Decision Analytics Journal 9 (2023) 100338

Contents lists available at ScienceDirect

Decision Analytics Journal

journal homepage: www.elsevier.com/locate/dajour

A novel Romberg integration method for neutrosophic valued functions

Sandip Moi a , Suvankar Biswas b ,∗, Smita Pal Sarkar a
a
Department of Mathematics, Indian Institute of Engineering Science and Technology, Shibpur, B. Garden, Howrah 711103, India
b
Discipline of Mathematics, School of Sciences, Indira Gandhi National Open University, New Delhi 110068, India

ARTICLE INFO ABSTRACT

Keywords: This study proposes a numerical integration technique to determine the approximate integral of the neu-
Neutrosophic integration trosophic valued function. Newton Cot’s method with a positive coefficient has been used for neutrosophic
Newton Cot’s formula integration, and then the suggested technique is used to find the approximate value of the integral. This method
Peano’s error representation
is called the Romberg integration method. In this method, the composite trapezoidal rule is applied initially,
Romberg integration
and Richardson’s extrapolation technique is used next to improve the solution. Richardson’s extrapolation is
Richardson extrapolation
the basis of Romberg’s integration. Formulas with high-order truncation errors are created using Richardson
extrapolation, which employs an averaging method. Richardson’s extrapolation helps the proposed method to
achieve high computational efficiency. The convergence of the numerical approximation is provided in terms
of theorems. A comparative analysis of the proposed method with other existing methods is demonstrated to
show the efficiency and reliability of the proposed method.

1. Introduction linear optimization method of a single-valued neutrosophic set. Also,

the author has described the sensitivity analysis of attribute weights.
In the field of uncertain mathematics, Smarandache [1] Introduced Hamido et al. [9] proposed a new concept of neutrosophic topological
the idea of the neutrosophic set. After that, many researchers have space. Also, they have proved that neutrosophic topological space is not
devoted their interest to work on the application of the neutrosophic the classical topological space. In [10], Ihsan et al. described the idea
set. Broumi et al. [2] introduced the idea of bipolar and complex of a single-valued neutrosophic hypersoft expert set. Also, they have
neutrosophic sets for multi-criteria decision-making problems. After proposed a new type of algorithm to solve decision-making problems.
that Nguyen et al. [3] used neutrosophic sets in different fields of After that Rahman et al. [11] introduced the idea of a neutrosophic
medical diagnoses. They have shown that all the incomplete and incon- parameterized Hypersoft set. In the same research work, they pro-
sistent information can be managed by a neutrosophic set for modeling
posed an algorithm to study neutrosophic decision-making problems.
patients’ medical situations. Mallick et al. [4] developed the idea of
Bhaumik et al. [12] developed matrix game theory to solve multi-
a penta-partitioned neutrosophic set, the modified version of single-
objective linguistic problems in a neutrosophic environment. They have
valued neutrosophic sets. This type of neutrosphic set can be applied to
shown that their research work can be applied to different tourism
solve multi-criteria decision-making problems. Dat et al. [5] proposed
management problems. In [13], Saeed et al. studied neutrosophic hy-
a single-valued linguistic complex neutrosophic set and an interval
linguistic complex neutrosophic set. Also, they have given the idea persoft set and neutrosophic hypersoft set mapping. Also, they have
of some set-theoretic operations for single-valued linguistic complex discussed the diagnosis of hepatitis with their related problems in a
neutrosophic set and interval linguistic complex neutrosophic set. In the neutrosophic environment. Broumi et al. [14] proposed the idea of
same article, Dat et al. [5] presented a new decision-making problem an n-valued interval neutrosophic set. In the same research work,
for a single-valued and interval linguistic complex neutrosophic set. they have proposed an efficient algorithm for solving the n-valued
In [6] Pratihar et al. studied a transportation problem in a neutro- neutrosophic multi-criteria decision-making problem. Ali et al. [15]
sophic environment. They used a modified northwest corner method developed a single valued neutrosophic multi-criteria decision-making
to find the basic feasible solution to a neutrosophic transportation problem. They have developed the idea of Hausdroff distance between
problem. Lu et al. [7] proposed a new emergency transport model single-valued neutrosophic sets. Luo et al. [16] described a novel
on a single-valued neutrosophic set. They have transformed the emer- distance between single-valued neutrosophic sets and they have applied
gency transshipment problem into a multi-attribute decision-making this distance to studying pattern recognition and medical diagnoses.
problem in a neutrosophic environment. Then Deli [8] developed a Antonysamy et al. [17] used neutrosophic set to determine the effect

∗ Corresponding author.
E-mail address: [email protected] (S. Biswas).

https://doi.org/10.1016/j.dajour.2023.100338
Received 3 August 2023; Received in revised form 4 October 2023; Accepted 5 October 2023
Available online 9 October 2023
2772-6622/© 2023 The Author(s). Published by Elsevier Inc. This is an open access article under the CC BY-NC-ND license
(http://creativecommons.org/licenses/by-nc-nd/4.0/).
S. Moi, S. Biswas and S.P. Sarkar Decision Analytics Journal 9 (2023) 100338

of corona virus on a patient. Recently, many researchers have been 1.2. Novelty
working on the application of neutrosophic set [18–22].
Neutrosophic calculus plays an important role in different scientific The objective of this article has been given below.
fields such as mathematical science, physical science, statistical sci-
• To propose a numerical integration technique to determine the
ence and engineering science. So, many authors have dedicated their
approximate integral of the neutrosophic valued function.
interests to work in neutrosophic calculus. In [23], Smarandache de-
• To use Newton Cot’s formula in the neutrosophic integral.
fined the neutrosophic integral with the help of neutrosophic measure.
• To introduce the Romberg integration method for neutrosophic
After that, Smarandache [24] proposed the notion of neutrosophic function.
derivatives, limits and integrals. Smarandache et al. [25] provided the • To use composite Trapezoidal rule for finding the approximation
concept of neutrosophic definite and indefinite integral. The definition of the given function.
of the definite integral is exactly same as the definition of the classical • To use the Richardson extrapolation method for finding the better
integral. The idea of neutrosophic thick function and its integration approximation of the given function.
was first presented by Alaswad [26]. In [27], Son et al. presented
the notion of granular fractional neutrosophic integration. They orig- 1.3. Arrangement of the article
inally proposed the concept of granular fractional Riemann–Liouville
and Caputo derivatives of neutrosophic-valued functions in the same This work is organized as follows: The preliminary idea of this work
publication. Also, they have proposed the definition of neutrosophic is given in Section 2. In Section 3, Newton Cot’s formula is discussed for
granular partial integrals. Then, Alhasan [28] studied neutrosophic neutrosophic function. Romberg integration for a neutrosophic function
integration by substitution method. In the same article, Alhasan [28] is provided in Section 4. Some numerical examples are examined in
defined neutrosophic trigonometric identities. Also, Alhasan [29] dis- Section 5. Finally, a brief concluding remark is given in Section 6.
cussed the neutrosophic integrals by parts method. In addition, he
introduced a tabular method to compute the neutrosophic integrals 2. Preliminary ideas
and he has shown that the tabular method is easier to implement
than by by-part method. Biswas et al. [30] defined neutrosophic in- Definition 2.1 ([40]). Let  = {⟨𝑛, 𝑇 (𝑛), 𝐼 (𝑛), 𝐹 (𝑛)⟩ ∶ 𝑛 ∈ 𝑋}.
tegration with the help of Riemann integration approach and they Then  is called single valued neutrosophic set over the universal set
also proposed some significant characteristics of neutrosophic numbers X, if 𝑇 ∶ 𝑋 → [0, 1], 𝐼 ∶ 𝑋 → [0, 1], 𝐹 ∶ 𝑋 → [0, 1] such that
and neutrosophic functions. In [31], Alhasan used partial fraction in 0 ≤ 𝑇 (𝑛) + 𝐼 (𝑛) + 𝐹 (𝑛) ≤ 3, ∀𝑛 ∈ 𝑋. Here 𝑇 (𝑛), 𝐼 (𝑛) and 𝐹 (𝑛)
neutrosophic integration. In the same work. Alhasan discussed neu- are denoted as truth, indeterminacy, and falsity membership function
trosophic proper rational function. Also, they introduced the concept respectively.
of the integral of neutrosophic improper rational functions. Recently,
̃ 𝑁̃ ∈  on 𝑌 = {𝑦1 , 𝑦2 , … , 𝑦𝑛 }. Then the
Definition 2.2 ([41]). Let 𝑀,
Biswas et al. [32] used Gaussian quadrature methods to compute
Hausdorff distance  between 𝑀 ̃ and 𝑁̃ on X defined as follows
the numerical integration of neutrosophic functions. Also, they have
1∑
𝑛
introduced a distance function for neutrosophic numbers. In the same ̃ 𝑁)
(𝑀, ̃ = max{|𝑇𝑀̃ (𝑦𝑖 ) − 𝑇𝑁̃ (𝑦𝑖 )|, |𝐼𝑀̃ (𝑦𝑖 ) − 𝐼𝑁̃ (𝑦𝑖 )|, |𝐹𝑀̃ (𝑦𝑖 ) − 𝐹𝑁̃ (𝑦𝑖 )|}
work, they have defined the continuity of neutrosophic functions on 𝑛 𝑖=1
closed-bounded intervals.
Neutrosophic differential and integral equation play an important Proposition 2.1 ([34]). Consider two neutrosophic numbers 𝑚̃ and 𝑛̃ such
role in the development of neutrosophic differential and integral calcu- that
lus. In recent years, many works [33–35] on neutrosophic differential
1. (𝑚⊙
̃ 𝑛) ̃ (𝛼,𝛽,𝛾) = 𝑚̃ (𝛼,𝛽,𝛾) ⊙𝑛̃(𝛼,𝛽,𝛾) , where ⊙ denotes any binary operation
and integral equation [36] are being published for further development ′ +′ ,′ −′ and ′ ×′ .
of this topic.
̃ (𝛼,𝛽,𝛾) = 𝜂 𝑚̃ (𝛼,𝛽,𝛾) , here 0 ≠ 𝜂 ∈ R.
2. (𝜂 𝑚)

1.1. Motivation Definition 2.3 ([25]). Let 𝑔 ∶ [𝑐, 𝑑] →  be a neutrosophic continuous

function on [c, d]. Then neutrosophic integral of 𝑔 on [c, d] is defined
It has been noted from the literature review that not much study as
has been done on the numerical integration of a neutrosophic function. 𝑑 ∑
𝑛
𝑑−𝑐
𝑔(𝑥)𝑑𝑥 = lim 𝑔(𝑃𝑖 )
Some authors have already used the Romberg integration method for ∫𝑐 𝑛→∞ 𝑛
𝑖=1
crisp and fuzzy valued functions [37,38]. Newton Cot’s methods have
here 𝑃𝑖 ∈ [𝑥𝑖−1 , 𝑥𝑖 ], for 𝑖 ∈ {1, 2, … , 𝑛} and 𝑐 = 𝑥0 < 𝑥1 < ⋯ < 𝑥𝑛−1 <
been used to integrate fuzzy functions by Tofigh Allahviranloo [39].
𝑥𝑛 = 𝑑 are the partitions of the interval [𝑐, 𝑑]. Here 𝑔(𝑃𝑖 ) may be a real
So, there is a huge opportunity to develop and modify this type of
set or may have some indeterminacy.
work in a neutrosophic environment. In recent years, researchers have
been working on the numerical solution of neutrosophic differential,
3. Newton Cot’s formula
integral, and integro-differential equations. So, it is necessary to work
on different types of numerical integration of neutrosophic valued
Consider 𝑔 to be a neutrosophic function. Then for any 𝑛 ∈ N,
functions for the development of neutrosophic differential, integral, and
where N be the set of all natural numbers, Newton Cot’s formula for
integro-differential equations. So, from there, we get the motivation
a neutrosophic valued function is given as
to develop a numerical integration method for neutrosophic valued
function. In addition, the concept of Richardson’s extrapolation method
𝑏 ∑
𝑛
𝑏−𝑎
𝑔(𝑥) = ℎ 𝑔𝑖 𝑤𝑖 + 𝐸, 𝑔𝑖 = 𝑔(𝑎 + 𝑖ℎ), ℎ = (1)
needs to be developed in a neutrosophic environment for further de- ∫𝑎 𝑛
𝑖=1
velopment of this topic. In the literature review, it has been observed 𝑏
These provides the approximate solutions of the integral ∫𝑎 𝑔(𝑥)𝑑𝑥.
that no work has been studied on Richardson’s extrapolation method
Now, the (𝛼, 𝛽, 𝛾)−cuts of the Eq. (1) are as follows :
in a neutrosophic environment. So, from there, we got the motivation
to explore this research work and write this article on the Romberg 𝑏 ∑
𝑛
𝑔𝑇𝑗 (𝑥; 𝛼)𝑑𝑥 = ℎ 𝑤𝑖 𝑔𝑇𝑗 (𝑥; 𝛼) + 𝐸(𝑔𝑇𝑗 ; 𝛼) (2)
integration method for neutrosophic valued function. ∫𝑎
𝑖=1

2
S. Moi, S. Biswas and S.P. Sarkar Decision Analytics Journal 9 (2023) 100338

𝑏 ∑
𝑛
3.1. Peano’s error representation
𝑔𝐼𝑗 (𝑥; 𝛽)𝑑𝑥 = ℎ 𝑤𝑖 𝑔𝐼𝑗 (𝑥; 𝛽) + 𝐸(𝑔𝑇𝑗 ; 𝛽) (3)
∫𝑎
𝑖=1
𝑏 ∑𝑛 From the Eqs. (11), (12) and (13) we have,
𝑔𝐹𝑗 (𝑥; 𝛾)𝑑𝑥 = ℎ 𝑤𝑖 𝑔𝐹𝑗 (𝑥; 𝛾) + 𝐸(𝑔𝑇𝑗 ; 𝛾) (4)
∫𝑎 𝐸(𝑔𝑇𝑗 ; 𝛼) = 𝐿(𝑔𝑇𝑗 ; 𝛼) − 𝐽 (𝑔𝑇𝑗 ; 𝛼) (14)
𝑖=1

where 𝑗 = 1, 2 denotes the left and right branches respectively. Here 𝐸(𝑔𝐼𝑗 ; 𝛽) = 𝐿(𝑔𝐼𝑗 ; 𝛽) − 𝐽 (𝑔𝐼𝑗 ; 𝛽) (15)
𝑤𝑖 , 𝑖 = 1, 2, … , 𝑛 be the weights and rational numbers with the property
∑𝑛 𝐸(𝑔𝐹𝑗 ; 𝛾) = 𝐿(𝑔𝐹𝑗 ; 𝛾) − 𝐽 (𝑔𝐹𝑗 ; 𝛾) (16)
that 𝑤𝑖 = 𝑛. This will hold when we take 𝑔𝑇𝑗 (𝑥; 𝛼) = 𝑔𝐼𝑗 (𝑥; 𝛽) =
𝑖=1 The error function 𝐸(𝑔) is linear operator so,
𝑔𝐹𝑗 (𝑥; 𝛾) = 1.
Now, the approximation error can be expressed as follows: 𝐸(𝜆𝑓𝑇𝑗 (𝑥; 𝛼) + 𝑔𝑇𝑗 (𝑥; 𝛼)) = 𝜆𝐸(𝑓𝑇𝑗 ; 𝛼) + 𝐸(𝑔𝑇𝑗 ; 𝛼) (17)

𝐸(𝑔𝑇𝑗 ; 𝛼) = ℎ 𝑝+1
.𝑘.𝑔𝑇(𝑝) (𝛿𝑇𝑗 ; 𝛼), 𝛿𝑇𝑗 ∈ (𝑎, 𝑏) (5) 𝐸(𝜆𝑓𝐼𝑗 (𝑥; 𝛽) + 𝑔𝐼𝑗 (𝑥; 𝛽)) = 𝜆𝐸(𝑓𝐼𝑗 ; 𝛽) + 𝐸(𝑔𝐼𝑗 ; 𝛽) (18)
𝑗

𝐸(𝑔𝐼𝑗 ; 𝛽) = ℎ 𝑝+1
.𝑘.𝑔𝐼(𝑝) (𝛿𝐼𝑗 ; 𝛽), 𝛿𝐼𝑗 ∈ (𝑎, 𝑏) (6) 𝐸(𝜆𝑓𝐹𝑗 (𝑥; 𝛾) + 𝑔𝐹𝑗 (𝑥; 𝛾)) = 𝜆𝐸(𝑓𝐹𝑗 ; 𝛾) + 𝐸(𝑔𝐹𝑗 ; 𝛾) (19)
𝑗

𝐸(𝑔𝐹𝑗 ; 𝛾) = ℎ𝑝+1 .𝑘.𝑔𝐹(𝑝) (𝛿𝐹𝑗 ; 𝛾), 𝛿𝐹𝑗 ∈ (𝑎, 𝑏) (7) for 𝑓 , 𝑔 ∈  and 𝜆 ∈ R where  is the linear neutrosophic function
𝑗
space.
Here the values of p and k depended on the value of n but not on The following theorem shows that the integral representation of the
the integrand g. Sometimes the value of 𝑤𝑖 becomes negative for a error 𝐸(𝑔) is a classical result due to Peano.
large value of n and the given formulas are not suitable for numerical
approximation as many terms are canceled to compute the sum. Theorem 3.2. Let us consider E(p)=0 holds for all 𝑝 ∈ P𝑛 , where P𝑛 be
Let the set of all polynomials whose degree is less than or equal to n. Then for
∑
𝑛
all neutrosophic functions in (𝛼, 𝛽, 𝛾)-cut form 𝑔𝑇1 (𝑥; 𝛼), 𝑔𝑇2 (𝑥; 𝛼), 𝑔𝐼1 (𝑥; 𝛽),
𝐽 (𝑔𝑇𝑗 ; 𝛼) = ℎ 𝑤𝑖 𝑔𝑇𝑗 (𝑥𝑖 ; 𝛼) (8) 𝑔𝐼2 (𝑥; 𝛽), 𝑔𝐹1 (𝑥; 𝛾), 𝑔𝐹2 (𝑥; 𝛾) ∈ 𝐶 𝑛+1 [𝑎, 𝑏],
𝑖=1
∑𝑛 𝑏
𝐽 (𝑔𝐼𝑗 ; 𝛽) = ℎ 𝑤𝑖 𝑔𝐼𝑗 (𝑥𝑖 ; 𝛽) (9) 𝐸(𝑔𝑇𝑗 ; 𝛼) = 𝑔𝑇𝑛+1 (𝑡; 𝛼)𝐾(𝑡)𝑑𝑡 (20)
∫𝑎 𝑗
𝑖=1
𝑏
∑𝑛
𝐸(𝑔𝐼𝑗 ; 𝛽) = 𝑔𝑇𝑛+1 (𝑡; 𝛽)𝐾(𝑡)𝑑𝑡 (21)
𝐽 (𝑔𝐹𝑗 ; 𝛾) = ℎ 𝑤𝑖 𝑔𝐹𝑗 (𝑥𝑖 ; 𝛾) (10) ∫𝑎 𝑗
𝑖=1 𝑏
Then from the Eqs. (8), (9) and (10), we have 𝐸(𝑔𝐹𝑗 ; 𝛾) = 𝑔𝐹𝑛+1 (𝑡; 𝛾)𝐾(𝑡)𝑑𝑡 (22)
∫𝑎 𝑗
𝑏
𝐿(𝑔𝑇𝑗 ; 𝛼) = 𝑔𝑇𝑗 (𝑥; 𝛼)𝑑𝑥 = 𝐽 (𝑔𝑇𝑗 ; 𝛼) + 𝐸(𝑔𝑇𝑗 ; 𝛼) (11) where
∫𝑎 {
1 (𝑥 − 𝑡)𝑛 , 𝑥≥𝑡
𝑏 𝐾(𝑡) = 𝐸 [(𝑥 − 𝑡)𝑛+ ], (𝑥 − 𝑡)𝑛+ =
𝐿(𝑔𝐼𝑗 ; 𝛽) = 𝑔𝐼𝑗 (𝑥; 𝛽)𝑑𝑥 = 𝐽 (𝑔𝐼𝑗 ; 𝛽) + 𝐸(𝑔𝐼𝑗 ; 𝛽) (12) 𝑛! 𝑥 0, 𝑥<𝑡
∫𝑎
𝑏 Here 𝐸𝑥 [(𝑥 − 𝑡)𝑛+ ] denotes the error (𝑥 − 𝑡)𝑛+ .
𝐿(𝑔𝐹𝑗 ; 𝛾) = 𝑔𝐹𝑗 (𝑥; 𝛾)𝑑𝑥 = 𝐽 (𝑔𝐹𝑗 ; 𝛾) + 𝐸(𝑔𝐹𝑗 ; 𝛾) (13)
∫𝑎
The following theorem shows that 𝐽 (𝑔𝑇𝑗 ; 𝛼), 𝐽 (𝑔𝐼𝑗 ; 𝛽), 𝐽 (𝑔𝐹𝑗 ; 𝛾) conver- Proof. Consider the Taylor expansions of 𝑔𝑇𝑗 (𝑥; 𝛼), 𝑔𝐼𝑗 (𝑥; 𝛽) and
gent to 𝐿(𝑔𝑇𝑗 ; 𝛼), 𝐿(𝑔𝐼𝑗 ; 𝛽) and 𝐿(𝑔𝐹𝑗 ; 𝛾) respectively, where 𝑗 = 1, 2 𝑔𝐹𝑗 (𝑥; 𝛾), ∈ [𝑎, 𝑏] at the point c are
denotes the left and right branch respectively. 𝑔 (𝑛) (𝑐; 𝛼)
𝑔𝑇𝑗 (𝑥; 𝛼) = 𝑔𝑇𝑗 (𝑐; 𝛼) + 𝑔𝑇′ (𝑐; 𝛼)(𝑥 − 𝑐) + ⋯ + (𝑥 − 𝑐)𝑛 + 𝑅𝑇𝑗 𝑛 (𝑥; 𝛼)
𝑗 𝑛!
Theorem 3.1. Let 𝑔(𝑡) be a neutrosophic valued continuous function (23)
in the metric D. Then 𝐽 (𝑔𝑇𝑗 ; 𝛼), 𝐽 (𝑔𝐼𝑗 ; 𝛽), 𝐽 (𝑔𝐹𝑗 ; 𝛾) converge uniformly to
𝑔 (𝑛) (𝑐; 𝛽)
𝐿(𝑔𝑇𝑗 ; 𝛼), 𝐿(𝑔𝐼𝑗 ; 𝛽) and 𝐿(𝑔𝐹𝑗 ; 𝛾) respectively. 𝑔𝐼𝑗 (𝑥; 𝛽) = 𝑔𝐼𝑗 (𝑐; 𝛽) + 𝑔𝐼′ (𝑐; 𝛽)(𝑥 − 𝑐) + ⋯ + (𝑥 − 𝑐)𝑛 + 𝑅𝐼𝑗 𝑛 (𝑥; 𝛽)
𝑗 𝑛!
𝑏
(24)
Proof. Since A converges to the integral ∫𝑎 𝑔(𝑡)𝑑𝑡 in metric D if
[ ] 𝑔 (𝑛) (𝑐; 𝛾)
𝑔𝐹𝑗 (𝑥; 𝛾) = 𝑔𝐹𝑗 (𝑐; 𝛾) + 𝑔𝐹′ (𝑐; 𝛾)(𝑥 − 𝑐) + ⋯ + (𝑥 − 𝑐)𝑛 + 𝑅𝐹𝑗 𝑛 (𝑥; 𝛾)
lim max (𝑡𝑖 − 𝑡𝑖−1 ) = 0
𝑗 𝑛!
𝑛→∞ 1≤𝑖≤𝑛 (25)
Now, the Eqs. (8), (9) and (10) can be represented in the form of So, the Remainder terms can be written as follows
𝐴𝑇1 (𝛼), 𝐴𝑇2 (𝛼), 𝐴𝐼1 (𝛽), 𝐴𝐼2 (𝛽), 𝐴𝐹1 (𝛾), and 𝐴𝐹2 (𝛾). 𝑥 𝑏
1 1
Then we have 𝑅𝑇𝑗 𝑛 (𝑥; 𝛼) = 𝑔 𝑛+1 (𝑡; 𝛼)(𝑥 − 𝑡)𝑛 𝑑𝑡 = 𝑔 𝑛+1 (𝑡; 𝛼)(𝑥 − 𝑡)𝑛+ 𝑑𝑡 (26)
{ 𝑛! ∫𝑎 𝑇𝑗 𝑛! ∫𝑎 𝑇𝑗
𝐷(𝐴, 𝐿(𝑔)) = sup max{|𝐴𝑇1 (𝛼) − 𝐿(𝑔𝑇1 ; 𝛼)|, |𝐴𝑇2 (𝛼) − 𝐿(𝑔𝑇2 ; 𝛼)|}+ 1
𝑥
1
𝑏
𝛼,𝛽,𝛾 𝑅𝐼𝑗 𝑛 (𝑥; 𝛽) = 𝑔 𝑛+1 (𝑡; 𝛽)(𝑥 − 𝑡)𝑛 𝑑𝑡 = 𝑔 𝑛+1 (𝑡; 𝛽)(𝑥 − 𝑡)𝑛+ 𝑑𝑡 (27)
𝑛! ∫𝑎 𝐼𝑗 𝑛! ∫𝑎 𝐼𝑗
max{|𝐴𝐼1 (𝛽) − 𝐿(𝑔𝐼1 ; 𝛽)|, |𝐴𝐼2 (𝛽) − 𝐿(𝑔𝐼2 ; 𝛽)|}+ 𝑥 𝑏
} 1 1
max{|𝐴𝐹1 (𝛾) − 𝐿(𝑔𝐹1 ; 𝛾)|, |𝐴𝐹2 (𝛾) − 𝐿(𝑔𝐹2 ; 𝛾)|} 𝑅𝐹𝑗 𝑛 (𝑥; 𝛾) = 𝑔 𝑛+1 (𝑡; 𝛾)(𝑥 − 𝑡)𝑛 𝑑𝑡 = 𝑔 𝑛+1 (𝑡; 𝛾)(𝑥 − 𝑡)𝑛+ 𝑑𝑡 (28)
𝑛! ∫𝑎 𝐹𝑗 𝑛! ∫𝑎 𝐹𝑗
Since lim 𝐷(𝐴, 𝐿(𝑔)) = 0, max (𝑡𝑖 − 𝑡𝑖−1 ) → 0 as 𝑛 → ∞. Applying the linear operator E to the Eqs. (23), (24) and (25) we have
𝑛→∞ 1≤𝑖≤𝑛 ( 𝑏 )
So, 𝐴𝑇1 (𝛼), 𝐴𝑇2 (𝛼), 𝐴𝐼1 (𝛽), 𝐴𝐼2 (𝛽), 𝐴𝐹1 (𝛾), and 𝐴𝐹2 (𝛾) converges uni- 1
formly to 𝐿(𝑔𝑇1 ; 𝛼), 𝐿(𝑔𝑇2 ; 𝛼), 𝐿(𝑔𝐼1 ; 𝛽), 𝐸(𝑔𝑇𝑗 ; 𝛼) = 𝐸(𝑅𝑇𝑗 𝑛 ; 𝛼) = 𝐸𝑥 𝑔𝑇𝑛+1 (𝑡; 𝛼)(𝑥 − 𝑡)𝑛+ 𝑑𝑡 (29)
𝑛! ∫𝑎 𝑗
𝐿(𝑔𝐼2 ; 𝛽), 𝐿(𝑔𝐹1 ; 𝛾) and 𝐿(𝑔𝐹2 ; 𝛾) respectively. ( 𝑏 )
1
Here 𝐽 (𝑔𝑇1 ; 𝛼), 𝐽 (𝑔𝑇2 ; 𝛼), 𝐽 (𝑔𝐼1 ; 𝛽), 𝐽 (𝑔𝐼2 ; 𝛽), 𝐽 (𝑔𝐹1 ; 𝛾), 𝐽 (𝑔𝐹2 ; 𝛾) are the 𝐸(𝑔𝐼𝑗 ; 𝛽) = 𝐸(𝑅𝐼𝑗 𝑛 ; 𝛽) = 𝐸𝑥 𝑔𝐼𝑛+1 (𝑡; 𝛽)(𝑥 − 𝑡)𝑛+ 𝑑𝑡 (30)
𝑛! ∫𝑎 𝑗
particular case of 𝐴𝑇1 (𝛼), 𝐴𝑇2 (𝛼), ( 𝑏 )
𝐴𝐼1 (𝛽), 𝐴𝐼2 (𝛽), 𝐴𝐹1 (𝛾), and 𝐴𝐹2 (𝛾). Since (𝑡𝑖 − 𝑡𝑖−1 ) ∶= ℎ𝑤𝑖 , 𝑖 = 1, … , 𝑛 1
𝐸(𝑔𝐹𝑗 ; 𝛾) = 𝐸(𝑅𝐹𝑗 𝑛 ; 𝛾) = 𝐸𝑥 𝑔 𝑛+1 (𝑡; 𝛾)(𝑥 − 𝑡)𝑛+ 𝑑𝑡 (31)
This completes the proof. □ 𝑛! ∫𝑎 𝐹𝑗

3
S. Moi, S. Biswas and S.P. Sarkar Decision Analytics Journal 9 (2023) 100338

Since 𝐸(𝑝) = 0 for 𝑝 ∈ P𝑛 . Now, we have to interchange the 𝐸𝑥 Now, 𝐽𝑘,1 (𝑔𝑇𝑗 ; 𝛼), 𝐽𝑘,1 (𝑔𝐼𝑗 ; 𝛽) and 𝐽𝑘,1 (𝑔𝐹𝑗 ; 𝛾) have been used to de-
operator with the integration. Then from [42], it has been shown that note the portion of the Eqs. (39), (40), (41), (42), (43) and (44) for
𝐸𝑥 commutes with integration. This Completes the proof. □ Trapezoidal approximation. Then,
ℎ [ ]
𝐽1,1 (𝑔𝑇𝑗 ; 𝛼) = 1 𝑔𝑇𝑗 (𝑥0 ; 𝛼) + 𝑔𝑇𝑗 (𝑥𝑛 ; 𝛼) (45)
4. Romberg integration 2
ℎ1 [ ]
𝐽1,1 (𝑔𝐼𝑗 ; 𝛽) = 𝑔𝐼𝑗 (𝑥0 ; 𝛽) + 𝑔𝐼𝑗 (𝑥𝑛 ; 𝛽) (46)
2
In Romberg integration, the composite Trapezoidal rule has been ℎ1 [ ]
used to find the approximation of the given function. Then the Richard- 𝐽1,1 (𝑔𝐹𝑗 ; 𝛾) = 𝑔𝐹𝑗 (𝑥0 ; 𝛾) + 𝑔𝐹𝑗 (𝑥𝑛 ; 𝛾) (47)
2
son extrapolation method has been used to find a better approximation Then,
ℎ [ ]
of the given function.
The Richardson extrapolation [43] can be used in the following 𝐽2,1 (𝑔𝑇𝑗 ; 𝛼) = 2 𝑔𝑇𝑗 (𝑥0 ; 𝛼) + 2𝑔𝑇𝑗 (𝑥0 + ℎ2 ; 𝛼) + 𝑔𝑇𝑗 (𝑥𝑛 ; 𝛼)
2
manner : 𝑥 − 𝑥0 [ ( 𝑥 − 𝑥0 )]
= 𝑛 𝑔𝑇𝑗 (𝑥0 ; 𝛼) + 𝑔𝑇𝑗 (𝑥𝑛 ; 𝛼) + 2𝑔𝑇𝑗 𝑥0 + 𝑛 ;𝛼
4 2
𝑈 − 𝑉 (ℎ) = 𝐶1 ℎ + 𝐶2 ℎ2 + ⋯ + 𝐶𝑛 ℎ𝑛 (32) [ ]
1
= 𝐽 (𝑔 ; 𝛼) + ℎ1 𝑔𝑇𝑗 (𝑥0 + ℎ2 ; 𝛼) (48)
2 1,1 𝑇𝑗
where 𝐶1 , 𝐶2 , … , 𝐶𝑛 are constants and 𝑉 (ℎ) is an approximation to the
Similarly,
unknown value U. The truncation in Eq. (32) is dominated by 𝐶1 ℎ,
[ ]
where h is small. In Richardson’s extrapolation formula, the averaging 1
𝐽2,1 (𝑔𝐼𝑗 ; 𝛽) = 𝐽 (𝑔 ; 𝛽) + ℎ1 𝑔𝐼𝑗 (𝑥0 + ℎ2 ; 𝛽) (49)
2 1,1 𝐼𝑗
technique has been used to find higher-order truncation errors. Also, [ ]
1
the extrapolation was used to find the approximation of the definite 𝐽2,1 (𝑔𝐹𝑗 ; 𝛾) = 𝐽 (𝑔 ; 𝛾) + ℎ1 𝑔𝐹𝑗 (𝑥0 + ℎ2 ; 𝛾) (50)
2 1,1 𝐹𝑗
neutrosophic integral. Now, the composite Trapezoidal rule is used to { [ ]}
1
approximate the definite neutrosophic integral on the interval 𝑎 = 𝑥0 < 𝐽3,1 (𝑔𝑇𝑗 ; 𝛼) = 𝐽2,1 (𝑔𝑇𝑗 ; 𝛼) + ℎ2 𝑔𝑇𝑗 (𝑥0 + ℎ3 ; 𝛼) + 𝑔𝑇𝑗 (𝑥0 + 3ℎ3 ; 𝛼)
2
𝑥1 < 𝑥2 < ⋯ < 𝑥𝑛 = 𝑏, where [𝑎, 𝑏] divided into 𝑛 sub-intervals. Then, (51)
[ ] { [ ]}
∑
𝑛−1 1
ℎ 𝐽3,1 (𝑔𝐼𝑗 ; 𝛽) = 𝐽2,1 (𝑔𝐼𝑗 ; 𝛽) + ℎ2 𝑔𝐼𝑗 (𝑥0 + ℎ3 ; 𝛽) + 𝑔𝐼𝑗 (𝑥0 + 3ℎ3 ; 𝛽)
𝐿(𝑔𝑇𝑗 ; 𝛼) = 𝑔𝑇𝑗 (𝑥0 ; 𝛼) + 2 𝑔𝑇𝑗 (𝑥𝑖 ; 𝛼) + 𝑔𝑇𝑗 (𝑥𝑛 ; 𝛼) + 𝐸(𝑔𝑇𝑗 ; 𝛼) (33) 2
2 𝑖=1 (52)
[ ] { [ ]}
ℎ ∑
𝑛−1 1
𝐿(𝑔𝐼𝑗 ; 𝛽) = 𝑔𝐼𝑗 (𝑥0 ; 𝛽) + 2 𝑔𝐼𝑗 (𝑥𝑖 ; 𝛽) + 𝑔𝐼𝑗 (𝑥𝑛 ; 𝛽) + 𝐸(𝑔𝐼𝑗 ; 𝛽) (34) 𝐽3,1 (𝑔𝐹𝑗 ; 𝛾) = 𝐽2,1 (𝑔𝐹𝑗 ; 𝛾) + ℎ2 𝑔𝐹𝑗 (𝑥0 + ℎ3 ; 𝛾) + 𝑔𝐹𝑗 (𝑥0 + 3ℎ3 ; 𝛾)
2 2
𝑖=1
[ ] (53)
ℎ ∑
𝑛−1
𝐿(𝑔𝐹𝑗 ; 𝛾) = 𝑔𝐹𝑗 (𝑥0 ; 𝛾) + 2 𝑔𝐹𝑗 (𝑥𝑖 ; 𝛾) + 𝑔𝐹𝑗 (𝑥𝑛 ; 𝛾) + 𝐸(𝑔𝐹𝑗 ; 𝛾) (35) In general
2 𝑖=1
⎧ 2∑𝑘−2 ⎫
where, 1⎪ ⎪
𝐽𝑘,1 (𝑔𝑇𝑗 ; 𝛼) = ⎨𝐽𝑘−1,1 (𝑔𝑇𝑗 ; 𝛼) + ℎ𝑘−1 𝑔𝑇𝑗 (𝑥0 + (2𝑗 − 1)ℎ𝑘 ; 𝛼)⎬ (54)
2⎪ ⎪
ℎ2 ⎩
𝑗=1
⎭
𝐸(𝑔𝑇𝑗 ; 𝛼) = − (𝑏 − 𝑎)𝑔𝑇(2) (𝛿𝑇𝑗 ; 𝛼) (36)
12 𝑗
⎧ ⎫
ℎ2 2∑𝑘−2
𝐸(𝑔𝐼𝑗 ; 𝛽) = − (𝑏 − 𝑎)𝑔𝐼(2) (𝛿𝐼𝑗 ; 𝛽) (37) 1⎪ ⎪
12 𝑗 𝐽𝑘,1 (𝑔𝐼𝑗 ; 𝛽) = ⎨𝐽𝑘−1,1 (𝑔𝐼𝑗 ; 𝛽) + ℎ𝑘−1 𝑔𝐼𝑗 (𝑥0 + (2𝑗 − 1)ℎ𝑘 ; 𝛽)⎬ (55)
2⎪ ⎪
ℎ2 (2) ⎩
𝑗=1
⎭
𝐸(𝑔𝐹𝑗 ; 𝛾) = − (𝑏 − 𝑎)𝑔𝐹 (𝛿𝐹𝑗 ; 𝛾) (38)
12 𝑗
⎧ 2∑𝑘−2 ⎫
At first, find the composite trapezoidal rule for finding the approxima- 1⎪ ⎪
𝐽𝑘,1 (𝑔𝐹𝑗 ; 𝛾) = ⎨𝐽𝑘−1,1 (𝑔𝐹𝑗 ; 𝛾) + ℎ𝑘−1 𝑔𝐹𝑗 (𝑥0 + (2𝑗 − 1)ℎ𝑘 ; 𝛾)⎬ (56)
tion with 𝑛1 = 1, 𝑛2 = 2, 𝑛3 = 4, … , 𝑛𝑚 = 2𝑚−1 , where 𝑚 ∈ N. Let ℎ𝑘 be 2⎪
𝑏−𝑎 𝑏−𝑎 𝑗=1 ⎪
the step size to 𝑛𝑘 , then ℎ𝑘 = = . ⎩ ⎭
𝑛𝑘 2𝑘−1
Then, from the Trapezoidal rule, where 𝑘 = 2, 3, … , 𝑛 and 𝑗 = 1, 2 denotes the left and right branch
respectively.
ℎ𝑘 ⎡⎢ ⎤
2𝑘−1
∑−1 Here the Richardson extrapolation has been used to show the speed
𝐿(𝑔𝑇𝑗 ; 𝛼) = 𝑔𝑇𝑗 (𝑥0 ; 𝛼) + 2 𝑔𝑇𝑗 (𝑥𝑖 ; 𝛼) + 𝑔𝑇𝑗 (𝑥𝑛 ; 𝛼)⎥ + 𝐸(𝑔𝑇𝑗 ; 𝛼) of the Convergence of the method. If 𝑔𝑇𝑗 , 𝑔𝐼𝑗 , 𝑔𝐹𝑗 ∈ 𝐶 ∞ [𝑎, 𝑏] then the
2 ⎢ ⎥
⎣ 𝑖=1 ⎦ composite Trapezoidal rule can be written as
(39) ∑
∞ ∑
∞
𝐿(𝑔𝑇𝑗 ; 𝛼) − 𝐽𝑘,1 (𝑔𝑇𝑗 ; 𝛼) = 𝐶𝑖𝑗 ℎ2𝑖 2
𝑘 = 𝐶11 ℎ𝑘 + 𝐶𝑖𝑗 ℎ2𝑖 (57)
ℎ ⎡ ⎤
2𝑘−1
∑−1 𝑘
𝐿(𝑔𝐼𝑗 ; 𝛽) = 𝑘 ⎢𝑔𝐼𝑗 (𝑥0 ; 𝛽) + 2 𝑔𝐼𝑗 (𝑥𝑖 ; 𝛽) + 𝑔𝐼𝑗 (𝑥𝑛 ; 𝛽)⎥ + 𝐸(𝑔𝐼𝑗 ; 𝛽) 𝑖=1 𝑖=2
2 ⎢ ⎥ ∑∞ ∑∞
⎣ 𝑖=1 ⎦ 𝐿(𝑔𝐼𝑗 ; 𝛽) − 𝐽𝑘,1 (𝑔𝐼𝑗 ; 𝛽) = 𝐶𝑖𝑗′ ℎ2𝑖 ′ 2
𝐶𝑖𝑗′ ℎ2𝑖 (58)
𝑘 = 𝐶11 ℎ𝑘 + 𝑘
(40) 𝑖=1 𝑖=2
∑∞ ∑∞
ℎ𝑘 ⎡⎢ ⎤
2𝑘−1
∑−1 𝐿(𝑔𝐹𝑗 ; 𝛾) − 𝐽𝑘,1 (𝑔𝐹𝑗 ; 𝛾) = 𝐶𝑖𝑗′′ ℎ2𝑖 ′′ 2
𝑘 = 𝐶11 ℎ𝑘 + 𝐶𝑖𝑗′′ ℎ2𝑖 (59)
𝐿(𝑔𝐹𝑗 ; 𝛾) = 𝑔𝐹𝑗 (𝑥0 ; 𝛾) + 2 𝑔𝐹𝑗 (𝑥𝑖 ; 𝛾) + 𝑔𝐹𝑗 (𝑥𝑛 ; 𝛾)⎥ + 𝐸(𝑔𝐹𝑗 ; 𝛾) 𝑘
2 ⎢ ⎥ 𝑖=1 𝑖=2
⎣ 𝑖=1 ⎦
where 𝐶𝑖𝑗 , 𝐶𝑖𝑗′ , 𝐶𝑖𝑗′′ are all independent of the value of ℎ𝑘 and the
(41)
values of 𝐶𝑖𝑗 , 𝐶𝑖𝑗′ , 𝐶𝑖𝑗′′ depends only on 𝑔𝑇2𝑖−1 (𝑎; 𝛼), 𝑔𝐼2𝑖−1 (𝑎; 𝛽), 𝑔𝐹2𝑖−1 (𝑎; 𝛾),
𝑗 𝑗 𝑗
where, 𝑔𝑇2𝑖−1 (𝑏; 𝛼), 𝑔𝐼2𝑖−1 (𝑏; 𝛽) and 𝑔𝐹2𝑖−1 (𝑏; 𝛾). So, in the form of the Composite
𝑗 𝑗 𝑗
ℎ2𝑘 Trapezoidal rule, the term involving ℎ2𝑘 can eliminate and replace ℎ𝑘
𝐸(𝑔𝑇𝑗 ; 𝛼) = − (𝑏 − 𝑎)𝑔𝑇(2) (𝛿𝑇𝑗 ; 𝛼) (42) ℎ
12 𝑗
by ℎ𝑘+1 = 𝑘 . Then,
ℎ2𝑘 2
𝐸(𝑔𝐼𝑗 ; 𝛽) = − (𝑏 − 𝑎)𝑔𝐼(2) (𝛿𝐼𝑗 ; 𝛽) (43) ∑
∞ ∑
∞
𝐶𝑖𝑗 ℎ2𝑖 𝐶11 ℎ2𝐾 ∑
∞
𝐶𝑖𝑗 ℎ2𝑖
12 𝑗 𝑘 𝑘
𝐿(𝑔𝑇𝑗 ; 𝛼) − 𝐽𝑘,1 (𝑔𝑇𝑗 ; 𝛼) = 𝐶𝑖𝑗 ℎ2𝑖 = = +
ℎ2𝑘 𝑘+1
22𝑖 4 4𝑖
𝐸(𝑔𝐹𝑗 ; 𝛾) = − (𝑏 − 𝑎)𝑔𝐹(2) (𝛿𝐹𝑗 ; 𝛾) (44) 𝑖=1 𝑖=1 𝑖=2
12 𝑗
(60)

4
S. Moi, S. Biswas and S.P. Sarkar Decision Analytics Journal 9 (2023) 100338

∑
∞ ∑∞ 𝐶 ′ ℎ2𝑖 ′ ℎ2
𝐶11 ∑∞ 𝐶 ′ ℎ2𝑖
𝑖𝑗 𝑘 𝐾 𝑖𝑗 𝑘 5. Comparative analysis, advantages and limitations of the pro-
𝐿(𝑔𝐼𝑗 ; 𝛽) − 𝐽𝑘,1 (𝑔𝐼𝑗 ; 𝛽) = 𝐶𝑖𝑗′ ℎ2𝑖
𝑘+1
= = +
𝑖=1 𝑖=1 2
2𝑖 4 𝑖=2
4𝑖 posed method
(61)
5.1. Comparative analysis
∑
∞ ∑
∞ 𝐶𝑖𝑗′′ ℎ2𝑖
𝑘
′′ ℎ2
𝐶11 𝐾
∑
∞ 𝐶𝑖𝑗′′ ℎ2𝑖
𝑘
𝐿(𝑔𝐹𝑗 ; 𝛾) − 𝐽𝑘,1 (𝑔𝐹𝑗 ; 𝛾) = 𝐶𝑖𝑗′′ ℎ2𝑖
𝑘+1
= = +
𝑖=1 𝑖=1 22𝑖 4 𝑖=2
4𝑖 In this Section, a Comparative study of the proposed method with
(62) other existing methods has been discussed to show the superiority,
reliability, and effectiveness of the proposed method.
Now, multiply the Eq. (60) by 4 and subtracting the Eq. (57) from this In Table 1, it has been seen that Wu [44] applied Riemann–Stieltjes
integral to obtain the integration of fuzzy valued function. This is not
we have
a numerical method and we cannot find the numerical integration
[ ] of the fuzzy valued function. Also, this method cannot be applied
𝐽𝑘+1,1 (𝑔𝑇𝑗 ; 𝛼) − 𝐽𝑘,1 (𝑔𝑇𝑗 ; 𝛼)
𝐿(𝑔𝑇𝑗 ; 𝛼) − 𝐽𝑘+1,1 (𝑔𝑇𝑗 ; 𝛼) + for neutrosophic valued functions as the falsity and indeterminacy
3 parts are absent here. However, our proposed method is applicable for
( )
∑
∞
𝐶𝑖𝑗 ℎ2𝑖
𝑘
neutrosophic valued functions. Pedro et al. [45] used fuzzy Riemann
= − ℎ2𝑖
𝑘 integration to find the integration of a fuzzy valued function. Also, this
𝑖=2
3 4𝑖−1
( ) method is not a numerical method. So, we cannot find the numerical
∑∞
𝐶𝑖 1 − 4𝑖−1
= ℎ2𝑖 (63) integration of a function with the help of this method. Also, this method
𝑘
𝑖=2
3 4𝑖−1 has been developed for fuzzy functions. So, we cannot apply this
[ ]
𝐽𝑘+1,1 (𝑔𝐼𝑗 ; 𝛽) − 𝐽𝑘,1 (𝑔𝐼𝑗 ; 𝛽) method to find the numerical integration of the neutrosophic valued
𝐿(𝑔𝐼𝑗 ; 𝛽) − 𝐽𝑘+1,1 (𝑔𝐼𝑗 ; 𝛽) + function as the falsity and indeterminacy parts are absent here. But our
3
( ) proposed method is a numerical method and we can apply this method
∑
∞ 𝐶𝑖𝑗′ ℎ2𝑖
𝑘 to find the numerical integration of neutrosophic valued functions. In
= − ℎ2𝑖
𝑘
𝑖=2
3 4𝑖−1 addition, Alhasan et al. [28] used the substitution method to compute
( ) the integration of the neutrosophic valued function. This method is an
∑∞
𝐶𝑖′ 1 − 4𝑖−1
= ℎ2𝑖
𝑘 (64) analytical method and we cannot apply this method to compute the
𝑖=2
3 4𝑖−1
[ ] numerical integration of neutrosophic valued function. However, our
𝐽𝑘+1,1 (𝑔𝐹𝑗 ; 𝛾) − 𝐽𝑘,1 (𝑔𝐹𝑗 ; 𝛾) method is applicable to compute the numerical integration of neutro-
𝐿(𝑔𝐹𝑗 ; 𝛾) − 𝐽𝑘+1,1 (𝑔𝐹𝑗 ; 𝛾) +
3 sophic valued functions. Therefore, this study shows the superiority,
∞ 𝐶 ′′
( ) reliability, and effectiveness of our proposed method.
∑ 𝑖𝑗 ℎ2𝑖
𝑘
= − ℎ2𝑖
𝑘
𝑖=2
3 4𝑖−1
5.2. Advantages of the proposed method
( )
∑∞
𝐶𝑖′′ 1 − 4𝑖−1
= ℎ2𝑖
𝑘 (65)
𝑖=2
3 4𝑖−1 1. This numerical method gives more accurate results than other
existing numerical methods such as the trapezoidal rule and
Now, simplify the notion then Simpson’s rule.
2. The Romberg integration method converges to the exact solu-
𝐽𝑘,1 (𝑔𝑇𝑗 ; 𝛼) − 𝐽𝑘−1,1 (𝑔𝑇𝑗 ; 𝛼) tion within a few computational steps. So, this method is more
𝐽𝑘,2 (𝑔𝑇𝑗 ; 𝛼) = 𝐽𝑘,1 (𝑔𝑇𝑗 ; 𝛼) + (66) efficient compared to the other existing methods.
3
𝐽𝑘,1 (𝑔𝐼𝑗 ; 𝛽) − 𝐽𝑘−1,1 (𝑔𝐼𝑗 ; 𝛽) 3. The Romberg integration method provides less error in every
𝐽𝑘,2 (𝑔𝐼𝑗 ; 𝛽) = 𝐽𝑘,1 (𝑔𝐼𝑗 ; 𝛽) + (67) computational step. This shows the reliability of the proposed
3
𝐽𝑘,1 (𝑔𝐹𝑗 ; 𝛾) − 𝐽𝑘−1,1 (𝑔𝐹𝑗 ; 𝛾) method.
𝐽𝑘,2 (𝑔𝐹𝑗 ; 𝛾) = 𝐽𝑘,1 (𝑔𝐹𝑗 ; 𝛾) + (68) 4. The Romberg integration method is easy to use and imple-
3 ment compared to other complex numerical techniques such as
adaptive quadrature methods.
where 𝑘 = 2, 3, … , 𝑛. Now, we apply the Richardson extrapolation
method and continue this process for 𝑘 = 2, 3, … , 𝑛 and 𝑚 = 2, 3, … , 𝑘. 5.3. Limitations of the proposed method
Then we have,
1. The major limitation of the proposed method is that the method
𝐽𝑘,𝑚−1 (𝑔𝑇𝑗 ; 𝛼) − 𝐽𝑘−1,𝑚−1 (𝑔𝑇𝑗 ; 𝛼) is difficult to implement for highly complex functions. Also, this
𝐽𝑘,𝑚 (𝑔𝑇𝑗 ; 𝛼) = 𝐽𝑘,𝑚−1 (𝑔𝑇𝑗 ; 𝛼) + (69)
4𝑚−1 − 1 method is domain-specific. If the domain is unbounded, then this
𝐽𝑘,𝑚−1 (𝑔𝐼𝑗 ; 𝛽) − 𝐽𝑘−1,𝑚−1 (𝑔𝐼𝑗 ; 𝛽) method cannot be applicable to find the numerical integration of
𝐽𝑘,𝑚 (𝑔𝐼𝑗 ; 𝛽) = 𝐽𝑘,𝑚−1 (𝑔𝐼𝑗 ; 𝛽) + (70) a function.
4𝑚−1 − 1
𝐽𝑘,𝑚−1 (𝑔𝐹𝑗 ; 𝛾) − 𝐽𝑘−1,𝑚−1 (𝑔𝐼𝑗 ; 𝛾)
𝐽𝑘,𝑚 (𝑔𝐹𝑗 ; 𝛾) = 𝐽𝑘,𝑚−1 (𝑔𝐹𝑗 ; 𝛾) + (71) 6. Examples
4𝑚−1 − 1

| Some numerical examples are provided in this part to support our

We generate approximations until the terms |𝐽𝑚−1,𝑚−1 (𝑔𝑇𝑗 ; 𝛼)−
| suggested methodology. In addition, some error analysis has been done
| | | |
𝐽𝑚,𝑚 (𝑔𝑇𝑗 ; 𝛼)|, |𝐽𝑚−1,𝑚−1 (𝑔𝐼𝑗 ; 𝛽) − 𝐽𝑚,𝑚 (𝑔𝐼𝑗 ; 𝛽)| and |𝐽𝑚−1,𝑚−1 (𝑔𝐹𝑗 ; 𝛾)−
| | | | of the proposed method with the help of the following convergence
|
𝐽𝑚,𝑚 (𝑔𝐹𝑗 ; 𝛾)| are within the tolerance. indicators.
|
| |
• The absolute error: |E𝑘,𝑚 (𝑔𝑀𝑖 ; 𝜇)| = |𝐽 𝑒𝑥𝑎𝑐𝑡 (𝑔𝑀𝑖 ; 𝜇) − 𝐽𝑘,𝑚 (𝑔𝑀𝑖 ; 𝜇)|
Not only the above terms but it has also been shown that the terms | |
𝑟𝑒𝑓
| | | | • The reference error : 𝑒𝑘,𝑚 (𝑔𝑀𝑖 ; 𝜇) = ‖E𝑘,𝑚 (𝑔𝑀𝑖 ; 𝜇)‖2 .
|𝐽𝑚−2,𝑚−2 (𝑔𝑇𝑗 ; 𝛼) − 𝐽𝑚−1,𝑚−1 (𝑔𝑇𝑗 ; 𝛼)|, |𝐽𝑚−2,𝑚−2 (𝑔𝐼𝑗 ; 𝛽) − 𝐽𝑚−1,𝑚−1 (𝑔𝐼𝑗 ; 𝛽)|
| | | |
| |
and |𝐽𝑚−2,𝑚−2 (𝑔𝐹𝑗 ; 𝛾) − 𝐽𝑚−1,𝑚−1 (𝑔𝐹𝑗 ; 𝛾)| are within the tolerance. Here 𝑀 = 𝑇 , 𝐼, 𝐹 and 𝜇 = 𝛼, 𝛽, 𝛾 where 𝑖 = 1, 2 respectively.
| |

5
S. Moi, S. Biswas and S.P. Sarkar Decision Analytics Journal 9 (2023) 100338

Table 1
Comparative analysis of the proposed method with other existing methods.
Methods Nature of the environment Description of the methods Remarks
Method described by Fuzzy Using Riemann–Stieltjes integral to find the This method is not applicable to find the
Wu [44]. integration of fuzzy valued function. numerical solution of a function. Also, this
method has been developed for fuzzy-valued
functions. This method cannot be applicable
for neutrosophic valued functions.
Method described by Fuzzy Using Frechet derivative and the Riemann This method cannot be applicable for
Pedro et al. [45] integral to find the integration of fuzzy neutrosophic valued function.
valued function.
Method described by Neutrosophic Using substitution method to compute the This is an analytical method. This method
Alhasan [28] integration of a neutrosophic valued cannot be applicable for computing the
function. numerical integration of neutrosophic
valued function.
Our method Neutrosophic Using Romberg integration method to find This is a numerical method. This method is
the numerical integration of neutrosophic applicable for computing the numerical
valued function. integration of neutrosophic valued functions.

Table 2
Absolute errors for truth membership function.
𝛼 |𝐸11 (𝑔𝑇1 ; 𝛼)| |𝐸21 (𝑔𝑇1 ; 𝛼)| |𝐸22 (𝑔𝑇1 ; 𝛼)| |𝐸11 (𝑔𝑇2 ; 𝛼)| |𝐸21 (𝑔𝑇2 ; 𝛼)| |𝐸22 (𝑔𝑇2 ; 𝛼)|
0 0.00E−0 0.00E−0 0.00E−0 6.15E−1 1.55E−1 1.16E−3
0.2 1.02E−1 2.58E−2 1.93E−4 5.13E−1 1.29E−1 9.66E−4
0.4 2.05E−1 6.15E−2 3.86E−4 4.10E−1 1.03E−1 7.72E−4
0.6 3.08E−1 7.73E−2 5.79E−4 3.08E−1 7.73E−2 5.79E−4

Example 1. Consider the following neutrosophic integration In Table 2, 3 and 4, 𝐸𝑘,𝑚 (𝑔𝑇𝑖 ; 𝛼), 𝐸𝑘,𝑚 (𝑔𝐼𝑖 ; 𝛽) and 𝐸𝑘,𝑚 (𝑔𝐹𝑖 ; 𝛾) have
1 been investigated respectively. Here 𝐸𝑘,𝑚 (𝑔𝑇𝑖 ; 𝛼), 𝐸𝑘,𝑚 (𝑔𝐼𝑖 ; 𝛽) and
̃ 𝑥 + 𝑥2 + 1) 𝑑𝑥
𝑛(𝑒 𝑤ℎ𝑒𝑟𝑒 𝑛̃ = ⟨(0, 1, 2); 0.6, 0.4, 0.2⟩. 𝐸𝑘,𝑚 (𝑔𝐹𝑖 ; 𝛾) denotes Absolute errors for truth indeterminacy and falsity
∫𝑜
of 𝑔 respectively. Also, from Figs. 1, 2, 3, 4, 5 and 6, it has been
Now, the exact solution of the above equation is given below seen that 𝐸𝑘,𝑚 (𝑔𝑇𝑖 ; 𝛼), 𝐸𝑘,𝑚 (𝑔𝐼𝑖 ; 𝛽) and 𝐸𝑘,𝑚 (𝑔𝐹𝑖 ; 𝛾) are decrease as (k,
( ) ⟨[ ] [ 5(1 − 𝛽) 5𝛽 + 1 ] [ 5(1 − 𝛾) 5𝛾 + 3 ]⟩ m) increase. Also, 𝑒𝑟𝑒𝑓 (𝑔 ; 𝛼), 𝑒𝑟𝑒𝑓 (𝑔 ; 𝛽) and 𝑒𝑟𝑒𝑓 (𝑔 ; 𝛾) have been
1 5𝛼 6 − 5𝛼 𝑘,𝑚 𝑇𝑖 𝑘,𝑚 𝐼𝑖 𝑘,𝑚 𝐹𝑖
𝑒+ , , , , ,
3 3 3 3 3 4 4 investigated in Table 5. Here 𝑒𝑟𝑒𝑓 (𝑔 ; 𝛼), 𝑒𝑟𝑒𝑓
𝑘,𝑚 𝑇𝑖
(𝑔 ; 𝛽) and 𝑒𝑟𝑒𝑓
𝑘,𝑚 𝐼𝑖
(𝑔 ; 𝛾)
𝑘,𝑚 𝐹𝑖
denotes Reference errors for truth indeterminacy and falsity of 𝑔
where 𝛼 ∈ [0, 0.6], 𝛽 ∈ [0.4, 1] and 𝛾 ∈ [0.2, 1]. Now, we are going to
respectively (see Table 5). From Table 5, it has been noticed that
determine the integral with the help of Romberg integration method.
𝑒𝑟𝑒𝑓 (𝑔 ; 𝛼), 𝑒𝑟𝑒𝑓 (𝑔 ; 𝛽) and 𝑒𝑟𝑒𝑓 (𝑔 ; 𝛾) decrease as (k, m) increases. The
̃ 𝑥 + 𝑥2 + 1). For ℎ = 1,
Let us consider 𝑔(𝑥) = 𝑛(𝑒 𝑘,𝑚 𝑇𝑖 𝑘,𝑚 𝐼𝑖 𝑘,𝑚 𝐹𝑖
pictorial representation of 𝑔(𝑥) has been shown in Figs. 7, 8, and 9. In
5(4 + 𝑒)𝛼 (4 + 𝑒)(6 − 5𝛼) 𝑔(𝑥), we have considered the parameters as a triangular neutrosophic
𝐽1,1 (𝑔𝑇1 ; 𝛼) = , 𝐽1,1 (𝑔𝑇2 ; 𝛼) = ,
6 6 number. Then, from Figs. 7, 8, and 9, it has been seen that 𝑔(𝑥) forms
5(4 + 𝑒)(1 − 𝛽) (4 + 𝑒)(5𝛽 + 1) a triangular neutrosophic number.
𝐽1,1 (𝑔𝐼1 ; 𝛽) = , 𝐽1,1 (𝑔𝐼2 ; 𝛽) = ,
6 6
5(4 + 𝑒)(1 − 𝛾) (4 + 𝑒)(5𝛾 + 3)
𝐽1,1 (𝑔𝐹1 ; 𝛾) = , 𝐽1,1 (𝑔𝐹2 ; 𝛾) = 7. Conclusion
8 8
1 In this research work, the concept of the Romberg integration
For ℎ = ,
2 method is developed for neutrosophic valued function. For the first
√
5(13 + 2𝑒 + 4 𝑒)𝛼 time, the Romberg integration method has been examined in a neu-
𝐽2,1 (𝑔𝑇1 ; 𝛼) = , trosophic environment. The Richardson extrapolation method has been
24
√ used in this numerical technique to find a better approximation of the
(13 + 2𝑒 + 4 𝑒)(6 − 5𝛼)
𝐽2,1 (𝑔𝑇2 ; 𝛼) = given function. Also, for the first time, the Richardson extrapolation
24
√ method is studied in a neutrosophic environment. In addition, from
5(13 + 2𝑒 + 4 𝑒)(1 − 𝛽)
𝐽2,1 (𝑔𝐼1 ; 𝛽) = , Theorem 3.1, it can be concluded that 𝐽 (𝑔𝑇𝑗 ; 𝛼), 𝐽 (𝑔𝐼𝑗 ; 𝛽), 𝐽 (𝑔𝐹𝑗 ; 𝛾)
24
√ converge uniformly to 𝐿(𝑔𝑇𝑗 ; 𝛼), 𝐿(𝑔𝐼𝑗 ; 𝛽) and 𝐿(𝑔𝐹𝑗 ; 𝛾) respectively.
(13 + 2𝑒 + 4 𝑒)(5𝛽 + 1) Additionally, it can be deduced from Theorem 3.2 that the integral
𝐽2,1 (𝑔𝐼2 ; 𝛽) =
24 form of the error 𝐸(𝑔) is a well-known result due to Peano. These
√
5(13 + 2𝑒 + 4 𝑒)(1 − 𝛾) conclusions demonstrate the usefulness of our suggested approach. An
𝐽2,1 (𝑔𝐹1 ; 𝛾) = ,
32 analysis of a numerical example along with a few convergence indica-
√
(13 + 2𝑒 + 4 𝑒)(5𝛾 + 3) tors demonstrates the applicability and effectiveness of the suggested
𝐽2,1 (𝑔𝐹2 ; 𝛾) = approach. From Tables 1 to 4, it can be inferred that absolute errors and
32
reference errors reduce as the value of (k, m) grows. In Figs. 1 to 6, this
Then we have,
√ √ finding is depicted graphically. Additionally, we have deduced from
5(9 + 𝑒 + 4 𝑒)𝛼 (9 + 𝑒 + 4 𝑒)(6 − 5𝛼) Figs. 7, 8, and 9 that the integration in Example 1 yields a triangular
𝐽2,2 (𝑔𝑇1 ; 𝛼) = , 𝐽2,2 (𝑔𝑇2 ; 𝛼) = ,
18 18 neutrosophic number if all the parameters are assumed to be triangular
√ √
5(9 + 𝑒 + 4 𝑒)(1 − 𝛽) (9 + 𝑒 + 4 𝑒)(5𝛽 + 1) neutrosophic numbers. Triangular neutrosophic numbers are not the
𝐽2,2 (𝑔𝐼1 ; 𝛽) = , 𝐽2,2 (𝑔𝐼2 ; 𝛽) = , only ones to which this result applies. Any neutrosophic number can
18 18
√ √ be included in its extension. This numerical example also demonstrates
(9 + 𝑒 + 4 𝑒)(1 − 𝛾) (9 + 𝑒 + 4 𝑒)(5𝛾 + 3)
𝐽2,2 (𝑔𝐹1 ; 𝛾) = , 𝐽2,2 (𝑔𝐹2 ; 𝛾) = the validity of the suggested approach. The researchers are encouraged
24 24
to continue working on the numerical integration for neutrosophic

6
S. Moi, S. Biswas and S.P. Sarkar Decision Analytics Journal 9 (2023) 100338

Fig. 1. Absolute error for left branch of truth membership function.

Fig. 2. Absolute error for right branch of truth membership function.

Fig. 3. Absolute error for left branch of indeterminacy membership function.

Table 3
Absolute errors for indeterminacy membership function.
𝛽 |𝐸11 (𝑔𝐼1 ; 𝛽)| |𝐸21 (𝑔𝐼1 ; 𝛽)| |𝐸22 (𝑔𝐼1 ; 𝛽)| |𝐸11 (𝑔𝐼2 ; 𝛽)| |𝐸21 (𝑔𝐼2 ; 𝛽)| |𝐸22 (𝑔𝐼2 ; 𝛽)|
0.4 3.08E−1 7.73E−2 5.79E−4 3.08E−1 7.73E−2 5.79E−4
0.6 2.05E−1 5.15E−2 3.86E−4 4.10E−1 1.03E−1 7.72E−4
0.8 1.02E−1 2.58E−2 1.93E−4 5.12E−1 1.29E−1 9.66E−4
1.0 0.00E−0 0.00E−0 0.00E−0 6.15E−1 1.55E−1 1.16E−3

7
S. Moi, S. Biswas and S.P. Sarkar Decision Analytics Journal 9 (2023) 100338

Table 4
Absolute errors for falsity membership function.
𝛾 |𝐸11 (𝑔𝐹1 ; 𝛾)| |𝐸21 (𝑔𝐹1 ; 𝛾)| |𝐸22 (𝑔𝐹1 ; 𝛾)| |𝐸11 (𝑔𝐹2 ; 𝛾)| |𝐸21 (𝑔𝐹2 ; 𝛾)| |𝐸22 (𝑔𝐹2 ; 𝛾)|
0.2 3.07E−1 7.73E−2 5.79E−4 3.07E−1 7.73E−2 5.79E−4
0.6 1.54E−1 3.86E−2 2.90E−4 4.61E−1 1.16E−1 8.69E−4
0.8 7.69E−2 1.93E−2 1.45E−4 5.38E−1 1.35E−1 1.01E−3
1.0 0.00E−0 0.00E−0 0.00E−0 6.15E−1 1.55E−1 1.16E−3

Fig. 4. Absolute error for right branch of indeterminacy membership function.

Fig. 5. Absolute error for left branch of falsity membership function.

Fig. 6. Absolute error for right branch of falsity membership function.

8
S. Moi, S. Biswas and S.P. Sarkar Decision Analytics Journal 9 (2023) 100338

Fig. 7. Graphical representation for truth membership function of g(x).

Fig. 8. Graphical representation for indeterminacy membership function of g(x).

Fig. 9. Graphical representation for falsity membership function of g(x).

valued functions by this research. In addition, Romberg integration has formula of the trapezoidal rule for integration. In an uncertain real-
some managerial implications. Romberg integration is an extrapolation world environment, many scientific and engineering problems such as

9
S. Moi, S. Biswas and S.P. Sarkar Decision Analytics Journal 9 (2023) 100338

Table 5 [14] Said Broumi, Irfan Deli, Florentin Smarandache, N-valued interval neutrosophic
Reference errors for truth, indeterminacy and falsity membership function. sets and their application in medical diagnosis, Collected Pap. Vol. XIV:
Error (𝑘, 𝑚) = (1, 1) (𝑘, 𝑚) = (2, 1) (𝑘, 𝑚) = (2, 2) Neutrosophics Topics (2022) 48.

𝑒𝑟𝑒𝑓 (𝑔 ; 0.4) 4.89E−1 1.27E−1 9.21E−4 [15] Mehboob Ali, Zahid Hussain, Miin-Shen Yang, Hausdorff distance and similarity
𝑘,𝑚 𝑇1
measures for single-valued neutrosophic sets with application in multi-criteria
𝑒𝑟𝑒𝑓 (𝑔 ; 0.4)
𝑘,𝑚 𝑇2
1.25E−0 5.24E−1 2.35E−3
decision making, Electronics 12 (1) (2022) 201.
𝑒𝑟𝑒𝑓 (𝑔 ; 0.6)
𝑘,𝑚 𝐼1
4.89E−0 1.23E−0 9.21E−4
𝑒𝑟𝑒𝑓 (𝑔 ; 0.6) 1.25E−0 3.15E−1 2.35E−3 [16] Minxia Luo, Guofeng Zhang, Lixian Wu, A novel distance between single valued
𝑘,𝑚 𝐼2
neutrosophic sets and its application in pattern recognition, Soft Comput. 26 (21)
𝑒𝑟𝑒𝑓 (𝑔 ; 0.8)
𝑘,𝑚 𝐹1
4.38E−0 1.10E−1 8.25E−4
(2022) 11129–11137.
𝑒𝑟𝑒𝑓 (𝑔 ; 0.8)
𝑘,𝑚 𝐹2
1.30E−0 3.28E−1 2.46E−3
[17] V. Antonysamy, M. Lellis Thivagar, S. Jafari, Abdulsattar Abdullah Hamad,
Neutrosophic sets in determining corona virus, Mater. Today: Proc. 49 (2022)
2654–2658.
[18] Sarah Qahtan, Aws Alaa Zaidan, Hassan Abdulsattar Ibrahim, Muhammet De-
diffraction problems, water waves, scattering in quantum mechanics,
veci, Weiping Ding, Dragan Pamucar, A decision modeling approach for smart
the Volterra population model, heat transfer, heat radiation, etc. are training environment with motor imagery-based brain computer interface under
determined by neutrosophic integral equations. In these real-world neutrosophic cubic fuzzy set, Expert Syst. Appl. 224 (2023) 119991.
problems, Romberg integration can be used to approximate the integral [19] Muhammet Deveci, Ilgin Gokasar, Dragan Pamucar, Aws Alaa Zaidan, Xin
and find the approximate solution of the equations. Wen, Brij B. Gupta, Evaluation of cooperative intelligent transportation system
scenarios for resilience in transportation using type-2 neutrosophic fuzzy VIKOR,
Transp. Res. Part A: Policy Pract. 172 (2023) 103666.
Declaration of competing interest
[20] Ilgin Gokasar, Vladimir Simic, Muhammet Deveci, Tapan Senapati, Alternative
prioritization of freeway incident management using autonomous vehicles in
The authors declare the following financial interests/personal rela- mixed traffic using a type-2 neutrosophic number based decision support system,
tionships which may be considered as potential competing interests: Eng. Appl. Artif. Intell. 123 (2023) 106183.
Mr. Sandip Moi reports financial support was provided by Council of [21] Ilgin Gokasar, Muhammet Deveci, Mehtap Isik, Tugrul Daim, Aws A. Zaidan,
Florentin Smarandache, Evaluation of the alternatives of introducing electric
Scientific and Industrial Research (CSIR), Government of India.
vehicles in developing countries using Type-2 neutrosophic numbers based RAFSI
model, Technol. Forecast. Soc. Change 192 (2023) 122589.
Data availability [22] Hui Ran, MABAC method for multiple attribute group decision making under
single-valued neutrosophic sets and applications to performance evaluation of
No data was used for the research described in the article. sustainable microfinance groups lending, PLoS One 18 (1) (2023) e0280239.
[23] Florentin Smarandache, Introduction to Neutrosophic Measure, Neutrosophic
Integral, and Neutrosophic Probability, Infinite Study, 2013.
Acknowledgments
[24] Florentin Smarandache, Neutrosophic Precalculus and Neutrosophic Calculus:
Neutrosophic Applications, Infinite Study, 2015.
Sandip Moi thanks the Council of Scientific and Industrial Research
[25] Florentin Smarandache, Huda E. Khalid, Neutrosophic Precalculus and
(CSIR), Government of India (File No.- 08/003(0135)/2019-EMR-I) for Neutrosophic Calculus, Infinite Study, 2018.
financial support of this research work. [26] Malath F. Alaswad, A study of the integration of neutrosophic thick function,
Int. J. Neutrosophic Sci. 6 (2) (2020) 97–106.
References [27] Nguyen Thi Kim Son, Nguyen Phuong Dong, Hoang Viet Long, Alireza Khastan,
et al., Linear quadratic regulator problem governed by granular neutrosophic
[1] Florentin Smarandache, Neutrosophic set-a generalization of the intuitionistic fractional differential equations, ISA Trans. 97 (2020) 296–316.
fuzzy set, in: 2006 IEEE International Conference on Granular Computing, IEEE, [28] Yaser Ahmad Alhasan, The neutrosophic integrals and integration methods,
2006, pp. 38–42. Neutrosophic Sets Syst. 43 (2021) 290–301.
[2] Said Broumi, Assia Bakali, Mohamed Talea, Florentin Smarandache, Prem Kumar [29] Yaser Ahmad Alhasan, The neutrosophic integrals by parts, Neutrosophic Sets
Singh, Vakkas Uluçay, Mohsin Khan, Bipolar complex neutrosophic sets and its Syst. 45 (2021) 306–319.
application in decision making problem, Fuzzy Multi-Criteria Decis.-Mak. Using [30] Suvankar Biswas, Sandip Moi, Smita Pal Sarkar, Neutrosophic Riemann
Neutrosophic Sets (2019) 677–710. integration and its properties, Soft Comput. (2021) 1–13.
[3] Gia Nhu Nguyen, Le Hoang Son, Amira S. Ashour, Nilanjan Dey, A survey of [31] Yaser Ahmad Alhasan, The neutrosophic integrals by partial fraction,
the state-of-the-arts on neutrosophic sets in biomedical diagnoses, Int. J. Mach. Neutrosophic Sets Syst. 49 (2022) 438–457.
Learn. Cybern. 10 (2019) 1–13.
[32] Suvankar Biswas, Sandip Moi, Smita Pal Sarkar, Numerical integration of
[4] Rama Mallick, Surapati Pramanik, Pentapartitioned Neutrosophic Set and Its
neutrosophic valued function by Gaussian quadrature methods, Arabian J. Math.
Properties, Vol. 36, Infinite Study, 2020.
11 (2) (2022) 189–211.
[5] Luu Quoc Dat, Nguyen Tho Thong, Mumtaz Ali, Florentin Smarandache,
Mohamed Abdel-Basset, Hoang Viet Long, et al., Linguistic approaches to [33] Ahmad Salama, Rasha Dalla, Malath Al Aswad, Rozina Ali, On some results about
interval complex neutrosophic sets in decision making, IEEE Access 7 (2019) the second order neutrosophic differential equations by using neutrosophic thick
38902–38917. function, J. Neutrosophic Fuzzy Syst. (JNFS) Vol. 4 (01) (2022) 30–40.
[6] Jayanta Pratihar, Ranjan Kumar, Arindam Dey, Said Broumi, Transportation [34] Sandip Moi, Suvankar Biswas, Smita Pal(Sarkar), Second-order neutrosophic
problem in neutrosophic environment, in: Neutrosophic Graph Theory and boundary-value problem, Complex Intell. Syst. 7 (2) (2021) 1079–1098.
Algorithms, IGI Global, 2020, pp. 180–212. [35] Ashish Acharya, Animesh Mahata, Supriya Mukherjee, Manajat Ali Biswas, Kr-
[7] Lin Lu, Xiaochun Luo, Emergency transportation problem based on single-valued ishna Pada Das, Sankar Prasad Mondal, Banamali Roy, A neutrosophic differential
neutrosophic set, Discrete Dyn. Nat. Soc. 2020 (2020) 1–8. equation approach for modeling glucose distribution in the bloodstream using
[8] İrfan Deli, Linear optimization method on single valued neutrosophic set and its neutrosophic sets, Decis. Anal. J. (2023) 100264.
sensitivity analysis, TWMS J. Appl. Eng. Math. 10 (1) (2020) 128–137. [36] Sandip Moi, Suvankar Biswas, Smita Pal Sarkar, An efficient method for solving
[9] Riad K. Al-Hamido, A new approach of neutrosophic topological space, Int. J. neutrosophic Fredholm integral equations of second kind, Granular Comput. 8
Neutrosophic Sci. 7 (2020) 55–61. (1) (2023) 1–22.
[10] Muhammad Ihsan, Atiqe Ur Rahman, Muhammad Saeed, Single valued neutro- [37] Tofigh Allahviranloo, Romberg integration for fuzzy functions, Appl. Math.
sophic hypersoft expert set with application in decision making, Neutrosophic Comput. 168 (2) (2005) 866–876.
Sets Syst. 47 (2021) 451–471.
[38] D.B. Hunter, The numerical evaluation of Cauchy principal values of integrals
[11] Atiqe Ur Rahman, Muhammad Saeed, Alok Dhital, Decision Making Application
by Romberg integration, Numer. Math. 21 (3) (1973) 185–192.
Based on Neutrosophic Parameterized Hypersoft Set Theory, Infinite Study, 2021.
[12] Ankan Bhaumik, Sankar Kumar Roy, Gerhard Wilhelm Weber, Multi-objective [39] Tofigh Allahviranloo, Newton Cot’s methods for integration of fuzzy functions,
linguistic-neutrosophic matrix game and its applications to tourism management, Appl. Math. Comput. 166 (2) (2005) 339–348.
J. Dyn. Games 8 (2) (2021). [40] Rıdvan Şahin, Ahmet Küçük, Subsethood measure for single valued neutrosophic
[13] Muhammad Saeed, Muhammad Ahsan, Muhammad Haris Saeed, Asad Mehmood, sets, J. Intell. Fuzzy Systems 29 (2) (2015) 525–530.
Thabet Abdeljawad, An application of neutrosophic hypersoft mapping to di- [41] Said Broumi, Florentin Smarandache, Extended hausdorff distance and similarity
agnose hepatitis and propose appropriate treatment, IEEE Access 9 (2021) measures for neutrosophic refined sets and their application in medical diagnosis,
70455–70471. J. New Theory 7 (2015) 64–78.

10
S. Moi, S. Biswas and S.P. Sarkar Decision Analytics Journal 9 (2023) 100338

[42] J. Bulirsh Stoer, R. Bulirsch, R. 1980, Introduction to Numerical Analysis, [44] Hsien-Chung Wu, The fuzzy Riemann-Stieltjes integral, Int. J. Uncertain.
[43] S. Abbasbandy, T. Allah Viranloo, Extrapolation method for improving the Fuzziness Knowl.-Based Syst. 6 (01) (1998) 51–67.
solution of fuzzy initial value problems, Math. Comput. Appl. 9 (2) (2004) [45] Francielle Santo Pedro, Estevao Esmi, Laecio Carvalho de Barros, Calculus for
205–214. linearly correlated fuzzy function using Fréchet derivative and Riemann integral,
Inform. Sci. 512 (2020) 219–237.

11