j),
⎪ if i ≥ j + k,
Lkλ =
i,j ⎪
⎩
0, otherwise,

where sλ (n, k) is the degenerate Stirling number of the first kind (22).

Proof. It suffices to proceed by induction on k, taking into account that for k > n, we have
Lkλ = 0.

Theorem 4. For every real numbers x, λ ∈ R, Pλ [ x ] = e xLλ .

119
Mathematics 2023, 11, 2731

Proof. By part (a) of Theorem 3, if x = 0, then e xLλ = In+1 = Pλ [ x ]. Now, assume that
x = 0 since Lkλ = 0 for k > n, the infinite series for e xLλ reduces to the finite sum

x2 2 xn n
e xLλ = In+1 + xLλ + Lλ + · · · + L . (23)
2 n! λ

Applying Lemma 1, we can now read off the entries in e xLλ . Clearly, it is a lower
triangular matrix, and the diagonal entries are all 1. Now suppose i > j, and let 0 ≤ k ≤
i − j. Then, using (22), we have that the (i, j)-th entry in the sum (23) is

    i− j 
xk  k 
i− j
i i
e xLλ
i,j
= ∑ L
k! λ i,j
= ∑
j k =0
sλ (i − j, k) x k =
j
( x )i− j,λ = pi,j,λ ( x ).
k =0

This completes the proof.

As a consequence of Lemma 1 and Theorem 4, we obtain the following Appell-

type property.

Corollary 5. The generalized degenerate Pascal matrix of the first kind Pλ [ x ] satisfies the following
differential equations:
Dxk Pλ [ x ] = Lkλ Pλ [ x ], 1 ≤ k ≤ n, (24)
where Dxk Pλ [ x ] is the matrix resulting from the k-th derivative with respect to x of each entry of
Pλ [ x ].

(α)
Definition 2. The generalized degenerate (n + 1) × (n + 1) Bernoulli matrix Bλ ( x ) of (real or
complex) order α is defined by the entries
⎧
⎪ i (α)
⎨ ( j)Bi− j,λ ( x ), i ≥ j,
(α)
Bi,j,λ ( x ) =
⎪
⎩
0, otherwise.

Remark 6.
(α)
(i) It is worth mentioning that the entries (2) of Bλ ( x ) coincide with the entries of the general-
(α)
ized degenerate Bernoulli matrix Bm [λ, x ] introduced in [8], when these matrices are the
same order.
(1)
(ii) We denote by Bλ ( x ) the degenerate Bernoulli matrix Bλ ( x ).

The following result was established in [8] (Theorem 4).

(α)
Theorem 5. The generalized degenerate Bernoulli matrices Bλ ( x ) satisfy the following product
formulas.

(α+ β) (α) ( β) ( β) (α)

Bλ ( x + y ) = Bλ ( x ) Bλ ( y ) = Bλ ( x ) Bλ ( y )
(α) ( β)
= Bλ ( y ) Bλ ( x ) . (25)

Definition 2 and the inversion-type Formula (12) lead to the following result:

Theorem 6. The generalized degenerate Pascal matrix of the first kind Pλ [ x ] can be factorized in
terms of Bλ ( x ) as follows:
Pλ [ x ] = Bλ ( x )Hλ , (26)

120
Mathematics 2023, 11, 2731

where Hλ is an (n + 1) × (n + 1) invertible matrix with entries

⎧  
⎪ i (1)i− j+1,λ
⎪
⎨ , i ≥ j,
i−j i−j+1
Hi,j,λ =
⎪
⎪
⎩
0, otherwise.

Proof. Let us consider n ∈ N0 and 0 ≤ i, j ≤ n such that i ≤ j. From Definition 2 and the
inversion-type Formula (12), we have

i
 i− j  
i j i−j+1
i − j + 1 k∑
pi,j,λ ( x ) = ( x )i− j,λ = (1)k+1,λ Bi− j−k,λ ( x )
j =0
k+1
i − j #  $#  $
i−j i (1)k+1,λ
= ∑ Bi− j−k,λ ( x ) . (27)
k =0
k i−j k+1

Since the right hand member of (27) is the (i, j)-th entry of matrix product Bλ ( x )Hλ ,
we conclude that (26) holds.

The following example shows the validity of Theorem 6.

Example 1. Let us consider n = 2. It follows from Definition 1, (26), and a simple computation that
⎡ ⎤
(00)( x )0,λ 0 0
⎢ ⎥
⎢ ⎥
⎢ ⎥
Pλ [ x ] = ⎢ (10)( x )1,λ (11)( x )0,λ 0 ⎥
⎢ ⎥
⎣ ⎦
2 2 2
(0)( x )2,λ (1)( x )1,λ (2)( x )0,λ
⎡ ⎤⎡ 0 ⎤
(0)B0,λ ( x ) 0 0 (0)(1)1,λ 0 0
⎢ 0 ⎥⎢ ⎥
⎢ ⎥⎢ ⎥
⎢ 1 ⎥ ⎢ 1 (1)2,λ ⎥
= ⎢ (0)B1,λ ( x ) (1)B0,λ ( x )
1
0 ⎥ ⎢ (1) 2 1
(0)(1)1,λ 0 ⎥
⎢ ⎥⎢ ⎥
⎣ ⎦⎣ ⎦
(20)B2,λ ( x ) (21)B1,λ ( x ) (22)B0,λ ( x ) (1) (1)
(22) 33,λ (21) 22,λ (20)(1)1,λ
- ./ 0- ./ 0
Bλ ( x ) Hλ

(α)
Definition 3. The generalized degenerate (n + 1) × (n + 1) Euler matrix Eλ ( x ) is defined by
the entries ⎧
⎪ i (α)
⎨ ( j)Ei− j,λ ( x ), i ≥ j,
(α)
Ei,j,λ ( x ) =
⎪
⎩
0, otherwise.
(1)
We denote by Eλ ( x ) the degenerate Euler matrix Eλ ( x ).

Definition 3 and the inversion-type Formula (17) lead to the following result:

Theorem 7. The generalized degenerate Pascal matrix of the first kind Pλ [ x ] can be factorized in
terms of Eλ ( x ) as follows:
Pλ [ x ] = Eλ ( x )Tλ , (28)

121
Mathematics 2023, 11, 2731

where Tλ is an (n + 1) × (n + 1) invertible matrix with entries

⎧  
⎪ i 1 + ai− j (λ) (1)i− j,λ
⎪
⎨ , i ≥ j,
Ti,j,λ = i − j 2
⎪
⎪
⎩
0, otherwise.

Proof. Let us consider n ∈ N0 and 0 ≤ i, j ≤ n such that i ≤ j. From Definition 3 and the
inversion-type Formula (17), we have
   i− j  
i 1 i i−j
pi,j,λ ( x ) =
j
( x )i− j,λ =
2 j k =0 ∑ k
(1 + ak (λ))(1)k,λ Ei− j−k,λ ( x )

i − j #  $#  $
i−j i (1 + ak (λ))(1)k,λ
= ∑ k
Ei− j−k,λ ( x )
j 2
. (29)
k =0

Since the right-hand member of (29) is the (i, j)-th entry of matrix product Eλ ( x )Tλ ,
we conclude that (28) holds.

Combining Theorems 6 and 7 gives the following connection formula.

Corollary 6. For any λ, x ∈ R, we have

Eλ ( x ) = Bλ ( x )Hλ Tλ−1 .

The next result is an immediate consequence of Definition 3 and the addition Formula (9).

(α)
Theorem 8. The generalized degenerate Euler matrices Eλ ( x ) satisfy the following product
formulas.

(α+ β) (α) ( β) ( β) (α)

Eλ ( x + y ) = Eλ ( x ) Eλ ( y ) = Eλ ( x ) Eλ ( y )
(α) ( β)
= Eλ ( y ) Eλ ( x ) . (30)

(α,β) (α) ( β)
Proof. Let Ci,j,λ ( x, y) be the (i, j)-th entry of the matrix product Eλ ( x ) Eλ (y), then, by
the addition Formula (9), we have
n    
(α,β) i (α) k ( β)
Ci,j,λ ( x, y) = ∑ Ei−k,λ ( x ) Ek− j,λ (y), n ≥ 0
k =0
k j
i   
i i−j (α) ( β)
= ∑ Ei−k,λ ( x )Ek− j,λ (y)
k= j
j i−k
  i− j  
i i−j (α) ( β)
=
j ∑ k
Ei− j−k,λ ( x )Ek,λ (y),
k =0
 
i (α+ β)
= Ei− j,λ ( x + y), for i ≥ j,
j

which implies the first equality of (30). The second and third equalities of (30) can be
derived in a similar way.

122
Mathematics 2023, 11, 2731

Corollary 7. Let ( x1 , . . . , xk ) ∈ Rk . For α j real or complex parameters, the generalized degenerate

(α)
Euler matrices Eλ ( x ) satisfy the following product formulas, j = 1, . . . , k.

k
(α1 +α2 +···+αk ) (α j )
Eλ ( x1 + x2 + · · · + x k ) = ∏ Eλ ( x j ).
j =1

Proof. The application of induction on k gives the desired result.

Taking x = x1 = x2 = · · · = xk and α = α1 = α2 = · · · = αk , we obtain the following

(α)
simple formula for the powers of the generalized degenerate Euler matrices Eλ ( x ).

(α)
Corollary 8. The generalized degenerate Euler matrices Eλ ( x ) satisfy the following identity.
 k
(α) (α)
Eλ ( x ) = Eλ (kx ), k ∈ N.

Remark 7. Analogously, the above corollaries hold, mutatis mutandis, for the generalized degen-
erate Bernoulli matrices. More precisely, from Theorem 5, and using the same assumptions as
Corollaries 7 and 8, we obtain
k
(α1 +α2 +···+αk ) (α j )
Bλ ( x1 + x2 + · · · + x k ) = ∏ Bλ ( x j ),
j =1
 k
(α) (α)
Bλ ( x ) = Bλ (kx ).

(α)
Theorem 9. The generalized degenerate Euler matrices Eλ ( x ) satisfy the following relations.

(α) (α) (α)

Eλ ( x + y ) = Eλ ( x ) Pλ [y] = Pλ [ x ] Eλ (y)
(α)
= Eλ (y) Pλ [ x ]. (31)

Proof. The substitution β = 0 into (30) yields

(α) (α) (0) (0) (α)

Eλ ( x + y ) = Eλ ( x ) Eλ ( y ) = Eλ ( x ) Eλ ( y )
(α) (0)
= Eλ ( y ) Eλ ( x ) .

(0)
Since Eλ ( x ) = Pλ [ x ], we obtain

(α) (α)
Eλ ( x + y) = Pλ [ x ]Eλ (y).

(α)
A similar argument allows us to show that E (α) ( x + y) = E (α) ( x ) Pλ [y] and Eλ ( x +
(α)
y) = Eλ (y) Pλ [ x ]. This completes the proof of (31).

4. Conclusions
The aim of our research was to determine novel properties of generalized degenerate
Bernoulli and Euler matrices. First, we focused our attention on some matrix-inversion
formulae involving these matrices. Secondly, we showed some analytic properties for the
generalized degenerate Pascal matrix of the first kind and provided some factorizations for
the generalized degenerate Euler polynomial matrix in terms of the generalized degenerate
Pascal matrix of the first kind.
Finally, it is worth mentioning that the use of the Cauchy product of the power series
is the technique behind some of our formulations. This approach is not a novelty; however,
it has been useful for generating new families of special polynomials (satisfying or not

123
Mathematics 2023, 11, 2731

Appell-type conditions), even very recently. In this regard, we refer the interested reader
to [25,26] and the references therein for a detailed exposition about very recent trends in
this broad field.

Author Contributions: Conceptualization, J.H. and Y.Q.; methodology, J.H., D.P. and Y.Q.; formal
analysis, J.H., D.P. and Y.Q.; investigation, J.H., D.P. and Y.Q.; writing—original draft preparation,
Y.Q.; writing—review and editing, J.H., D.P. and Y.Q.; supervision, Y.Q.; project administration, Y.Q.;
funding acquisition, J.H. and Y.Q. All authors have read and agreed to the published version of the
manuscript.
Funding: The research of J. Hernández has been partially supported by the Fondo Nacional de Inno-
vación y Desarrollo Científico y Tecnológico (FONDOCYT), Dominican Republic, under grant 2020-
2021-1D1-135. The research of Y. Quintana has been partially supported by the grant CEX2019-000904-
S funded by MCIN/AEI/10.13039/501100011033, and by the Madrid Government (Comunidad
de Madrid-Spain) under the Multiannual Agreement with UC3M in the line of Excellence of Uni-
versity Professors (EPUC3M23), in the context of the Fifth Regional Programme of Research and
Technological Innovation (PRICIT).
Data Availability Statement: Data sharing is not applicable to this article.
Acknowledgments: The authors would like to thank the reviewers for their valuable comments and
suggestions.
Conflicts of Interest: The authors declare no conflict of interest.

References
1. Kim, T.; Kim, D.S. Degenerate Bernstein polynomials. Rev. R. Acad. Cienc. Exactas Fís. Nat. Ser. A Math. RACSAM 2019, 113,
2913–2920. [CrossRef]
2. Kim, T.; Kim, D.S.; Hyeon, S.-H.; Park, J.-W. Some new formulas of complete and incomplete degenerate Bell polynomials. Adv.
Differ. Equ. 2021, 2021, 326. [CrossRef]
3. Quintana, Y.; Ramírez, W.; Urieles, A. Generalized Apostol-type polynomial matrix and its algebraic properties. Math. Rep. 2019,
21, 249–264.
4. Quintana, Y.; Ramírez, W.; Urieles, A. Euler matrices and their algebraic properties revisited. Appl. Math. Inf. Sci. 2020, 14,
583–596. [CrossRef]
5. Bayat, M.; Teimoori, H. The linear algebra of the generalized Pascal functional matrix. Linear Algebra Appl. 1999, 295, 81–89.
[CrossRef]
6. Brawer, R.; Pirovino, M. The linear algebra of the Pascal matrix. Linear Algebra Appl. 1992, 174, 13–23. [CrossRef]
7. Call, G.S.; Velleman, D.J. Pascal’s Matrices Am. Math. Mon. 1993, 100, 372–376. [CrossRef]
8. Can, M.; Cihat-Dağli, M. Extended Bernoulli and Stirling matrices and related combinatorial identities. Linear Algebra Appl. 2014,
444, 114–131. [CrossRef]
9. Yang, Y.; Micek, C. Generalized Pascal functional matrix and its applications. Linear Algebra Appl. 2007, 423, 230–245. [CrossRef]
10. Zhang, Z. The linear algebra of the generalized Pascal matrix. Linear Algebra Appl. 1997, 250, 51–60. [CrossRef]
11. Kim, T.; Kim, D.S. Identities involving degenerate Euler numbers and polynomials arising from non-linear differential equations.
J. Nonlinear Sci. Appl. 2016, 9, 2086–2098. [CrossRef]
12. Kim, T.; Kim, D.S. Identities of symmetry for degenerate Euler polynomials and alternating generalized falling factorial sums.
Iran. J. Sci. Technol. Trans. Sci. 2017, 41, 939–949. [CrossRef]
13. Kim, T.; Kim, D.S.; Jang, L.-C.; Lee, H.; Kim, H. Representations of degenerate Hermite polynomials. Adv. Appl. Math. 2022, 139,
102359. [CrossRef]
14. Kim, T.; Kim, D.S.; Kim, H.K. On generalized degenerate Euler-Genocchi polynomials. Appl. Math. Sci. Eng. 2022, 31, 2159958.
[CrossRef]
15. Kim, T.; Kim, D.S. On some degenerate differential and degenerate difference operators. Russ. J. Math. Phys. 2022, 29, 37–46.
[CrossRef]
16. Carlitz, L. A degenerate Staudt-Clausen theorem. Arch. Math. 1956, 7, 28–33. [CrossRef]
17. Carlitz, L. Degenerate Stirling, Bernoulli and Eulerian numbers. Util. Math. 1979, 15, 51–88.
18. Chandragiri, S.; Shishkina, O.A. Generalized Bernoulli numbers and polynomials in the context of the Clifford analysis. J. Sib.
Fed. Univ.-Math. Phys. 2018, 11, 127–136.
19. Grigoriev, A.A.; Leinartas, E.K.; Lyapin, A.P. Summation of functions and polynomial solutions to a multidimensional difference
equation. J. Sib. Fed. Univ.-Math. Phys. 2023, 16, 153–161.
20. Leinartas, E.K.; Shishkina, O.A. The discrete analog of the Newton-Leibniz formula in the problem of summation over simplex
lattice points. J. Sib. Fed. Univ.-Math. Phys. 2019, 12, 503–508. [CrossRef]

124
Mathematics 2023, 11, 2731

21. Navas, L.; Ruiz, F.J.; Varona, J.L. Existence and reduction of generalized Apostol-Bernoulli, Apostol-Euler and Apostol-Genocchi
polynomials. Arch. Math. 2019, 55, 157–165. [CrossRef]
22. Nørlund, N.E. Vorlesungen über Differenzenrechnung; Springer: Berlin, Germany, 1924.
23. Howard, F.T. Degenerate weighted Stirling numbers. Discrete Math. 1985, 57, 45–58. [CrossRef]
24. Sándor, J.; Crstici, B. Handbook of Number Theory II; Kluwer Academic Publishers: Dordrecht, The Netherlands, 2004.
25. Albosaily, S.; Quintana, Y.; Iqbal, A.; Khan, W. Lagrange-based hypergeometric Bernoulli polynomials. Symmetry 2022, 14, 1125. .
[CrossRef]
26. Quintana, Y. Generalized mixed type Bernoulli-Gegenbauer polynomial. Kragujev. J. Math. 2023, 47, 245–257. [CrossRef]

Disclaimer/Publisher’s Note: The statements, opinions and data contained in all publications are solely those of the individual
author(s) and contributor(s) and not of MDPI and/or the editor(s). MDPI and/or the editor(s) disclaim responsibility for any injury to
people or property resulting from any ideas, methods, instructions or products referred to in the content.

125
 mathematics

Article
Sequentially Ordered Sobolev Inner Product and
Laguerre–Sobolev Polynomials
Abel Díaz-González 1 , Juan Hernández 2 and Héctor Pijeira-Cabrera 3, *

1 Department of Computer Science, Vanderbilt University, Nashville, TN 37240, USA;

[email protected]
2 Escuela de Matemáticas, Universidad Autónoma de Santo Domingo,
Santo Domingo 10105, Dominican Republic; [email protected]
3 Departamento de Matemáticas, Universidad Carlos III de Madrid, 28911 Leganés, Madrid, Spain
* Correspondence: [email protected]

Abstract: We study the sequence of polynomials {Sn }n≥0 that are orthogonal with respect to the gen-
) dj ( k ) ( c ) g ( k ) ( c ),
eral discrete Sobolev-type inner product f , g s = f ( x ) g( x )dμ( x ) + ∑ N j=1 ∑k=0 λ j,k f j j
where μ is a finite Borel measure whose support supp(μ) is an infinite set of the real line, λ j,k ≥ 0,
and the mass points ci , i = 1, . . . , N are real values outside the interior of the convex hull of supp(μ)
(ci ∈ R \ Ch(supp(μ))◦ ). Under some restriction of order in the discrete part of ·, · s , we prove that
Sn has at least n − d∗ zeros on Ch(supp(μ))◦ , being d∗ the number of terms in the discrete part of
·, · s . Finally, we obtain the outer relative asymptotic for {Sn } in the case that the measure μ is the
classical Laguerre measure, and for each mass point, only one order derivative appears in the discrete
part of ·, · s .

Keywords: orthogonal polynomials; Sobolev orthogonality; zeros location; asymptotic behavior

MSC: 41A60; 42C05; 33C45; 33C47

Citation: Díaz-González, A.;

1. Introduction
Hernández, J.; Pijeira-Cabrera, H. Let μ be a positive finite Borel measure with finite moments, whose support Δ ⊂ R
Sequentially Ordered Sobolev Inner contains infinitely many points. We will denote by Ch( A) the convex hull of a set A and by
Product and Laguerre–Sobolev A◦ its interior.
Polynomials. Mathematics 2023, 11, Let { Pn }n≥0 be the monic orthogonal polynomial sequence with respect to the inner
1956. https://doi.org/10.3390/ product 
math11081956
f, g μ = f ( x ) g( x )dμ( x ).
Academic Editor: Carsten Schneider
Δ

Received: 16 March 2023

An inner product is called standard if the multiplication operator is symmetric with
Revised: 17 April 2023 respect to the inner product. Obviously, x f , g μ = f , xg μ , i.e., ·, · μ is standard. Signif-
Accepted: 18 April 2023 icant parts of the applications of orthogonal polynomials in mathematics and particular
Published: 20 April 2023 sciences are based on the following three consequences of this fact.
1. The polynomial Pn has exactly n real simple zeros in Ch(Δ)◦ . Moreover, there is a zero
of Pn−1 between any two consecutive zeros of Pn .
2. The three-term recurrence relation
Copyright: © 2023 by the authors.
Licensee MDPI, Basel, Switzerland.
xPn ( x ) = Pn+1 ( x ) + β n Pn ( x ) + γn2 Pn−1 ( x ); P0 ( x ) = 1, P−1 ( x ) = 0,
This article is an open access article
distributed under the terms and

where γn =  Pn μ / Pn−1 μ for n ≥ 1, β n = Pn , xPn μ /  Pn μ
2 and  · μ = ·, · μ
conditions of the Creative Commons
Attribution (CC BY) license (https:// denotes the norm induced by ·, · μ .
creativecommons.org/licenses/by/ 3. For the kernel polynomials
4.0/).

Mathematics 2023, 11, 1956. https://doi.org/10.3390/math11081956 126 https://www.mdpi.com/journal/mathematics

Mathematics 2023, 11, 1956

n
Pk ( x ) Pk (y)
Kn ( x, y) = ∑  Pk 2μ
, (1)
k =0

we have the Christoffel–Darboux identities

⎧ P ( x) P (y)− P (y) P ( x)
⎨ n +1 n n +1
 Pn 2μ ( x −y)
n
, if x = y,
Kn ( x, y) = P ( x) Pn ( x)− ( )  (x) (2)
⎩ n +1 P n + 1 x Pn
, if x = y.
 P 2 n μ

These identities play a fundamental role in the treatment of Fourier expansions with
respect to a system of orthogonal polynomials (see [1], Section 2.2). For a review of the
use of (1) and (2) in the spectral theory of orthogonal polynomials, we refer the reader
to [2]. In addition, see the usual references [3–5], for a basic background on these and other
properties of { Pn }n≥0 .
Let ( a, b) = Ch(supp(μ))◦ , N, d j ∈ Z+ , λ j,k ≥ 0, for j = 1, . . . , N, k = 0, 1, . . . , d j ,
{c1 , c2 , . . . , c N } ⊂ R \( a, b), where ci = c j if i = j and I+ = {( j, k) : λ j,k > 0}. We consider
the following Sobolev-type (or discrete Sobolev) inner product
 N dj
f, g s = f ( x ) g( x )dμ( x ) + ∑ ∑ λ j,k f (k) (c j ) g(k) (c j )
j =1 k =0

= f ( x ) g( x )dμ( x ) + ∑ λ j,k f (k) (c j ) g(k) (c j ), (3)
( j,k)∈ I+

where f (k) denotes the k-th derivative of the function f . Without loss of generality, we also
assume {( j, d j )} N
j=1 ⊂ I+ and d1 ≤ d2 ≤ · · · ≤ d N . For n ∈ Z+ , we shall denote by Sn the
monic polynomial of the lowest degree satisfying

x k , Sn s = 0, for k = 0, 1, . . . , n − 1. (4)

It is easy to see that for all n ≥ 0, there exists such a unique polynomial Sn of degree
n. This is deduced by solving a homogeneous linear system with n equations and n + 1
unknowns. Uniqueness follows from the minimality of the degree for the polynomial
solution. We refer the reader to [6,7] for a review of this type of non-standard orthogonality.
Clearly, (3) is not standard, i.e., xp, q s = p, xq s , for some p, q ∈ P. It is well known
that the properties of orthogonal polynomials with respect to standard inner products differ
from those of the Sobolev-type polynomials. In particular, the zeros of the Sobolev-type
polynomials can be complex, or if real, they can be located outside the convex hull of the
support of the measure μ, as can be seen in the following example.

Example 1 (Zeros outside the convex hull of the measures supports). Set
 ∞
f, g s = f ( x ) g( x )e− x dx + 2 f  (−1) g (−1),
0

then the corresponding second-degree

√ monic√Sobolev-type orthogonal polynomial is S2 (z) = z2 − 2,
whose zeros are z1,2 = ± 2. Note that − 2 ∈ [−1, ∞).

Let { Qn }n≥0 be the sequence of monic orthogonal polynomials with respect to the
inner product

d j +1 d j +1
f, g μρ = f ( x ) g( x ) dμρ ( x ), where ρ( x ) = ∏ x − cj ∏ cj − x
cj ≤a c j ≥b

and dμρ ( x ) = ρ( x )dμ( x ).

127
Mathematics 2023, 11, 1956

Note that ρ is a polynomial of degree d = ∑ N j=1 ( d j + 1), which is positive on ( a, b ).

If n > d, from (4), {Sn } satisfies the following quasi-orthogonality relations with respect
to μρ 
Sn , f μρ = Sn , ρ f μ = Sn ( x ) f ( x )ρ( x )dμ( x ) = Sn , ρ f s = 0,

for f ∈ Pn−d−1 , where Pn is the linear space of polynomials with real coefficients and the
degree at most n ∈ Z+ . Hence, the polynomial Sn is quasi-orthogonal of order d with respect to
μρ and by this argument, we obtain that Sn has at least (n − d) changes of sign in ( a, b).
The results obtained for measures μ with bounded support (see [8], (1.10)) suggest
that the number of zeros located in the interior of the support of the measure is closely
related to d∗ = | I+ |, the number of terms in the discrete part of ·, · s (i.e., λ j,k > 0), instead
of this greater quantity d.
Our first result, Theorem 1, goes in this direction for the case when the inner product is
sequentially ordered. This kind of inner product is introduced in Section 2 (see Definition 1).

Theorem 1. If the discrete Sobolev inner product (3) is sequentially ordered, then Sn has at least
n − d∗ changes of sign on ( a, b), where d∗ is the number of positive coefficients λ j,k in (3).

Previously, this result was obtained for more restricted cases in ([9], Th. 2.2) and ([10],
Th. 1). In ([9], Th. 2.2), the authors proved this result for the case N = 1. In ([10], Th. 1),
the notion of a sequentially ordered inner product is more restrictive than here, because it
did not include the case when the Sobolev inner product has more than one derivative
order at the same mass point.
In the second part of this paper, we focus our attention on the Laguerre–Sobolev-type
polynomials (i.e., dμ = x α e− x dx, with α > −1). In the case of the inner product, (3) takes
the form
 ∞ N
f, g s =
0
f ( x ) g( x ) x α e− x dx + ∑ λ j f ( d j ) ( c j ) g ( d j ) ( c j ), (5)
j =1

where λ j := λ j,d j > 0, c j < 0, for j = 1, 2, . . . , N, we obtain the outer relative asymptotic of
the Laguerre–Sobolev-type polynomials.

Theorem 2. Let { Lαn }n≥0 be the sequence of monic Laguerre polynomials and let {Sn }n≥0 be the
monic orthogonal polynomials with respect to the inner product (5). Then,
⎛√  ⎞
Sn ( x ) N − x − |c j |
⇒ ∏⎝ √  ⎠, K ⊂ C \ R+ . (6)
Lαn ( x ) j =1 − x + |c | j

Throughout this paper, we use the notation f n ⇒ f , K ⊂ U when the sequence of

functions f n converges to f uniformly on every compact subset K of the region U.
Combining this result with Theorem 1, we obtain that the Sobolev polynomials Sn ,
orthogonal with respect to a sequentially ordered inner product in the form (5), have at
least n–N zeros in (0, ∞) and, for sufficiently large n, each one of the other N zeros are
contained in a neighborhood of each mass point c j (j = 1, . . . , N). Then, we have located all
zeros of Sn and we obtain that for a sufficiently large n, they are simple and real, as in the
Krall case (see [11]) or the Krall–Laguerre-type orthogonal polynomial (see [12]). This is
summarized in the following corollary.

Corollary 1. Let μ = μα be the classical Laguerre measure (dμα ( x ) = x α e− x dx) and (5) a
sequentially ordered discrete Sobolev inner product. Then, the following statements hold:
1. Every point c j attracts exactly one zero of Sn for sufficiently large n, while the remaining n–N
zeros are contained in (0, ∞). This means:

128
Mathematics 2023, 11, 1956

For every r > 0, there exists a natural value N such that if n ≥ N , then the n zeros of Sn
{ξ i }in=1 satisfy

ξ j ∈ B(c j , r ) for j = 1, . . . , N and ξ i ∈ (0, ∞) for i = N + 1, N + 2, . . . , n.

2. The zeros of Sn are real and simple for large-enough values of n.

3. The zeros of {Sn }∞n=1 are at a finite distance from (0, ∞ ). This means that there exists a
positive constant M such that if ξ is a zero of Sn , then

d(ξ, (0, ∞)) := inf {| x − ξ |} < M.

x >0

Section 2 is devoted to introducing the notion of a sequentially ordered Sobolev inner

product and to prove Theorem 1. In Section 3, we summarize some auxiliary properties
of Laguerre polynomials to be used in the proof of Theorem 2. Some results about the
asymptotic behavior of the reproducing kernels are given. The aim of the last section is to
prove Theorem 2 and some of its consequences stated in Corollary 2.

2. Sequentially Ordered Inner Product

Definition 1 (Sequentially ordered Sobolev inner product). Consider a discrete Sobolev inner
product in the general form (3) and assume d1 ≤ d2 ≤ · · · ≤ d N without loss of generality. We say
that a discrete Sobolev inner product is sequentially ordered if the conditions
 ◦
Δk ∩ Ch ∪ik=−01 Δi = ∅, k = 1, 2, . . . , d N ,

hold, where 
Chsupp(μ) ∪ {c j  : λ j,0 > 0} , if k = 0,
Δk = (7)
Ch {c j : λ j,k > 0} , if 1 ≤ k ≤ d N .

Note that Δk is the convex hull of the support of the measure associated with the k-th
order derivative in the Sobolev inner product (3). Let us see two examples.

Example 2 (Sequentially ordered inner product).

Set
 ∞
f, g s = f ( x ) g( x )e− x dx + 10 f (−1) g(−1) + 5 f  (−3) g (−3)
0
+ 5 f  (−9) g (−9) + 20 f  (−10) g (−10),

then the corresponding fifth-degree Sobolev orthogonal polynomial has the following exact expression

380961336355365 4 1836311881214045 3 7830454972601355 2

S5 ( x ) = x 5 + x + x − x
16894750106161 16894750106161 16894750106161
36972053870326650 22386262325875230
− x− ,
16894750106161 16894750106161
whose zeros are approximately ξ 1 ≈ 4.46, ξ 2 ≈ −0.74, ξ 3 ≈ −2.8, ξ 4 ≈ −11.74 + 2.51i and
ξ 5 ≈ −11.74 − 2.51i. Note that four of them are outside of (0, ∞) and two are even complex.

Example 3 (Non-sequentially
 ordered inner product).
∞
Set f , g = f ( x ) g( x ) e− x dx + f  (−15) g (−15) + f  (−9) g (−9), then the corresponding
0
fifth-degree Sobolev orthogonal polynomial has the following exact expression

129
Mathematics 2023, 11, 1956

55079160 4 5053767275 3 40953207555 2

S5 ( x ) = x 5 + x − x + x
21682477 21682477 21682477
98030649090 42523040550
− x+ ,
21682477 21682477
whose zeros are approximately ξ 1 ≈ 0.55, ξ 2 ≈ 3.36, ξ 3 ≈ 6.66 + 3.02i, ξ 4 ≈ 6.66 − 3.02i and
ξ 5 ≈ −19.77. Note that, in spite of Theorem 1, d∗ = 2 and three of the zeros of S5 are outside of
(0, ∞), with two of them as not even real.

In the sequentially ordered example (Example 2), S5 has exactly 1 = 5 − 4 = n − d∗

simple zeros on the interior of the convex hull of the support of the Laguerre measure
(0, ∞), and thus, the bound of Theorem 1 is sharp. In addition, this example shows that the
remaining d∗ zeros might even be complex, although Corollary 1 shows that this does not
happen when n is sufficiently large.
On the other hand, in the non-sequentially ordered example (Example 3), this condition
is not satisfied, since S5 has only 2 < 3 = 5 − 2 = n − d∗ zeros on (0, ∞), showing that the
sequential order plays a main role in the localization of the zeros of Sn , at least to obtain
this property for every value of n.
Throughout the remainder of this section, we will consider inner products of the
form (3) that are sequentially ordered. The next lemma is an extension of ([13], Lemma 2.1)
and ([10], Lemma 3.1).

Lemma 1. Let { Ii }im=0 be a set of m + 1 intervals on the real line and let P be a polynomial with
real coefficients of degree ≥ m. If
 ◦
Ik ∩ Ch ∪ik=−01 Ii = ∅, k = 1, 2, . . . , m, (8)

then
m    
Nz ( P; J ) + N◦ ( P; I0 \ J ) + ∑ N◦ P(i) ; Ii ≤ Nz P(m) ; J
i =1
 
+ N◦ P(m) ; Ch(∪im=0 Ii ) \ J + m, (9)

for every closed subinterval J of I0 ◦ (both empty set and unitary sets are assumed to be intervals).
Here, given a real set A and a polynomial P, N◦ ( P; A) denotes the number of values where the
polynomial P vanishes on A (i.e., zeros of P on A without counting multiplicities), and Nz ( P; A)
denotes the total number of zeros (counting multiplicities) of P on A.

Proof. First, we point out the following consequence of Rolle’s Theorem. If I is a real
interval and J is a closed subinterval of I ◦ , then

Nz ( P; J ) + N◦ ( P; I \ J ) ≤ Nz P ; J + N◦ P ; I ◦ \ J + 1. (10)

It is easy to see that (9) holds for m = 0. We now proceed by induction on m. Suppose that
we have m + 1 intervals { Ii }im=0 satisfying (8); thus, the first m intervals { Ii }im=−01 also satisfy
(8), and we obtain (9) by induction hypothesis (taking m − 1 instead of m). Then
m  
Nz ( P; J ) + N◦ ( P; I0 \ J ) + ∑ N◦ P(i) ; Ii ,
i =1
       
( m −1)
≤ Nz P ; J + N◦ P(m−1) ; Ch ∪im=−01 Ii \ J + m − 1 + N◦ P(m) ; Im ,
    ◦   
≤ Nz P(m) ; J + N◦ P(m) ; Ch ∪im=−01 Ii \ J + m + N◦ P(m) ; Im ,
   
≤ Nz P(m) ; J + N◦ P(m) ; Ch(∪im=0 Ii ) \ J + m,

130
Mathematics 2023, 11, 1956

where in the second inequality, we have used (10).

As an immediate consequence of Lemma 1, the following result is obtained.

Lemma 2. Under the assumptions of Lemma 1, we have

m  
Nz ( P; J ) + N◦ ( P; I0 \ J ) + ∑ N◦ P(i) ; Ii ≤ deg P (11)
i =1

for every J closed subinterval of I0 ◦ . In particular, for J = ∅, we obtain

m  
∑ N◦ P(i) ; Ii ≤ deg P. (12)
i =0

Lemma 3. Let {(ri , νi )}iM

=1 ⊂ R × Z+ be a set of M ordered pairs. Then, there exists a unique
monic polynomial U M of minimal degree (with 0 ≤ deg U M ≤ M), such that
(ν )
U Mi (ri ) = 0, i = 1, 2, . . . , M. (13)

Furthermore, if the intervals Ik = Ch({ri : νi = k}), k = 0, 1, 2, . . . , νM , satisfy (8), then U M

has degree u M = min I M − 1, where

I M = {i : 1 ≤ i ≤ M and νi ≥ i } ∪ { M + 1}.

Proof. The existence of a nonidentical zero polynomial with degree ≤ M satisfying (13)
reduces to solving a homogeneous linear system with M equations and M + 1 unknowns
(its coefficients). Thus, a non-trivial solution always exists. In addition, if we suppose that
there exist two different minimal monic polynomials U M and U 4 M , then the polynomial
U* M = UM − U 4 M is not identically zero, it satisfies (13), and deg U* M < deg U M . Thus,
if we divide U* M by its leading coefficient, we reach a contradiction.
The rest of the proof runs by induction on the number of points M. For M = 1,
the result follows taking 
x − r1 , if ν1 = 0,
U1 ( x ) =
1, if ν1 ≥ 1.
Suppose that, for each sequentially ordered sequence of M ordered pairs, the corre-
sponding minimal polynomial U M has degree u M .
M −1
Let {(ri , νi )}iM
=1 be a set of M ordered pairs satisfying (8). Obviously, {(ri , νi )}i =1 also
satisfies (8) and U M satisfies (13) for i = 1, 2, . . . , M − 1; thus, we obtain deg U M−1 = u M−1
and deg U M ≥ deg U M−1 . Now, we divide the proof into two cases:
1. If u M = M, then for all 1 ≤ i ≤ M we have νi < i, which yields
deg U M ≥ deg U M−1 = u M−1 = M − 1 ≥ νM .

Since {(ri , νi )}iM

=1 satisfies (8), from (12) we obtain
νM  
(i )
M ≤ ∑ N◦ U M ; Ii ≤ deg U M ,
i =0

which implies that deg U M = M = u M .

2. If u M ≤ M − 1, then there exists a minimal j (1 ≤ j ≤ M), such that νj ≥ j, and νi < i
for all 1 ≤ i ≤ j − 1. Therefore, u M = j − 1 = u M−1 . From the induction hypothesis,
we obtain
deg U M−1 = u M−1 = j − 1 ≤ νj − 1 ≤ νM − 1,
(ν )
−1 ≡ 0. Hence, U M ≡ U M−1 and, consequently, we obtain
which gives U MM
deg U M = deg U M−1 = u M−1 = u M .

131
Mathematics 2023, 11, 1956

Note that, in Lemma 3, condition (8) is necessary for asserting that the polynomial U M
has degree u M . If we consider {(−1, 0), (1, 0), (0, 1)}, whose corresponding convex hulls
I0 = [−1, 1] and I1 = {0} do not satisfy (8), we obtain U3 ( x ) = x2 − 1 and u3 = 3 = deg U3 .
Now we are able to prove the zero localization theorem for sequentially ordered
discrete Sobolev inner products.

Proof of Theorem 1. Let ξ 1 < ξ 2 < · · · < ξ η be the points on ( a, b) = Ch(supp(μ))◦ where
Sn changes sign and suppose that η < n − d∗ . Consider the set of ordered pairs
d∗ +η η
{(ri , νi )}i=1 = {(ξ i , 0)}i=1 ∪ {(c j , k) : η j,k > 0, j = 1, 2, . . . , N, k = 1, . . . , d j }.

Since ·, · s is sequentially ordered, the intervals Ik = Δk for k = 0, 1, . . . , νN

(see (7)) satisfy (8) (we can assume without loss of generality that ν1 ≤ ν2 ≤ · · · ≤ νd∗ +η ).
Consequently, from Lemma 3, there exists a unique monic polynomial Ud∗ +η of minimal
degree, such that
Ud∗ +η (ξ i ) = 0; for i = 1, . . . , η,
(k)
Ud∗ +η (c j ) = 0; for each ( j, k) : η j,k > 0, (14)

and deg Ud∗ +η = min Id∗ +η − 1 ≤ d∗ + η, where

Id∗ +η = {i : 1 ≤ i ≤ d∗ + η and νi ≥ i } ∪ {d∗ + η + 1}. (15)

Now, we need to consider the following two cases.

1. If deg Ud∗ +η = d∗ + η, from (15), we obtain deg Ud∗ +η ≥ νη +d∗ + 1. Thus, taking the
closed interval J = [ξ 1 , ξ η ] ⊂ ( a, b) in (11), we obtain
νd∗ +η      
(k)
d∗ + η ≤ ∑ N◦ Ud∗ +η ; Ik ≤ Nz Ud∗ +η ; [ξ 1 , ξ η ] + N◦ Ud∗ +η ; I0 \ [ξ 1 , ξ η ]
k =0
νd∗ +η  
(k)
+ ∑ N◦ Ud∗ +η ; Ik ≤ deg Ud∗ +η = d∗ + η.
k =1

2. If deg Ud∗ +η < d∗ + η, from (15), there exists 1 ≤ j ≤ d∗ + η such that deg Ud∗ +η =
j − 1, νj ≥ j and νi ≤ i − 1 for i = 1, 2, . . . , j − 1. Hence,

νj−1 + 1 ≤ j − 1 = deg Ud∗ +η

and, again, from (11) we have

νj − 1      
(k)
j−1 ≤ ∑ N◦ Ud∗ +η ; Ik ≤ Nz Ud∗ +η ; [ξ 1 , ξ η ] + N◦ Ud∗ +η ; I0 \ [ξ 1 , ξ η ]
k =0
νj−1  
(k)
+ ∑ N◦ Ud∗ +η ; Ik ≤ deg Ud∗ +η = j − 1.
k =1

In both cases, we obtain that Ud∗ +η has no other zeros in I0 than those given by
   
construction, and from N◦ Ud∗ +η ; [ξ 1 , ξ η ] = Nz Ud∗ +η ; [ξ 1 , ξ η ] , all the zeros of Sn on I ◦
are simple. Thus, in addition to (14), we obtain that Sn Ud∗ +η does not change sign on I ◦ .
Now, since deg Ud∗ +η ≤ d∗ + η < n, we arrive at the contradiction
 N dj
(k) (k)
0 = Sn , Ud∗ +η = Sn ( x )Ud∗ +η ( x )dμ( x ) + ∑ ∑ λ j,k Sn (c j )Ud∗ +η (c j )
j =1 k =0
 b
= Sn ( x )Ud∗ +η ( x )dμ( x ) = 0.
a

132
Mathematics 2023, 11, 1956

3. Auxiliary Results
The family of Laguerre polynomials is one of the three very well-known classical
orthogonal polynomials families (see [3–5]). It consists of the sequence of polynomials
(α)
{ Ln } that are orthogonal with respect to the measure dμ = x α e− x dx, x ∈ (0, ∞), for
(−1)n
α > −1, and that are normalized by taking n! as the leading coefficient of the n-th
degree polynomial of the sequence. Laguerre polynomials play a key role in applied
mathematics and physics, where they are involved in the solutions of the wave equation of
the hydrogen atom (c.f. [14]).
Some of the structural properties of this family are listed in the following proposition
in order to be used later.
(α)
Proposition 1. Let { Ln }n≥0 (note the brackets in parameter α) be the sequence of Laguerre
polynomials and let { Lαn }n≥0 be the monic sequence of Laguerre polynomials. Then, the following
statements hold.
1. For every n ∈ N,
(α) (−1)n α
Ln ( x ) = L n ( x ). (16)
n!

2. Three-term recurrence relation. For every n ≥ 1,

xLαn ( x ) = Lαn+1 ( x ) + (2n + α + 1) Lαn ( x ) + n(n + α) Lαn−1 ( x )

(α) (α) (α) (α)
xLn ( x ) = −(n + 1) Ln+1 ( x ) + (2n + α + 1) Ln ( x ) − (n + α) Ln−1 ( x )

(α) (α)
with L−1 ≡ Lα−1 = 0, and L0 ≡ L0α ≡ 1.
3. Structure relation. For every n ∈ N,

(α) ( α +1) ( α +1)

Ln ( x ) = Ln ( x ) − L n −1 ( x ).

4. For every n ∈ N,
 
(α) n+α Γ ( α + n + 1)
|| Ln ||2μ = Γ(α + 1) = . (17)
n n!

In addition, we have
|| Lαn ||2μ = n!Γ(n + α + 1)
5. Hahn condition. For every n ∈ N,
(α) ( α +1)
[ L n ]  ( x ) = − L n −1 ( x ). (18)

6. Outer strong asymptotics (Perron’s asymptotics formula on C \ R+ ). Let α ∈ R. Then

 
1/2 p −1
(α) e x/2 nα/2−1/4 e2(−nx)
Ln ( x ) = ∑ Ck ( x )n−k/2 + O(n− p/2 ) . (19)
2π 1/2 (− x )α/2+1/4 k =0

p −1
Here, {Ck ( x )}k=0 are certain analytic functions of x independent of n, with C0 ≡ 1. This
relation holds for x in the complex plane with a cut along the positive part of the real axis.
The bound for the remainder holds uniformly in every closed domain with no points in common
with x ≥ 0 (see [5], Theorem 8.22.3).

Now, we summarize some auxiliary lemmas to be used in the proof of Theorem 2 (see
([15], Lem. 1) and ([16], Prop. 6)).

133
Mathematics 2023, 11, 1956

Lemma 4. For z ∈ C \ [0, ∞), α, β ∈ R and j, k ≥ −n we have

⎧ √    3
(α+ β)
Ln+ j (z ) ⎨1 + ( j−k√) −z + α2 − 14 − z ( j−2 k) ( j−n k) + Oz n− 2 if β = 0
n
=  √ β    (20)
(α)
Ln+k (z ) ⎩ √n 1 + Oz n−1/2 if β = 0.
−z

where Oz (n− j ) denotes some analytic function sequence { gn (z)}∞

n=1 such that { n gn } is uniformly
j

bounded on every compact subset of C \ [0, ∞).

To study the outer relative asymptotic between the standard Laguerre polynomials
and the Laguerre–Sobolev orthogonal polynomials (see Formula (6)), we need to compute
the behavior of the Laguerre kernel polynomials and their derivatives when n approaches
infinity. To this end, we prove the following auxiliary result, which is an extension of ([17],
Ch. 5, Th. 16).

Lemma 5. Let G and G  be two open subsets of the complex plane and f n : G × G  −→ C be
a sequence of functions that are analytic with respect to each variable separately. If { f n }∞
n=1 is a
uniformly bounded sequence on each set in the form K × K  , where K ⊂ G and K  ⊂ G  are compact
sets, then any of its partial derivative sequences are also uniformly bounded on each set in the form
K × K .

Proof. Note that it is sufficient to prove this for the first derivative order with respect
to any of the variables and then proceed by induction. Let K ⊂ G and K  ⊂ G  be two
compact sets. Denote G c = C \ G, d(K, G c ) = inf c |z − w|, r = d(K, G c )/2 > 0 and
z∈K,w∈ G
5
B(z, r ) = {ζ ∈ C : |z − ζ | < r }. Take K ∗ as the closure of z∈K B(z, r ); thus, K ∗ is a compact
subset of G. Thus, there exists a positive constant M > 0 such that | f n (z, w)| ≤ M for all
z ∈ K ∗ , w ∈ K  and n ∈ N. Hence, for all z ∈ K, w ∈ K  and n ∈ N, we obtain
    6 7
 ∂ fn   1  f n (ξ, w)  V (c(z, r )) | f n (ξ, w)|
 ( z, w ) = dξ ≤ max
 ∂z   2πi c(z,r) (ξ − z)2  2π ξ ∈c(z,r ) | ξ − z |2
2πr M
= max {| f n (ξ, w)|} ≤ ,
2πr2 ξ ∈c(z,r) r

where c(z, r ) denotes the circle with center at z, radius r and length V (c(z, r )).

From the Fourier expansion of Sn in terms of the basis { Lαn }n0 we obtain
n Liα ( x ) n −1 Liα ( x )
Sn ( x ) = ∑ Sn , Liα μ1
α α
12 = L n ( x ) + ∑ Sn , L i μ1 1
i =0 1L 1α
i = 0 1 L α 12
i μ i μ
⎛ ⎞
n −1 Lα ( x )
(k)
= Lαn ( x ) + ∑⎝ Sn , Liα s − ∑ λ j,k Sn (c j )( Liα )(k) (c j )⎠ 1 i 12
i =0 ( j,k)∈ I+
1 Lα 1
i μ
(k)
(k)
n −1 Liα ( x ) Liα (c j )
= Lαn ( x ) − ∑ λ j,k Sn (c j ) ∑ 1 α 12
( j,k)∈ I+ i =0 1L 1
i μ
(k) (0,k)
= Lαn ( x ) − ∑ λ j,k Sn (c j )Kn−1 ( x, c j ), (21)
( j,k)∈ I+

∂ j+k Kn ( x, y)
( j,k)
where we use the notation Kn ( x, y) = to denote the partial derivatives of
∂ j x∂k y
the kernel polynomials defined in (1). Differentiating Equation (21) -times and evaluating
then at x = ci for each ordered pair (i, ) ∈ I+ , we obtain the following system of d∗ linear
(k)
equations and d∗ unknowns Sn (c j ).

134
Mathematics 2023, 11, 1956

  dj
(,) () (,k) (k)
( Lαn )() (ci ) = 1 + λi, Kn−1 (ci , ci ) Sn (ci ) + ∑ λ j,k Kn−1 (ci , c j )Sn (c j ). (22)
( j,k)∈ I+
( j,k)=(i,)

Lemma 6. The Laguerre kernel polynomials and their derivatives satisfy the following behavior
when n approaches infinity for x, y ∈ C \ [0, ∞)

(y)  
( α +i ) (α+ j)
(i,j) ∂ i + j K n −1 Ln ( x ) Ln i+ j −1/2
Kn−1 ( x, y) = ( x, y ) = √ √ (− 1 ) + O x,y ( n ) , i, j ≥ 0,
∂i x∂ j y nα− 2 ( − x + − y )
1

where O x,y (n−k ) denotes some sequence of functions { gn ( x, y)}∞n=1 that are holomorphic with
respect to each variable and whose sequence {nk gn } is uniformly bounded on every set K × K  , such
that K and K  are compact subsets of C \ R+ .

Proof. The proof is by induction on k = i + j. First, suppose k = 0 (i.e., i = j = 0) and split

the proof into two cases according to whether x = y or not. If x = y, from (2), (16), (18)
and (20), we obtain
(α)
 Ln−1 2μ (α) (α) (α) (α)
Kn−1 ( x, x ) = Ln ( x )( Ln−1 ) ( x ) − ( Ln ) ( x ) Ln−1 ( x )
n
( α +1) (α) ( α +1) (α)
= L n −1 ( x ) L n −1 ( x ) − L n −2 ( x ) L n ( x )
⎛ ⎞
( α +1) (α)
( α +1) (α) L n −1 ( x ) Ln ( x ) ⎠
= L n −2 ( x ) L n −1 ( x ) ⎝ − (α)
( α +1)
L n −2 ( x ) L n −1 ( x )
# √ # $
( α +1) (α) −x α+1 1 x 1
= L n −2 ( x ) L n −1 ( x ) 1 + √ + − − + O x (n−3/2 )
n 2 4 2 n
 √ # $ $
−x α 1 x 1
− 1+ √ + − − + O x (n−3/2 )
n 2 4 2 n
 
( α +1) (α) 1
= L n −2 ( x ) L n −1 ( x ) + O x (n−3/2 )
2n
(α) (α)  √  
L ( x ) Ln ( x ) n
= n √ 1 + O x (n−1/2 )
2n −x
Ln ( x ) Ln ( x )  
(α) (α)
= √ √ 1 + O x (n−1/2 ) .
2 n −x

On the other hand, if x = y, from (2) and (20) we obtain

(α) (α) (α) (α) (α)

 Ln−1 2μ L ( x ) L n ( y ) − L n ( x ) L n −1 ( y )
Kn−1 ( x, y) = n−1
n x−y
⎛ ⎞
(α) (α)
Ln−1 ( x ) Ln−1 (y) L(nα) (y) (α)
Ln ( x ) ⎠
= ⎝ − (α)
x−y L
(α)
(y) L ( x )
n −1 n −1
(α) (α) √ √ 
L ( x ) L n −1 ( y ) −y − − x
= n −1 √ + O x,y (n−1 )
x−y n
(α) (α)  
L ( x ) L n −1 ( y ) 1
= n√−1 √ √ + O x,y (n−1 )
− x + −y n
(α)
L ( x ) Ln (y)
(α)  
= √ n√ √ 1 + O x,y (n−1/2 ) .
n( − x + −y)

135
Mathematics 2023, 11, 1956

From (17) and ([18], Appendix, (1.14))

(α) Γ(n + α)
 Ln−1 2μ = = nα (1 + O(n−1 )),
Γ(n)

which proves the case k = 0. Now, we assume that the theorem is true for i + j = k and
we will prove it for i + j = k + 1. By the symmetry of the formula, the proof is analogous
when any of the variables increase its derivative order; thus, we only will prove it when
the variable y does.
 ( α +i )
(y)  
(α+ j)
∂ k +1 K n −1 ∂ Ln ( x ) Ln
+
( x, y) = √ √ (−1)k + O x,y (n−1/2 )
∂x ∂ y
i j 1 ∂y nα− 2 ( − x + −y)
1

 
Ln
( α +i )
(x) ∂ L
(α+ j)
(y)  
= √ n √ (−1)k + O x,y (n−1/2 )
n 2α − 1
∂y − x + − y

(y) ∂  
(α+ j)
Ln −
+√ √ (−1) + O x,y (n
k 1/2
)
− x + −y ∂y
⎡ √
( α +i ) √ ( α + j +1) (α+ j)
L ( x ) ⎣ −( − x + −y) Ln−1 (y) + 12 Ln (y)(−y)−1/2
= n √ √
nα− 2
1
( − x + − y )2

  L
(α+ j)
(y)
· (−1)k + O x,y (n−1/2 ) + √ n √ O x,y ( n −1/2
)
− x + −y
⎡⎛ √
−y
⎞
Ln
( α +i ) ( α + j +1)
( x ) L n −1 (y) √ + O x,y ( n−1 )
= ⎣ ⎝ −1 + √ n
√ √ ⎠
1 √ √
nα− 2 ( − x + − y ) 2 −y( − x + −y)
   √−y  $
· (−1)k + O x,y (n−1/2 ) + √ + O x,y (n−1 ) O x,y (n−1 )
n
( α +i ) ( α + j +1)   
Ln ( x ) L n −1 (y)
= 1 √ √ −1 + O x,y (n−1/2 ) (−1)k + O x,y (n−1/2 )
α −
n 2 ( − x + −y)
+O x,y (n−3/2 )
( α +i ) ( α + j +1)
Ln ( x ) Ln (y)
= √ √ (−1)k+1 + O x,y (n−1/2 ) ,
nα− 2 (
1
− x + −y)
∂ −1
where in the third equality we use Lemma 5 to guarantee that ∂y O x,y ( n ) = O x,y (n−1 ),
and in the fourth equality, we use (20).

4. Proof of Theorem 2 and Consequences

(α)
Proof of Theorem 2. Without loss of generality, we will consider the polynomials Ln =
(−1)n /n! Lαn and S*n = (−1)n /n! Sn , instead of the monic polynomials Lαn and Sn .
Multiplying both sides of (21) by (−1)n /n!, we obtain
N
(α) (d j ) (0,d )
S*n ( x ) = Ln ( x ) − ∑ λ j S*n (c j )Kn−1j ( x, c j ), (23)
j =1

(α)
Dividing by Ln ( x ) on both sides of (23), we obtain
(0,d j )
S*n ( x ) N
(d ) Kn−1 ( x, c j )
(α)
= 1 − ∑ λ j S*n j (c j ) (α)
. (24)
Ln ( x ) j =1 Ln ( x )

Recall that we are considering the Laguerre–Sobolev polynomials {S*n } that are or-
thogonal with respect to (5). In this case, the consistent linear system (22) becomes

136
Mathematics 2023, 11, 1956

    N
(α) (dk ) (d ,d ) (d ) (d ,d ) (d )
Ln (ck ) = 1 + λk Kn−k 1 k (ck , ck ) S*n k (ck ) + ∑ λ j Kn−k 1 j (ck , c j )S*n j (c j ),
(25)
j =1
j=k

for k = 1, 2, . . . , N. Let us define

(0,d j )
(d j ) Kn−1 ( x, c j )
α
Pn,j ( x ) := −λ j S*n (c j ) (α)
and Pjα ( x ) := lim Pn,j
α
( x ).
n→∞
Ln ( x )

From (24), in order to prove the existence of the limit (6), we need to figure out the
values of Pjα ( x ). Note that
(α) (α)
(d j ) Ln ( x ) Pn,j ( x )
S*n (c j ) = − (0,d j )
.
λ j Kn−1 ( x, c j )

If we replace these expressions in (25), then we obtain the following linear system in
the unknowns Pn,j ( x )
⎛ ⎞⎛ α (x) ⎞ ⎛ ⎞
a1,1 (n, x ) a1,2 (n, x ) ··· a1,N (n, x ) Pn,1 −1
⎜ a2,1 (n, x ) a2,2 (n, x ) ··· a1,N (n, x ) ⎟⎜ α (x)
Pn,2 ⎟ ⎜ −1 ⎟
⎜ ⎟⎜ ⎟ ⎜ ⎟
⎜ .. .. .. .. ⎟⎜ .. ⎟=⎜ .. ⎟, (26)
⎝ . . . . ⎠⎝ . ⎠ ⎝ . ⎠
a N,1 (n, x ) a N,2 (n, x ) · · · a N,N (n, x ) α
Pn,N ( x ) −1

where
⎧ (dk ,d j )
⎪
⎪
(α)
Ln ( x )Kn−1 (ck ,c j )
⎪
⎪   , j = k,
⎪
⎨ L(nα) (dk )(ck )Kn(0,d j)
 −1 ( x,c j ) 
ak,j (n, x ) = (α) (d ,d )
⎪
⎪ Ln ( x ) λ1 +Kn−k1 k (ck ,ck )
⎪
⎪ k
j = k.
⎪
⎩ L(α) (dk )(c )K(0,dk )( x,c ) ,
 
n k n −1 k

Now, we will find the behavior of the coefficients ak,j (n, x ) when n approaches infinity.
If k = j, we have
 
(α) (d ,d )
Ln ( x ) λ1 + Kn−k 1 k (ck , ck )
ak,k (n, x ) = 
k

(α) (dk ) (0,d )
Ln (ck )Kn−1k ( x, ck )
 (α+dk ) (α+dk )  
(α) (c ) L (c )
Ln ( x ) λ1 + n α− 1 √ k n √ k (−1)dk +dk + O(n−1/2 )
L
k
n 2 −ck + −ck
= (α+dk ) (α)
(α+dk )
(−1) k Ln−d (ck ) Lαn− 1 √(ck ) L√
d n (x)
(−1)dk + O x (n−1/2 )
k
n 2( − x + −ck )
  
α− 12 (α+dk )
√ √ n
(α+dk ) + Ln (ck ) 1 + O(n−1/2 )
− x + −ck λk Ln (ck )
= √
2 −ck (α+dk )
Ln−d (ck ) 1 + O x (n−1/2 )
⎛ k
⎞
⎜ α− 12 ⎟
⎝ 
n
2 + 1 + O(n−1/2 )⎠
√ √ λk
(α+d )
Ln k (ck )
− x + −ck
= √
2 −ck 1 + O x (n−1/2 )
√ √ −
− x + −ck 1 + O(n 1/2 )
= √ ,
2 −ck 1 + O x (n−1/2 )

where in the last equality we use Perron’s Asymptotic Formula (19) to obtain

137
Mathematics 2023, 11, 1956

nα− 2 4πnα− 2 (−ck )α+dk + 2

1 1 1
1
(α+dk )
= √ √ O(1) = √ √ O(1),
e c k +4 −ck n
nα+dk − 2 −ck n
1
( Ln (ck ))2 n d k e4

which has exponential decay (ck < 0). On the other hand, if k = j, we obtain
(α) (dk ,d j )
L n ( x ) K n −1 ( c k , c j )
ak,j (n, x ) =  
(α) (dk ) (0,d )
Ln (ck )Kn−1j ( x, c j )
(α+dk )  
(ck ) dk +d j −1/2 )
L
√n √ (− 1 ) + O( n
−ck + −c j
=
Ln−d k (ck )  
(α+d )

(−1)dk √− x+k √−c (−1)d j + O(n−1/2 )

j
 
√  −1/2 )
− x + −c j 1 + O(n
= √  .
−ck + −c j 1 + O(n−1/2 )

Hence,
⎧√ √ 
⎪ √
⎨ √ − x+ √−c j , if j = k −x + |c j |
−ck + −c j
lim ak,j (n, x ) = √ √ =   .
n→∞ ⎪
⎩ − x + −ck
|ck | + |c j |
√
2 −ck
, if j = k

Next, taking limits on both sides of (26) when n approaches ∞, we obtain

⎛ √ √ √ √ √ √ ⎞⎛ ⎞ ⎛ ⎞
− x + | c1 |
⎜ √|c1 |+√|c1 | √− x+ √|c2 | ··· √− x+ √|c N | ⎟⎜ P1α ( x ) ⎟ ⎜ −1 ⎟
⎜ √ |c1 |+ |c2 | |c1 |+ |c N | ⎟⎜ ⎟ ⎜ ⎟
⎜ √ √ √ √ √ ⎟⎜ ⎟ ⎜ ⎟
⎜ √− x+ √|c1 | √ − x + | c2 |
√ ··· √ − x + |c N |
√ ⎟⎜ Pα ( x ) ⎟ ⎜ −1 ⎟
⎜ ⎟⎜ 2 ⎟ ⎜ ⎟
⎜ |c2 |+ |c1 | |c2 |+ | c2 | |c2 |+ |c N | ⎟⎜ ⎟=⎜ ⎟
⎜ .. .. .. ⎟⎜ .. ⎟ ⎜ . ⎟.
⎜ .. ⎟⎜ ⎟ ⎜ .. ⎟
⎜ . . . . ⎟⎜ . ⎟ ⎜ ⎟
⎜ √ √ √ ⎟⎜ ⎟ ⎜ ⎟
⎝ √ √ √ ⎠⎝ ⎠ ⎝ ⎠
√ − x+ √|c1 | √ − x+ √|c2 | ··· √ − x+ √|c N | α (x)
PN −1
|c N |+ | c1 | |c N |+ | c2 | |c N |+ |c N |

Using Cauchy determinants, it is not difficult to prove that the N solutions of the
above linear system are
 ⎛  ⎞
−2 | c j | N |c j | + |cl |
α
Pj ( x ) = √  ∏  ⎝  ⎠.
− x + | c j | l =1 |c j | − |cl |
l = j

Now, from (24), we obtain

⎛  ⎞ 
S*n ( x ) |c j | N N |c j | + |cl |
2
lim
n→∞ (α)
= 1+ ∑ √  ∏⎝   ⎠.
Ln ( x ) j =1 − x + | c j | l =1 |c j | − |cl |
l = j

√
 the change of variable z =
If we consider − x and for simplicity we also consider the
notation t j = |c j |, then we obtain the following partial fraction decomposition

N 2t j N t j + tl
1+ ∑
z + t j l∏
.
j =1 =1
t j − tl
l = j

Thus, we only have to prove that this is the partial fraction decomposition of

138
Mathematics 2023, 11, 1956


N z − tj
∏ z + tj
.
j =1

Let PN (z) = ∏ N j=1 ( z − t j ) and Q N ( z ) = ∏ j=1 ( z + t j ), then

N

N z − tj PN (z) PN (z) − Q N (z) N Aj
∏ z + t j = Q N (z) = 1 + Q N ( z )
= 1+ ∑
z + tj
,
j =1 j =1

where
PN (z) − Q N (z) PN (−t j ) − Q N (−t j )
A j = lim (z + t j ) =
z→−t j Q N (z) QN (−t j )
N N
∏(−t j − tl ) − ∏(−t j + tl ) (−1) N 2t j N

t j + tl N

t j + tl
l =1 l =1
= N
=
(−1) N ∏ t j − tl
= 2t j ∏
t j − tl
,
l =1 l =1
∏(−t j + tl ) l = j l = j
l =1
l = j

which completes the proof.

Obviously, the inner product (5) and the monic polynomial Sn depend on the param-
eter α > −1, so that in what follows, we will denote Snα = Sn . Formula (6) allows us to
obtain other asymptotic formulas for the polynomials Snα . Three of them are included in the
following corollary.

Corollary 2. Let α, β > −1, n ∈ Z+ and k ≥ −n. Under the hypotheses of Theorem 2, we obtain
⎛√  ⎞
α+ β √ − β −x − |c j |
Sn + k ( z ) N
(1)
nk+ β/2 Lαn (z)
⇒ (−1) k
−z ∏⎝ √  ⎠, K ⊂ C \ R+ . (27)
j =1 −x + |c j |
α+ β − β √
Sn + k ( z )
(2) ⇒ (−1)k− z , K ⊂ C \ R+ . (28)
nk+ β/2 Snα (z)
⎛√  ⎞
(Snα (z))(ν) N − x − |c j |
(3) ( ν )
⇒ ∏⎝ √  ⎠, K ⊂ C \ R+ . (29)
( Lαn (z)) j =1 − x + |c j |

Proof. Formulas (27) and (28) are direct consequences of Theorem 2 and Lemma 4.
The proof of (29) is by induction on ν. Of course, (6) is (29) for ν = 0. Assume that (29)
is true for ν = κ ≥ 0. Note that
 
(Snα (z))(κ +1) ( Lαn (z))(κ ) (Snα (z))(κ ) (Snα (z))(κ )
= +
( Lαn (z))(κ +1) ( Lαn (z))(κ +1) ( Lαn (z))(κ ) ( Lαn (z))(κ )

From (16), (18) and Lemma 4

( α +κ )
( Lαn (z))(κ ) L n −κ
= ⇒ 0, K ⊂ C \ R+ .
( Lαn (z))(κ +1) ( α +κ +1)
L n −κ −1

Hence, from Theorem 2, we obtain (29) for ν = κ + 1.

Author Contributions: All authors have contributed equally. All authors have read and agreed to
the published version of the manuscript.

139
Mathematics 2023, 11, 1956

Funding: The research of J. Hernández was partially supported by the Fondo Nacional de Innovación
y Desarrollo Científico y Tecnológico (FONDOCYT), Dominican Republic, under grant 2020-2021-
1D1-135.
Data Availability Statement: Not applicable.
Conflicts of Interest: The authors declare no conflict of interest.

References
1. Osilenker, B. Fourier Series in Orthogonal Polynomials; World Scientific: Singapore, 1999.
2. Simon, B. The Christoffel-Darboux kernel. Perspectives in PDE, Harmonic Analysis and Applications: A volume in honor of VG
Maz’ya’s 70th birthday. Proc. Sympos. Pure Math. Am. Math. Soc. 2008, 79, 295–335.
3. Chihara, T.S. An Introduction to Orthogonal Polynomials; Gordon and Breach: New York, NY, USA, 1978.
4. Freud, G. Orthogonal Polynomials; Pergamon Press: Oxford, UK, 1971.
5. Szegő, G. Orthogonal Polynomials, 4th ed.; American Mathematical Society Colloquium Publications Series; American Mathematical
Society: Providence, RI, USA, 1975; Volume 23.
6. Marcellán, F.; Xu, Y. On Sobolev orthogonal polynomials. Expo. Math. 2015, 33, 308–352. [CrossRef]
7. Martinez-Finkelshtein, A. Analytic properties of Sobolev orthogonal polynomials revisited. J. Comput. Appl. Math. 2001, 127,
255–266. [CrossRef]
8. Lagomasino, G.L.; Marcellán, F.; Assche, W.V. Relative asymptotics for orthogonal polynomials with respect to a discrete Sobolev
inner product. Constr. Approx. 1995, 11, 107–137.
9. Alfaro, M.; Lagomasino, G.L.; Rezola, M.L. Some properties of zeros of Sobolev-type orthogonal polynomials. J. Comput. Appl.
Math. 1996, 69, 171–179. [CrossRef]
10. Díaz-González, A.; Pijeira-Cabrera, H.; Pérez-Yzquierdo, I. Rational approximation and Sobolev-type orthogonality. J. Approx.
Theory 2020, 260, 105481. [CrossRef]
11. Littlejohn, L.L. The Krall polynomials: A new class of orthogonal polynomials. Quaest. Math. 1982, 5, 255–265. [CrossRef]
12. Huertas, E.J.; Marcellán, F.; Pijeira-Cabrera, H. An electrostatic model for zeros of perturbed Laguerre polynomials. Proc. Am.
Math. Soc. 2014, 142, 1733–1747. [CrossRef]
13. Lagomasino, G.L.; Pijeira-Cabrera, H.; Pérez, I. Sobolev orthogonal polynomials in the complex plane. J. Comput. Appl. Math.
2001, 127, 219–230. [CrossRef]
14. Schatz, G.C.; Ratner, M.A. Quantum Mechanics in Chemistry; Dover Publications: Mineola, NY, USA, 2002.
15. Due nas, H.; Huertas, E.; Marcellán, F. Asymptotic properties of Laguerre-Sobolev type orthogonal polynomials. Numer.
Algorithms 2012, 26, 51–73. [CrossRef]
16. Marcellán, F.; Zejnullahu, R.; Fejzullahu, B.; Huertas, E. On orthogonal polynomials with respect to certain discrete Sobolev inner
product. Pac. J. Math. 2012, 257, 167–188. [CrossRef]
17. Ahlfors, L. Complex Analysis, 3rd ed.; McGraw-Hill Book Co.: New York, NY, USA, 1979.
18. Rusev, P. Classical Orthogonal Polynomials and Their Associated Functions in Complex Domain; Marin Drinov Academic Publishing
House: Sofia, Bulgaria, 2005.

Disclaimer/Publisher’s Note: The statements, opinions and data contained in all publications are solely those of the individual
author(s) and contributor(s) and not of MDPI and/or the editor(s). MDPI and/or the editor(s) disclaim responsibility for any injury to
people or property resulting from any ideas, methods, instructions or products referred to in the content.

140
 mathematics

Article
On Apostol-Type Hermite Degenerated Polynomials
Clemente Cesarano 1, *,† , William Ramírez 1,2, *,† , Stiven Díaz 2,† , Adnan Shamaoon 3,†
and Waseem Ahmad Khan 4,†

1 Section of Mathematics, Universitá Telematica Internazionale Uninettuno, 00186 Rome, Italy

2 Departamento de Ciencias Naturales y Exactas, Universidad de la Costa, Barranquilla 00928-1345, Colombia
3 Syed Babar Ali School of Science and Engineering, Lahore University of Management Sciences,
Lahore 54792, Pakistan
4 Department of Mathematics and Natural Sciences, Prince Mohammad Bin Fahd University,
P.O. Box 1664, Al Khobar 31952, Saudi Arabia
* Correspondence: [email protected] (C.C.); [email protected] (W.R.)
† These authors contributed equally to this work.

Abstract: This article presents a generalization of new classes of degenerated Apostol–Bernoulli,

Apostol–Euler, and Apostol–Genocchi Hermite polynomials of level m. We establish some algebraic
and differential properties for generalizations of new classes of degenerated Apostol–Bernoulli
polynomials. These results are shown using generating function methods for Apostol–Euler and
Apostol–Genocchi Hermite polynomials of level m.

Keywords: Hermite polynomials; Apostol-type polynomials; degenerate Apostol-type polynomials

MSC: 11B68; 11B83; 11B39; 05A19

1. Introduction
In this document, the customary conventions of mathematical notation are employed,
where N := {1, 2, . . .}; N0 := {0, 1, 2, . . .}; Z refers to a set of integers; R refers to a set of
real numbers; and C refers to a set of complex numbers.
Citation: Cesarano, C.; Ramírez, W.; There have been numerous studies in the literature that have focused on Apostol–
Díaz, S.; Shamaoon, A.; Khan, W.A. Bernoulli, Apostol–Euler, and Apostol–Genocchi Hermite polynomials, as well as their
On Apostol-Type Hermite extensions and relatives. These studies include works in [1–15]. In recent years, several
Degenerated Polynomials. researchers have explored degraded versions of well-known polynomials, such as Bernoulli,
Mathematics 2023, 11, 1914. https:// Euler, falling factorial, and Bell polynomials, by utilizing generating functions, umbral
doi.org/10.3390/math11081914 calculus, and p-adic integrals. Examples of such studies can be found in [16–18].
Academic Editor: Valery Karachik The generalization of two-variable Hermite polynomials introduced by Kampé de
Fériet is given by [19]:
Received: 15 March 2023 [ ω2 ]
η ν ξ ω −2ν
Revised: 10 April 2023 Hω (ξ, η ) = ω! ∑ ν!(ω − 2ν)!
.
Accepted: 17 April 2023 ν =0
Published: 18 April 2023
It is to be noted that [20]
Hω (2ξ, −1) = Hω (ξ ).
These polynomials satisfy the following generating equation:
Copyright: © 2023 by the authors.
∞
Licensee MDPI, Basel, Switzerland. τω
eξτ +ητ =
2

This article is an open access article ∑ Hω (ξ, η )

ω!
. (1)
ω =0
distributed under the terms and
conditions of the Creative Commons
Attribution (CC BY) license (https://
creativecommons.org/licenses/by/
4.0/).

Mathematics 2023, 11, 1914. https://doi.org/10.3390/math11081914 141 https://www.mdpi.com/journal/mathematics

Mathematics 2023, 11, 1914

Two-variable degenerate Hermite polynomials Hn (ξ, η; μ) ([21], p. 65) are defined by

means of the generating function
∞
ξ η τω
(1 + μτ ) μ (1 + μτ 2 ) μ = ∑ Hω (ξ, η; μ)
ω!
. (2)
ω =0

We note that
lim Hω (ξ, η; μ) = Hω (ξ, η ).
μ →0

The first and second kind of Stirling numbers are given, respectively, by (see [22]):
∞
1 τω
[ln(1 + τ )]ν = ∑ S(ω, ν)
ν! ω =ν ω!

and
∞
1 τ τω
( e − 1) ν = ∑ S(ω, ν) .
ν! ω =ν ω!
The generalized falling factorial (ξ |μ)ω with increment μ is defined by (see [18],
Definition 2.3):
ω −1
(ξ |μ)ω = ∏ (ξ − μν),
ν =0

for positive integer ω, with the convention (ξ |μ)0 = 1. Furthermore,

ω
(ξ |μ)ω = ∑ S(ω, ν)μω−ν ξ ν .
ν =0

From the Binomial Theorem, we have

∞
ξ τω
(1 + μτ ) μ = ∑ (ξ |mu)ω ω! .
ω =0

Khan [14] introduced degenerate Hermite–Bernoulli polynomials of the second kind,

defined by
1
∞
log(1 + μτ ) μ ξ η τω
1
(1 + μτ ) μ (1 + μτ 2 ) μ = ∑ H Bω ( ξ, η; μ )
ω!
.
(1 + μτ ) − 1 μ ω =0

For λ, u ∈ C, and α ∈ N, with u = 1, the generalized degenerate Apostol-type Frobe-

nius Euler–Hermite polynomials of order α are given by a generating function (see [15],
p. 569):
⎛ ⎞α
∞
⎝ 1−u ξ
⎠ (1 + μτ ) μ (1 + μτ 2 ) μ =
η τω
1 ∑ H hω ( ξ, η; μ; λ; u )
ω!
. (3)
λ(1 + μτ ) − u
μ ω =0

Taking u = −1 and α = 1 in (3), the degenerate Hermite–Euler polynomials are

obtained (see [7], p. 3, Equation (17)):
∞
2 ξ η τω
1
(1 + μτ ) μ (1 + μτ 2 ) μ = ∑ H Eω ( ξ, η; μ; λ )
ω!
.
λ(1 + μτ ) + 1 μ ω =0

142
Mathematics 2023, 11, 1914

Clemente et al. [23] introduced and studied new families of Apostol-type degenerated
polynomials by means of the following generating functions:
∞
ξ
[m−1,α] τω
τ mα [σ (λ; μ, b; τ )]α (1 + μτ ) μ = ∑ Bω (ξ; μ, b; λ) , (4)
ω =0
ω!

∞
ξ
[m−1,α] τω
2mα [ψ(λ; μ, b; τ )]α (1 + μτ ) μ = ∑ Eω (ξ; μ, b; λ) (5)
ω =0
ω!
and
∞
ξ
[m−1,α] τω
(2τ )mα [ψ(λ; μ, b; τ )]α (1 + μτ ) μ = ∑ Gω (ξ; μ, b; λ) (6)
ω =0
ω!
where,
 −1
m −1
1 (τ log b)l
σ(λ; a, b; τ ) = λ(1 + μτ ) μ − ∑ l!
l =0

and 
m −1 −1
1 (τ log b)l
ψ(λ; μ, b; τ ) = λ(1 + μτ ) μ + ∑ l!
.
l =0

If ξ = 0, in (4)–(6), we obtain the Apostol-type degenerated numbers of order α and

level m:
∞
[m−1,α] τω
τ mα [σ (λ; μ, b; τ )]α = ∑ Bω (ξ; μ, b; λ) ,
ω =0
ω!
∞
[m−1,α] τω
2mα [ψ(λ; μ, b; τ )]α = ∑ Eω (ξ; μ, b; λ) ,
ω =0
ω!
∞
[m−1,α] τω
(2τ )mα [ψ(λ; μ, b; τ )]α = ∑ Gω (ξ; μ, b; λ) .
ω =0
ω!
The past few years have seen significant advancements in the generalizations of special
functions used in mathematical physics. These developments provide an analytical foun-
dation for many exact solutions to problems in mathematical physics and have practical
applications in various fields. One important area of development is the introduction of one-
and double-variable special functions, which have been recognized for their significance
in both pure mathematical and applied contexts. Multi-index and multi-variable special
functions are also necessary for solving problems in several branches of mathematics, such
as partial differential equations and abstract group theory. Hermite polynomials, devel-
oped by Hermite [24–27], are an example of such special functions, which are important
in combinatorics, numerical analysis, and physics. They are associated with the quantum
harmonic oscillator and are utilized in solving the Schrödinger equation for the oscillator.
This article aims to introduce new families of Hermite–Apostol-type degenerated polyno-
mials. Some algebraic properties and relations for these polynomials are derived. These
results extend certain relations and identities of the related polynomials.

2. Generalizations of New Classes of Degenerated Apostol–Bernoulli, Apostol–Euler,

and Apostol–Genocchi Hermite Polynomials of Level m
In this section, based on (2) and (4)–(6), we define new families of Hermite–Apostol-
type degenerated polynomials.

Definition 1. For arbitrary real or complex parameter α and for μ, b ∈ R+ , the generalizations
[m−1,α]
degenerate the Apostol–Bernoulli Hermite polynomials H Bω (ξ, η; μ, b; λ), the generalizations
[m−1,α]
degenerate Apostol–Euler Hermite polynomials H Eω (ξ, η; μ, b; λ), and the generalizations
[m−1,α]
degenerate Apostol–Genocchi Hermite polynomials H Gω (ξ, η; μ, b; λ), m ∈ N, λ ∈ C of

143
Mathematics 2023, 11, 1914

order α and level m, are defined, in a suitable neighborhood of t = 0, by means of the generating
functions:
∞ ω
ξ η
[m−1,α] τ
τ mα [σ (λ; μ, b; τ )]α (1 + aτ ) μ (1 + μτ 2 ) μ = ∑ H Bω (ξ, η; μ, b; λ) , (7)
ω =0
ω!

∞ ω
ξ η
[m−1,α] τ
2mα [ψ(λ; μ, b; τ )]α (1 + μτ ) μ (1 + μτ 2 ) μ = ∑ H Eω (ξ, η; μ, b; λ) (8)
ω =0 ω!
and
∞ ω
ξ η
[m−1,α] τ
(2τ )mα [ψ(λ; μ, b; τ )]α (1 + μτ ) μ (1 + μτ 2 ) μ = ∑ H Gω (ξ, η; μ, b; λ) , (9)
ω =0
ω!

where 
m −1 −1
1 (τ log b)l
σ (λ; μ, b; τ ) = λ(1 + μτ ) μ − ∑ l!
l =0

and 
m −1 −1
1 (τ log b)l
ψ(λ; μ, b; τ ) = λ(1 + μτ ) μ + ∑ l!
.
l =0

Note that for α = 1, λ = 1, and b = e in (7), we have

⎛ ⎞α
∞
τω ⎜ ⎟
τm
[m−1,α] ⎜ ⎟ ξ η

∑ lim Bω (ξ, η; μ, b; λ)
ω!
= lim ⎜
μ →0⎝
⎟ (1 + μτ ) μ (1 + μτ 2 ) μ
(τ log b)l ⎠m −1
ω =0 μ →0
1
λ(1 + μτ ) − ∑ μ

l =0 l!
⎛ ⎞
⎜ ⎟
τm
⎜ ⎟ ξτ +ητ2
=⎜ ⎟e
⎝ τ l ⎠
m −1
τ
e − ∑
l =0 l!
∞
[ m −1] τω
= ∑ Bω (ξ, η )
ω!
,
ω =0

[ m −1]
where Bω (ξ ) are called generalized Hermite–Bernoulli polynomials (see [28], Equation (6)).
Analogously,
∞ ∞
[m−1,α] τω [ m −1] τω
∑ lim Eω (ξ, η; μ, b; λ)
ω!
= ∑ Eω (ξ, η )
ω!
,
ω =0 μ →0 ω =0
∞ ∞
[m−1,α] τω [ m −1] τω
∑ lim Gω (ξ, η; μ, b; λ)
ω!
= ∑ Gω (ξ, η )
ω!
.
ω =0 μ →0 ω =0

[m−1,α] [m−1,α]
where Eω (ξ ) and Eω (ξ ) are called generalized Hermite–Euler polynomials and generalized
Hermite–Genocchi polynomials, respectively.
If ξ = 0 and η = 0, in Definition 1, we obtain the generalizations of degenerate Apostol–
Bernoulli Hermite numbers, generalizations of degenerate Apostol–Euler Hermite numbers, and
generalizations of degenerate Apostol–Genocchi Hermite numbers of order α and level m.
∞ ω
[m−1,α] τ
τ mα [σ(λ; μ, b; τ )]α = ∑ H Bω (ξ; μ, b; λ) ,
ω =0
ω!

∞ ω
[m−1,α] τ
2mα [ψ(λ; μ, b; τ )]α = ∑ H Eω (ξ; μ, b; λ) ,
ω =0
ω!
∞ ω
[m−1,α] τ
(2τ )mα [ψ(λ; μ, b; τ )]α = ∑ H Gω (ξ; μ, b; λ) .
ω =0
ω!

144
Mathematics 2023, 11, 1914

Continuation will show the standard notation for several sub-classes of polynomials, with
parameters λ ∈ C, μ, b ∈ R+ , order α ∈ N, and level m ∈ N (see [12,29–31] and the references
therein).

[ m −1] [m−1,1]
ω-th generalized Bernoulli polynomial of level m Bω (ξ ) := lim+ H Bω (ξ, 0; μ, e; 1)
μ →0
(α) [m−1,1]
ω-th generalized Euler polynomial of level m Eω (ξ ) := lim H Bω (ξ, 0; μ, e; 1)
μ → 0+
(α) [m−1,1]
ω-th generalized Genocchi polynomial of level m Gω (ξ ) := lim H Gω (ξ, 0; μ, e; 1)
μ → 0+
(α) [0,α]
ω-th generalized Apostol–Genocchi Hermite polynomial Gω (ξ; λ) := lim+ H Gω (ξ, 0; μ, b; λ)
μ →0
[0,1]
ω-th Apostol–Bernoulli polynomial Bω (ξ; λ) := lim H Bω ( ξ, 0; μ, b; λ )
μ → 0+
[0,1]
ω-th Apostol–Euler polynomial Eω (ξ; λ) := lim+ H Eω (ξ, 0; μ, b; λ)
μ →0
[0,1]
ω-th Apostol–Genocchi Hermite polynomial Gω (ξ; λ) := lim+ H Gω (ξ, 0; μ, b; λ)
μ →0
(α) [0,α]
ω-th generalized Bernoulli polynomial Bω (ξ ) := lim H Bω ( ξ, 0; μ, b; 1)
μ → 0+
(α) [0,α]
ω-th generalized Euler polynomial Eω (ξ ) := lim H Eω ( ξ, 0; μ, b; 1)
μ → 0+
(α) [0,α]
ω-th generalized Genocchi polynomial Gω (ξ ) := limμ→0+ H Gω ( ξ, 0; μ, b; 1)
[0,1]
ω-th Bernoulli polynomial Bω (ξ ) := lim H Bω ( ξ, 0; μ, b; 1)
μ → 0+
[0,1]
ω-th Euler polynomial Eω (ξ ) := lim H Eω ( ξ, 0; μ, b; 1)
μ → 0+
[0,1]
ω-th Genocchi polynomial Gω (ξ ) := lim H Gω ( ξ, 0; μ, b; 1)
μ → 0+

Theorem 1. For m ∈ N and the new families of Hermite–Apostol-type degenerated polynomials in invariable
x, with parameters λ ∈ C and μ ∈ Z, order α ∈ N0 and level m, the following relationship holds
ω  
[m−1,α] ω [m−1,α]
H Bω (ξ + γ, η + w; μ, b; λ) = ∑ B
k H ω −k
(ξ, η; μ, b; λ) Hk (γ, w; μ), (10)
k =0
  ω
[m−1,α] ω [m−1,α]
H Eω (ξ + γ, η + w; μ, b; λ) = ∑ E
k H ω −k
(ξ, η; μ, b; λ) Hk (γ, w; μ), (11)
k =0
ω  
[m−1,α] ω [m−1,α]
H Gω (ξ + γ, η + w; μ, b; λ) = ∑ G (ξ, η; μ, b; λ) Hk (γ, w; μ). (12)
k =0
k H ω −k

Proof. By (7) and (2), we have

∞
[m−1,α] τω ξ +γ η +w

∑ H Bω (ξ + γ, η + w; μ, b; λ) = τ mα [σ(λ; μ, b; τ )]α (1 + μτ ) μ (1 + μτ 2 ) μ
ω =0
ω!
ξ η γ w
= τ mα [σ(λ; μ, b; τ )]α (1 + μτ ) μ (1 + μτ 2 ) μ (1 + μτ ) μ (1 + μτ 2 ) μ
 
∞ ∞
[m−1,α] τω τω
= ∑ H Bω (ξ, η; μ, b; λ) ∑ Hω (γ, w; μ) ω!
ω =0
ω! ω =0

∞ ω  
ω [m−1,α] τω
= ∑ ∑ H Bω − ν (ξ, η; μ, b; λ) Hk (γ, w; μ) .
ω =0 ν =0
ν ω!

In view of the above equation, we get the result (10). The proofs of (11) and (12) are given
analogously.

145
Mathematics 2023, 11, 1914

Theorem 2. For m ∈ N and the new families of Hermite–Apostol-type degenerated polynomials in invariable
x, with parameters λ ∈ C and μ ∈ Z, order α ∈ N0 and level m, the argument addition theorem holds
ω  
[m−1,α+ β] ω [m−1,β]
H Bω (ξ + η, γ + w; μ, b; λ) = ∑ B
ν H ν
(η, w; μ, b; λ) (13)
ν =0
[m−1,α]
× H Bω − k (ξ, γ; μ, b; λ),

ω  
[m−1,α+ β] ω [m−1,β]
H Eω (ξ + η, γ + w; μ, b; λ) = ∑ E
ν H ν
(η, w; μ, b; λ) (14)
ν =0
[m−1,α]
× H Eω −ν (ξ, γ; μ, b; λ),
ω  
[m−1,α+ β] ω [m−1,β]
H Gω (ξ + η, γ + w; μ, bλ) = ∑ ν H Gν (η, w; μ, b; λ) (15)
ν =0
[m−1,α]
× H Gω − ν (ξ, γ; μ, b; λ).

Proof. Observe that,

∞
[m−1,α+ β]) τω ξ +η γ+w
∑ H Bω (ξ + η, γ + w; μ, b; λ) = (τ m σ(λ; μ, b; τ ))α+ β (1 + μτ ) μ (1 + μτ 2 ) μ

ω =0
ω!

∞ ω
[m−1,α] τ
= ∑ H Bω (ξ, γ; μ, b; λ)
ω!
ω =0

∞ ω
[m−1,β] τ
× ∑ H Bω (η, w; μ, b; λ)
ω!
ω =0
 
∞ ω
ω [m−1,α]
= ∑ ∑ B
ν H ω −ν
(ξ, γ; μ, b; λ)
ω =0 ν =0
 τω
[m−1,β]
× H Bν (η, w; μ, b; λ) .
ω!
Therefore, by the above equation, we obtain result (13). The proofs of (14) and (15) are given
analogously.

Theorem 3. For m ∈ N and the new families of Hermite–Apostol-type degenerated polynomials in invariable
x, with parameters λ ∈ C and μ ∈ Z, order α ∈ N0 and level m, the following relationships are obeyed:
[m−1,α] [m−1,α] [m−1,α]
H Bω (ξ, η; μ; λ) = H Bω (ξ + μ, η; μ, b; λ) − μω H Bω −1 (ξ, η; μ, b; λ), (16)
[m−1,α] [m−1,α] [m−1,α]
H Eω (ξ, η; μ; λ) = H Eω (ξ + μ, η; μ, b; λ) − μω H Eω −1 (ξ, η; μ, b; λ), (17)
[m−1,α] [m−1,α] [m−1,α]
H Gω (ξ, η; μ; λ) = H Gω (ξ + μ, η; μ, b; λ) − μω H Gω −1 (ξ, η; μ, b; λ). (18)

Proof. From generating function (8), we have

∞ ω
ξ +μ η [m−1,α] τ
(τ m σ(λ; μ, b; τ ))α (1 + μτ ) μ (1 + μτ 2 ) a = (1 + μτ ) ∑ H Eω (ξ, η; μ, b; λ)
ω =0
ω!
∞ ∞
[m−1,α] τω [m−1,α] τ ω
∑ H Eω (ξ + μ, η; μ, b; λ)
ω!
= ∑ H Eω (ξ, η; μ, b; λ)
ω!
ω =0 ω =0
∞ ω
[m−1,α] τ
+μτ ∑ H Eω (ξ, η; μ, b; λ)
ω!
.
ω =0

Then,
∞ ∞
[m−1,α] τω [m−1,α] τ ω
∑ H Eω (ξ + μ, η; μ, b; λ)
ω!
= ∑ H Eω (ξ, η; μ, b; λ)
ω!
ω =0 ω =0
∞
[m−1,α] τ μω
+ ∑ ω H Eω −1 (ξ, η; μ, b; λ)
ω!
.
ω =0

146
Mathematics 2023, 11, 1914

Thus, we have
∞ ∞
[m−1,α] τω [m−1,α]
∑ H Eω (ξ + μ, η; μ, b; λ)
ω!
= ∑ H Eω (ξ, η; μ, b; λ)
ω =0 ω =0
[m−1,α] τω
+μω H Eω −1 (ξ, η; μ, b; λ) .
ω!
In view of the above equation, the result is
[m−1,α] [m−1,α] [m−1,α]
H Eω (ξ, η; μ, b; λ) = H Eω (ξ + μ, η; μ, b; λ) − μω H Eω −1 (ξ, η; μ, b; λ).

Therefore, we obtain (17). The proofs of (16) and (18) are analogous to the previous proce-
dure.

Theorem 4. For m ∈ N, the new families of Hermite–Apostol-type degenerated polynomials in invariable x,

with parameters λ ∈ C and μ ∈ Z, order α ∈ N0 and level m comply with the following relationships:
[m−1,α] [m−1,α] [m−1,α]
H Bω (ξ, η; μ; λ) = H Bω (ξ, η + μ; μ, b; λ) − μω (ω − 1) H Bω −2 (ξ, η; μ, b; λ), (19)

[m−1,α] [m−1,α] [m−1,α]

H Eω (ξ, η; μ; λ) = H Eω (ξ, η + μ; μ, b; λ) − μω (ω − 1) H Eω −2 (ξ, η; μ, b; λ), (20)

[m−1,α] [m−1,α] [m−1,α]

H Gω (ξ, η; μ; λ) = H Gω (ξ, η + μ; μ, b; λ) − μω (ω − 1) H Gω −2 (ξ, η; μ, b; λ). (21)

Proof. From generating function (9), we have

∞ ω
ξ η +μ
[m−1,α] τ
((2τ )m σ(λ; μ, b; τ ))α (1 + μτ ) μ (1 + μτ 2 ) μ = (1 + μτ 2 ) ∑ H Gω (ξ, η; μ, b; λ)
ω =0
ω!
∞ ∞
[m−1,α] τω [m−1,α] τ ω
∑ H Gω (ξ, η + μ; μ, b; λ)
ω!
= ∑ H Gω (ξ, η; μ, b; λ)
ω!
ω =0 ω =0
∞ ω
[m−1,α] τ
+μτ 2 ∑ H Gω (ξ, η; μ, b; λ)
ω!
.
ω =0

Then,
∞ ∞
[m−1,α] τω [m−1,α] τ ω
∑ H Gω (ξ, η + μ; μ, b; λ)
ω!
= ∑ H Gω (ξ, η; μ, b; λ)
ω!
ω =0 ω =0
∞
[m−1,α] τω
+ ∑ H Gω −2 (ξ, η; μ, b; λ)μω (ω − 1) .
ω!
ω =0

Thus, we have
∞ ∞
[m−1,α] τω [m−1,α]
∑ H Gω (ξ, η + μ; μ, b; λ)
ω!
= ∑ H Gω (ξ, η; μ, b; λ)
ω =0 ω =0
[m−1,α] τω
+μω (ω − 1) H Gω −2 (ξ, η; μ, b; λ) .
ω!
Comparing the coefficients of τ ω on both sides of the equation, we obtain the result (21). The
proofs of (19) and (20) are analogous to the previous procedure.

Theorem 5. For m ∈ N, for the new families of Hermite–Apostol-type degenerated polynomials in invariable
x, with parameters λ ∈ C and μ ∈ Z, order α ∈ N0 and level m, the following properties are maintained:

[m−1,α]  
∂ H Bω (ξ, η; μ, b; λ) ω −1 k! ω−1 [m−1,α]
= ∑ ω (−1)k μk H Bω −1−k ( ξ, η; μ, b; λ ), (22)
∂ξ, k =0
k+1 k
[m−1,α]  
∂ H Eω (ξ, η; μ, b; λ) ω −1 k! ω−1 [m−1,α]
= ∑ ω (−1)k μk H Eω −1−k ( ξ, η; μ, b; λ ), (23)
∂ξ, k =0
k+1 k

147
Mathematics 2023, 11, 1914

[m−1,α]  
∂ H Gω (ξ, η; μ, b; λ) ω −1 k! ω−1 [m−1,α]
= ∑ ω (−1)k μk H Gω −1−k ( ξ, η; μ, b; λ ). (24)
∂ξ, k =0
k+1 k

Proof. Partially differentiating (7) with respect to ξ, we have

∞
∂ [m−1,α] τω ∂ ξ η

∑ H Bω (ξ, η; μ, b; λ) = τ mα [σ(λ; μ, b; τ )]α (1 + μτ ) μ (1 + μτ 2 ) μ ,

ω =0
∂ξ ω! ∂ξ
ξ η 1
= τ mα [σ(λ; μ, b; τ )]α (1 + μτ ) μ (1 + μτ 2 ) μ ln(1 + μτ )
μ

∞
[m−1,α] τω
= ∑ H Bω (ξ, η; μ, b; λ)
ω!
ω =0

∞
(−1)ω ω +1 ω +1 1
× ∑ μ τ
ω =0
ω+1 μ
∞ ω
[m−1,α]
= ∑ ∑ H Bω − k (ξ, η; μ, b; λ)
ω =0 k =0
 
ω k! τ ω +1
×(−1)k μk .
k k + 1 ω!

Thus, we have
∞ ∞ ω −1
∂ [m−1,α] τω [m−1,α]
∑ ∂ξ H Bω (ξ, η; μ, b; λ)
ω!
= ∑ ∑ H Bω −1−k ( ξ, η; μ, b; λ )
ω =0 ω =0 k =0
 
ω−1 k! τ ω
×(−1)k μk ω .
k k + 1 ω!

Comparing the coefficients of τ ω on both sides of the equation, the result is

[m−1,α]  
∂ H Bω (ξ, η; μ, b; λ) ω −1 k! ω−1 [m−1,α]
= ∑ ω (−1)k μk H Bω −1−k ( ξ, η; μ, b; λ ).
∂ξ k =0
k + 1 k

The proofs of (23) and (24) are analogous to (22).

Theorem 6. For m ∈ N, for the new families of Hermite–Apostol-type degenerated polynomials in invariable
x, with parameters λ ∈ C and μ ∈ Z, order α ∈ N0 and level m, the following properties are maintained:

[m−1,α]  
∂ H Bω (ξ, η; μ, b; λ) ω −k 2k! ω − 2 [m−1,α]
= ∑ ω (ω − 1)(−1)k μk H Bω −2k−2 ( ξ, η; μ, b; λ ), (25)
∂η k =0
k + 1 2k

[m−1,α]  
∂ H Eω (ξ, η; μ, b; λ) ω −k 2k! ω − 2 [m−1,α]
= ∑ ω (ω − 1)(−1)k ak H Eω −2k −2 ( ξ, η; μ, b; λ ), (26)
∂η k =0
k+1 2k
[m−1,α]  
∂ H Gω (ξ, η; μ, b; λ) ω −k 2k! ω − 2 [m−1,α]
= ∑ ω (ω − 1)(−1)k ak H Gω −2k −2 ( ξ, η; μ, b; λ ). (27)
∂η k =0
k+1 2k

148
Mathematics 2023, 11, 1914

Proof. Partially differentiating (7) with respect to η, we have

∞
∂ [m−1,α] τω ξ ∂ η

∑ H Bω (ξ, η; μ, b; λ) = τ mα [σ(λ; μ, b; τ )]α (1 + μτ ) μ (1 + μτ 2 ) μ

ω =0
∂η ω! ∂η
ξ η 1
= τ mα [σ(λ; μ, b; τ )]α (1 + μτ ) μ (1 + μτ 2 ) μ ln(1 + μτ 2 )
μ
 
∞ ∞
[m−1,α] τω (−1)ω ω +1 2n+2 1
= ∑ H Bω (ξ, η; μ, b; λ) ∑ μ τ
ω =0
ω! ω =0
ω+1 μ
∞ ω
[m−1,α] (−1)k k τ ω +k+2
= ∑ ∑ H Bω − k (ξ, η; μ, b; λ)
k+1
μ
(ω − k)!
.
ω =0 k =0

Thus, we have
∞ ∞ ω −k
∂ [m−1,α] τω [m−1,α]
∑ ∂η H Bω (ξ, η; μ, b; λ)
ω!
= ∑ ∑ H Bω −2−2k ( ξ, η; μ, b; λ )
ω =0 ω =0 k =0
 
ω − 2 2k! τ ω
×(−1)k μk ω (ω − 1) .
2k k + 1 ω!

Comparing the coefficients of τ ω on both sides of the equation, the result is

[m−1,α]  
∂ H Bω (ξ, η; μ, b; λ) ω −k 2k! ω − 2 [m−1,α]
= ∑ ω (ω − 1)(−1)k μk H Bω −2k−2 ( ξ, η; μ, b; λ ).
∂η k =0
k+1 2k

The proofs of (26) and (27) are analogous to (25).

Theorem 7. For m ∈ N, the new families of Hermite–Apostol-type degenerated polynomials in invariable x,

with parameters λ ∈ C and μ ∈ Z, order α ∈ N0 and level m comply with the following relationships:
ω ω  
[m−1,α] [m−1,α] ω [m−1,α]
∑ H Bω − k (ξ, η; μ, b; λ) H Bk (ξ, η; μ, b; λ) = ∑ B
k H ω −k
(μ, b; λ) (28)
k =0 k =0
[m−1,α]
× H Bk (2ξ, 2η; μ, b; λ),

ω ω  
[m−1,α] [m−1,α] ω [m−1,α]
∑ H Eω − k (ξ, η; μ, b; λ) H Ek (ξ, η; μ, b; λ) = ∑ E
k H ω −k
(μ, b; λ) (29)
k =0 k =0
[m−1,α]
× H Ek (2ξ, 2η; μ, b; λ),

ω ω  
[m−1,α] [m−1,α] ω [m−1,α]
∑ H Gω − k (ξ, η; μ, b; λ) H Gk (ξ, η; μ, b; λ) = ∑ G
k H ω −k
(μ, b; λ) (30)
k =0 k =0
[m−1,α]
× H Gk (2ξ, 2η; μ, b; λ).

Proof. Consider the following expressions:

∞ ω
ξ η
[m−1,α] τ
τ mα [σ (λ; μ, b; τ )]α (1 + μτ ) μ (1 + μτ 2 ) μ = ∑ H Bω (ξ, η; μ, b; λ) (31)
ω =0
ω!

and
∞
ξ η
[m−1,α] τω
τ mα [σ (λ; μ, b; τ )]α (1 + μτ ) μ (1 + μτ 2 ) μ = ∑ H Br (ξ, η; μ, b; λ) . (32)
r =0
ω!

149
Mathematics 2023, 11, 1914

From (31) and (32), we have

2ξ 2η
[τ mα [σ(λ; μ, b; τ )]]2α (1 + μτ ) μ (1 + μτ 2 ) μ =
∞ ω ∞
[m−1,α] τ [m−1,α] τω
∑ H Bω (ξ, η; μ, b; λ)
ω! r∑ H Br (ξ, η; μ, b; λ)
ω!
ω =0 =0

∞ ω ∞
[m−1,α] τ [m−1,α] τω
∑ H Bω (μ, b; λ)
ω! r∑ H Br (2ξ, 2η; μ, b; λ)
ω!
=
ω =0 =0
∞ ω ∞
[m−1,α] τ [m−1,α] τω
∑ H Bω (ξ, η; μ, b; λ)
ω! r∑ H Br (ξ, η; μ, b; λ)
ω!
ω =0 =0

∞ ω  
ω [m−1,α] [m−1,α] τω
∑ ∑ k H Bω − k (μ, b; λ) H Bk (2ξ, 2η, μ, b; λ)
ω!
=
ω =0 k =0
∞ ω  
ω [m−1,α] [m−1,α] τω
∑ ∑ k H Bω−k (ξ, η; μ, b; λ) H Bk (ξ, η; μ, b; λ) .
ω!
ω =0 k =0

Hence, we get contention (28).

The proofs of (29) and (30) are comparable to (28).

Theorem 8. For m ∈ N, the new families of Hermite–Apostol-type degenerated polynomials in invariable x,

with parameters λ ∈ C and μ ∈ Z, order α ∈ N0 and level m comply with the following relationships:

[m−1,α] (−1)α ω! [m−1,α]

H Bω (ξ, η; μ, b; −λ) = E (ξ, η; μ, b; λ), (33)
(2)mα (ω − mα)! H ω −mα
[m−1,α] (−2) ω!
mα
[m−1,α]
H Eω (ξ, η; μ, b; −λ) = B (ξ, η; μ, b; λ). (34)
(n + mα)! H n+mα

Proof. Proof of (33). Considering the generating function (7):

∞ ω
ξ η
[m−1,α] τ
τ mα [σ (−λ; μ, b; τ )]α (1 + μτ ) μ (1 + μτ 2 ) μ = ∑ H Bω (ξ, η; μ, b; −λ)
ω =0
ω!
∞
(−1)α 2mα mα ξ η
[m−1,α] τ ω
τ [ψ(λ; μ, b; τ )]α (1 + μτ ) μ (1 + μτ 2 ) μ = ∑ H Bω (ξ, η; μ, b; −λ) ,
2mα ω =0
ω!

we have
∞ ω ∞ n+mα
[m−1,α] τ (−1)α [m−1,α] τ
∑ H Bω (ξ, η; μ, b; −λ)
ω!
=
2mα ∑ H Eω (ξ, η; μ, b; λ)
ω!
ω =0 ω =0
∞ ∞
[m−1,α] τω (−1)α [m−1,α] τω
∑ H Bω (ξ, η; μ, b; −λ)
ω!
=
2mα ∑ H Eω (ξ, η; μ, b; λ)
(ω − mα)!
.
ω =0 ω =0

Therefore, by the above equation, we obtain the result.

Proof. Proof of (34). Considering the generating function (8):

∞ ω
ξ η
[m−1,α] τ
2mα [ψ(λ; μ, b; τ )]α (1 + μτ ) μ (1 + μτ 2 ) μ = ∑ H Eω (ξ, η; μ, b; λ)
ω =0
ω!
∞
(−1)α 2mα mα ξ η
[m−1,α] τ ω
τ [σ(λ; μ, b; τ )]α (1 + μτ ) μ (1 + μτ 2 ) μ = ∑ H Eω (ξ, η; μ, b; −λ) ,
τ mα ω =0
ω!

150
Mathematics 2023, 11, 1914

we have
∞ ∞
[m−1,α] τ ω
[m−1,α] τ (ω −mα)
∑ H Eω (ξ, η; μ, b; −λ)
ω!
= (−2)mα ∑ Bω (ξ, η; μ, b; λ)
ω!
ω =0 ω =0
∞ ∞
[m−1,α] τω [m−1,α] τω
∑ H Eω (ξ, η; μ, b; −λ)
ω!
= (−2)mα ∑ H Bn+mα ( ξ, η; μ, b; λ )
(n + mα)!
.
ω =0 ω =0

In view of the above equation, we obtain the result.

Theorem 9. For m ∈ N, the new families of Hermite–Apostol-type degenerated polynomials in invariable x,

with parameters λ ∈ C and μ ∈ Z, order α ∈ N0 and level m comply with the following relationships:
[m−1,α] [m−1,α]
H Gω (ξ, η; μ, b; −λ) = (−2)mα H Bω (ξ, η; μ, b; λ), (35)
[m−1,α] ω! [m−1,α]
H Gω (ξ, η; μ, b; λ) = E (ξ, η; μ, b; λ). (36)
(ω − mα)! H ω −mα

Proof. Proof of (35). Taking into account the generating function (7), we can observe that
∞ ω
ξ η
[m−1,α] τ
τ mα [σ (λ; μ, b; τ )]α (1 + μτ ) μ (1 + μτ 2 ) μ = ∑ H Bω (ξ, η; μ, b; λ)
ω =0
ω!
∞ ω
α
ξ η [m−1,α] τ
2 mα mα
τ [ψ(−λ; μ, b; τ )] (1 + μτ ) (1 + μτ )
μ 2 a = (−2m )α ∑ H Bω (ξ, η; μ, b; λ) . (37)
ω =0
ω!

Therefore, from (9) and (37), we obtain

∞ ω ∞
[m−1,α] τ [m−1,α] τω
∑ H Gω ( x; μ, b; −λ)
ω!
= ∑ (−2)mα H Bω ( x; μ, b; λ)
ω!
.
ω =0 ω =0

In view of the above equation, we obtain the result.

Proof. Proof of (36). From (9), we have:

∞ ω
ξ η [m−1,α] τ
2mα τ mα [ψ(λ; μ, b; τ )]α (1 + μτ ) μ (1 + μτ 2 ) a = ∑ H Gω (ξ, η; μ, b; λ)
ω =0
ω!
∞ n+mα ∞ ω
[m−1,α] τ [m−1,α] τ
∑ H Eω (ξ, η; μ, b; λ)
ω!
= ∑ H Gω (ξ, η; μ, b; λ)
ω!
,
ω =0 ω =0

then,
∞ ∞
[m−1,α] τω [m−1,α] τ ω
∑ H Eω −mα ( ξ, η; μ, b; λ )
(ω − mα)!
= ∑ H Gω (ξ, η; μ, b; λ)
ω!
.
ω =0 ω =0

Therefore, by the above equation, we obtain the result.

3. Conclusions
In recent years, Apostol-type polynomials have become the subject of intensive research due
to their diverse range of applications, while Bernoulli, Euler, Genocchi, and Hermite polynomials
are well-known families of polynomials with many applications in areas such as numerical analysis,
asymptotic approximation, and special function theory, which have led to a wide range of uses in
engineering and applied sciences [20]. Due to the importance of these application areas, many exten-
sions of Apostol-type polynomials have been studied, such as degenerate Apostol-type polynomials
in [19], Hermite-based Apostol-type polynomials in [2], Laguerre-based Apostol-type polynomials
in [3,24,32], and truncated-exponential-based Apostol-type polynomials, especially in the last decade.
In the literature, extensions of several structures are considered essential if the extension unifies
existing structures. Unification focuses researchers on investigating advanced properties rather than
just studying modified families that have similar properties to the existing area.
The objective of this paper is to examine new families of Hermite–Apostol-type degenerated
polynomials, specifically the Apostol–Bernoulli, Apostol–Euler, and Apostol–Genocchi Hermite
polynomials of level m. These polynomials have significant applications in the areas of applied
mathematics, physics, and engineering. The properties of these polynomials are established based on

151
Mathematics 2023, 11, 1914

classical special functions. The theorems presented in this study demonstrate the usefulness of the
series rearrangement technique for the treatment of special functions theory.
Author Contributions: C.C., W.R., S.D., A.S. and W.A.K. developed the theory and performed the
computations. C.C., W.R., S.D., A.S. and W.A.K. discussed the results. All authors have read and
agreed to the published version of the manuscript.
Funding: This research received no external funding.
Data Availability Statement: Not applicable.
Acknowledgments: The authors acknowledge and appreciate the assistance provided by the Uni-
versidad Telemática Internacional Uninettuno (Italy) and the Universidad de la Costa (Colombia) in
conducting this study.
Conflicts of Interest: The authors declare no conflict of interest.

References
1. Apostol, T. On the Lerch Zeta-function. Pacific J. Math. 1951, 1, 161–167. [CrossRef]
2. Bedoya, D.; Cesarano, C.; Díaz, S.; Ramírez, W. New Classes of Degenerate Unified Polynomials. Axioms 2023, 12, 21. [CrossRef]
3. Bedoya, D.; Ortega, M.; Ramírez, W.; Urieles, A. New biparametric families of Apostol-Frobenius–Euler polynomials of level m.
Mat. Stud. 2021, 55, 10–23. [CrossRef]
4. Castilla, L.; Ramírez, W.; Urieles, A. An Extended Generalized -Extensions for the Apostol Type Polynomials. Abstr. Appl. Anal.
2018, 2018, 2937950. [CrossRef]
5. Cesarano, C. Operational Methods and New Identities for Hermite Polynomials. Math. Model. Nat. Phenom. 2017, 12, 44–50.
[CrossRef]
6. Cesarano, C.; Cennamo, G.M.; Placidi, L. Operational methods for Hermite polynomials with applications. WSEAS Trans. Math.
2014, 13, 925–931.
7. Cesarano, C.; Ramírez, W.; Khan, S. A new class of degenerate Apostol-type Hermite polynomials and applications. Dolomites
Res. Notes Approx. 2022, 15, 10.
8. Cesarano, C. Integral representations and new generating functions of Chebyshev polynomials. Hacet. J. Math. Stat. 2015, 44,
541–552. [CrossRef]
9. Cesarano, C. Generalized Chebyshev polynomials. Hacet. J. Math. Stat. 2014, 43, 731–740.
10. Dattoli, G.; Cesarano, C. On a new family of Hermite polynomials associated with parabolic cylinder functions. Appl. Math.
Comput. 2003, 141, 143–149. [CrossRef]
11. Liu, H.; Wang, W. Some identities on the Bernoulli, Euler and Genocchi polynomials via power sums and alternate power sums.
Discrete Math. 2009, 309, 3346–3363. [CrossRef]
12. Natalini, P.; Bernardini, A. A generalization of the Bernoulli polynomials. J. Appl. Math. 2003, 3, 155–163. [CrossRef]
13. Srivastava, H.M.; Choi, J. Series Associated with the Zeta and Related Functions; Springer: Dordrecht, The Netherlands, 2001.
14. Khan, W.A. Degenerate Hermite-Bernoulli Numbers and Polynomials of the second kind. Prespacetime J. 2016, 7, 1200–1208.
15. Khan, W.A. A new class of degenerate Frobenius Euler–Hermite polynomials. Adv. Stud. Contemp. Math. 2018, 28, 567–576.
16. Burak, K. Explicit relations for the modified degenerate Apostol-type polynomials. BalıKesir üNiversitesi Fen Bilim. EnstitüSü Derg.
2018, 20, 401–412.
17. Lim, D. Some identities of degenerate Genocchi polynomials. Bull. Korean Math. Soc. 2016, 53, 569–579. [CrossRef]
18. Subuhi, K.; Tabinda, N.; Mumtaz, R. On degenerate Apostol-type polynomials and applications. Bol. Soc. Mat. Mex. 2019, 25,
509–528.
19. Appell, P.; Kampé de Fériet, J. Fonctions Hypergéométriques et Hypersphériques Polynomes d’Hermite; Gautier Villars: Paris, France,
1926.
20. Andrews, L.C. Special functions for Engineers and Applied Mathematicians; Macmillan: New York, NY, USA, 1985.
21. Khan, W.A. A note on degenerate Hermite poly-Bernoulli numbers and polynomials. J. Class. Anal. 2016, 8, 65–76. [CrossRef]
22. Srivastava, H.M.; Choi, J. Zeta and q-Zeta Functions and Associated Series and Integrals; Elsevier: London, UK, 2012.
23. Cesarano, C.; Ramírez, W. Some new classes of degenerated generalized Apostol-Bernoulli, Apostol–Euler and Apostol-Genocchi
polynomials. Carpathian Math. Publ. 2022, 14, 354–363.
24. Böck, C.; Kovács, P.; Laguna, P.; Meier, J.; Huemer, M. ECG Beat Representation and Delineation by means of Variable. IEEE Trans.
Biomed. Eng. 2021, 68, 2997–3008. [CrossRef]
25. Dózsa, T.; Radó, J.; Volk, J.; Kisari, A.; Soumelidis, A.; Kovócs, P. Road abnormality detection using piezoresistive force sensors
and adaptive signal models. IEEE Trans. Instrum. Meas. 2022, 71, 9509211. [CrossRef]
26. Kovács, P.; Böck, C.; Dózsa, T.; Meier, J.; Huemer, M. Waveform Modeling by Adaptive Weighted Hermite Functions. In
Proceedings of the 44th IEEE International Conference on Acoustics, Speech and Signal Processing (ICASSP), Brighton, UK, 12–17
May 2019; pp. 1080–1084.

152
Mathematics 2023, 11, 1914

27. Kovács, P.; Bognár, G.; Huber, C.; Huemer, M. VPNET: Variable Projection Networks. Int. J. Neural Syst. 2021, 32, 2150054.
[CrossRef] [PubMed]
28. Pathan, M. A new class of generalized Hermite-Bernoulli polynomials. Georgian Math. J. 2012, 19, 559–573. [CrossRef]
29. Quintana, Y.; Ramírez, W.; Urieles, A. On an operational matrix method based on generalized Bernoulli polynomials of level m.
Calcolo 2018, 55, 30. [CrossRef]
30. Tremblay, R.; Gaboury, S.; Fugére, B.-J. Some new classes of generalized Apostol–Euler and Apostol-Genocchi polynomials. Int. J.
Math. Math. Sci. 2012, 2012, 182785. [CrossRef]
31. Tremblay, R.; Gaboury, S.; Fugère, B.-J. A further generalization of Apostol-Bernoulli polynomials and related polynomials.
Honam Math. J. 2012, 34, 311–326. [CrossRef]
32. Ramírez, W.; Cesarano, C.; Díaz, S. New Results for Degenerated Generalized Apostol–bernoulli, Apostol–euler and Apos-
tol–genocchi Polynomials. WSEAS Trans. Math. 2022, 21, 604–608. [CrossRef]

Disclaimer/Publisher’s Note: The statements, opinions and data contained in all publications are solely those of the individual
author(s) and contributor(s) and not of MDPI and/or the editor(s). MDPI and/or the editor(s) disclaim responsibility for any injury to
people or property resulting from any ideas, methods, instructions or products referred to in the content.

153
 mathematics

Article
The Cauchy Exponential of Linear Functionals on the Linear
Space of Polynomials
Francisco Marcellán 1, *,† and Ridha Sfaxi 2,†

1 Departamento de Matemáticas, Universidad Carlos III de Madrid, Avenida de la Universidad 30,

28911 Leganés, Spain
2 Faculty of Sciences of Gabes, University of Gabes, 6072 Gabes, Tunisia
* Correspondence: [email protected]
† These authors contributed equally to this work.

Abstract: In this paper, we introduce the notion of the Cauchy exponential of a linear functional on
the linear space of polynomials in one variable with real or complex coefficients using a functional
equation by using the so-called moment equation. It seems that this notion hides several properties
and results. Our purpose is to explore some of these properties and to compute the Cauchy exponen-
tial of some special linear functionals. Finally, a new characterization of the positive-definiteness of a
linear functional is given.

Keywords: cauchy power of linear functional; cauchy exponential of linear functional; weakly-
regular linear functional; regular linear functional; positive-definite linear functional; orthogonal
polynomial sequence; Du -Laguerre–Hahn operator

MSC: 33C45; 42C05; 46F10

1. Introduction
We start with a brief overview of some basic notions and results about the linear
space of polynomials in one variable PK := K[ x ], where K = R or C. Let PK  be the algebraic

Citation: Marcellán, F.; Sfaxi, R. The dual space of PK , i.e., the set of all linear functionals from PK to K. Here, u, p is the action
of u ∈ PK on p ∈ P . We denote by ( u ) : = u, x n , n ≥ 0, the moment of order n of the
Cauchy Exponential of Linear K n
Functionals on the Linear Space of linear functional u ∈ PK  . In the sequel, we recall some useful operations in P and some
K
of their properties. For u and v in PK  , f ( x ) = m a x ν in P , a, b and c in K, with a  = 0,
Polynomials. Mathematics 2023, 11, ∑ ν =0 ν K
1895. https://doi.org/10.3390/ let Du = u , f u, uv, ( x − c)−1 u, ha (u), tb (u) and σ (u) be the linear functionals defined by
math11081895 duality [1–4].
Academic Editor: Clemente Cesarano
- The derivative of a linear functional
Received: 22 February 2023
Revised: 13 April 2023 u , p := − u, p , p ∈ PK .
Accepted: 14 April 2023
Published: 17 April 2023
Its moments are (u )n = −n(u)n−1 , n ≥ 0, (u)−1 = 0.
- The left-multiplication of a linear functional by a polynomial f ( x ) = ∑m k
k =0 a k x .

f u, p : = u, f p , p ∈ PK .
Copyright: © 2023 by the authors.
Licensee MDPI, Basel, Switzerland. The corresponding moments are ( f u)n = ∑m
ν=0 aν ( u )n+ν , n ≥ 0.
This article is an open access article
distributed under the terms and - The Cauchy product of two linear functionals.
conditions of the Creative Commons
Attribution (CC BY) license (https:// uv, p := u, vp , p ∈ PK ,
creativecommons.org/licenses/by/
4.0/).

Mathematics 2023, 11, 1895. https://doi.org/10.3390/math11081895 154 https://www.mdpi.com/journal/mathematics

Mathematics 2023, 11, 1895

where the right-multiplication of v by p is a polynomial given by

xp( x ) − yp(y)
(vp)( x ) := vy , , p ∈ Pc .
x−y

Its moments are (uv)n = ∑nν=0 (u)ν (v)n−ν , n ≥ 0.

- The Dirac delta linear functional at a point c.
Given c ∈ K, δc is the Dirac linear functional at point c, defined by

δc , p := p(c), p ∈ PK .

In the sequel, we denote δ = δ0 . Notice that δ is the unit element for the Cauchy
product of linear functionals.
- The division of a linear functional by a polynomial of first degree.

p( x ) − p(c)
( x − c)−1 u, p := u, θc ( p) = u, , p ∈ PK .
x−c
n −1
Its moments are ( x − c)−1 u n
= ∑ c ν ( u ) n −1− ν , n ≥ 0.
ν =0

- The dilation of a linear functional.

h a ( u ), p : = u, ha ( p) = u, p( ax ) , p ∈ PK .

The corresponding moments are ha (u) n

= an (u)n , n ≥ 0.
- The shift of a linear functional.

tb ( u ), p : = u, t−b ( p) = u, p( x + b) , p ∈ PK .
 
n
Its moments are (tb (u))n = ∑nν=0 bν (u)n−ν , n ≥ 0.
ν
- The σ-transformation of a linear functional.

σ (u), p := u, σ( p) = u, p( x2 ) , p ∈ PK .

Its moments are σ (u) n

= (u)2n , n ≥ 0.

As usual, u(n) will denote the nth derivative of u ∈ PK  , with the convention u(0) = u.

By referring to [3], u ∈ PK has an inverse for the Cauchy product, denoted by u−1 , i.e.,

uu−1 = u−1 u = δ, if and only if (u)0 = 0.

Recall that u ∈ PK is said to be symmetric if ( u )
2n+1 = 0, for all n ≥ 0. Moreover, u is
symmetric if and only if σ( xu) = 0, or, equivalently, h−1 u = u.

 is said to be weakly-regular if φu = 0, where

Definition 1 ([5]). A linear functional u ∈ PK
φ ∈ PK , then φ ≡ 0.

Definition 2 ([1,3]). A linear functional u ∈ PK  is said to be regular (quasi-definite, according

to [6]), if there exists a sequence of monic polynomials { Bn ( x )}n≥0 in PK , deg Bn = n, n ≥ 0,

such that u, Bn Bm = rn δn.m , n, m ≥ 0, where rn ∈ K, rn = 0, n ≥ 0, (δn,m is the Kronecker
delta).

In this case, { Bn ( x )}n≥0 is said to be a monic orthogonal polynomial sequence with

respect to u (in short, MOPS). Any regular linear functional on polynomials is weakly-
regular. The converse is not true; see [5].

155
Mathematics 2023, 11, 1895

Definition 3 ([1,6,7]). A linear functional u ∈ PR is said to be positive (resp. positive-definite),

if u, p2 ≥ 0, (resp. u, p2 > 0), for all p ∈ PR , p ≡ 0.

 . The following statements are equivalent.

Proposition 1 ([1,6,7]). Let u ∈ PR
(i) u is positive-definite.
(ii) There exists a MOPS { Bn ( x )}n≥0 in PR such that u, Bn Bm = rn δn,m , for every n, m ≥ 0,
where rn > 0, for all n ≥ 0.

This contribution aims to introduce the analog of the exponential function in the
framework of linear functionals and then provide some of its properties. First of all, we
must specify that the Cauchy exponential of a linear functional is also a linear functional.
We will denote it as eu . On the other hand, it satisfies

eλδ = eλ δ, λ ∈ PK .
eu+v = eu ev , u, v ∈ PK

.

Here, eu ev is the Cauchy product of eu and ev . The Cauchy exponential of a linear

functional on the linear space of polynomials can be defined in several equivalent ways.
The easiest one, which fits best with the theory of linear functionals on the linear space
of polynomials, is based on its moments. Indeed, the moments of eu can be defined in an
iterate way as follows:

n −1
( e u ) 0 = e ( u )0 , n(eu )n = ∑ (n − ν)(eu )ν (u)n−ν , n ≥ 1.
ν =0

Once defined, we highlight several formulas and properties satisfied by the Cauchy
exponential map as a function from PK  to P , and to compute the Cauchy exponential of
K
some classical linear functionals (see [6,8,9]).

e2δ = B(1/2) : Bessel linear functional with parameter α = 1/2.

e−(1/8)δ = B[0] : Symmetric D −semiclassical linear functional of class 1.
eαδ−2 = t−1 J (α, −1 − α) : Shifted Jacobi linear functional.

Among others, the following formulas: are deduced.

h a ( e u ) = eh a u ,
−1
δb−1 tb (eu ) = eδb tb ( u )
,
1 σ(u)
σ(e ) = e
u 2 ,
 and every a, b in K, where a  = 0.
for every u in PK
The manuscript is structured as follows. In Section 2, we first introduce the notion of
the Cauchy exponential of a linear functional on the linear space of polynomials. Second,
we establish several formulas and properties satisfied by the Cauchy exponential map.
In Section 3, we compute the Cauchy power of some special linear functionals by using
some properties of the Cauchy exponential map. In Section 4, we give necessary and
sufficient conditions on a given linear functional on the linear space of polynomials for
its Cauchy exponential will be weakly-regular. In Section 5, we establish a necessary and
sufficient condition on a given linear functional in the linear space of polynomials so that its
Cauchy exponential will be positive-definite. This enables us to give a new characterization
of the positive-definite of a linear functional on the linear space of polynomials. Finally,
some open problems concerning orthogonal polynomials associated with the Cauchy
exponential function of a linear functional are stated.

156
Mathematics 2023, 11, 1895

2. The Cauchy Exponential of a Linear Functional on the Linear Space of Polynomials

2.1. Definition and Basic Properties
For any u ∈ PK , let M( u ) be the linear functional in P that is the solution of the
K
following functional equation:

M(u) 0
= e ( u )0 , x M(u) = ( xu) M(u). (1)

Equivalently, the sequence of moments {(M(u))n }n≥0 satisfies the following recur-
rence relation:
n −1
M(u) 0
= e ( u )0 , n M(u) n
= ∑ (n − ν) M(u) ν (u)n−ν , n ≥ 1. (2)
ν =0

To list some properties of M, we need the following formulas.

 , any f ∈ P , and any a, c in K with a  = 0, we have

Lemma 1 ([2,3]). For any u, v in PK K

( x − c) ( x − c)−1 u = u, (3)
( x − c)−1 ( x − c)u = u − (u)0 δc , (4)
uv = vu, δu = u, (5)
(uv) = u v + uv + x −1 (uv), (6)
( f u) = f u + f  u, (7)
x −1 (uv) = ( x −1 u)v = u( x −1 v). (8)

Following (3), where c = 0, (1) is equivalent to

M(u) 0
= e ( u )0 , M(u) = − x −1 M(u) + x −1 ( xu) M(u). (9)

 , any τ ∈ K, and any non-negative integer n, we have the

Proposition 2. For any u, v in PK
following properties
(i) M(τδ) = eτ δ.
(ii) M(u + v) = M(u)M(v).
n
(iii) M(u) = M(nu).


Proof. From (1) taken with u = τδ, where τ ∈ K, we get M(τδ) 0 = eτ and x M(τδ) =
0. Thus, x M(τδ) = 0. Then, M(τδ) = M(τδ) 0 δ = eτ δ, according to (4) when c = 0.
Hence, (i) holds.
 . Putting v = M( u ), v = M( v ), w = M( u + v ) and w = v v . From
Let u, v in PK 1 2 1 2 1 2
(9), we have

(v1 )0 = e(u)0 , v1 = − x −1 v1 + x −1 ( xu) v1 , (10)

(v2 )0 = e(v)0 , v2 = − x −1 v2 + x −1 ( xv) v2 , (11)
( u + v )0 
( w1 ) 0 = e , w1 = −x −1
w1 + x −1
x ( u + v ) w1 . (12)

Clearly, (w2 )0 = (v1 v2 )0 = (v1 )0 (v2 )0 = e(u)0 e(v)0 = e(u+v)0 .

From (6), (8), (10) and (11), we obtain

w2 = (v1 v2 ) = v1 v2 + v1 v2 + x −1 (v1 v2 )

= − x −1 v1 + x −1 ( xu) v1 v2 + − x −1 v2 + x −1 ( xv) v2 v1 + x −1 (v1 v2 )
 

= − x −1 v1 v2 + x −1 x ( u + v ) v1 v2 .

157
Mathematics 2023, 11, 1895

Therefore,
 

(w2 )0 = e(u+v)0 , w2 = − x −1 v1 v2 + x −1 x ( u + v ) v1 v2 . (13)

From (12), (13), and by the definition of the operator M, we infer that w1 = w2 , i.e.,
M(u + v) = M(u)M(v). Hence, (ii) holds.
The property (iii) is a straightforward consequence of (i) and (ii).

In a natural way, it is convenient to use the following notation


eu := M(u), for every u ∈ PK . (14)

Definition 4. For any u ∈ PK  , the Cauchy exponential of u, that we denote by eu , is the unique

linear functional in PK that satisfies

eu 0
= e ( u )0 , xeu = ( xu) eu .

By an iteration process, we deduce

( e u ) 1 = e ( u )0 ( u ) 1 ,
1
(eu )2 = e(u)0 (u)21 + (u)2 ,
2
( u )0 1
e 3=e
u
(u)31 + (u)1 (u)2 + (u)3 .
6
From Proposition 2 and Definition 4, the following formulas hold.

eτδ = eτ δ, (15)
u+v
e = eu ev , (16)
u n
(e ) = enu , (17)
 , any τ ∈ K and any non-negative integer n.
for any u, v in PK

2.2. Some Properties of the Cauchy Exponential Map

The linear functional Cauchy exponential induces a map in the algebraic dual space
 as follows
PK
 
ExpP : PK −→ PK
K

u −→ ExpP (u) = eu .

K

 , the following properties hold.

Proposition 3. For any u, v in PK
(i) When K = C, then eu = ev if and only if there exists an integer k such that u = v + (2kπi )δ,
where i2 = −1.
(ii) When K = R, then eu = ev if and only u = v.
(iii) ExpP is an isomorphism of Abelian groups from (PR  , +) to (P + , .), where P + = { v ∈
R R R
 |( v ) > 0}.
PR 0

 are such that eu = ev . Then,

Proof. Assume that u, v in PC

( e u ) 0 = e ( u )0 , (eu ) = − x −1 eu + x −1 ( xu) eu ,
( e v ) 0 = e ( v )0 , (ev ) = − x −1 ev + x −1 ( xv) ev .

158
Mathematics 2023, 11, 1895

Since e(u)0 = e(v)0 in C, then there

 exists an integer
 k such that (u)0 = (v)0 + 2kπi, i2 = −1.
 
Moreover, we can see that x −1 x (u − v) eu = 0. Thus, x (u − v) eu = 0, according to

(3) for c = 0. However, since eu is invertible, (eu )0 = 0, then x (u − v) = 0. This requires
that, x (u − v) = 0. Thus, u − v = (u)0 − (v)0 δ = (2kπi )δ, on account of (4) taken with
c = 0.
Conversely, assume that u and v are in PC  such that u = v + (2kπi ) δ. From (15) and

(16), we get e = e
u v +( 2kπi ) δ =e e
v ( 2kπi ) δ = e (e
v 2kπi δ) = ev .
Hence, (i) holds.
The property (ii) is a straightforward consequence of (i).
For any v ∈ PR + , let u be the unique linear functional defined by

n −1
(u)0 = ln (v)0 , n ( u ) n ( v )0 = n ( v ) n − ∑ (n − ν)(u)n−ν (v)ν , n ≥ 1. (18)
ν =1

Equivalently,
( v ) 0 = e ( u )0 , ( xv) = ( xu) v. (19)
By Definition 4, we infer that v = eu . This concludes the proof of (iii).

Furthermore, we need the following formulas.

 , any f ∈ P , and any a, c in K with a  = 0, we have the

Lemma 2 ([2,3]). For any u, v in PK K
following formulas.

ha (u ) = a ha (u) , (20)
x −1 h a ( u ) = a −1 h a ( x −1 u ), (21)
−1
ha ( f u) = f ( a x )ha u, (22)
ha (uv) = ha (u)ha (v), (23)

tb ( u  ) = tb ( u ) , (24)
tb ( f u ) = tb ( f )tb ( u ), (25)
tb (uv) = tb (u)tb (v)δb−1 , (26)
f (uv) = ( f u)v + x (uθ0 f )( x )v, (27)
σ f ( x 2 ) u = f ( x ) σ ( u ), (28)
 
σ (u ) = 2 σ ( xu) , (29)
 
2 σ(u) = σ ( xu) , (30)
σ (uv) = σ(u)σ (v), if either u or v is symmetric. (31)

Proposition 4. For any a, b in K, where a = 0, we have

(i) ExpP ◦ ha = ha ◦ ExpP .
K K
−1
(ii) tb (eu )δb−1 = eδb tb (u) , for all u ∈ PK
 .

(iii) σ e = e
u σ ( u )  .
, for all symmetric u ∈ PK
u
(iv) e is symmetric if and only if u is symmetric.

 . Putting w = eha (u) , then

Proof. Let a ∈ K, with a = 0, and u in PK 1


(w1 )0 = e(ha (u))0 = e(u)0 , w1 = − x −1 w1 + x −1 xha (u) w1 . (32)

159
Mathematics 2023, 11, 1895

Using (20)–(23) and (27), we can derive

ha−1 w1 = a−1 (ha−1 w1 ) ,

h a −1 ( x −1 w1 ) = a −1 x −1 h a −1 w1 ,

ha−1 ( xha u) = a−1 ha−1 ( xha u) = ( xu) ,
 
ha−1 ( xha u) w1 = ( xu) ha−1 w1 ,
   
 
h a −1 x −1 xha (u) w1 = a−1 x −1 ha−1 xha (u) w1 = a−1 x −1 ( xu) ha−1 w1 .

Applying the operator ha−1 in both sides of (32), it follows that

( h a −1 w1 ) 0 = e ( u )0 , (ha−1 w1 ) = − x −1 ha−1 w1 + x −1 ( xu) ha−1 w1 .

From the uniqueness of the solution of the last equation, we can say that ha−1 w1 = eu
and, then, w1 = ha (eu ). Hence, (i) holds.
Assume that b ∈ K and u in PK  . Let first establish the following formula

tb (vδ−b ) = tb (v)δb−1 , v ∈ PK

. (33)

Indeed, by (26), tb (vδ−b ) = tb (v)tb (δ−b )δb−1 . Since tb (δ−b ) = δ, then we have tb (vδ−b ) =
tb (v)δb−1 . Setting w = tb (eu )δb−1 . Clearly, tb (u)δb−1 0 = (u)0 and (w)0 = (eu )0 = e(u)0 . On
the other hand, by (25), (33) and (27),

xw = x tb (eu )δb−1
= xtb (eu δ−b )
 
= tb ( x + b) eu δ−b
 
= tb ( x + b)δ−b eu + x δ−b θ0 ( x + b) ( x )eu .

However, from ( x + b)δ−b = 0 and δ−b θ0 ( x + b) ( x ) = 1, we get xw = tb ( xeu ). From

Definition 4, and while using (24), (26) and (33), we obtain

 

( xw) = tb ( xeu
= tb ( xu) eu
= tb ( xu) tb eu δb−1 = tb ( xu) w

= tb ( xu) w.

From (27), we have xu = ( x + b)δ−b u + x δ−b θ0 ( x + b) u = ( x + b)(uδ−b ). By (25)

and (26), we deduce

tb ( xu) = tb ( x + b)(δ−b u)
= xtb (δ−b u)
= xtb (δ−b )tb (u)δb−1
= xtb (u)δb−1 .
−1 
Accordingly, we have (w)0 = e(tb (u)δb )0
and ( xw) = x tb (u)δb−1 w. From the
−1
uniqueness of the solution of the last equation, we get w = etb (u)δb and, as a consequence,
−1
tb eu δb−1 = etb (u)δb . Hence, (ii) holds.

160
Mathematics 2023, 11, 1895

Next, assume that u is a symmetric linear functional, i.e., σ ( xu) = 0. If w2 = eu , then

( w2 ) 0 = e ( u )0 , ( xw2 ) = −( xu) w2 . (34)

Since u is symmetric, then ( xu) is also symmetric. By (31), (29), and (28), it follows
that

σ ( xu) w2 = σ ( xu) σ w2

= 2 σ ( x2 u) σ w2

= 2 xσ(u) σw2 .

Therefore, if we apply the operator σ in both hand sides of (34), then

 
(σw2 )0 = e(u)0 , xσ (w2 ) = − xσ(u) σ(w2 ).

The uniqueness of the solution of the last equation yields σw2 = eσu .
Hence, (iii) holds.
Assume that u is symmetric, i.e., h−1 u = u. By (i), taken with a = −1, we obtain
h−1 (eu ) = eh−1 (u) = eu . Thus, eu is also symmetric.
Conversely, assume that eu is symmetric, i.e., h−1 (eu ) = eu . Again by (i), when a = −1,
we deduce eh−1 (u) = eu . Notice that

( e u ) 0 = e ( u )0 , ( xeu ) = ( xu) eu .
h−1 ( u ) ( u )0
(e )0 = e , ( xeu ) = ( xh−1 (u)) eu .

This implies ( xu) eu = ( xh−1 (u)) eu . If we multiply both hand sides of the last equa-
tion by e−u , then ( xu) = ( xh−1 (u)) , and so that xu = xh−1 (u). Since (h−1 u)0 = (u)0 ,
then h−1 u = u, by (4) taken with c = 0. Hence, u is symmetric. Thus, the statement (iv) is
proved.

3. Cauchy Power of a Linear Functional

We start recalling the following formulas.

 , any f ∈ P and any a, c in K where a  = 0, we have

Lemma 3 ([2,3,10]). For any u, v in PK K

(u−1 ) = −u−2 u − 2x −1 u−1 . (35)

 and any arbitrary non-negative integer number n, we can define the

For any u in PK
Cauchy power of order n of u, denoted by un , as follows

un = u...u
-./0 , u = δ.
0

n−times

When (u)0 = 0, recall that u is invertible. In such a case, we can extend the definition
−1
of un to negative integer numbers n as follows un = u - ./...u−01 .
(−n)−times
In [11], we have deduced that (u2 ) = 2uu + x −1 u2 . More generally, we have

 , the following properties hold.

Proposition 5. For any u ∈ PK
(i) For every positive integer number n we have

(un ) = nun−1 u + (n − 1) x −1 un .

161
Mathematics 2023, 11, 1895

(ii) If (u)0 = 0, then for every integer number n,

(un ) = nun−1 u + (n − 1) x −1 un .

Proof. We proceed by induction. If n = 1, then u = δu . Therefore, the statement is true.

We assume that the statement is true for n = k, i.e., (uk ) = kuk−1 u + (k − 1) x −1 uk . From
the previous Lemma, we get

( u k +1 )  = ( u k u ) 
= ( u k )  u + u k u  + x −1 u k +1
= kuk−1 u + (k − 1) x −1 uk u + uk u + x −1 uk+1
= ( k + 1) u k u  + ( k ) x −1 u k +1 .

Thus, if the statement is true for n = k, then it also holds for n = k + 1. Hence, (i)
holds.
Assume that (u)0 = 0. Then u is invertible and uu−1 = u−1 u = δ. Clearly, the state-
ment (ii) is true, for n = 0, it comes back to δ = − x −1 δ. Let n be a negative integer number
n. By (i) and Lemma 3, we have

( u n )  = ( u −1 ) − n
= − n ( u −1 ) − n −1 ( u −1 )  − ( n + 1) x −1 ( u −1 ) − n
= −nun+1 (u−1 ) − (n + 1) x −1 un
= −nun+1 (−u−2 u − 2x −1 u−1 ) − (n + 1) x −1 un
= nun−1 u + (n − 1) x −1 un .

Hence, (ii) holds.

First application. Recall that the moments of the classical Bessel linear functional
B(1/2), with parameter α = 12 , are (B(1/2))n = (−n!2) , n ≥ 0. Equivalently, see [7–9],
n


B(1/2) 0
= 1, B(1/2) − ( x + 2)B(1/2) = 0.

Proposition 6. For any integer number m and λ ∈ K, λ = 0, we have

 
(i) h− λ e−2 δ = eλδ .
2
m
(ii) B(1/2) = hm B(1/2) .
  
Proof. We start by showing that e−2 δ = B( 12 ). Indeed, observe that ( xe−2 δ ) + δ e−2 δ = 0.

If we compute the first moments of e−2 δ and multiply the last equation by x, after using
   
(27) and an easy computation, we find e 2 δ 0 = 1, x2 e−2 δ − ( x + 2)e−2 δ = 0. By the
−

uniqueness of the solution of the last equation, e−2 δ = B( 12 ). By Proposition 4, (i), we get
h (−2δ )
h− λ e −2 δ =e −λ
2 . Since h− λ (−2δ ), p = −2δ , p(− λ2 x ) = −λp (0), p ∈ PK , then
2 2
 
h− λ (−2δ ) = λδ . Thus, h− λ (e−2 δ ) = eλδ . Hence, (i) holds.
2 2
m
Let m be a non-zero integer. By (17) and the last property (i), we get B( 12 ) =
 
( e −2 δ ) m = e−2m δ = hm B( 12 ) . Hence, (ii) holds.

162
Mathematics 2023, 11, 1895

Second application. Let first recall that the moments of the generalized Bessel linear
functional B[0] with parameter ν = 0, a symmetric D—semi-classical linear functional of
class one, see [8,9], are

(−1)n
(B[0])2n+1 = 0, (B[0])2n = , n ≥ 0.
22n n!
Equivalently, B[0] satisfies the Pearson equation:

 1
x3 B[0] − (2x2 + )B[0] = 0, where (B[0])0 = 1 and (B[0])1 = 0.
2

Proposition 7. For any integer number m and λ ∈ K, λ = 0, we have

1  
(i) h2i√2λ e 4 δ = eλδ .
m
(ii) B[0] = h√m B[0] .

δ  
Proof. First, let us show that e− 8 = B[0]. Indeed, we have ( xe− 8 δ ) − 14 δ e− 8 δ =
1 1 1

1 
0. If we compute the first moments of e− 8 δ and then multiply the last equation by
1   1 
x2 , we get after using (27) and an easy computation, x3 e− 8 δ − (2x2 + 12 )e− 8 δ =
δ δ
)0 = 1, and (e− 8
1 1
0, with (e 4 )1 = 0. By the uniqueness of the solution of this equation,
δ 1  √ (δ )
= B[0]. By Proposition 4, (i), we get h2i√2λ (e− 8 δ ) = e− 8 h2i
1
we get e − 18 2λ . However,
1  
since h2i√2λ (δ ) = −8λδ , it follows that h2i√2λ e− 8 δ = eλδ . Hence, (i) holds.
m
Let m be a non-zero integer number. By (17) and the last property (i), we get B[0] =
m
(e − 1 δ
8
m 
) = e− 2 δ = h√m B[0] . Hence, (ii) holds.

Third application. Recall that the moments with respect to the sequence {( x − 1)n }n≥0
of the classical Jacobi linear functional J (α, −1 − α) with parameter α, a non-integer
number, are

Γ(n − α)Γ(α)
J (α, −1 − α) = J (α, −1 − α), ( x − 1)n = (−2)n , n ≥ 0.
n,1 Γ(−α)n!

Equivalently, (see [1,7,8])


J (α, −1 − α) 0
= 1, ( x2 − 1)J (α, −1 − α) + (− x + 2α + 1)J (α, −1 − α) = 0.

Notice that the shifted linear functional w = t−1 J (α, −1 − α) satisfies


(w)0 = 1, x ( x + 2)w + (− x + 2α)w = 0.

Proposition 8. For any non-zero complex number c and any positive integer number n, we have

(i) For any non-integer complex number α such nα is a non-integer number, t−1 J (α, −1 −
n
α) = t−1 J (nα, −1 − nα). Equivalently,
 n
J (α, −1 − α) = J (nα, −1 − nα) δ1n−1 .

(ii) For any pair of non-integer complex numbers (α, γ) such that α + γ is a non-integer number,
  
t−1 J (α, −1 − α) t−1 J (γ, −1 − γ) = t−1 J (α + γ, −1 − α − γ).

Equivalently, J (α, −1 − α)J (γ, −1 − γ) = J (α + γ, −1 − α − γ) δ1 .

163
Mathematics 2023, 11, 1895

Proof. Let α be a fixed non-integer complex number. First, let’s show that eαδ−2 =
t−1 J (α, −1 − α). Indeed, if we put w = eαδ−2 , then (w)0 = 1, w − x −1 (δ−2 w) = x −1 w.
Since, δ−2 w = (w)0 δ − 2( x + 2)−1 w = δ − 2( x + 2)−1 w, then (w)0 = 1, w − x −1 w − 2( x +
2)−1 w = x −1 w. If we multiply both hand sides of the last equation by x ( x + 2), we get

x ( x + 2)w + x + 2(α + 1) w = 0, i.e., x ( x + 2)w + (− x + 2α)w = 0. This implies that
α h− c δ−2
w = t−1 J (α, −1 − α). By Proposition 4, (i), h− c (eα δ−2 ) = e 2 . Since, h− c (δ−2 ) = δc ,
2 2
then h− c (e
α δ−2 )
= eα δc .
Hence, the first statement in (i) holds.
2
Let n be a non-zero integer number and α be a non-integer complex number such that
nα is a non-integer number. From (17) and the previous property (i), we get t−1 J (α, −1 −
n n
α) = enα δ−2 = t−1 J (nα, −1 − nα). Therefore, t−1 J (α, −1 − α) = t−1 J (nα, −1 − nα).
By applying the operator t1 and using (26), we get J (α, −1 − α) δ1−n+1 = J (nα, −1 −
n
n n −1
nα). This yields J (α, −1 − α) = J (nα, −1 − nα) δ1 .
Hence, the second statement in (i) holds.
Let (α, γ) be a pair of non-integer complex numbers such that α + γ is a non-integer
number. We can write

t−1 J (α, −1 − α) t−1 J (γ, −1 − γ) = eα δ−2 eγ δ−2

= e(α+γ) δ−2
= t−1 J (α + γ, −1 − α − γ).

Finally, if we apply the operator t1 and we use (26), we find

J (α, −1 − α)J (γ, −1 − γ) = J (α + γ, −1 − α − γ) δ1 .

Hence, (ii) holds.

4. Weak-Regularity Property
We start with the following Lemma.

 , if ( xu ) is weakly-regular, then eu is also weakly-regular.

Lemma 4. For any u ∈ PK

Proof. Assume that u ∈ PK  is such that ( xu ) is weakly-regular. Suppose that there exists

φ ∈ PK , φ ≡ 0 such that φeu = 0. Necessarily, deg(φ) ≥ 1. Indeed, if we suppose that

deg(φ) = 0, then 0 = (φeu )0 = φe(u)0 . This is a contradiction, because φ = 0 and e(u)0 = 0.
From (7), (27) and the definition of Cauchy exponential of a linear functional, we obtain

0 = (φxeu )
= φ ( xeu ) + φ( xeu )
= φ ( xeu ) + φ ( xu) eu
= φ ( xeu ) + (φeu )( xu) + x eu θ0 φ ( x )( xu)
= φ ( xeu ) + x eu θ0 φ ( x )( xu) .

Multiplying both hand sides of the last equation by φ and assuming φeu = 0, we get
xφ(eu θ0 φ)( x )( xu) = 0. This is a contradiction, taking into account ( xu) is weakly-regular
and the fact that deg(φ) ≥ 1, (eu )0 = 0 and so that deg(eu θ0 φ) ≥ 0.

 , the following statements are equivalent.

Proposition 9. For any u in PK
(i) eu is weakly-regular.
(ii) ( xu) is weakly-regular. Otherwise, we must have

min{deg( A) | A ∈ PK , A ≡ 0 and A( xu) = 0} ≥ 2.

164
Mathematics 2023, 11, 1895

Proof. (i ) ⇒ (ii ). Assume that eu is weakly-regular. Suppose that ( xu) is not weakly-
regular. Then there exists A ∈ PK , A ≡ 0, with minimum degree, such that A( xu) = 0 and
deg A ≥ 2. We have to treat two cases.
First case: deg( A) = 0. In such a situation ( xu) = 0, and then u = (u)0 δ. In this case,
e = e(u)0 eδ = e(u)0 δ and then xeu = 0. This contradicts the assumption eu is weakly-regular.
u

Second case: deg( A) = 1. Therefore, there exists c ∈ K such that ( x − c)( xu) = 0. Thus,
( xu) = (( xu) )0 δ = 0 and so that u = (u)0 δ. This is a contradiction.
Hence, min{deg( A) | A ∈ PK , A ≡ 0 and A( xu) = 0} ≥ 2.
(ii ) ⇒ (i ). By Lemma 4, if ( xu) is weakly-regular, eu is also weakly-regular. Assume
that min{deg( A)| A ∈ PK , A ≡ 0 and A( xu) = 0} ≥ 2. Then, there exists A ∈ PK ,
deg( A) ≥ 2, with minimum degree that satisfies A( xu) = 0. We have

A( xeu ) = A ( xu) eu
= A( xu) eu + x ( xu) θ0 A eu
= x ( xu) θ0 A eu .

Equivalently,  
( Axeu ) − A ( x ) + ( xu) θ0 A ( x ) xeu = 0.

The last equation can not be simplified. Otherwise, suppose that it can be simplified
by x − c, where A(c) = 0. Then,
  
( x − c)θc ( A) xeu − ( xu) θ0 ( x − c)θc ( A) ( x ) xeu = 0.

Notice that
( x − c)θc ( A)( x ) − (y − c)θc ( A)(y)
( xu) θ0 ( x − c)θc ( A) ( x ) = (yu) ,
x−y
θ ( A )( x ) − θ c ( A )( y )
= (yu) , ( x − c)
c
+ θc ( A)(y)
x−y
θc ( A)( x ) − θc ( A)(y)
= ( x − c) (yu) , + (yu) , θc ( A)(y) .
x−y
 

Then, ( x − c) θc ( A) xeu − ( xu) θ0 θc ( A) xeu − (yu) , θc ( A)(y) xeu = 0. The sim-
plification by ( x − c) requires the two following conditions:
6
θc ( A)( xeu ) − ( xu) θ0 θc ( A) ( x ) xeu , 1 = 0,
(yu) , θc ( A)(y) = 0.

The simplification gives θc ( A) xeu − ( xu) θ0 θc ( A) ( xeu ) = 0. By the definition of
the Cauchy exponential, θc ( A)( xu) eu − ( xu) θ0 θc ( A)( xeu ) = 0. By (27), it follows that
θc ( A)( xu) eu = 0. If we multiply both hand sides of the last equation by e−u and we use
the property e−u eu = eu e−u = δ, we get θc ( A)( xu) = 0. This contradicts the fact that A is
of minimum degree such that A( xu) = 0.
If V = xeu , then it satisfies ( AV ) − A + ( xu) θ0 A V = 0, where deg A ≥ 2, which
can not be simplified. Moreover, V = 0. Indeed, if V = 0, then eu = e(u)0 δ. This implies
( xu) = 0. This is a contradiction. For the sequel, notice that V is weakly-regular if and
only if eu is weakly-regular. Indeed, suppose that there exists a non-zero polynomial Φ
with a minimal degree such that ΦV = 0. Thus, we have

AV  = ( xu) θ0 A V, (36)
ΦV  = −Φ V. (37)

165
Mathematics 2023, 11, 1895

Since the pseudo-class (see [11]) of V is equal to deg( A), then A divides Φ. So, there
exists Q ∈ PK such that Φ = AQ. From (36) and (37), we have

QAV  = −( QA) V, (38)

Q ( xu) θ0 A V = −( QA) V.

(39)

So, BV = 0, where B = Q ( xu) θ0 A + ( QA) . Since deg( A) ≥ 2, then deg( B) =

deg( Q) + deg( A) − 1 ≥ deg( Q) + 1. Moreover, deg( B) < deg(Φ). This contradicts the
fact that Φ is of minimal degree such that ΦV = 0. Thus, V is weakly-regular and then eu is
also weakly-regular.

5. A Du -Laguerre–Hahn Property
  = { u ∈ P | ( u )  = − n, for all integer n ≥ 1}. For any u
In what follows, let PK K 0
 
in PK , the non-singular lowering operator Du on the linear space of polynomials is defined
by [10,11]

p( x ) − p(y)
Du ( p)( x ) := p ( x ) + uθ0 p( x ) = p ( x ) + uy , , p ∈ PK . (40)
x−y

Let us give some fundamental properties satisfied by the non-singular lowering

operator Du .
Linearity: Du (αp + βq) = αDu ( p) + βDu (q), p, q ∈ PK , α, β ∈ K.
Lowering of degrees:

n −2 −1
Du ( x n )( x ) = n + ( u )0 x n −1 + ∑ ( u ) n − ν −1 x ν , n ≥ 1, ( ∑ = 0),
ν =0 ν =0
D u (1) = 0.

Under the condition (u)0 = −n, for all integer n ≥ 1, we can see that deg(Du ( p)) =
deg( p) − 1, for all p ∈ PK .
Symmetry:
When u is symmetric, i.e., (u)2n+1 = 0, n ≥ 0, and the MPS { Bn ( x )}n≥0 is symmetric,
then the polynomial sequence { Qn ( x )}n≥0 defined by Qn ( x ) = Du ( Bn+1 )( x ), n ≥ 0, is
also symmetric.
The product rule:

Du ( f g) = Du ( f ) g + f Du ( g) + uθ0 ( f g) − (uθ0 f ) g − (uθ0 g) f , f , g ∈ PK . (41)

In particular, we have

Du ( x f )( x ) = xDu ( f )( x ) + f ( x ) + u, f , f ∈ PK . (42)

By transposition of the operator Du , we obtain

t
D u ( w ), p = w, Du ( p)
= w, p + uθ0 p
= −w + x −1 wu, p , 
p ∈ PK , w ∈ PK .

Then, t Du (w) = −w + x −1 (wu), w ∈ PK

 . If we set D : = −t D , we have
u u

Du (w) = w − x −1 (uw), w ∈ PK

, (43)

and we can write

Du (w), p = − w, Du ( p) , p ∈ PK . (44)

166
Mathematics 2023, 11, 1895

The following product rule is a straightforward consequence of the previous defini-

tions and formulas

Du ( f w) = Du ( f )w + f Du w + (wθ0 f )u − (uθ0 f )w, f ∈ PK , w ∈ PK . (45)

 , let S = S ( u ) be the unique linear functional defined by [2]
For any u ∈ PK
6
(S)0 = 1,
(46)
Du (S) = − (u)0 + 1 x −1 S.

Equivalently,
6
(S)0 = 1,
(47)
S − x −1 (uS) = − (u)0 + 1 x −1 S.

i.e.,
6
(S)0 = 1,
(48)
( xS) − u − (u)0 δ S = 0.

Let {en ( x; u)}n≥0 be the sequence of monic polynomials defined by

en := en ( x; u) = S−1 x n , n ≥ 0, (49)

where S is given by (46). Observe that

Du (en ) = (n + (u)0 )en−1 , n ≥ 0. (50)

Clearly, {en ( x; u)}n≥0 is an Appell sequence with respect to Du . In addition, the poly-
nomial sequence {en ( x )}n≥0 can be characterized by

e0 ( x ) = 1, en+1 ( x ) = xen ( x ) + (S−1 )n+1 , n ≥ 0. (51)

 , we have
Proposition 10. For any v ∈ PK

D( xv) (ev ) = − x −1 ev . (52)

 and recall that ev is defined by

Proof. Assume that v ∈ PK

( e v ) 0 = e ( v )0 , ( xev ) = ( xv) ev . (53)

Observe that ( xv) ∈  ,

PK because ( xv) = 0 = −n, n ≥ 1. From (48) taken with
0
u = ( xv) , we have

S ( xv) 0
= 1, xS ( xv) − ( xv) S ( xv) = 0. (54)

By the uniqueness of the solution of each of (53) and (54), we deduce

ev = e(v)0 S ( xv) . (55)

This yields the desired result, according to (46), where u = ( xv) .

167
Mathematics 2023, 11, 1895

−1 n
Setting ẽn ( x ) = en x; ( xv) = S ( xv) x , n ≥ 0. According to (49) and (50), we
can say that

ẽn ( x ) = e(v)0 e−v x n , n ≥ 0. (56)

Du (ẽn ) = nẽn−1 , n ≥ 0. (57)
ẽ0 ( x ) = 1, ẽn+1 ( x ) = x ẽn ( x ) + e(v)0 (e−v )n+1 , n ≥ 0. (58)

From (56), observe that

ev , ẽn = e(v)0 δn,0 , n ≥ 0. (59)

Lemma 5. For any v ∈ PK  , the monic polynomial sequence { ẽ ( x )}

n n≥0 defined by ẽn ( x ) =
( v ) 0 −
e e x , n ≥ 0, satisfies
v n

x ẽn ( x ) + ( xv) ẽn ( x ) = nẽn ( x ), n ≥ 0. (60)

Proof. Assume that v ∈ PK  . Notice that (57) can be rewritten as ẽ ( x ) + ( xv ) θ ẽ ( x ) =

n 0 n
nẽn−1 ( x ), n ≥ 0. If we multiply both hand sides of the last equation by x and we use (58),
then we obtain

x ẽn ( x ) + x ( xv) θ0 ẽn ( x ) ( x ) = n ẽn − e(v)0 (e−v )n , n ≥ 0. (61)

However, from (ye−v ) = −(yv) e−v and while taking into account (56), we get

x ẽn ( x ) − yẽn (y)

x ( xv) θ0 ẽn ( x ) = (yv) , − ẽn (y)
x−y
= ( xv) ẽn ( x ) − (yv) , ẽn (y)
= ( xv) ẽn ( x ) − e(v)0 (yv) e−v , yn
= ( xv) ẽn ( x ) + e(v)0 (ye−v ) , yn
= ( xv) ẽn ( x ) − ne(v)0 e−v )n , n ≥ 0.

Then, (61) gives x ẽn + ( xv) ẽn − ne(v)0 e−v )n = n ẽn − e(v)0 (e−v )n = nẽn , n ≥ 0.
Hence, the desired result.

6. A New Characterization of Positive-Definiteness

We start with the two following technical Lemmas.

Lemma 6 ([5]). For any w ∈ PR  , the following statements are equivalent.

(i) w is positive-definite.
(ii) w is weakly-regular and positive.

Lemma 7. For any g ∈ PK , there exists p ∈ PK , with deg( p) = deg( g), such that

g( x ) − e−(v)0 ev , g = xp ( x ) + ( xv) p ( x ). (62)

168
Mathematics 2023, 11, 1895

Proof. Assume that g ∈ PK . We always have g = ∑νN=0 θν ẽν , where θν ∈ K, 0 ≤ ν ≤ N.

From (59) and (60), we have
N
g( x ) − e−(v)0 ev , g = ∑ θν ẽν ( x ) − e−(v)0 ev , ẽν
ν =0
N
= ∑ θν ẽν ( x ) − e−(v)0 e(v)0 δν,0
ν =0
N
= ∑ θν ẽν (x)
ν =1
N
θν
= ∑ x ẽν ( x ) + ( xv) ẽν ( x )
ν =1
ν
= xp ( x ) + ( xv) p ( x ),
θν
where p( x ) = ∑νN=1 ν ẽν ( x ).

Theorem 1. For any linear functional v ∈ PR  such that ev is weakly-regular, the following
statements are equivalent.
(i) ev is positive-definite.
(ii) For any p ∈ PR , deg( p) = 2l, l ≥ 1, the polynomial xp ( x ) + ( xv) p ( x ) has at least one
real zero.

Proof. (i ) ⇒ (ii ). Let v ∈ PR  such that ev is positive-definite. Suppose that there exists
p ∈ PR , deg( p) = 2l, l ≥ 1, and such that xp ( x ) + ( xv) p ( x ) has not real zeros. Clearly,
deg xp + ( xv) p = 2l. Without loss of generality, we can suppose that the leading
coefficient of p is positive. Then xp ( x ) + ( xv) p ( x ) is a positive polynomial. Under the
assumption ev is positive-definite, then we get ev , xp + ( xv) p > 0. This is a contradiction,
because ev , xp + ( xv) p = −( xev ) + ( xv) ev , p = 0, by the definition of ev . Thus,
xp ( x ) + ( xv) p ( x ) must have at least one real zero.
(ii ) ⇒ (i ). Let g ∈ PR , p ≡ 0 and g ≥ 0. Let deg( g) = 2l, l ≥ 0.
If l = 0, i.e., g( x ) = m > 0, then we have ev , g = me(v)0 > 0.
If l ≥ 1, there exists p ∈ PR , deg( p) = 2l, such that g( x ) − e−(v)0 ev , g = xp ( x ) +
( xv) p ( x ), by virtue of Lemma 7. By the assumption, there exists c ∈ R, such that
g(c) − e−(v)0 ev , g = 0. Then, ev , g = e(v)0 g(c) ≥ 0. Thus, ev is a positive linear functional.
Since ev is weakly-regular, it follows that ev is positive-definite, according to Lemma 6.
+
 , the following statements
Corollary 1. For any weakly-regular linear functional w ∈ PR
are equivalent.
(i) w is positive-definite.
(ii) For any p ∈ PR , deg( p) = 2l, l ≥ 1, the polynomial w−1 x (wp) ( x ) has at least one real
zero.
+
 . By Proposition 3, (iii), there exists a unique v ∈ P  such that w = ev .
Proof. Let w ∈ PR R
By Lemma 7, Theorem 1, and under the assumption w is weakly-regular, we infer that
w is positive-definite, if and only if xp ( x ) + ( xv) p ( x ) has at least one real zero, for all
p ∈ PR , where deg( p) = 2l, l ≥ 1. Let p ∈ PR , deg( p) = 2l, l ≥ 1. We always have
p( x ) = ∑2l (v)0 e−v x n , n ≥ 0. Then,
ν=0 θν ẽν ( x ), where ẽn ( x ) = e

169
Mathematics 2023, 11, 1895

2l
xp ( x ) + ( xv) p ( x ) = ∑ θν x ẽν ( x ) + ( xv) ẽν ( x )
ν =0
2l
= ∑ νθν ẽν (x)
ν =0
2l
= e ( v )0 e − v ∑ νθν xν
ν =0
2l

= e ( v )0 e − v x ∑ θν e−(v)0 ev xν
ν =0
2l

= w −1 x w ∑ θν x ν
ν =0
= w−1 x (wp) ( x ). 

This concludes the proof.

7. Concluding Remarks
In this contribution, the Cauchy exponential of a linear functional in the linear space
of polynomials with either real or complex coefficients has been introduced. Some analytic
and algebraic properties are studied. The Cauchy power of a linear functional is defined.
Some illustrative examples of Jacobi and Bessel’s classical linear functionals are discussed.
A characterization of the weak regularity of the Cauchy exponential of a linear functional is
given. A characterization of the positive definiteness of the Cauchy exponential of a linear
functional is presented.
As further work, we are dealing with the following problems.
(i) Given a regular linear functional u such that its Cauchy exponential eu is also a
regular linear functional there exists a connection formula between the corresponding
sequences of orthogonal polynomials?
(ii) Assuming u is a D—semiclassical linear functional, see [3], is eu a D −semiclassical
linear functional?
(iii) Can do you define other analytic functions of linear functionals in a natural way, by
using the corresponding Taylor expansions?

Author Contributions: Conceptualization, F.M. and R.S.; Methodology, F.M. and R.S.; Validation,
F.M.; Investigation, R.S.; Writing—original draft, F.M. and R.S.; Writing—review & editing, F.M. and
R.S. All authors have read and agreed to the published version of the manuscript.
Funding: The research of R.S. has been supported by the Faculty of Sciences of Gabes, University
of Gabes, City Erriadh 6072 Zrig, Gabes, Tunisia. The research of Francisco Marcellán has been
supported by FEDER/Ministerio de Ciencia e Innovación—Agencia Estatal de Investigación of
Spain, grant PID2021-122154NB-I00, and the Madrid Government (Comunidad de Madrid-Spain)
under the Multiannual Agreement with UC3M in the line of Excellence of University Professors,
grant EPUC3M23 in the context of the V PRICIT (Regional Program of Research and Technological
Innovation).
Acknowledgments: We thank the careful revision by the referees. Their comments and suggestions
have improved the presentation of the manuscript).
Conflicts of Interest: The authors declare that there is no conflict of interest regarding the publication
of this paper.

References
1. Garcia-Ardila, J.C.; Marcellán, F.; Marriaga, M.E. Orthogonal Polynomials and Linear Functionals—An Algebraic Approach and
Applications; EMS Series of Lectures in Mathematics; EMS Press: Berlin, Germany, 2001. [CrossRef]

170
Mathematics 2023, 11, 1895

2. Maroni, P. Sur quelques espaces de distributions qui sont des formes linéaires sur l’espace vectoriel des polynômes. In Orthogonal
Polynomials and Applications (Bar-le-Duc, 1984); Lecture Notes in Math 1171; Springer: Berlin/Heidelberg, Germany, 1985;
pp. 184–194. [CrossRef]
3. Maroni, P. Une théorie algébrique des polynômes orthogonaux. Application aux polynômes orthogonaux semi-classiques.
Orthogonal Polynomials Appl. 1991, 9, 98–130.
4. Trèves, F. Topological Vector Spaces, Distributions and Kernels; Academic Press: New York, NY, USA, 1967.
5. Marcellán, F.; Sfaxi, R. A characterization of weakly regular linear functionals. Rev. Acad. Colomb. Cienc. 2007, 31, 285–295.
6. Chihara, T.S. An Introduction to Orthogonal Polynomials; Gordon and Breach: New York, NY, USA, 1978.
7. Ismail, M.E.H. Classical and quantum orthogonal polynomials in one variable. In Encyclopedia of Mathematics and its Applications;
Cambridge University Press: Cambridge, UK, 2005; Volume 98 . [CrossRef]
8. Maroni, P. Fonctions eulériennes. Polynômes orthogonaux classiques. Tech. L’IngÉnieur A 1994, 154, 1–30. [CrossRef]
9. Maroni, P. An integral representation for the Bessel form. Proceedings of the Fourth International Symposium on Orthogonal
Polynomials and their Applications (Evian-Les-Bains, 1992). J. Comput. Appl. Math. 1995 , 57, 251–260. [CrossRef]
10. Sfaxi, R. On the Laguerre-Hahn Intertwining Operator and Application to Connection Formulas. Acta Appl. Math. 2011, 113,
305–321. [CrossRef]
11. Marcellán, F.; Sfaxi, R. Lowering operators associated with D-Laguerre-Hahn polynomials. Integral Transform.s Spec. Funct. 2011,
22, 879–893. [CrossRef]

Disclaimer/Publisher’s Note: The statements, opinions and data contained in all publications are solely those of the individual
author(s) and contributor(s) and not of MDPI and/or the editor(s). MDPI and/or the editor(s) disclaim responsibility for any injury to
people or property resulting from any ideas, methods, instructions or products referred to in the content.

171
 mathematics

Article
Integral Inequalities Involving Strictly Monotone Functions
Mohamed Jleli † and Bessem Samet *,†

Department of Mathematics, College of Science, King Saud University, P.O. Box 2455,
Riyadh 11451, Saudi Arabia; [email protected]
* Correspondence: [email protected]
† These authors contributed equally to this work.

Abstract: Functional inequalities involving special functions are very useful in mathematical analysis,
and several interesting results have been obtained in this topic. Several methods have been used by
many authors in order to derive upper or lower bounds of certain special functions. In this paper,
we establish some general integral inequalities involving strictly monotone functions. Next, some
special cases are discussed. In particular, several estimates of trigonometric and hyperbolic functions
are deduced. For instance, we show that Mitrinović-Adamović inequality, Lazarevic inequality, and
Cusa-Huygens inequality are special cases of our obtained results. Moreover, an application to
integral equations is provided.

Keywords: integral inequalities; strictly monotone functions; functional inequalities

MSC: 26D15; 26D05; 33B10

1. Introduction
The use of integral inequalities is very frequent in various branches of mathematics
such as differential and partial differential equations, numerical analysis, stability analysis
and measure theory. Due to this fact, the study of integral inequalities is of particular
importance.
Several results related to the development of integral inequalities involving monotone
functions have been published. One of the most useful inequalities in analysis is due to
Citation: Jleli, M.; Samet, B. Integral
Bellman [1]: Let ι, τ, κ ∈ C ([α, β]), α, β ∈ R, α < β, ι > 0 and τ, κ ≥ 0. If ι is monotonic
Inequalities Involving Strictly
nondecerasing, and 
Monotone Functions. Mathematics x
2023, 11, 1873. https://doi.org/ τ ( x ) ≤ ι( x ) + κ (s)τ (s) ds
α
10.3390/math11081873
for all x ∈ [α, β], then  
Academic Editor: Yamilet Quintana x
τ ( x ) ≤ ι( x ) exp κ (s) ds
Received: 11 March 2023 α
Revised: 31 March 2023
for all x ∈ [α, β]. Another important inequality is due to Chebyshev (see e.g., [2]). This
Accepted: 10 April 2023
inequality can be stated as follows. Let ωi ∈ L1 ([α, β]), i = 1, 2, ωi is decreasing for all i, or
Published: 14 April 2023
ωi is increasing for all i. Let ϑ ∈ L1 ([α, β]) and ϑ > 0. Then

2  β
  β
 β

∏ α
ωi ( x )ϑ ( x ) dx ≤
α
ϑ ( x ) dx
α
ω1 ( x )ω2 ( x )ϑ ( x ) dx . (1)
Copyright: © 2023 by the authors. i =1
Licensee MDPI, Basel, Switzerland.
This article is an open access article An extension of the above inequality to higher dimensions have been derived in [3]. In [4–7],
distributed under the terms and reversed inequalities of Hölder, Hardy and Poincaré type have been proved. Some results
conditions of the Creative Commons related to integral inequalities for operator monotonic functions can be found in [8]. Other
Attribution (CC BY) license (https:// integral inequalities involving monotone functions can be found in [9–13].
creativecommons.org/licenses/by/ In [14], using inequality (1), Qi, Cui and Xu established several inequalities involving
4.0/). trigonometric functions and other inequalities involving the integral of sinx x . Motivated by

Mathematics 2023, 11, 1873. https://doi.org/10.3390/math11081873 172 https://www.mdpi.com/journal/mathematics

Mathematics 2023, 11, 1873

the above mentioned contribution and also by the importance of trigonometric inequalities
in real analysis, we establish in this paper new integral inequalities for strictly monotone
functions, which can be useful for obtaining several functional inequalities involving
trigonometric and hyperbolic functions.
Before stating our main results, let us fix some notations:
• N: The set of positive integers.
• a, b ∈ R, a < b.
• f ∈ V ([ a, b]) means that f : [ a, b] → R is C1 , f ([ a, b]) ⊂ [0, +∞[ and

f  (] a, b[) ⊂]0, +∞[ or f  (] a, b[) ⊂] − ∞, 0[.

We present below our results.

Theorem 1. Let σ ∈ R\{0, −2}, f ∈ V ([ a, b]), w ∈ C ([ a, b[) and w(] a, b[) ⊂]0, +∞[. Then,
for every n ∈ N and x ∈] a, b[, it holds that
 x   
f 2 ( a) f 2 (t) 2 f σ +2 ( a ) x
( x − t ) n −1 f σ ( t ) − w(t) dt < ( x − t)n−1 w(t) dt. (2)
a σ σ+2 σ ( σ + 2) a

Theorem 2. Let σ ∈ R\{0, −2}, f ∈ V ([ a, b]), w ∈ C (] a, b]) and w(] a, b[) ⊂]0, +∞[. Then,
for every n ∈ N and x ∈] a, b[, it holds that
 b   
f 2 (b) f 2 (t) 2 f σ +2 ( b ) b
( t − x ) n −1 f σ ( t ) − w(t) dt < (t − x )n−1 w(t) dt. (3)
x σ σ+2 σ ( σ + 2) x

Theorem 3. Let f ∈ C1 ([ a, b]). Assume that f  (] a, b[) ⊂] − ∞, 0[. Then, for every n ∈ N, n ≥ 2,
and x ∈] a, b[, it holds that
 x  x
( x − t)n−1 f (t) dt > (n − 1) ( x − t)n−2 (t − a) f (t) dt. (4)
a a

In the case when f  (] a, b[) ⊂]0, +∞[, we have the following result.

Theorem 4. Let f ∈ C1 ([ a, b]). Assume that f  (] a, b[) ⊂]0, +∞[. Then, for every n ∈ N, n ≥ 2,
and x ∈] a, b[, it holds that
 x  x
( x − t)n−1 f (t) dt < (n − 1) ( x − t)n−2 (t − a) f (t) dt.
a a

Theorem 5. Let f ∈ C1 ([ a, b]). Assume that f  (] a, b[) ⊂] − ∞, 0[. Then, for every n ∈ N, n ≥ 2,
and x ∈] a, b[, it holds that
 b  b
(t − x )n−1 f (t) dt < (n − 1) (t − x )n−2 (b − t) f (t) dt.
x x

Theorem 6. Let f ∈ C1 ([ a, b]). Assume that f  (] a, b[) ⊂]0, +∞[. Then, for every n ∈ N, n ≥ 2,
and x ∈] a, b[, it holds that
 b  b
(t − x )n−1 f (t) dt > (n − 1) (t − x )n−2 (b − t) f (t) dt
x x

for all integer n ≥ 2 and a < x < b.

The proofs of The above theorems are given in Section 2. Next, some special cases are
discussed in Section 3. Finally, in Section 4, an application to integral equations is provided.

173
Mathematics 2023, 11, 1873

2. The Proofs
Proof of Theorem 1. Let
 
F ( t ) = − f  ( t ) f σ −1 ( t ) f 2 ( t ) − f 2 ( a )

for all t ∈] a, b[. Due to the assumptions on f and f  , we have two possible cases:

f  (t) < 0, 0 ≤ f (b) < f (t) < f ( a), a<t<b

or
f  (t) > 0, 0 ≤ f ( a) < f (t) < f (b), a < t < b.
Observe that in both cases, we have

f (] a, b[) ⊂]0, +∞[, F (] a, b[) ⊂] − ∞, 0[.

Then, for all s ∈] a, b[, it holds that

 s
F (t) dt < 0,
a

which is equivalent to
 s 
− f  (t) f σ+1 (t) + f 2 ( a) f  (t) f σ−1 (t) dt < 0. (5)
a

On the other hand, we have

 s 
− f  (t) f σ+1 (t) + f 2 ( a) f  (t) f σ−1 (t) dt
a
# $s
f σ +2 ( t ) f 2 ( a) f σ (t)
= − +
σ+2 σ t= a
f σ +2 ( s ) f 2 ( a) f σ (s) f σ +2 ( a ) f σ +2 ( a )
=− + + −
σ+2 σ σ+2 σ
 2 
σ f ( a) f 2 (s) 2 σ +2
= f (s) − − f ( a ),
σ σ+2 σ ( σ + 2)

which implies by (5) that

 
f 2 ( a) f 2 (s) 2
f σ (s) − < f σ +2 ( a ).
σ σ+2 σ ( σ + 2)

Multiplying by w and integrating over ] a, x [, where x ] a, b[, we obtainr

 x    x
f 2 ( a) f 2 (s) 2
f σ (s) − w(s) ds < f σ +2 ( a ) w(s) ds,
a σ σ+2 σ ( σ + 2) a

which shows that (2) holds for n = 1.

Let us now assume that (2) holds for some p ∈ N, that is,
 y   
f 2 ( a) f 2 (t) 2 f σ +2 ( a ) y
( y − t ) p −1 f σ ( t ) − w(t) dt < (y − t) p−1 w(t) dt
a σ σ+2 σ ( σ + 2) a

174
Mathematics 2023, 11, 1873

for all y ∈] a, b[. Integrating over ] a, x [, where x ∈] a, b[, we obtain

 x  y  2  
f ( a) f 2 (t)
( y − t ) p −1 f σ ( t ) − w(t) dt dy
a a σ σ+2
  y  (6)
2 f σ +2 ( a ) x
< (y − t) p−1 w(t) dt dy.
σ ( σ + 2) a a

On the other hand, by Fubini’s theorem, we have

 x  y  2  
f ( a) f 2 (t)
( y − t ) p −1 f σ ( t ) − w(t) dt dy
a a σ σ+2
 x  2 2 (t)   x 
f ( a ) f
= f σ (t) − w(t) (y − t) p−1 dy dt (7)
a σ σ+2 t
  2 2 (t) 
1 x f ( a ) f
= ( x − t) p f σ (t) − w(t) dt.
p a σ σ+2

Similarly, we have
 x  y 
(y − t) p−1 w(t) dt dy
a a
 x  x 
= w(t) (y − t) p−1 dy dt (8)
a t

1 x
= ( x − t) p w(t) dt.
p a

Thus, it follows from (6)–(8) that

 x   
f 2 ( a) f 2 (t) 2 f σ +2 ( a ) x
( x − t) p f σ (t) − w(t) dt < ( x − t) p w(t) dt,
a σ σ+2 σ ( σ + 2) a

which shows that (2) holds for p + 1. Thus, by induction, (2) holds for every n ∈ N.

Proof of Theorem 2. Let

G (t) = − f  (t) f σ−1 (t)( f 2 (b) − f 2 (t))

for all t ∈] a, b[. Due to the assumptions on f and f  , we have

f (] a, b[) ⊂]0, +∞[, G (] a, b[) ⊂] − ∞, 0[.

Then, for every s ∈] a, b[, there holds

 b
G (t) dt < 0,
s

which is equivalent to
 b 
− f  (t) f σ−1 (t) f 2 (b) + f  (t) f σ+1 (t) dt < 0. (9)
s

On the other hand, we have

 b   2 
f (b) f 2 (s) 2 f σ +2 ( b )
− f  (t) f σ−1 (t) f 2 (b) + f  (t) f σ+1 (t) dt = f σ (s) − − ,
s σ σ+2 σ ( σ + 2)

which implies by (9) that

 
f 2 (b) f 2 (s) 2 f σ +2 ( b )
f σ (s) − < , a < s < b.
σ σ+2 σ ( σ + 2)

175
Mathematics 2023, 11, 1873

Multiplying the above inequality by w(s), we get

 
f 2 (b) f 2 (s) 2 f σ +2 ( b )
f σ (s) − w(s) < w ( s ), a < s < b.
σ σ+2 σ ( σ + 2)

Integrating the above inequality over ] x, b[, where a < x < b, we obtain
 b   
f 2 (b) f 2 (s) 2 f σ +2 ( b ) b
f σ (s) − w(s) ds < w(s) ds,
x σ σ+2 σ ( σ + 2) x

which shows that (3) holds for n = 1.

The rest of the proof is similar to that of the previous theorem.

Proof of Theorem 3. We provide two different proofs of Theorem 3. The second proof was
suggested by one of the referees of the paper.

Proof 1. Let
H (t) = −(t − a) f  (t)
for all t ∈] a, b[. Due to the assumption on f  , we have

H (] a, b[) ⊂]0, +∞[,

which implies that  s

H (t) dt > 0 (10)
a
for every s ∈] a, b[. Integrating by parts, we get
 s  s
H (t) dt = − (t − a) f  (t) dt
a a
  s 
= − [(t − a) f (t)]st=a − f (t) dt
a
  s 
= − (s − a) f (s) − f (t) dt
a
 s
= −(s − a) f (s) + f (t) dt,
a

which implies by (10) that  s

f (t) dt > (s − a) f (s).
a
Integrating over ] a, x [, where x ∈] a, b[, we obtain
 x  s   x
f (t) dt ds > (s − a) f (s) ds. (11)
a a a

Furthermore, an integration by parts yields

 x  s  #  s
$x  x
f (t) dt ds = s f (t) dt − s f (s) ds
a a a s= a a
 x  x
= x f (t) dt − s f (s) ds,
a a

that is,  x  s   x
f (t) dt ds = ( x − t) f (t) dt,
a a a

176
Mathematics 2023, 11, 1873

which implies by (11) that

 x  x
( x − t) f (t) dt > (t − a) f (t) dt, a < x < b.
a a

This shows that (4) holds for n = 2.

Let us now assume that (4) is satisfied for some p ∈ N, p ≥ 2, that is,
 y  y
(y − t) p−1 f (t) dt > ( p − 1) (y − t) p−2 (t − a) f (t) dt
a a

for all y ∈] a, b[. Integrating over ] a, x [, where x ∈] a, b[, we obtain

 x  y   x  y 
(y − t) p−1 f (t) dt dy > ( p − 1) (y − t) p−2 (t − a) f (t) dt dy. (12)
a a a a

On the other hand, by Fubini’s theorem, we have

 x  y 
(y − t) p−1 f (t) dt dy
a a
 x  x 
= f (t) (y − t) p−1 dy dt (13)
a t

1 x
= ( x − t) p f (t) dt
p a

and  x  y 
(y − t) p−2 (t − a) f (t) dt dy
a a
 x  x 
= (t − a) f (t) (y − t) p−2 dy dt (14)
a t
 x
1
= ( x − t) p−1 (t − a) f (t) dt.
p−1 a

Thus, it follows form (12)–(14) that

 x  x
( x − t) p f (t) dt > p ( x − t) p−1 (t − a) f (t) dt,
a a

which shows that (4) holds for p + 1. Hence, by induction, (4) holds for all n ∈ N, n ≥ 2.

Proof 2. Observe first that (4) is equivalent to

 x
( x − t)n−2 ( x − nt + (n − 1) a) f (t) dt > 0. (15)
a

On the other hand, we have

 x
( x − t)n−2 ( x − nt + (n − 1) a) f (t) dt
a
 x−a + a

( x − t)n−2 ( x − nt + (n − 1) a) f (t) dt
n
= (16)
a
 x
+ x−a + a
( x − t)n−2 ( x − nt + (n − 1) a) f (t) dt.
n

Observe that
x−a
x − nt + (n − 1) a > 0, a<t< +a
n

177
Mathematics 2023, 11, 1873

and
x−a
x − nt + (n − 1) a < 0, + a < t < x.
n
Then, since f  (t) < 0 for all a < t < b, we have
 x−a + a

( x − t)n−2 ( x − nt + (n − 1) a) f (t) dt
n

a
   x−a + a (17)
x−a
( x − t)n−2 ( x − nt + (n − 1) a) dt
n
> f +a
n a

and  x
( x − t)n−2 ( x − nt + (n − 1) a) f (t) dt
x−a + a

n
 x (18)
x−a
> f +a x−a + a
( x − t)n−2 ( x − nt + (n − 1) a) dt.
n n

Thus, (16)–(18) yield

 x
( x − t)n−2 ( x − nt + (n − 1) a) f (t) dt
a
  x (19)
x−a
> f +a ( x − t)n−2 ( x − nt + (n − 1) a) dt.
n a

On the other hand, an integration by parts yields

 x
( x − t)n−2 ( x − nt + (n − 1) a) dt
a
 x
1 x n
=− ( x − nt + (n − 1) a)( x − t)n−1 − ( x − t)n−1 dt
n−1 t= a n−1 a
( x − a)n ( x − a)n
= − = 0.
n−1 n−1
Hence, by (19), we obtain (15).

Proof of Theorem 4. Applying inequality (4) with − f instead of f , we obtain the re-
sult.

Proof of Theorem 5. Introducing the function

I (t) = −(b − t) f  (t)

for all t ∈] a, b[, and proceeding as in the proof of Theorem 3, the desired inequality
follows.

Proof of Theorem 6. Applying Theorem 5 with − f instead of f , we obtain the desired

inequality.

3. Some Special Cases

Functional inequalities involving special functions are very useful in mathematical
analysis, and several interesting results have been obtained in this topic. See e.g., [2,15–25].
Here, some estimates involving trigonometric and hyperbolic functions are deduced
from our main results.
 
Corollary 1. Let σ ∈
 R\{ 0, −2}, w ∈ C ([0, π2 )) and w(t) > 0 for every t ∈ 0, π2 . Then, for
evry n ∈ N and x ∈ 0, π2 , it holds that
 x    x
1 cos2 (t) 2
( x − t)n−1 cosσ (t) − w(t) dt < ( x − t)n−1 w(t) dt. (20)
0 σ σ+2 σ ( σ + 2) 0

178
Mathematics 2023, 11, 1873

Proof. Let
f (t) = cos t
 
for all t ∈ 0, . It can be easily seen that f ∈ V ([ a, b]) with a = 0 and b = π2 . Then,
π
2
the functions f and w verify the assumptions of Theorem 1, and (2) holds for all n ∈ N,
σ ∈ R\{0, −2} and 0 < x < π2 . Namely, we have
 x   
cos2 (0) cos2 (t) 2 cosσ+2 (0) x
( x − t)n−1 cosσ (t) − w(t) dt < ( x − t)n−1 w(t) dt,
0 σ σ+2 σ ( σ + 2) 0

which yields (20).

Taking w = 1 in the above result, we obtain the following

π
Corollary 2. Let σ ∈ R\{0, −2}. Then, for all n ∈ N and 0 < x < 2, we have
 x  
1 1 cos2 (t) 2
( x − t)n−1 cosσ (t) − dt < . (21)
xn 0 σ σ+2 nσ(σ + 2)

The following inequality derived by Mitrinović and Adamović [15] is a special case of
Corollary 2.

π
Corollary 3. For all 0 < x < 2, we have
 3
sin x
> cos x. (22)
x

Proof. Taking n = 1 and σ = − 43 in (21), we obtain

 x  
1 3 3 cos2 (t) 9
cos− 3 (t) − −
4
dt < − ,
x 0 4 2 4

that is,  
 x
3 3 cos2 (t) 9x
cos− 3 (t)
4
+ dt > . (23)
0 4 2 4
On the other hand, for all 0 < t < x, we have
   
3 3 cos2 (t) 3 3 3 cos2 (t)
cos− 3 (t) = cos− 3 (t)
4 4
+ cos2 t + sin2 t +
4 2 4 4 2
9 3
cos 3 (t) + cos− 3 (t) sin2 t
2 4
=
4 4 
9 1
cos (t) + cos− 3 (t) sin2 t
2 4
= 3
4 3
 
d 9 −1
= sin t cos 3 (t) ,
dt 4

which yields
 x  
3 3 cos2 (t) 9 −1
cos− 3 (t)
4
+ dt = sin x cos 3 ( x ). (24)
0 4 2 4
Finally, (22) follows from (23) and (24).

Corollary 4. Let σ ∈ R\{0, −2} and w ∈ C (R) be such that

w(t) > 0

179
Mathematics 2023, 11, 1873

for every t > 0. Then, for all n ∈ N and x > 0, it holds that
 x
  x
n −1 1 cosh2 (t) 2
( x − t) σ
cosh (t) − w(t) dt < ( x − t)n−1 w(t) dt. (25)
0 σ σ+2 σ ( σ + 2) 0

Proof. Let b > 0 and

f (t) = cosh t
for every t ∈ [0, b]. It can be easily seen that f ∈ V ([ a, b]), where a = 0. Then, the
functions f and w verify the assumptions of Theorem 1, and (2) is satusfied for every n ∈ N,
σ ∈ R\{0, −2} and x > 0 (since b > 0 is arbitrary chosen). Namely, we obtain
 x
 
n −1 cosh2 (0) cosh2 (t) 2 coshσ+2 (0) x
( x − t) σ
cosh (t) − w(t) dt < ( x − t)n−1 w(t) dt,
0 σ σ+2 σ ( σ + 2) 0

which yields (25).

Taking w = 1 in the above result, we deduce the following inequality.

Corollary 5. Let σ ∈ R\{0, −2}. Then, for all n ∈ N and x > 0, we have
 x

1 n −1 σ 1 cosh2 (t) 2
( x − t) cosh (t) − dt < . (26)
xn 0 σ σ+2 nσ (σ + 2)

The following result due to Lazarevic [16] is a special case of Corollary 5.

Corollary 6. We have
 3
sinh x
> cosh x (27)
x
for every x = 0.

Proof. Without restriction of the generality, we may suppose that x > 0. Taking n = 1 and
σ = − 43 in (26), we obtain

 x

1 3 3 cosh2 (t) 9
cosh− 3 (t) − −
4
dt < − ,
x 0 4 2 4

that is, 
 x
− 43 3 3 cosh2 (t) 9x
cosh (t) + dt > . (28)
0 4 2 4

On the other hand, for all 0 < t < x, we have

 
− 43 3 3 cosh2 (t) − 43 3 3 3 cosh2 (t)
cosh (t) + = cosh (t) cosh2 t − sinh2 t +
4 2 4 4 2
9 3
cosh 3 (t) − cosh− 3 (t) sinh2 t
2 4
=
4 4 
9 1
cosh 3 (t) − cosh− 3 (t) sinh2 t
2 4
=
4 3
 
d 9 −1
= sinh t cosh 3 (t) ,
dt 4

180
Mathematics 2023, 11, 1873

which yields
 x

− 43 3 3 cosh2 (t) 9 −1
cosh (t) + dt = sinh x cosh 3 ( x ). (29)
0 4 2 4

Finally, (27) follows from (28) and (29).

From Theorem 3, we deduce the following inequality.

π
Corollary 7. For all n ∈ N, n ≥ 2 and 0 < x < 2, we have
 x
( x − nt)( x − t)n−2 cos t dt > 0. (30)
0

Proof. Let
f (t) = cos t, t ∈ R.
π
Let a = 0 and b = 2. One has

f  (t) = − sin t < 0, a < t < b.

Then, the function f satisfies the assumptions of Theorem 3. Hence, using (4), we ob-
tain (30).

From Corollary 7, we deduce the following Cusa-Huygens inequality (see [2]).

π
Corollary 8. For all 0 < x < 2, we have

sin x 2 + cos x
< . (31)
x 3

Proof. Taking n = 3 in (30), we obtain that

 x
( x − t)( x − 3t) cos t dt > 0 (32)
0

π
for all 0 < x < 2. A double integration by parts shows that
 x
2 + cos x sin x
( x − t)( x − 3t) cos t dt = − . (33)
0 3 x
Hence, (31) follows from (32) and (33).

Similarly, taking f (t) = cosh(t), t > 0, in Theorem 4, we obtain the following result.

Corollary 9. For all n ∈ N, n ≥ 2 and x > 0, we have

 x
( x − nt)( x − t)n−2 cosh(t) dt < 0. (34)
0

Taking n = 3 in (34), we obtain the following hyperbolic version of inequality (31)

(see [16]).

Corollary 10. We have

sinh x 2 + cosh x
< , x = 0.
x 3

From Theorem 1, we deduce the following inequality.

181
Mathematics 2023, 11, 1873

Corollary 11. We have

 
ln(tan x + sec x ) 20 2 tan x π
> − sec3 x − sec x − 4, 0<x< . (35)
x 3 3 x 2

Proof. Using Theorem 1 with f (t) = cos t, w(t) = cos−5 t, a = 0, b = π

2, σ = 3 and n = 1,
we obtain  x   
1 cos2 t 2 x
cos−2 t − dt < cos−5 t dt (36)
0 3 5 15 0
π
for all 0 < x < 2. Moreover, we have
 x  
1 cos2 t tan x x
cos−2 t − dt = − (37)
0 3 5 3 5

and  
 x sec4 x sin x + 3 1
sec x tan x + 12 ln(tan x + sec x )
2
cos−5 t dt = . (38)
0 4
Using, (36)–(38), we obtain (35).

4. An Application
Our aim is to investigate the the existence and uniqueness of solutions to
 x

n −1 σ 1 cosh2 (t)
u( x ) = ( x − t) cosh (t) − F (t, u(t)) dt, 0 ≤ x ≤ h, (39)
0 σ σ+2

where h > 0, σ > 0, n ∈ N and F : [0, h] × R → R is a continuous function. Namely, using

Corollary 5, we shall establish the following result.

Theorem 7. Assume that there exists α > 0 such that

| F (t, y) − F (t, z)| ≤ α|y − z| (40)

for all 0 < t < h and y, z ∈ R. If

 1 ( 
nσ (σ + 2) n
−1 2
0 < h < min , cosh 1+ , (41)
2α σ

then (39) admits a unique solution u∗ ∈ C ([0, h]). Moreover, for any u0 ∈ C ([0, h]), the Picard
sequence {u p } ⊂ C ([0, h]) defined by

 x

n −1 σ 1 cosh2 (t)
u p +1 ( x ) = ( x − t) cosh (t) − F (t, u p (t)) dt, 0≤x≤h
0 σ σ+2

converges uniformly to u∗ .

Proof. Let us equip C ([0, h]) with the norm

u = max |u( x )|, u ∈ C ([0, h]).

0≤ x ≤ h

It is well-known that (C ([0, h]),  · ) is a Banach space. We introduce the mapping

T : C ([0, h]) → C ([0, h])

182
Mathematics 2023, 11, 1873

defined by
 x

n −1 σ 1 cosh2 (t)
( Tu)( x ) = ( x − t) cosh (t) − F (t, u(t)) dt, 0 ≤ x ≤ h, u ∈ C ([0, h]).
0 σ σ+2

Observe that u ∈ C ([0, h]) is a solution to (39) if and only if u is a fixed point of the mapping
T (i.e., Tu = u). On the other hand, for all u, v ∈ C ([0, h]) and 0 ≤ x ≤ h, we have

|( Tu)( x ) − ( Tv)( x )|
 x
 
1 cosh2 (t) 

≤ ( x − t)n−1 coshσ (t) − | F (t, u(t)) − F (t, v(t))| dt.
0 σ σ+2 

On the other hand, by (41), we have

(
2
0 < h < cosh−1 1+ ,
σ

which implies that

1 cosh2 (t)
− ≥ 0, 0 ≤ t ≤ h.
σ σ+2
Hence, it holds that

|( Tu)( x ) − ( Tv)( x )|
 x

1 cosh2 (t)
≤ ( x − t)n−1 coshσ (t) − | F (t, u(t)) − F (t, v(t))| dt.
0 σ σ+2

Making use of (40), we obtain

|( Tu)( x ) − ( Tv)( x )|
 x

n −1 1 cosh2 (t)
σ
≤α ( x − t) cosh (t)
− |u(t) − v(t)| dt
0 σ σ+2
 x

1 cosh2 (t)
≤ αu − v ( x − t)n−1 coshσ (t) − dt.
0 σ σ+2

Furthermore, using Corollary 5, we get

2αx n
|( Tu)( x ) − ( Tv)( x )| ≤ xn u − v
nσ(σ + 2)
2αhn
≤  u − v .
nσ(σ + 2)

Consequently, we deduce that

 Tu − Tv ≤ ku − v, u, v ∈ C ([0, h]),

where
2αhn
k= .
nσ(σ + 2)
On the other hand, due to (41), one has
 1
nσ (σ + 2) n
0<h< ,
2α

183
Mathematics 2023, 11, 1873

which yields
0 < k < 1.
Thus, from Banach contraction principle (see e.g., [26]), we deduce that T admits a unique
fixed point u∗ ∈ C ([0, h]), and the Picard sequence {u p } defined by u p+1 = Tu p converges
to u∗ with respect to the norm  · . This completes the proof of Theorem 7.

5. Conclusions
Some integral inequalities involving strictly monotone functions are provided. We
shown that the obtained inequalities can be useful for deriving several functional inequal-
ities involving trigonometric and hyperbolic functions. For instance, Theorem 1 unifies
and generalizes Mitrinović-Adamović [15] and Lazarevic [16] inequalities, and Theorem 3
generalizes Cusa-Huygens inequality [2]. By applying Theorem 1, we also obtained a new
ln(tan x + sec x )
inequality (see Corollary 11) that provides a lower bound of the function .
x
Further inequalities can also be obtained by considering other functions f in Theorems 1–6. We
also shown that our obtained results are useful for studying the existence and uniqueness
of solutions to integral equations.

Author Contributions: Investigation, M.J. and B.S. All authors have read and agreed to the published
version of the manuscript.
Funding: The second author is supported by Researchers Supporting Project number (RSP2023R4),
King Saud University, Riyadh, Saudi Arabia.
Data Availability Statement: No datasets were generated or analyzed during the current research.
Conflicts of Interest: The authors declare no conflict of interest.

References
1. Bellman, R.; Cooke, K.L. Differential-Difference Equations; Academic Press: New York, NY, USA, 1963.
2. Mitrinović, D.S.; Vasic, P.M. Analytic Inequalities; Springer: Berlin/Heidelberg, Germany, 1970; Volume 61.
3. Barza, S.; Persson, L.E.; Soria, J. Sharp weighted multidimensional integral inequalities of Chebyshev type. J. Math. Anal. Appl.
1999, 236, 243–253. [CrossRef]
4. Bergh, J.; Burenkov, V.I.; Persson, L.E. On some sharp reversed Hölder and Hardy type inequalities. Math. Nachr. 1994, 169, 19–29.
[CrossRef]
5. Bergh, J.; Burenkov, V.I.; Persson, L.E. Best constants in reversed Hardy’s inequalities for quasimonotone functions. Acta Sci.
Math. 1994, 59, 221–239.
6. Barza, S.; Pecarić, J.; Persson, L.E. Reversed Hölder type inequalities for monotone functions of several variables. Math. Nachr.
1997, 186, 67–80. [CrossRef]
7. Benguria, R.D.; Depassier, C. A reversed Poincaré inequality for monotone functions. J. Inequal. Appl. 2000, 5, 91–96. [CrossRef]
8. Dragomir, S.S. Some integral inequalities for operator monotonic functions on Hilbert spaces. Spec. Math. 2020, 8, 172–180.
[CrossRef]
9. Mond, B.; Pecarić, J.; Peric, I. On reverse integral mean inequalities. Houst. J. Math. 2006, 32, 167–181.
10. Chandra, J.; Fleishman, B.A. On a generalization of the Gronwall-Bellman lemma in partially ordered Banach spaces. J. Math.
Anal. Appl. 1970, 31, 668–681. [CrossRef]
11. Gogatishvili, A.; Stepanov, V.D. Reduction theorems for weighted integral inequalities on the cone of monotone functions. Russ.
Math. Surv. 2013, 68, 597–664. [CrossRef]
12. Rahman, G.; Aldosary, S.F.; Samraiz, M.; Nisar, K.S. Some double generalized weighted fractional integral inequalities associated
with monotone Chebyshev functionals. Fractal Fract. 2021, 5, 275. [CrossRef]
13. Heinig, H.; Maligranda, L. Weighted inequalities for monotone and concave functions. Stud. Math. 1995, 116, 133–165.
14. Qi, F.; Cui, L.-H.; Xu, S.-L. Some inequalities constructed by Tchebysheff’s integral inequality. Math. Inequal. Appl. 1999, 2, 517–528.
[CrossRef]
15. Mitrinović, D.S.; Adamović, D.D. Sur une inégalité élementaire où interviennent des fonctions trigonométriques. Univ. Beogr.
Publ. Elektrotehnickog Fak. Ser. Mat. Fiz. 1965, 149, 23–34.
16. Lazarevic, I. Neke nejednakosti sa hiperbolickim funkcijama. Univ. Beogr. Publ. Elektrotehnickog Fak. Ser. Mat. Fiz. 1966, 170,
41–48.
17. Neuman, E.; Sáandor, J. On some inequalities involving trigonometric and hyperbolic functions with emphasis on the Cusa-
Huygens, Wilker and Huygens inequalities. Math. Inequal. Appl. 2010, 13, 715–723. [CrossRef]

184
Mathematics 2023, 11, 1873

18. Qian, G.; Chen, X.D. Improved bounds of Mitrinović-Adamović-type inequalities by using two-parameter functions. J. Inequal.
Appl. 2023, 2023, 25. [CrossRef]
19. Nishizawa, Y. Sharp exponential approximate inequalities for trigonometric functions. Results Math. 2017, 71, 609–621. [CrossRef]
20. Zhu, L.; Nenezic, M. New approximation inequalities for circular functions. J. Inequal. Appl. 2018, 2018, 313. [CrossRef]
21. Nishizawa, Y. Sharpening of Jordan’s type and Shafer-Fink’s type inequalities with exponential approximations. Appl. Math.
Comput. 2015, 269, 146–154. [CrossRef]
22. Bhayo, B.A.; Sandor, J. On Jordan’s, Redheffer’s and Wilker’ inequality. Math. Inequal. Appl. 2016, 19, 823–839. [CrossRef]
23. Stojiljković, V.; Radojević S.; Cetin, E.; Cavić, V.S.; Radenović, S. Sharp bounds for trigonometric and hyperbolic functions with
application to fractional calculus. Symmetry 2022, 14, 1260. [CrossRef]
24. Thool, S.B.; Bagul, Y.J.; Dhaigude, R.M.; Chesneau, C. Bounds for quotients of inverse trigonometric and inverse hyperbolic
functions. Axioms 2022, 11, 262. [CrossRef]
25. Mortici, C.; Srivastava, H.M. Estimates for the arctangent function related to Shafer’s inequality. Colloq. Math. 2014, 136, 263–270.
[CrossRef]
26. Agarwal, P.; Jleli, M.; Samet, B. Fixed Point Theory in Metric Spaces: Recent Advances and Applications; Springer: Berlin, Germany, 2018.

Disclaimer/Publisher’s Note: The statements, opinions and data contained in all publications are solely those of the individual
author(s) and contributor(s) and not of MDPI and/or the editor(s). MDPI and/or the editor(s) disclaim responsibility for any injury to
people or property resulting from any ideas, methods, instructions or products referred to in the content.

185
 mathematics

Article
A New Discretization Scheme for the Non-Isotropic
Stockwell Transform
Hari M. Srivastava 1,2,3,4, *, Azhar Y. Tantary 5 and Firdous A. Shah 5

1
Department of Mathematics and Statistics, University of Victoria, Victoria, BC V8W 3R4, Canada
2
Center for Converging Humanities, Kyung Hee University, 26 Kyungheedae-ro, Dongdaemun-gu,
Seoul 02447, Republic of Korea
3
Department of Medical Research, China Medical University Hospital, China Medical University,
Taichung 40402, Taiwan
4
Department of Mathematics and Informatics, Azerbaijan University, 71 Jeyhun Hajibeyli Street,
AZ1007 Baku, Azerbaijan
5
Department of Mathematics, University of Kashmir, South Campus, Anantnag 192101, India;
[email protected] (F.A.S.)
* Correspondence: [email protected]

Abstract: To avoid the undesired angular expansion of the sampling grid in the discrete non-isotropic
Stockwell transform, in this communication we propose a scale-dependent discretization scheme that
controls both the radial and angular expansions in unison. Based on the new discretization scheme,
we derive a sufficient condition for the construction of Stockwell frames in L2 (R2 ).

Keywords: stockwell transform; two-dimensional fourier transform; discretization; frame

MSC: 42B10; 42C40; 42C15; 65R10

1. Introduction
For an efficient representation of non-transient signals, R.G. Stockwell [1] introduced
a hybrid time-frequency tool by combining the merits of the classical short-time Fourier
Citation: Srivastava, H.M.; Tantary, and wavelet transforms. For any finite energy signal f ∈ L2 (R), the Stockwell transform
A.Y.; Shah, F.A. A New Discretization with respect to a window function ψ ∈ L2 (R) is defined by
Scheme for the Non-Isotropic 
Stockwell Transform. Mathematics Sψ f (ω, b) = |ω | f (t) ψ ω (t − b) e−2πitω dt, b ∈ R, ω ∈ R \ { 0 } , (1)
2023, 11, 1839. https:// R
doi.org/10.3390/math11081839
where b and ω denote the time and spectral localization parameters, respectively. The Stock-
Academic Editor: Yamilet Quintana well transform (1) offers the absolutely referenced phase information of the given sig-
nal f by fixing the modulating sinusoids with respect to the time axis while translat-
Received: 18 March 2023
ing and dilating the window function ψ. Thus, the Stockwell transform provides a
Revised: 10 April 2023
Accepted: 10 April 2023
frequency-dependent resolution while maintaining a direct relationship with the Fourier
Published: 12 April 2023
spectrum [2–5]. These unique features of the Stockwell transform are apt for diversified
applications to different branches of science and engineering, including geophysics, optics,
quantum mechanics, signal and image processing, and so on [5–12].
To harness the merits of the Stockwell transform in higher dimensions, we have
Copyright: © 2023 by the authors. recently introduced the notion of non-isotropic angular Stockwell transform in [11].
Licensee MDPI, Basel, Switzerland. The essence of such a non-isotropic Stockwell transform lies in the fact that the underlying
This article is an open access article window functions are directionally tunable, which enhances the potency for resolving
distributed under the terms and geometric features in two-dimensional signals. For any f ∈ L2 (R2 ), the non-isotropic
conditions of the Creative Commons angular Stockwell transform with respect to the window function Ψ ∈ L2 (R2 ) is defined as
Attribution (CC BY) license (https://

f (t) Ψ Rθ Aw (t − b) e−2πit
creativecommons.org/licenses/by/ Tw
SΨ f (w, b, θ ) = | det Aw | dt, (2)
4.0/). R2

Mathematics 2023, 11, 1839. https://doi.org/10.3390/math11081839 186 https://www.mdpi.com/journal/mathematics

Mathematics 2023, 11, 1839

where t = (t1 , t2 ) T ∈ R2 , b = (b1 , b2 ) T ∈ R2 , w = (ω1 , ω2 ) T ∈ R2 with ω1 , ω2 = 0 and

θ ∈ [0, 2π ). The matrix Aw ∈ GL(2, R) and the rotation matrix Rθ appearing in (2) are
given by
   
ω1 0 cos θ − sin θ
Aw = and Rθ = , (3)
0 ω2 sin θ cos θ

respectively. Furthermore, in the same article [11], we have also presented a discrete
analogue of (2) by adopting the following procedure:
(i). The frequency variable w = (ω1 , ω2 ) T is discretized by choosing w j = (λ j , λ j ) T ,
where λ > 1 and j ∈ Z. Consequently, the matrix Aw given by (3) takes the form:

λj 0
Aj = .
0 λj

(ii). The angular parameter θ is discretized by sub-dividing the interval [0, 2π ) into
L-equally spaced3 angles by taking θ =  θ0 , where θ0 = 2π/L and  ∈ Z L =
2
0, 1, 2, . . . , L − 1 .
(iii). For m = (m0 , m1 ) T ∈ Z2 and α0 , α1 > 0, the translation parameter b is discretized
by taking into consideration both of the preceding discretizations of w and θ and
j,
choosing bm = A− j R−θ m0 α0 , m1 α1 .
However, much to the dismay, the aforementioned discretization process suffers
from a couple of severe limitations: first, the discretization of the frequency variable w is
non-parabolic in nature; second, the discretization of the angular variable θ is completely
independent of the scale λ, which results in an uncontrollable angular expansion of the
grid at higher values of j (see Figure 1), thereby limiting the directional selectivity at
higher frequencies. In this communication, our goal is to circumvent these limitations
by proposing a new scale-dependent discretization scheme for the discrete non-isotropic
angular Stockwell transform. Under the new discretization scheme, the frequency di-
lation is always doubly effective in one fixed direction as in the orthogonal direction.
Moreover, at each higher level of resolution, the split in the angular region is increased
proportionally, thereby preventing the undesired angular expansion of the sampling grid
and enhancing the directional selectivity at high frequencies.
The rest of the article is organized as follows: Section 2 serves as the pedestal and
deals with the formal aspects of the novel discretization scheme. In Section 3, we derive a
sufficient condition for the non-isotropic Stockwell frames in L2 (R2 ). Finally, a conclusion
together with an impetus to the future research work is extracted in Section 4.

2. Discourse on the New Discretization Scheme

This section is solely devoted to the formulation of a new discretization scheme for the
non-isotropic angular Stockwell transform (2). We reiterate that the proposed discretization
scheme is not only based on the parabolic scaling law but also prevents the undesired
angular expansion of the underlying sampling grid. A detailed exposition of the formal
discrete scheme is given below:
(i). The discretization of the frequency variable w = (ω1 , ω2 ) T is achieved via the
parabolic scaling law by choosing w j = (λ j , λ j/2 ) T , where λ > 1 is a fixed inte-
ger and j ∈ Z determines the level of resolution. Consequently, the anisotropy matrix
is given by

λj 0
Aj = , (4)
0 λ j/2

187
Mathematics 2023, 11, 1839

and the discretized frequency variable w j can be expressed via the matrix A j as

w j = (λ j , λ j/2 ) T = A j (1, 1) T . (5)

(ii) For fixed L0 ∈ Z, the rotation parameter θ is sampled into L0 equi-spaced pieces as

2π  2 3
θ = , where  ∈ Z L0 = 0, 1, 2, . . . , L0 − 1 . (6)
L0

To prevent the expansion of the angular region at higher values of j, it is desirable to

make the spacing between the consecutive angles scale-dependent. As such, we choose
L0 = λ j/2 , where  j/2 denotes the integral part of j/2.
Consequently, the scale-dependent angular discretization is given below:

2π   
θ j = , where  ∈ Zλ j/2 = 0, 1, 2, . . . , λ j/2 − 1 . (7)
λ j/2

(iii) The discretization of the spatial variable b is carried out by taking into consider-
ation both the previous discretizations of frequency and angular variables. For
m = (m1 , m2 ) T ∈ Z2 and β > 0, the spatial variable b is sampled as
 
j,
bm : = A− j R−θ βm . (8)
j

In view of the above discretization scheme, the novel sampling grid associated with the
discrete non-isotropic angular Stockwell transform takes the following form:
6  7
2π 
Λ= A j (1, 1) T , A− j R−θ ( βm), θ j : j ∈ Z, m ∈ Z2 ,  ∈ Zλ j/2 , θ j =  j/2 . (9)
j λ

In order to appreciate the nuances between the existing and the newly proposed dis-
cretization schemes, we depict the respective sampling grids separately in Figures 1 and 2.
For plotting the sampling grid associated with the discretization scheme proposed in [11],
we choose λ = 2, m = (1, 1) T and then partition the angular variable θ = 2π /L,  ∈ Z
in two ways by taking L = 8 and L = 16. Since the existing discretization is not scale-
dependent in the angular variable, with increased levels of resolution the angular expansion
is uncontrollable, as shown in Figure 1.

Figure 1. Basic discrete sampling grid for j = 0, 2, 4, 6, 8 with L = 8 (left) and L = 16 (right) [11].

188
Mathematics 2023, 11, 1839

In contrast to this, the sampling grid (9) efficiently prevents the angular expansion at
higher scales because the new discretization scheme is completely scale-dependent, and
the split in the angular region is increased at each next level of resolution. For a pictorial
illustration of the aforementioned fact, we choose λ = 2, β = 1 in (9) and vary the level
of resolution j over the set {0, 2, 4, 6, 8, 10, . . . }. Then, we observe that for j = 0, there is
no partition in the angular region. Additionally, for j = 2 there are two partitions in the
angular region determined by the points θ02 = 0 and θ12 = π, and the corresponding
partition in the spatial variable is determined by the points b2,0 m = ( A−2 R−θ0 ) m and 2

m = ( A−2 R−θ12 ) m. Furthermore, for j = 4 the angular region attains quadruple partition
b2,1
at the points θ04 = 0, θ14 = π/2, θ24 = π, and θ34 = 3π/2, and consequently the spatial
m = ( A−4 R−θ0 ) m, bm = ( A−4 R−θ1 ) m, bm = ( A−4 R−θ2 ) m
region is partitioned at b4,0 4,1 4,2
4 4 4

m = ( A−4 R−θ34 ) m. In a similar fashion, we can show that for j = 6, 8, 10, . . . both
and b4,3
the angular and spatial regions are partitioned into 8, 16, 32, . . . equispaced regions. Thus,
we infer that at higher values of j, the partition points of the angular region are increased
proportionally; as such, the angular expansion of sampling grid (9) can be efficiently
controlled, as shown in Figure 2.

Figure 2. Refined discrete sampling grid (9) at j = 0, 2, 4, 6, 8, 10.

3. The Non-Isotropic Stockwell Frames

This section is completely devoted to demonstrating that the new discretization scheme
proposed in Section 2 is also helpful for the construction of Stockwell frames in L2 (R2 ).
For A j (1, 1) T , A− j R−θ ( βm), θ j ∈ Λ, we define a quadruple of fundamental op-
j
erators, viz, translation (T( A− j R−θ )( βm) ), dilation (D A j ), rotation (Rθ ), and modulation
j j

(M A j (1,1)T ) operators acting on Ψ ∈ L2 (R2 ) as :

⎫
T ( A− j R−θ )( βm) Ψ ( t )
= Ψ(t − ( A− j R−θ )( βm)) ⎪
⎪
j j ⎪
⎪
  ⎪
⎪
 
D A j Ψ(t) = det A j Ψ( A j t) ⎪
⎪
⎬
. (10)
Rθ Ψ ( t ) = Ψ j ( t ) : = Ψ Rθ t ⎪
⎪
⎪
⎪
⎪
j j
 ⎪
⎪
⎪
⎭
M A j (1,1)T Ψ(t) = Ψ(t) exp 2πi t A j (1, 1)
T T

189
Mathematics 2023, 11, 1839

Upon joint application of the elementary operators defined in (10), we obtain a discrete
collection of analyzing functions Ψ j,m, (t) as

Ψ j,m, (t) = M A j (1,1)T Rθ T( A− j R−θ )( βm) D A j Ψ ( t )

j j
   
= det A j Ψ j A j t − A− j R−θ βm exp 2πi t T A j (1, 1) T (11)
j
   
= det A j Ψ j A j t − R−θ βm exp 2πi t T A j (1, 1) T .
j

Moreover, the two-dimensional Fourier transform of the analyzing functions (11) can
be computed as follows:

Ψ j,m, (t) e−2πi t
Tw
F Ψ j,m, (w) = dt
R2
   
= det A j  Ψ j A j t − R−θ βm exp 2πi t T A j (1, 1) T e−2πit w dt
T

R2 j
   T    T 
= Ψ j (z) exp 2πi A− j z + A− j R−θ βm A j (1, 1) T exp − 2πi A− j z + A− j R−θ βm w dz
R2 j j
6 7    
= exp 2πi ( βm) T Rθ (1, 1) T − A− j w Ψ j (z) exp 2πi z T (1, 1) T exp −2πi z T A− j w dz
j R2
6 7
= exp 2πi ( βm) Rθ (1, 1) − A− j w F Φ j A− j w ,
T T
j

where Φ is the modulated version of the given window function Ψ and is given by
 
Φ j (t) = Ψ j (t) exp 2πi t T (1, 1) T . (12)

Based on the refined sampling grid (9) and the family of analyzing functions con-
structed in (11), we define the novel discrete non-isotropic Stockwell system Γ(Ψ, Λ) as

 
Γ(Ψ, Λ) := Ψ j,m, (t) = M A j (1,1)T Rθ T( A− j R−θ )( βm) D A j Ψ ( t ) : j ∈ Z, m ∈ Z2 ,  ∈ Zλ j/2 . (13)
j j

Then, our main goal is to demonstrate that the system Γ(Ψ, Λ) constitutes a frame for
L2 (R2 ). To facilitate the motive, below we recall the fundamental notion of a frame in a
separable Hilbert space [3]:
2 3
Definition 1. Given a separable Hilbert space H, a sequence of elements f i in H is said to be a
frame for H, if there exists constants 0 < C1 ≤ C2 < ∞, such that
1 1 ; < 2 1 1
1 1   1 1
C1 1 f 1 ≤ ∑ f , f i  ≤ C2 1 f 1 , ∀ f ∈ H. (14)
H i 2 H

The constants C1 and C2 appearing in (14) are called as the lower and upper frame bounds,
respectively. In case C1 = C2 = C > 1, the frame is said to be tight, and if C = 1, the frame is
called a Parseval’s frame.

In the following theorem, we shall derive a sufficient condition for the system Γ(Ψ, Λ)
to be a frame for L2 (R2 ). Prior to that, for any Φ(t) as given by (12), we set
⎛ ⎞
  
 − −   − − 
H (ξ 1 , ξ 2 ) = ess. sup ⎝ ∑ ∑ F Φj λ ω1 , λ ω2 F Φj λ ω1 + ξ 1 , λ ω2 + ξ 2 ⎠. (15)
j j/2 j j/2
ω1 ,ω2 ∈R j∈Z  ∈Z  j/2
λ

190
Mathematics 2023, 11, 1839

Theorem 1. Let Ψ ∈ L2 (R2 ) be any window function and Φ be the corresponding modulated
version given by (12) such that
 2
 
C1 ≤ ∑ ∑ F Φ j λ− j ω1 , λ− j/2 ω2  ≤ C2 , (16)
j∈Z  ∈Z  j/2
λ

almost everywhere ω1 , ω2 ∈ R, with 0 < C1 ≤ C2 < ∞. Then, for fixed β > 0 the system (13)
constitutes a frame for L2 (R2 ) if the function H ( x, y) given by (15) satisfies:
1/2
∑ ∑ H β−1 r, β−1 s H − β−1 r, − β−1 s = C3 < C1 . (17)
0=r ∈Z 0=s∈Z
   
C1 −C3 C2 +C3
Moreover, in that case the lower and upper frame bounds are given by β2
and β2
,
respectively.

Proof. For any f ∈ L2 (R2 ), the implication of Plancheral theorem for the two-dimensional
Fourier transform yields
; < 2
 
∑ ∑ ∑  f , Ψ j,m, 
2
j∈Z m∈Z2  ∈Z  j/2
λ
 2
 
= ∑ ∑ ∑  
 R2 F f (w) F Ψ j,m, (w) dw
j∈Z m∈Z2  ∈Z  j/2
λ
 6 7 2
 
= ∑ ∑ ∑  A− j w exp −2πi ( βm) T Rθ (1, 1) T − A− j w dw
 R2 F f ( w ) F Φ  j j
j∈Z m∈Z2  ∈Z  j/2
λ
 β−1 λ j  β−1 λ j/2
6 7

= ∑ ∑ ∑ λ3j/2  exp −2πi ( βm) T Rθ (1, 1) T − A− j w
0 0 j
j∈Z m∈Z2  ∈Z  j/2
λ
 2

× ∑ ∑ F f ω1 + β−1 λ j n1 , ω2 + β−1 λ j/2 n2 F Φ j λ− j ω1 + β−1 n1 , λ− j/2 ω2 + β−1 n2 dω1 dω2 
n1 ∈Z n2 ∈Z
 β−1 λ j  β−1 λ j/2  #
1 
= ∑ ∑  ∑ ∑ F f ω1 + β −1 n 1 , ω2 + β −1 n 2 (18)
β2 j∈Z  ∈Z  j/2 0 0 n1 ∈Z n2 ∈Z
λ
$2

× F Φ j λ− j ω1 + β−1 λ j n1 , λ− j/2 ω2 + β−1 λ j/2 n2  dω1 dω2
 ∞  ∞ #
1
=
β2 ∑ ∑∑ ∑ −∞ −∞
F f ω1 , ω2 F f ω1 + β−1 λ j r, ω2 + β−1 λ j/2 s
j∈Z r ∈Z s∈Z  ∈Z  j/2
λ
$
× F Φ j λ− j ω1 , λ− j/2 ω2 F Φ j λ− j ω1 + β−1 r, λ− j/2 ω2 + β−1 s dω1 dω2
⎧ ⎫
 
1 ∞ ∞  2 ⎨


 −j − j/2
2 ⎬

F f ω1 , ω2  ∑
⎩ j∈Z  ∈Z∑
= 2 F Φ j λ ω1 , λ ω2  dω1 dω2
β −∞ −∞ ⎭
λ  j/2 
 ∞  ∞ #
1
+ 2 ∑ ∑ ∑
β j∈Z 0=r∈Z 0=s∈Z  ∈Z∑
F f ω1 , ω2 F f ω1 + β−1 λ j r, ω2 + β−1 λ j/2 s
−∞ −∞
λ  j/2 
$
−
× F Φ  j λ ω1 , λ
j − j/2 ω2 F Φ j λ− j ω1 + β−1 r, λ− j/2 ω2 + β−1 s dω1 dω2

= P (principle term) + R (residue term).

191
Mathematics 2023, 11, 1839

Note that the principle term is the product between the power of the input function
and the sum of the spectral powers of the analyzers. Therefore, in view of (16), it follows
that the lower and upper bounds for the principal term are given by
 1 1  1 1
C1 1 12 C2 1 12
1f1 ≤ P ≤ 1f1 . (19)
β 2 2 β2 2

The residue term captures the interference effect among the analyzing functions and
can be computed by invoking the Cauchy–Schwarz inequality twice successively in the
following fashion:
  ∞  ∞ #
1
R =  2 ∑ ∑ ∑ ∑ F f ω1 , ω2 F f ω1 + β−1 λ j r, ω2 + β−1 λ j/2 s
β j∈Z 0=r ∈Z 0=s∈Z  ∈Z  j/2 −∞ −∞
λ
$ 

× F Φ j λ− j ω1 , λ− j/2 ω2 F Φ j λ− j ω1 + β−1 r, λ− j/2 ω2 + β−1 s dω1 dω2 
# ∞  ∞  2  
1    
≤ ∑ ∑ ∑ ∑ F f ω1 , ω2  F Φ j λ− j ω1 , λ− j/2 ω2  (20)
β2 j∈Z 0=r ∈Z 0=s∈Z  ∈Z  j/2 −∞ −∞
λ
  $1/2
 
× F Φ j λ− j ω1 + β−1 r, λ− j/2 ω2 + β−1 s dω1 dω2
# ∞  ∞  2  
   
F f ω1 + β−1 λ j r, ω2 + β−1 λ j/2 s  F Φ j λ− j ω1 , λ− j/2 ω2 
−∞ −∞
  $1/2
 
× F Φ j λ− j ω1 + β−1 r, λ− j/2 ω2 + β−1 s dω1 dω2 .

Making use of the substitutions ω1 + β−1 λ j r = ξ 1 and ω2 + β−1 λ j/2 s = ξ 2 in the

post-factor on the R.H.S of inequality (20), we obtain
# 
∞  ∞  2  
1    
R≤ ∑ ∑ F f ω1 , ω2  ∑ ∑ F Φ j λ− j ω1 , λ− j/2 ω2 
β2 0=r ∈Z 0=s∈Z −∞ −∞ j∈Z  ∈Z  j/2
λ
  $1/2
 
× F Φ j λ− j ω1 + β−1 r, λ− j/2 ω2 + β−1 s  dω1 dω2

# 
∞  ∞  2  
   
F f ξ1, ξ2  ∑ ∑ F Φ j λ− j ξ 1 , λ− j/2 ξ 2  (21)
−∞ −∞ j∈Z  ∈Z  j/2
λ
  $1/2
 
× F Φ j λ− j ξ 1 − β−1 r, λ− j/2 ξ 2 − β−1 s  dξ 1 dξ 2

1 1 1
1 12 1/2
≤ 2 1f1 ∑ ∑ H β−1 r, β−1 s H − β−1 r, − β−1 s ,
β 2 0=r ∈Z 0=s∈Z

Consequently, the infimum and supremum of the power output are given by
⎛ ⎞  ⎛ ⎞
1 1 −2 ; < 2  2
1 1   1  
inf
f ∈ L2 (R2 ), f =0
⎝1 f 1 ∑ ∑
2 j∈Z
∑  f , Ψ j,m, 2  ⎠ ≥ β2 ω1inf
,ω2 ∈S
⎝∑ ∑ F Φj λ− j ω1 , λ− j/2 ω2  ⎠ (22)
m∈Z2  ∈Zλ j/2 j∈Z  ∈Z  j/2
λ

1/2
− ∑ ∑ H β−1 r, β−1 s H − β−1 r, − β−1 s .
0=r ∈Z 0=s∈Z

and

192
Mathematics 2023, 11, 1839

⎛ ⎞  ⎛ ⎞
1 1 −2 ; < 2  2
sup ⎝1 1
1f1 ∑ ∑ ∑

 f , Ψ j,m, 
 ⎠ 1
≥ 2 sup ⎝ ∑ ∑

F Φ j −j
λ ω1 , λ − j/2  ⎠
ω2  (23)
f ∈ L2 (R2 ), f =0 2 j∈Z m∈Z2  ∈Z  j/2 2 β ω1 ,ω2 ∈R j∈Z  ∈Z
λ λ j/2

1/2
−1 −1 −1 −1
+ ∑ ∑ H β r, β s H −β r, − β s .
0=r ∈Z 0=s∈Z

By virtue of the estimates (22) and (23), it follows that

  1 < 2  C + C  1 12
C1 − C3 1 1 12
;
  3 1 1
1f1 ≤ ∑ ∑ ∑
2
 f , Ψ j,m,  ≤ 1f1 .
β 2 2 j∈Z m∈Z2  ∈Z 2 β 2 2
λ j/2

This completes the proof of Theorem 1.

Towards the end of the ongoing section, we aim to formulate a simple condition under
which the hypothesis (17) is satisfied. More explicitly, we shall demonstrate that if the
function (12) is band-limited to a certain closed ball B∞ (t0 , r ) centered at t0 ∈ R2 with
radius r > 0, then the system (13) constitutes a frame for L2 (R2 ) provided the sampling
constant β > 0 is chosen to be small enough.
 
Corollary 1. Let Φ ∈ L2 (R2 ) be as given in (12) and 0 < β < 1/2r. If supp F Φ (w) ⊂
B∞ (0, r ), the closed ball centered about 0 = (0, 0) T ∈ R2 having radius r, and
 2
 
C1 ≤ ∑ ∑ F Φ j λ− j ω1 , λ− j/2 ω2  ≤ C2 , (24)
j∈Z  ∈Z  j/2
λ

almost everywhere ω1 , ω2 ∈ R, with 0 < C1 ≤ C2 < ∞, then the system (13) constitutes a frame
for L2 (R2 ) with the lower and upper frame bounds as β−2 C1 and β−2 C2 , respectively. In particular,
if C1 = C2 = C, then the system (13) turns to be a tight frame with the frame bound as β−2 C.

Proof. According to the hypothesis, the window function Ψ is so chosen thatthe  corre-
sponding modulated version Φ given
  by (12) is band-limited in the sense that F Φ (w) ⊂
B∞ (0, r ). Therefore, we have F Φ Rθ A− j w = 0 if and only if Rθ A− j w ∈ B∞ (0, r ).
j j

Consequently, for ξ = (ξ 1 , ξ 2 ) T ∈ R2 we obtain

 
 
F Φ Rθ A− j w + ξ  = 0 ⇐⇒ Rθ A− j w ∈ B∞ (−ξ, r ). (25)
  j   j

Clearly, if ξ ∈ R2 is such that B∞ (10, r1) ∩ B∞ (−ξ, r ) = ϕ, then in view of (15) we have
H (ξ ) = 0. Indeed, this is the case if 1ξ 1∞ > 2r. Hence, we conclude that

1/2
∑ ∑ H β−1 r, β−1 s H − β−1 r, − β−1 s = 0, ∀ β < 1/2r. (26)
0=r ∈Z 0=s∈Z

This evidently completes the proof of Corollary 1.

Remark 1. Since modulation in the spatial domain corresponds to a simple shift in the frequency
domain; therefore, in view of (12) it suffices to verify the conditions
2 (16) and3(17) for the function
Ψ j (t) instead of the modulated version Φ j (t) = Ψ j (t) exp 2πi t T (1, 1) T . Moreover, it is also
quite conspicuous that the argument of Corollary 1 holds in case the function Ψ isband-limited to the
closed ball centered about 1 = (1, 1) T ∈ R2 and having radius r; that is, F Ψ (w) ⊂ B∞ (1, r ).

193
Mathematics 2023, 11, 1839

4. Conclusions and Future Work

In this communication, we introduced a scale-dependent discretization scheme for the
non-isotropic Stockwell transform. Under the refined discretization procedure, one can effi-
ciently control both the radial and angular expansions simultaneously. As an endorsement
to the undertaken problem, we also demonstrated that the novel discretization scheme
allows for the construction of Stockwell frames in L2 (R2 ). Nevertheless, as a future re-
search aspect, it is lucrative to numerically compute the frame bounds for several classes of
two-dimensional functions, particularly the Gabor functions, so that general results can be
made regarding tightness of the frame with an increase in the number of frequency, spatial,
and orientation sampling steps. Based on the numerical outcomes, certain experimental
results concerning the image representation and reconstruction processes can be executed.
Moreover, in view of the fact that the two-dimensional Gabor functions play an important
role in many computer vision applications and modelling biological vision, the study can
further be extended in that direction.

Author Contributions: Conceptualization, F.A.S. and A.Y.T.; methodology, A.Y.T.; software, A.Y.T.;
validation, H.M.S.; formal analysis, A.Y.T.; investigation, F.A.S.; and writing, F.A.S. and A.Y.T. All
authors have read and agreed to the published version of the manuscript.
Funding: This research received no external funding.
Data Availability Statement: Not applicable.
Conflicts of Interest: The authors declare no conflict of interest.

References
1. Stockwell, R.G.; Mansinha, L.; Lowe, R.P. Localization of the complex spectrum: The S-transform. IEEE Trans. Signal. Process.
1996, 44, 998–1001. [CrossRef]
2. Gabor, D. Theory of communications. J. Inst. Electr. Eng. 1946, 93, 429–457. [CrossRef]
3. Debnath, L.; Shah, F.A. Wavelet Transforms and Their Applications; Birkhäuser: New York, NY, USA, 2015.
4. Debnath, L.; Shah, F.A. Lecture Notes on Wavelet Transforms; Birkhäuser: Boston, MA, USA, 2017.
5. Shah, F.A.; Tantary, A.Y. Wavelet Transforms: Kith and Kin; Chapman and Hall/CRC: Boca Raton, FL, USA, 2022.
6. Stockwell, R.G. A basis for efficient representation of the S-transform. Digit. Signal Process. 2007, 17, 371–393. [CrossRef]
7. Du, J.; Wong, M.W.; Zhu, H. Continuous and discrete inversion formulas for the Stockwell transform. Integral Transform. Spec.
Funct. 2007, 50, 537–543. [CrossRef]
8. Drabycz, S.; Stockwell, R.G.; Mitchell, J.R. Image texture characterization using the discrete orthonormal S-transform. J. Digit.
Imaging 2009, 22, 696–708. [CrossRef] [PubMed]
9. Moukadem, A.; Bouguila, Z.; Abdeslam, D.O.; Dieterlen, A. A new optimized Stockwell transform applied on synthetic and real
non-stationary signals. Digit. Signal Process. 2015, 46, 226–238. [CrossRef]
10. Shah, F.A.; Tantary, A.Y. Linear canonical Stockwell transform. J. Math. Anal. Appl. 2020, 484, 123673. [CrossRef]
11. Shah, F.A.; Tantary, A.Y. Non-isotropic angular Stockwell transform and the associated uncertainty principles. Appl. Anal. 2021,
100, 835–859. [CrossRef]
12. Soleimani, M.; Vahidi, A.; Vaseghi, B. Two-dimensional Stockwell transform and deep convolutional neural network for multi-
class diagnosis of pathological brain. IEEE Tran. Neural Sys. Rehab. Engn. 2021, 29, 163–172. [CrossRef] [PubMed]

Disclaimer/Publisher’s Note: The statements, opinions and data contained in all publications are solely those of the individual
author(s) and contributor(s) and not of MDPI and/or the editor(s). MDPI and/or the editor(s) disclaim responsibility for any injury to
people or property resulting from any ideas, methods, instructions or products referred to in the content.

194
 mathematics

Article
Redheffer-Type Bounds of Special Functions
Reem Alzahrani † and Saiful R. Mondal *,†

Department of Mathematics and Statistics, College of Science, King Faisal University,

Al Ahsa 31982, Saudi Arabia;
* Correspondence: [email protected]
† These authors contributed equally to this work.

Abstract: In this paper, we aim to construct inequalities of the Redheffer type for certain functions
defined by the infinite product involving the zeroes of these functions. The key tools used in our
proofs are classical results on the monotonicity of the ratio of differentiable functions. The results
are proved using the nth positive zero, denoted by bn (ν). Special cases lead to several examples
involving special functions, namely, Bessel, Struve, and Hurwitz functions, as well as several other
trigonometric functions.

Keywords: Redheffer inequality; Bessel functions; Struve functions; Dini functions; Lommel func-
tions; q-Bessel functions

MSC: 33B10; 33C10; 26D07; 26D05

1. Introduction
Several famous inequalities for real functions have been proposed in the literature.
One of them is the Redheffer inequality, which states that

sin( x ) π 2 − x2
≥ 2 , for all x ∈ R. (1)
x π + x2
Inequality (1) was proposed by Redheffer [1] and proved by Williams [2]. This work
Citation: Alzahrani, R.; Mondal, S.R. motivated many researchers, regarding its generalization, refinement, and applications.
Redheffer-Type Bounds of Special A new (but relatively difficult) proof of (1) using the Lagrange mean value theorem in
Functions. Mathematics 2023, 11, 379. combination with induction was given in [3]. In 2015, Sándor and Bhayo [4] offered two
https://doi.org/10.3390/ new interesting proofs and established two converse inequalities. They also pointed out a
math11020379
hyperbolic analog. Other notable works related to the Redheffer inequality include [5–10].
Academic Editor: Yamilet Quintana Motivated by the inequality (1), C.P. Chen, J.W. Zhao, and F. Qi [8], using mathematical
induction and infinite product representations of cos( x ), sinh( x ), cosh( x )
Received: 30 November 2022
# $ # $
Revised: 4 January 2023 4x2 4x2
Accepted: 8 January 2023 cos( x ) = ∏ 1−
(2n − 1)2 π 2
, cosh( x ) = ∏ 1+
(2n − 1)2 π 2
, (2)
Published: 11 January 2023 n ≥1 n ≥1

and
 
sinh( x ) x2
Copyright: © 2023 by the authors.
x
= ∏ 1+
n2 π 2
, (3)
Licensee MDPI, Basel, Switzerland.
n ≥1

This article is an open access article

respectively, established the following Redheffer-type inequalities:
distributed under the terms and
conditions of the Creative Commons
π 2 − 4x2 π 2 + 4x2 π
Attribution (CC BY) license (https:// cos( x ) ≥ and cosh( x ) ≤ , for all |x| ≤ . (4)
π 2 + 4x2 π 2 − 4x2 2
creativecommons.org/licenses/by/
4.0/).

Mathematics 2023, 11, 379. https://doi.org/10.3390/math11020379 195 https://www.mdpi.com/journal/mathematics

Mathematics 2023, 11, 379

A hyperbolic analog of inequality (1) has also been established [8], by proving that

sinh( x ) π 2 + x2
≤ 2 , for all | x | < π. (5)
x π − x2
In [6], inequalities (1) and (4) were extended and sharpened, and a Redheffer-type
inequality for tan( x ) was also established, as follows:
(i) Let 0 < x < π. Then,
 β  α
π 2 − x2 sin( x ) π 2 − x2
≤ ≤ (6)
π 2 + x2 x π 2 + x2

hold if and only if α ≤ π 2 /12 and β ≥ 1.

(ii) Let 0 ≤ x ≤ π/2. Then,
 β  α
π 2 − 4x2 π 2 − 4x2
≤ cos( x ) ≤ (7)
π 2 + 4x2 π 2 + 4x2

hold if and only if α ≤ π 2 /16 and β ≥ 1.

(iii) Let 0 < x < π/2. Then,
 α  β
π 2 + 4x2 tan( x ) π 2 + 4x2
≤ ≤ (8)
π 2 − 4x2 x π 2 − 4x2

hold if and only if α ≤ π 2 /24 and β ≥ 1.

(iv) Let 0 < x < r. Then,
 α  β
r2 + x2 sinh( x ) r2 + x2
≤ ≤ (9)
r2 − x2 x r2 − x2

hold if and only if α ≤ 0 and β ≥ r2 /12.

(v) Let 0 < x < r. Then,
 α  β
r2 + x2 r2 + x2
≤ cosh( x ) ≤ (10)
r2 − x2 r2 − x2

hold if and only if α ≤ 0 and β ≥ r2 /4.

(vi) Let 0 < x < r. Then,
 β  α
r2 − x2 tanh( x ) r2 − x2
≤ ≤ (11)
r2 + x2 x r2 + x2

hold if and only if α ≤ 0 and β ≥ r2 /6.

The Bessel function Jν of order ν is the solution of the differential equation:

x2 y ( x ) + xy ( x ) + ( x2 − ν2 )y( x ) = 0. (12)

196
Mathematics 2023, 11, 379

The function Iν ( x ) = −iJν (ix ) is known as the modified Bessel function. It is well
known that trigonometric functions are connected with Bessel and modified Bessel func-
tions, as follows
( (
πx πx
sin( x ) = J ( x ), cos( x ) = J ( x ),
2 1/2 2 −1/2
( (
πx πx
sinh( x ) = I ( x ), cosh( x ) = I ( x ).
2 1/2 2 −1/2

Based on the relationship between trigonometric and Bessel functions as stated above, and
as Bessel and modified Bessel functions have infinite product representations involving
their zeros, the Redheffer inequality (1) has been generalized for modified Bessel functions
in [7], and sharpened in [9]. There are several other special functions, such as Struve and
q-Bessel functions, which have infinite product representations and are also related to
trigonometric functions.
Motivated by the above facts, the aim of this study was to address the following
problem:

Problem 1. Construct the class of functions f that can be represented by an infinite product with
the factors involving the zeroes of f , such that f exhibits a Redheffer-type inequality.

To answer Problem 1, we consider a sequence {bn (ν)}ν∈R,n≥1 , such that

∞
1
∑ b 2 (ν)
 → l( ν )
n =1 n

for ν ∈ I ⊂ R and the infinite product

∞  
x2
∏ 1−
bn ( ν )
2
n =1

is also absolutely convergent to a function of x for x ∈ Ix ⊂ R.

We study several properties of functions that are members of the following two classes:
 
∞ 
x2
F ν : = ην ( x ) = ∏ 1 − 2 , (13)
n =1 bn ( ν )

  
∞
x2
Gν : = χν ( x ) = ∏ 1+
bn ( ν )
2
. (14)
n =1

It is easy to check that, for a fixed ν, {b1 (ν), b2 (ν), . . . , bn (ν), . . .} is a set of zeroes of the
functions in the class Fν . Unless mentioned otherwise, throughout the article, we denote
by bn (ν) the nth positive zero of the functions in the √ class Fν . For λν ∈ Gν and ην ∈ Fν , it
immediately follows that λν ( x ) = ην (ix ), where i = −1.
Using a similar concept as in [7,9], we derived the Redheffer inequality for the func-
tions from both classes, Fν and Gν . We also investigate the increasing/decreasing, log
convexity, and convexity nature of the functions (or their products) from the above two
classes. The main results are discussed in Section 2, while Section 3 provides several
examples based on the main result in Section 2. In Section 4, we compare the obtained
result with known results; especially the results given in [7,9–11].
The following lemma is required in the following.

197
Mathematics 2023, 11, 379

Lemma 1 ([12]). Suppose f ( x ) = ∑∞ ∞

k =0 ak x and g ( x ) = ∑k =0 bk x , where ak ∈ R and bk > 0
k k

for all k. Furthermore, suppose that both series converge on | x | < r. If the sequence { ak /bk }k≥0 is
increasing (or decreasing), then the function x → f ( x )/g( x ) is also increasing (or decreasing) on
(0, r ).

Lemma 2 (Lemma 2.2 in [13]). Suppose that −∞ < a < b < ∞ and p, q : [ a, b) → ∞ are
differentiable functions, such that q ( x ) = 0 for x ∈ ( a, b). If p /q is increasing (or decreasing) on
( a, b), then so is ( p( x ) − p( a))/(q( x ) − q( a)).

2. Main Results
Theorem 1. Suppose that λν ∈ Gν and ην ∈ Fν . Then, the following assertions are true:
1. The function x → λν ( x ) is increasing on (0, ∞).
2. The function x → λν ( x ) is strictly log-convex on Iν = (−b1 (ν), b1 (ν)) and strictly geomet-
ric convex on (0, ∞).
3. The function x → λν ( x ) satisfies the sharp exponential Redheffer-type inequality
 aν  bν
b12 (ν) + x2 b12 (ν) + x2
≤ λν ( x ) ≤ (15)
b12 (ν) − x2 b12 (ν) − x2

on Iν . Here, aν = 0 and bν = b12 (ν)l(ν)/2 are the best possible constants. 

4. The function x → λν ( x )ην ( x ) is increasing on (−b1 (ν), 0] and decreasing on 0, b1 (ν)
5. The function x → λν ( x )/ην ( x ) is strictly log-convex on Iν .
6. The function x → ην ( x ) satisfies the sharp Redheffer-type inequality.

 aν  bν
b12 (ν) − x2 b12 (ν) − x2
≤ ην ( x ) ≤ (16)
b12 (ν) b12 (ν)

on Iν . Here, bν = 1 and aν = b12 (ν)l(ν) are the best possible constants.

Proof. As λν ∈ Gν , from (14), it follows that

∞  
x2
λν ( x ) = ∏ 1+
bn2 (ν)
. (17)
n =1

Similarly, as ην ∈ Gν , from (13), it follows that

∞  
x2
ην ( x ) = ∏ 1−
bn2 (ν)
. (18)
n =1

1. Logarithmic differentiation of (17) leads to

∞
λν ( x ) 2x
(log(λν ( x ))) = = ∑ >0 (19)
λν ( x ) b
n =1 n
2 (ν) + x2

for x ∈ (0, ∞). This implies that log(λν ( x )) is increasing and, consequently, λν ( x ) is
also increasing.

198
Mathematics 2023, 11, 379

2. Let x ∈ Iν . Differentiation of both sides of (19) gives


∞
2 4x2
(log(λν ( x ))) = ∑ −
n =1 bn2 (ν) + x2 (bn2 (ν) + x2 )
2

∞ 2 bn2 (ν) − x2
= ∑ 2
> 0,
n =1 (bn2 (ν) + x2 )

for x ∈ Iν This is equivalent to the function x −→ λν ( x ) being log-convex on Iν .

From (19), we also have
  ∞  
xλν ( x ) bn2 (ν)
λν ( x )
= ∑ 2−
bn ( ν ) + x 2
2
n =1
∞
2xbn2 (ν)
= ∑ ( b 2 ( ν ) + x 2 )2
.
n =1 n

This implies that x −→ xλν ( x )/λν ( x ) is increasing on x ∈ (0, ∞) and, as a conse-
quence, we have that x −→ λν ( x ) is geometrically convex on (0, ∞).

3. Consider the function

log(λν ( x ))
hν ( x ) := .
log(b12 (ν) + x2 ) − log(b12 (ν) − x2 )

For x ∈ [0, ∞), define

p( x ) = log(λν ( x )), q( x ) = log(b12 (ν) + x2 ) − log(b12 (ν) − x2 ).

From the calculation along with (19), it follows that

λν ( x )
∞
p ( x ) λν ( x ) λν ( x ) b14 (ν) − x4 1 b14 (ν) − x4
q ( x )
= = .
2xλν ( x ) 2b12 (ν)
= 2 ∑ .
+ 2b1 (ν) n = 1 bn ( ν ) + x
2x 2x 2 2
b12 (ν)+ x2 b12 (ν)− x2

Then,
  ∞
d p ( x ) 1 −4x3 (bn2 (ν) + x2 ) − 2x (b14 (ν) − x4 )
dx q ( x )
=
2b12 (ν)
∑ (bn2 (ν) + x2 )2
n =1
∞ 2x2 bn2 (ν) + x4 + b12 (ν)
x
=−
b12 (ν) n=1
∑ (bn2 (ν) + x2 )2
≤0

on x ∈ [0, ∞). Thus, p ( x )/q ( x ) is decreasing and, hence,

p( x ) p ( x ) − p (0)
hν ( x ) = =
q( x ) q ( x ) − q (0)

is also decreasing on [0, b1 (ν)]. Finally,

lim hν ( x ) < hν ( x ) < lim hν ( x ),

x →b1 (ν) x →0

199
Mathematics 2023, 11, 379

where
p( x ) p ( x )
aν := lim hν ( x ) = lim = lim = 0,
x →b1 (ν) x →b1 (ν) q( x ) x →b1 (ν) q ( x )
p( x ) p ( x ) b2 ( ν )
bν := lim hν ( x ) = lim = lim  = 1 l (ν)
x →0 x →0 q( x ) x →0 q ( x ) 2

are the best possible constants and

∞
1
l (ν) = ∑ bn2 (ν)
.
n =1

4. As λν ∈ Gν and ην ∈ Fν , from (13) and (14), it follows that

∞  
x4
λ ν ( x ) ην ( x ) = ∏ 1−
bn4 (ν)
.
n =1

Logarithmic differentiation yields

(λν ( x )ην ( x )) ∞

4x3
=−∑ 4 ,
λ ν ( x ) ην ( x ) n =1 n ) − x
b ( ν 4

which is negative for x ∈ (0, b1 (ν)) and positive for x ∈ (−b1 (ν), 0). Hence, the result
follows.
5. From part (2), it follows that x −→ λν ( x ) is strictly log-convex on Iν . Now, consider
the function x −→ (ην ( x ))−1 . From (2), it follows that
   ∞
2x
log (ην ( x ))−1 = ∑ 2 (ν) − x2
n =1 b n

and
   ∞
bn2 (ν) + x2
log (ην ( x ))−1 =2 ∑ > 0.
n = 1 ( bn ( ν ) − x )
2 2 2

This implies that x −→ (ην ( x ))−1 is strictly log-convex on Iν . Finally, being the
product of two strictly log-convex functions, x −→ λν ( x )/ην ( x ) is strictly log-convex
on Iν .
6. To prove this result, we first need to set up a Rayleigh-type function for the Lommel
function. Define the function
∞
(2m)
αn (ν) := ∑ bn−2m (ν), m = 1, 2, . . . . (20)
n =1

Logarithmic differentiation of χν ( x ) yields

∞ ∞   −1 ∞ ∞
xχν ( x ) x2 x2 x2 x2 x2m
= −2 ∑ 2 = ∑ 2 1− 2 = ∑ ∑ .
χν ( x ) n = 1 bn ( ν ) − x
2
n = 1 bn ( ν ) bn ( ν ) n = 1 bn ( ν ) m = 0
2 bn2m (ν)

Interchanging the order of the summation, it follows that

∞ ∞ ∞
xχν ( x ) x2m+2 (2m)
= −2 ∑ ∑ 2m+2 = −2 ∑ αn (ν) x2m . (21)
χν ( x ) m = 0 n = 1 bn (ν) m =1

200
Mathematics 2023, 11, 379

Consider the function

log(χν ( x )) pμ ( x )
ϕμ ( x ) :=   = . (22)
2 qμ ( x )
log 1 − b2x(ν)
1

The binomial series, together with (21), gives the ratio of pμ and qμ as

xχν ( x ) (2m)
pμ ( x ) χν ( x ) ∑∞
m =1 α n (ν) x2m
=   −1 = . (23)
qμ ( x ) −2x2 x2
∞ −2m
∑m=1 b1 (ν) x2m
b1 (ν)2
1− b1 (ν)2

(2m)
Denote dm = b12m (ν)αn (ν). Then,

(2m+2) (2m)
dm+1 − dm = b12m+2 (ν)αn n(ν) − b12m (ν)αn n(ν)

∞ b2m ( ν ) b2 ( ν )
= ∑ 12m 1
− 1 < 0.
n = 1 bn ( ν ) bn ( ν )
2

This is equivalent to saying that the sequence {dm } is decreasing. Hence, by Lemma 1,
it follows that the ratio pμ /qμ is decreasing. In view of Lemma 2, we have that
τμ = pμ /qμ is decreasing.
From (22) and (23), it can be shown that

pμ ( x ) pμ ( x ) pμ ( x ) (2)

lim τμ ( x ) = lim = lim = lim = b12 (ν)αn (ν), (24)
x →0 x →0 qμ ( x ) x →0 qμ ( x ) x →0 qμ ( x )

and
pμ ( x ) ∞ b12 (ν) − x2
lim τμ ( x ) = lim
qμ ( x )
= lim ∑ = 1. (25)
n =1 n ( ν ) − x
x →b1 (ν) b 2 2
x →b12 (ν) x →b12 (ν)

(2)
It is easy to see that b12 (ν)αn (ν) = b12 (ν)l(ν) = bν .
This completes the proof of all of the results.

In the next result, by approaching a similar proof as in Theorem 1, we prove a sharper

upper bound for λν , compared to that presented in Theorem 1 (Part 3).

Theorem 2. If r > 0 and | x | < r, then the following inequality

  aν   bν
r2 − x2 r2 − x2
≤ λν ( x ) ≤ (26)
r2 r2

holds, where aν = 0 and bν = −r2 l(ν) are the best possible constants.

Proof. Due to symmetry, it is sufficient to show the result for [0, r ). Define Ψ : [0, r ) −→ R as
 
r2
Ψ( x ) := log(λν ( x )) − r2 l(ν)log .
r2 − x2

201
Mathematics 2023, 11, 379

Then,
∞ ∞
λν ( x ) 2xr2 2x 2xr2
Ψ ( x ) = − 2 l (ν) = ∑ − ∑
λν ( x ) r − x2 n = 1 bn ( ν ) + x
2 2
n = 1 ( r − x ) bn ( ν )
2 2 2

∞
2x (r2 − x2 )bn2 (ν) − 2xr2 (bn2 (ν) + x2 )
= ∑ bn2 (ν)(r2 − x2 )(bn2 (ν) + x2 )
n =1
∞
bn2 (ν) + r2
= −2x3 ∑ ≤ 0,
n=1 bn ( ν )(r − x )( bn ( ν ) + x )
2 2 2 2 2

for x ∈ [0, r ). This implies that Ψ is decreasing, and Ψ( x ) ≤ Ψ(0) = 0. This is equivalent to
  r 2 l( ν )   −r 2 l( ν )
r2 r2 − x2
log(λν ( x )) ≤ log =⇒ λν ( x ) ≤ .
r2 − x2 r2

This completes the proof. Now, to show the bν = −r2 l(ν) is the best possible constant,
consider
log(λν ( x ))
δν :=  2 2 .
r −x
log r2

Then, using the Bernoulli–L’Hôpital rule, we have

log(λν ( x ))
lim δν ( x ) = lim  2 
x 0 x 0log r2 − r
x2
  
λν ( x ) r2 − x2
= lim −
x 0 λ ν ( x ) 2x
∞ ∞
−(r2 − x2 ) r2
= lim ∑ bn2 (ν) + x2
=−∑ 2 = − r 2 l ( ν ) = bν .
x 0 n =1 n =1 n ( ν )
b

Thus, bν is the best possible constant.

3. Application Examples
As stated before, the primary aim of this work is to find a Redheffer-type inequality
for functions that are combinations of well-known functions. By constructing examples,
we show that Theorem 1 not only covers known results but also covers a wide range of
functions. We list each case as an example.

3.1. Example Involving Trigonometric Functions

Our very first example involves the well-known function f ( x ) = sinc( x ). In math-
ematics, physics, and engineering, there are two forms of the sinc( x ) function; namely,
non-normalized and normalized sinc functions. In mathematics, the non-normalized sinc
function is defined, for x = 0, as:

sin( x )
sinc( x ) := .
x
On the other hand, in digital and communication systems, the normalized form is defined as:

sin(πx )
sinc( x ) := , x = 0.
πx
The scaling of the independent variable (the x-axis) by a factor of π is the only dis-
tinction between the two definitions. In both scenarios, it is assumed that the limit value 1

202
Mathematics 2023, 11, 379

corresponds to the function’s value at the removable singularity at zero. The sinc function
is an entire function, as it is analytic everywhere.
The normalized sinc has the following infinite product representation:
∞  
sin(πx ) x2
πx
= ∏ 1−
n2
. (27)
n =1

It is well known that the infinite series ∑∞

n =1 n
−2 is convergent and

∞
1 π2
∑ n 2
=
6
.
n =1

We can conclude that sinc( x ) ∈ Fν . From Theorem 1, it follows that

(1 − x2 ) aν ≤ sinc( x ) ≤ (1 − x2 )bν

with | x | < 1, bν = 1, and aν = π 2 /6.

Now, replacing x with ix in (27), we have
∞  
sinh(πx ) x2
πx
= ∏ 1+
n 2
. (28)
n =1

Clearly, sinh(πx )/πx ∈ Fν . Hence, by Theorem 1 (part 3), it follows that

 τν  δν
1 + x2 sinh(πx ) 1 + x2
≤ ≤
1 − x2 πx 1 − x2

for | x | < 1. Here, τν = 0 and δν = π 2 /6 are the best possible values of the constants.
On the other hand, from Theorem 2, it follows that
 δν
sinh(πx ) r
≤
πx r2 − x2

for | x | < r, where δν = π 2 /6 is the best possible constant.

Next, we consider the infinite product
∞  
x2
∏ 1−
n2 π 2 − ν2
, |ν| < π. (29)
n =1

Using the Mathematica software, we find that

√ 
∞   ν csc(ν) sin ν2 + x 2
x2
∏ 1− 2 2
n π − ν2
= √
ν2 + x 2
(30)
n =1

and
∞
1 1 − ν cot(ν)
∑ n 2 π 2 − ν2
=
2ν2
. (31)
n =1
√  √
Clearly, ν csc(ν) sin ν2 + x2 / ν2 + x2 ∈ Fν , and we have the following result, accord-
ing to Theorem 1.

203
Mathematics 2023, 11, 379

Corollary 1. Let 0 = ν ∈ (−π, π ). Then, the following inequality

√ 
 2 a ν csc(ν) sin ν2 + x 2  2 b
π − ν2 + x 2 ν π − ν2 + x 2 ν
≤ √ ≤
π 2 − ν2 − x 2 ν2 + x 2 π 2 − ν2 − x 2

holds for | x | < π 2 − ν2 . Here, aν = 0 and bν = (1 − ν cot(ν))/4ν2 (π 2 − ν2 ) are the best possible
constants.

3.2. Examples Involving Hurwitz Zeta Functions

The Hurwitz zeta functions are zeta functions defined for the complex variable s, with
Re(s) > 0 and ν = −1, −2, −3, . . ., defined by
∞
1
ζ (s, ν) := ∑ ( n + ν)s
. (32)
n =0

This series is absolutely convergent for given values of s and ν, and can be extended to
meromorphic functions defined for all s = 1. In particular, the Riemann zeta function is
given by ζ (s, 1). For our study in this section, we consider s = m ∈ N \ {1} and ν > −1.
Now, consider the infinite product
∞  
x2
χm,ν ( x ) := ∏ 1−
(n + ν) m
, m ≥ 2 and ν > −1, (33)
n =1

for which the product is convergent. In the closed form of the product, we consider
m = 2, 3, 4. Then, χm,ν ( x ) have the forms

Γ ( ν +1)2
χ2,ν ( x ) = Γ(− x +ν+1)Γ( x +ν+1)
Γ ( ν +1)3
χ3,ν ( x ) = √ √
Γ(− x2/3 +ν+1)Γ( 12 ((1−i 3) x2/3 +2(ν+1)))Γ( 12 ((1+i 3) x2/3 +2(ν+1)))
Γ ( ν +1)4
χ4,ν ( x ) = √ √ √ √ .
Γ ( ν − − x +1) Γ ( ν + − x +1) Γ ( ν − x +1) Γ ( ν + x +1)

Next, we state a result related to the inequalities involving χm,ν ( x ). Although the
result is a direct consequence of Theorem 1 (Part 6), taking bn (ν) = (n + ν)m/2 for m ≥ 2
and ν > −1, we state it as a theorem due to its independent interest. Clearly,
∞ ∞
1 1 1
∑ b 2 (ν)
= ∑ ( n + ν ) m
= ζ (m, ν) − m .
ν
n =1 n n =1

Theorem 3. If m ≥ 2, ν > −1 and | x | < (n + ν)m , then the following sharp exponential
inequality holds:
  am,ν  bm,ν
(1 + ν ) m − x 2 (1 + ν ) m − x 2
≤ χm,ν ( x ) ≤ , (34)
(1 + ν ) m (1 + ν ) m

with the best possible constants as bm,ν = 1 and am,ν = (1 + ν)m (ζ (m, ν) − ν−m ).

Taking ν = 1 in (34), it follows that

  am,1  bm,1
x2 x2
1− ≤ χm,1 ( x ) ≤ 1− . (35)
2m 2m

204
Mathematics 2023, 11, 379

Now, by choosing m = 2, 3, 4, 5, 6 in (35), we have the following special cases of

Theorem 3:
  a2,1  b2,1
x2 x2 2π 2
(i ) 1− ≤ χ2,1 ( x ) ≤ 1− with a2,1 = ,
4 4 3
  a3,1  b3,1
x2 x2
(ii ) 1− ≤ χ3,1 ( x ) ≤ 1− with a3,1 = 8ζ (3, 1) = 9.61646,
8 8
 2  a4,1  2 b4,1
x x 8π 4
(iii ) 1− ≤ χ4,1 ( x ) ≤ 1− with a4,1 = ,
16 16 45
  a5,1  b5,1
x2 x2
(iv) 1− ≤ χ5,1 ( x ) ≤ 1− with a5,1 = 32 ζ (5, 1) = 33.1817,
16 16
 2  a6,1  2 b6,1
x x 64π 6
(v) 1− ≤ χ6,1 ( x ) ≤ 1− with a6,1 =
16 16 945

where, in each of the cases (m = 2, 3, 4, 5, 6), the best values of bm,1 = 1 and χm,1 ( x ) are
listed below
sin(πx )
χ2,1 ( x ) = ,
πx − πx3
1
χ3,1 ( x ) = −   √     √  ,
− 1) Γ 1 −
( x2 Γ x2/3 1
2 1 − i 3 x2/3 + 1 Γ 12 1 + i 3 x2/3 + 1
√ √
sin π x sinh π x
χ4,1 ( x ) = −
π 2 ( x3 − x )
χ5,1 ( x ) = √ 1
,
(1− x2 )Γ(1− x2/5 )Γ( 5 −1x2/5 +1)Γ(1−(−1)2/5 x2/5 )Γ((−1)3/5 x2/5 +1)Γ(1−(−1)4/5 x2/5 )
√  √ √ √ 
sin π 3 x cos π 3 x − cosh 3π 3 x
χ6,1 ( x ) = .
2π 3 x ( x2 − 1)

3.3. Examples Involving Bessel Functions

In this part, we discuss the generalization of the Redheffer type bound in terms of
Bessel and modified Bessel functions. In this regard, we consider the very first result given
by Baricz [7], and later by Khalid [9], as well as Baricz and Wu [10].
From ([14], p. 498), it is known that the Bessel function Jν has the infinite product
 
x2
Jν ( x ) = 2ν Γ(ν + 1) x −ν Jν ( x ) = ∏ 1− 2
(36)
n ≥1 jν,n

for arbitrary x and ν = −1, −2, −3, . . .. It is also well known that ([14], P. 502)
∞
1 1
∑ 2
jn,ν
=
4( ν + 1)
.
n =1

This implies Jν ∈ Fν . Similarly, Iν ( x )—the normalized form of the modified Bessel

function Iν —can be expressed as
 
x2
Iν ( x ) = 2ν Γ(ν + 1) x −ν Iν ( x ) = ∏ 1+ 2
, (37)
n ≥1 jν,n

which indicates that Iν ∈ Gν . Now, from Theorem 1 (3) and Theorem 2, we have the
following results.

Theorem 4. Consider ν > −1 and Iν ∈ Gν .

205
Mathematics 2023, 11, 379

1. For | x | < jν,1 , we have

 aν  bν
2 + x2
jν,1 2 + x2
jν,1
2 − x2
≤ Iν ( x ) ≤ 2 − x2
, (38)
jν,1 jν,1

with the best possible constants as aν = 0 and bν = jν,1

2 /8( ν + 1).

2. For any r > 0 and | x | < r, we have

  aν   bν
r2 − x2 r2 − x2
≤ Iν ( x ) ≤ , (39)
r2 r2

with the best possible constants as aν = 0 and bν = −r2 /4(ν + 1).

Now, from Theorem 1 (6), the following inequality holds for normalized Bessel func-
tions.

Theorem 5. Consider ν > −1 and Jν ∈ Fν . For | x | < jν,1 , we have

 aν  bν
2 − x2
jν,1 2 − x2
jν,1
2
≤ Jν ( x ) ≤ 2
, (40)
jν,1 jν,1

with the best possible constants as bν = 1 and aν = jν,1

2 /4( ν + 1).

3.4. Examples Involving Struve Functions

One of the most well-known special functions is the solution to the non-homogeneous
Bessel differential equation

z2 y (z) + zy (z) + (z2 − ν2 )y(z) = zμ+1 ,

called the Struve functions, Sν . If hν,n denotes the nth positive zero of Sν , then, for |ν| ≤ 1/2,
the function Sν can be expressed as (see [15])
∞  
z ν +1 z2
Sν ( z ) =
2 ν
√
πΓ ν + 3 ∏ 1−
hν,n
2
. (41)
2 n =1

From [16] (Theorem 1), it is useful to note that hν,n > hν,1 > 1 for |ν| < 1/2. From (41),
consider the normalized form
  ∞  
√ 3 −ν z2
Sν (z) := π2ν Γ ν + z Sν ( z ) = ∏ 1 − 2 . (42)
2 n =1 hν,n

From [17], it follows that for |ν| ≤ 1/2,

1 1
∑ h2v,n
=
3(2v + 3)
.
n ≥1

Consider the modified form of the Struve function

∞  
z2
Lν (z) = Sν (iz) = ∏ 1 + 2 .
n =1 hν,n

Clearly, Sν ∈ Fν and Lν ∈ Gν .
Now, from Theorem 1 (3) and Theorem 2, we have the following results.

Theorem 6. Consider |ν| < 1/2 and Lν ∈ Gν .

206
Mathematics 2023, 11, 379

1. For | x | < hν,1 , we have

 aν  bν
h2ν,1 + x2 h2ν,1 + x2
≤ Lν ( x ) ≤ , (43)
h2ν,1 − x2 h2ν,1 − x2

with the best possible constants as aν = 0 and bν = h2ν,1 /6(2ν + 3).

2. For any r > 0 and | x | < r, we have
  aν   bν
r2 − x2 r2 − x2
≤ Lν ( x ) ≤ , (44)
r2 r2

with the best possible constants as aν = 0 and bν = −r2 /3(2ν + 3).

Now, from Theorem 1 (6), the following inequality holds for normalized Bessel func-
tions.

Theorem 7. Consider ν > −1 and Sν ∈ Fν . For | x | < hν,1 , we have

 aν  bν
h2ν,1 − x2 h2ν,1 − x2
≤ Sν ( x ) ≤ , (45)
h2ν,1 h2ν,1

with the best possible constants as bν = 1 and aν = h2ν,1 /3(2ν + 3).

3.5. Examples Involving Dini Functions

The Dini function dν : Ω  C −→ C is defined by

dν (z) = (1 − v)Jν (z) + zJν (z) = Jν (z) − zJν+1 (z).

The modified Bessel functions are related to the Bessel functions by Iν (z) = i−ν Jν (iz), which
gives the modified Dini function

ξ ν = Ω  C −→ C,

defined by

ξ ν (z) = i−ν dν (iz) = (1 − ν) Iν (z) + zIν (z) = Iν (z) − zIν+1 (z).

For an integer ν, the domain Ω can be taken as the whole complex plane, while Ω is the
whole complex plane minus an infinite slit from the origin if ν is not an integer.
In view of the Weierstrassian factorization of dν (z)
 
zν z2
dν (z) = ν
2 Γ ( ν + 1) ∏ 1−
α2ν,n
, (46)
n ≥1

where ν > −1 and the formula ξ (z) = i−1 dν (iz), we have the following Weierstrassian
factorization of ξ ν (z) for all ν > −1 and z ∈ Ω:
 
zν z2
ξ ν (z) =
2ν Γ ( ν + 1 ) ∏ 1+
α2ν,n
, (47)
n ≥1

where the infinite product is uniformly convergent on each compact subset of the complex
plane, where αν,n is the nth positive zero of the Dini function dν . The principal branches of
dν (z) and ξ ν (z) correspond to the principal value of (z/2)ν , and are analytic in the z-plane

207
Mathematics 2023, 11, 379

cut along the negative real axis from 0 to infinity; that is, the half line (∞, 0]. Now for
ν > −1, define the function Λν : R −→ [1, ∞) as
 
x2
Λ ν ( x ) = 2ν Γ ( ν + 1 ) x − ν ξ ν ( x ) = ∏ 1+ . (48)
n ≥1 α2ν,n

Furthermore, for ν > −1, let us define the function Dν : R −→ R

 
x2
D ν ( x ) = 2ν Γ ( ν + 1 ) x − ν d ν ( x ) = ∏ 1 − 2 . (49)
n ≥1 αν,n

From [18], it follows that

∞
1 3
∑ α2ν,n
=
4( ν + 1)
.
n =1

Comprehensive details of the properties of Dini functions can be found in [11,18] and
the references therein.
From the definition of the classes Fν and Gν , it is clear that Λν ∈ Gν and Dν ∈ Gν .
Thus, we have the following results, by Theorems 1 and 2.

Theorem 8. Consider ν > −1 and Λν ∈ Gν .

1. For | x | < αν,1 , we have
 aν  bν
α2ν,1 + x2 α2ν,1 + x2
≤ Λν ( x ) ≤ , (50)
α2ν,1 − x2 α2ν,1 − x2

with the best possible constants as aν = 0 and bν = 3α2ν,1 /8(ν + 1).

2. For any r > 0 and | x | < r, we have
  aν   bν
r2 − x2 r2 − x2
≤ Λν ( x ) ≤ , (51)
r2 r2

with the best possible constants as aν = 0 and bν = −3r2 /4(ν + 1).

Further, Theorem 1 (6) gives the following result.

Theorem 9. For ν > −1 and | x | < αν,1 , we have

 aν  bν
α2ν,1 − x2 α2ν,1 − x2
≤ Dν ( x ) ≤ , (52)
α2ν,1 α2ν,1

with the best possible constants as bν = 1 and aν = 3α2ν,1 /4(ν + 1).

3.6. Examples Involving q-Bessel Functions

This section considers the Jackson and Hahn–Exton q-Bessel functions, respectively
(2) (3)
denoted by Jν (z; q) and Jν (z; q). For z ∈ C, ν > −1 and q ∈ (0, 1), both functions are
defined by the series
2n+ν
(2) q ν +1 ; q ∞ (−1)n 2z
Jν (z; q) := ∑ ν +1 ; q )
qn(n+ν) (53)
(q; q)∞ n≥0 ( q; q )n ( q n

(3) q ν +1 ; q ∞ (−1)n z2n+ν n ( n +1)

Jν (z; q) :=
(q; q)∞ ∑ ( q; q ) ( q ν +1 ; q )
q 2 . (54)
n ≥0 n n

208
Mathematics 2023, 11, 379

Here,
n    
( a; q)0 = 1, ( a; q)n = ∏ 1 − aqk−1 , and ( a; q)∞ = ∏ 1 − aqk−1
k =1 k ≥1

are known as the q-Pochhammer symbol. For a fixed z and q → 1, both of the above
(2)
q-Bessel functions relate to the classical Bessel function Jν as Jν ((1 − z)q; q) → Jν (z)
(3)
and Jν ((1 − z)q; q) → Jν (2z). The q-extension of Bessel functions has been studied
by several authors, notably, references [19–24] and the various references therein. The
geometric properties of q-Bessel functions have been discussed in [25]. It is worth noting
that abundant results are available in the literature, regarding the q-extension of Bessel
functions; however, we limit ourselves to the requirements of this article. For this purpose,
we recall the Hadamard factorization for the normalized q-Bessel functions:

(2) (2) (3) (3)

z → Jν (z; q) = 2ν cν (q)z−ν Jν (z; q) and z → Jν (z; q) = cν (q)z−ν Jν (z; q),

where cν (q) = (q; q)∞ / qν+1 ; q ∞

.

(2) (3)
Lemma 3 ([25]). For ν > −1, the functions z → Jν (z; q) and z → Jν (z; q) are entire
functions of order zero, which have Hadamard factorization of the form
   
(2) z2 (3) z2
Jν (z; q) = ∏ 1− 2 (q)
jν,n
, Jν (z; q) = ∏ 1− 2 (q)
lν,n
, (55)
n ≥1 n ≥1

(2) (3)
where jν,n (q) and lν,n (q) are the nth positive zeros of the functions Jν (.; q) and Jν (.; q),
respectively.

We recall that, from [25], the q-extension of the first Rayleigh sum for Bessel functions
of the first kind is
∞ ∞
1 1 1 q ν +1
∑ =
4( ν + 1)
, is ∑ =
4(q − 1)(qν+1 − 1)
. (56)
n=1 jν,n ( q )
2 2
n=1 jν,n

(3)
The series form of Jν (z; q) is

n ( n +1)
∞
(3) (−1)n z2n q 2
Jν (z; q) = ∑ (q, q)n (qν+1 , q)n . (57)
n =0

Comparing the coefficients of z2 in (55) and (57), it follows that

∞
1 q
∑ =
(q − 1)(qν+1 − 1)
. (58)
n=1 lν,n ( q )
2

(i )
The above facts imply that Jν (z; q) ∈ Fν for i = {1, 2}. For i = {1, 2} and ν > −1, denote
(i )
the nth zero of Jν (z; q) by bi,n (ν). From (56) and (58), it follows that
⎧
⎪ q ν +1
∞ ⎪
⎨ 4(q−1)(qν+1 −1) i = 1,
1
li ( ν ) : = ∑ 2 =
n=1 bi,n ( ν )
⎪
⎪
⎩ q
i = 2.
(q−1)(qν+1 −1)

Now, we have the following result, by Theorem 1 (6).

209
Mathematics 2023, 11, 379

(i )
Theorem 10. The function x → Jν (z; q) ∈ Fν for i = {1, 2} satisfies the sharp Redheffer-type
inequality

 2 (ν) − x2 aν  2 (ν) − x2 bν
bi,1 (i ) bi,1
2 (ν)
≤ Jν (z; q) ≤ 2 (ν)
(59)
bi,1 bi,1

on Iν . Here, bν = 1 and aν = bi,1

2 ( ν )l ( ν ) are the best possible constants.
i

4. Conclusions
In this article, we defined two classes of functions on the real domain, using the infinite
products of factors involving the positive zeroes of the function. We assume that the infinite
product is uniformly convergent, and it is also assumed that the sum of the square of zeroes
is convergent. We illustrate several examples that ensure that these classes are non-empty.
Functions starting from the most fundamental trigonometric functions (i.e., sin, cos) to
special functions, such as Bessel and q-Bessel functions, Hurwitz functions, Dini functions,
and their hyperbolic forms, are included in the classes. In conclusion, it follows that the
results obtained in Section 2 are similar to the results available in the literature for each of the
individual functions listed above. For example, Redheffer-type inequalities for Bessel and
modified functions, as stated in Theorem 5 and Theorem 4, form part of the results given
previously in [7,9,10], while the inequality obtained in Theorem 8 has also been obtained in
([11], Theorem 7). From Theorem 1 (part 4), it follows that the function x → Λν ( x )Dν ( x ) is
increasing on (−αν,n , 0) and decreasing on (0, αν,n ), which has also been obtained in ([11],
Theorem 8 (i)). To the best of our knowledge, Theorems 3 and 10 have not been published
in the existing literature. We finally conclude that the Redheffer-type inequalities obtained
in this study cover a wide range of functions, regarding Theorems 1 and 2. Using the
Rayleigh concepts provided in [26], more investigations into the zeroes of special functions
may lead to more examples related to the work in this study, and we intend to follow this
line of research for future investigations.

Author Contributions: Conceptualization, R.A. and S.R.M.; methodology, R.A. and S.R.M.; valida-
tion, R.A. and S.R.M.; formal analysis, R.A. and S.R.M.; investigation, R.A. and S.R.M.; resources, R.A.
and S.R.M.; writing—original draft preparation, R.A. and S.R.M.; writing—review and editing, R.A.
and S.R.M.; supervision, S.R.M. All authors have read and agreed to the published version of the
manuscript.
Funding: This work was supported by the Deanship of Scientific Research, Vice Presidency for
Graduate Studies and Scientific Research, King Faisal University, Saudi Arabia (grant project no. 1734).
Institutional Review Board Statement: Not applicable.
Informed Consent Statement: Not applicable.
Data Availability Statement: Not applicable.
Conflicts of Interest: The authors declare no conflict of interest.

References
1. Redheffer, R. Problem 5642. Am. Math. Mon. 1969, 76, 422. [CrossRef]
2. Williams, J. Solution of problem 5642. Am. Math. Mon. 1969, 10, 1153–1154.
3. Li, L.; Zhang, J. A new proof on Redheffer-Williams’ inequality. Far East J. Math. Sci. 2011, 56, 213–217.
4. Sándor, J.; Bhayo, B.A. On an inequality of Redheffer. Miskolc Math. Notes 2015, 16, 475–482. [CrossRef]
5. Zhu, L. Extension of Redheffer type inequalities to modified Bessel functions. Appl. Math. Comput. 2011, 217, 8504–8506.
[CrossRef]
6. Zhu, L.; Sun, J. Six new Redheffer-type inequalities for circular and hyperbolic functions. Comput. Math. Appl. 2008, 56, 522–529.
[CrossRef]
7. Baricz, Á. Redheffer type inequality for Bessel functions. J. Inequal. Pure Appl. Math. 2007, 8, 6.

210
Mathematics 2023, 11, 379

8. Chen, C.-P.; Zhao, J.; Qi, F. Three inequalities involving hyperbolically trigonometric functions. RGMIA Res. Rep. Coll. 2003, 6,
437–443.
9. Mehrez, K. Redheffer type inequalities for modified Bessel functions. Arab J. Math. Sci. 2016, 22, 38–42. [CrossRef]
10. Baricz, Á.; Wu, S. Sharp exponential Redheffer-type inequalities for Bessel functions. Publ. Math. Debrecen 2009, 74, 257–278.
[CrossRef]
11. Baricz, Á.; Ponnusamy, S.; Singh, S. Modified Dini functions: Monotonicity patterns and functional inequalities. Acta Math.
Hungar. 2016, 149, 120–142. [CrossRef]
12. Biernacki, M.; Krzyż, J. On the monotonity of certain functionals in the theory of analytic functions. Ann. Univ. Mariae
Curie-Skłodowska. Sect. A 1955, 9, 135–147.
13. Anderson, G.D.; Vamanamurthy, M.K.; Vuorinen, M. Inequalities for quasiconformal mappings in space. Pac. J. Math. 1993, 160,
1–18. [CrossRef]
14. Watson, G.N. A Treatise on the Theory of Bessel Functions; Cambridge University Press: Cambridge, UK, 1944.
15. Baricz, Á.; Ponnusamy, S.; Singh, S. Turán type inequalities for Struve functions. J. Math. Anal. Appl. 2017, 445, 971–984. [CrossRef]
16. Baricz, Á.; Szász, R. Close-to-convexity of some special functions and their derivatives. Bull. Malays. Math. Sci. Soc. 2016, 39
427–437. [CrossRef]
17. Baricz, Á.; Kokologiannaki, C.G.; Pogány, T.K. Zeros of Bessel function derivatives. Proc. Am. Math. Soc. 2018, 146, 209–222.
[CrossRef]
18. Baricz, Á.; Pogány, T.K.; Szász, R. Monotonicity properties of some Dini functions. In Proceedings of the 9th IEEE International
Symposium on Applied Computational Intelligence and Informatics, Timisoara, Romania, 15–17 May 2014; pp. 323–326.
19. Abreu, L.D. A q-sampling theorem related to the q-Hankel transform. Proc. Am. Math. Soc. 2005, 133, 1197–1203. [CrossRef]
20. Annaby, M.H.; Mansour, Z.S.; Ashour, O.A. Sampling theorems associated with biorthogonal q-Bessel functions. J. Phys. A 2010,
43, 295204. [CrossRef]
21. Ismail, M.E.H. The zeros of basic Bessel functions, the functions Jν+ ax ( x ), and associated orthogonal polynomials. J. Math. Anal.
Appl. 1982, 86, 1–19. [CrossRef]
22. Ismail, M.E.H.; Muldoon, M.E. On the variation with respect to a parameter of zeros of Bessel and q-Bessel functions. J. Math.
Anal. Appl. 1988, 135, 187–207. [CrossRef]
23. Koelink, H.T.; Swarttouw, R.F. On the zeros of the Hahn-Exton q-Bessel function and associated q-Lommel polynomials. J. Math.
Anal. Appl. 1994, 186, 690–710. [CrossRef]
24. Koornwinder, T.H.; Swarttouw, R.F. On q-analogs of the Fourier and Hankel transforms. Trans. Am. Math. Soc. 1992, 333, 445–461.
25. Baricz, Á.; Dimitrov, D.K.; Mező, I. Radii of starlikeness and convexity of some q-Bessel functions. J. Math. Anal. Appl. 2016, 435,
968–985. [CrossRef]
26. Kishore, N. The Rayleigh function. Proc. Am. Math. Soc. 1963, 14, 527–533. [CrossRef]

Disclaimer/Publisher’s Note: The statements, opinions and data contained in all publications are solely those of the individual
author(s) and contributor(s) and not of MDPI and/or the editor(s). MDPI and/or the editor(s) disclaim responsibility for any injury to
people or property resulting from any ideas, methods, instructions or products referred to in the content.

211
 MDPI AG
Grosspeteranlage 5
4052 Basel
Switzerland
Tel.: +41 61 683 77 34

Mathematics Editorial Office

E-mail: [email protected]
www.mdpi.com/journal/mathematics

Disclaimer/Publisher’s Note: The statements, opinions and data contained in all publications are
solely those of the individual author(s) and contributor(s) and not of MDPI and/or the editor(s).
MDPI and/or the editor(s) disclaim responsibility for any injury to people or property resulting from
any ideas, methods, instructions or products referred to in the content.
Academic Open
Access Publishing

mdpi.com ISBN 978-3-7258-1854-9