Malaysian Journal of Mathematical Sciences 17(3): 305–332 (2023)

https://doi.org/10.47836/mjms.17.3.05

Malaysian Journal of Mathematical Sciences

Journal homepage: https://mjms.upm.edu.my

Variable-order Implicit Fractional Differential Equations based on the

Kuratowski MNC Technique

Bouazza, Z.1 , Souid, M. S.2 , Hussin, C. H. C.∗3 , Mandangan, A.4 , and Sabit, S.5
1
Department of Informatics, University of Tiaret, Algeria
2
Department of Economic Sciences, University of Tiaret, Algeria
3
Preparatory Centre for Science and Technology, Universiti Malaysia Sabah, Malaysia
4
Faculty of Science and Natural Resources, Universiti Malaysia Sabah, Malaysia
5
Laboratoire Matériaux et Structures, Department of Mathematics, University of Tiaret, Algeria

E-mail: [email protected]
∗
Corresponding author
Received: 11 February 2023
Accepted: 30 May 2023

Abstract

In this manuscript, we examine the existence and the stability of solutions to the boundary value
problem of Riemann-Liouville fractional differential equations of variable order. The obtained
new results are based on the fixed point theorem of Darbo and Kuratowski’s metric of non-
compactness (MNK) with the help of piece-wise constant functions. In addition, the derived
fundamental results are proven suitable because they satisfy the Ulam-Hyers Rassias stability
sufficient conditions. Several numerical examples were discussed too to demonstrate the rea-
sonableness and effectiveness of the observed results.
Keywords: fractional differential equations of variable order; boundary value problem; Darbo’s
fixed point theorem; measure of non-compactness; Ulam Hyers Rassias stability;
Green function.
Bouazza et al. Malaysian J. Math. Sci. 17(3): 305–332(2023) 305 - 332

1 Introduction

Fractional calculus is a branch of mathematical analysis that studies the possibility of taking
real number powers or complex number powers of the differentiation and integration operators
[9]. In the last thirty years, fractional calculus has contributed to a multitude of significant discov-
eries in pure and applied mathematics and various domains,such as chemistry [9], physics [16],
biology [1], control theory [17], economics [12], biophysics [10], signal [24] and image process-
ing, etc [8, 7].

The study of fractional differential equations has attracted the attention of several researchers,
resulting in the publication of numerous papers that address various in this field. For example,
Baleanu et al. [2] used the Caputo and Fabrizio fractional derivative to express the model of
HIV and solve the equation. Ahmad et al. [17] presented the existence of solutions for nonlinear
neutral stochastic fractional differential systems. Shah et al. [23] obtained sufficient conditions for
the existence of solutions to the coupled system of nonlinear boundary value problems. Tuan et al.
[25] presented a mathematical model for the transmission of COVID-19 by the Caputo fractional-
order derivative.

The existence of solutions to variable-order problems is rarely discussed in the literature. Souid
et al. [6, 18] presented the existence, uniqueness and stability of solutions to many different prob-
lems (implicit, resonance). Stability theory is very significant, as each praticable control system is
designed to be stable. The analysis of solution stability has captivated many researchers due to its
promising potential (see [14, 22]).

Motivated by the previous research, we deal with the following boundary value problems
(BVP):  φ(s)  
 D0+ ξ(s) + h s, ξ(s), D0φ(s)
+ ξ(s) = 0, s ∈ Υ := [0, K],
(1)

ξ(0) = 0, ξ(K) = 0.
For 1 < φ(s) ≤ 2, the function h : Υ × Λ × Λ → Λ is continuous, Λ is a real (or complex) Banach
φ(s)
space and D0+ is the Riemann-Liouville fractional derivative of order variable φ(s).

2 Auxiliary Notions

Let the Banach space C(Υ, Λ) the set of the real-valued continuous functions defined on the
intervale Υ, equipped with the usual norm:

∥κ∥ = Sup{∥κ(s)∥ : s ∈ Υ}.

Definition 2.1. ([20, 21]) For −∞ < τ1 < τ2 < +∞ and φ(s) : [τ1 , τ2 ] → (0, +∞), the Riemann
Liouville fractional integral FI(RL) of variable order φ(s) for function υ is defined by:
s
(s − α)φ(s)−1
Z
φ(s)
Iτ + υ(s) =
υ α) dα, s > τ1 , (2)
1 τ1 Γ φ(s)

where Γ is the gamma function.

306
Bouazza et al. Malaysian J. Math. Sci. 17(3): 305–332(2023) 305 - 332

Definition 2.2. ([20, 21, 26]) For −∞ < τ1 < τ2 < +∞ and φ(s) : [τ1 , τ2 ] → (n − 1, n), n ∈ N, the
Riemann Liouville fractional derivative FD(RL) of variable-order φ(s) for function υ is defined by:
 d n  d n Z s (s − α)n−φ(s)−1
φ(s) n−φ(s)
Dτ + υ(s) = Iτ + υ(s) =
υ(α) dα, s > τ1 . (3)
1 ds 1 ds τ1 Γ n − φ(s)

As expected, FI(RL) and FD(RL) correspond with the usual Riemann Liouville fractional in-
tegral and Riemann Liouville derivative: [20, 15]. Consider the following essential observations.
Lemma 2.1. ([15]) Suppose ϱ, ϖ > 0, τ1 > 0 and υ ∈ L(τ1 , τ2 ), Dτϱ+ υ ∈ L(τ1 , τ2 ). Then, the differential
1
equation
ϱ
Dτ + υ = 0,
1

has solution
ι=n
X
υ(s) = ηι (s − τ1 )ϱ−ι ,
ι=1

and
ι=n
X
Iτϱ+ Dτϱ+ υ(s) = υ(s) + ηι (s − τ1 )ϱ−ι ,
1 1
ι=1

where ηι ∈ R, ι = 1, 2, ..., n with n − 1 < ϱ ≤ n. Furthermore,

Dτϱ+ Iτϱ+ υ(s) = υ(s),

1 1

and
Iτϱ+ Iτϖ+ υ(s) = Iτϖ+ Iτϱ+ υ(s) = Iτϱ+ϖ
+ υ(s).
1 1 1 1 1

Remark([29, 27, 32]) Note that the general functions ϱ(s) and ϖ(s) do not satisfy the semigroup
condition, i.e.,
ϱ(s) ϖ(s) ϱ(s)+ϖ(s)
Iτ + Iτ + υ(s) ̸= Iτ + υ(s).
1 1 1

Example 2.1: Let


s  1, s ∈ [0, 1],
ϱ(s) = , s ∈ [0, 3], ϖ(s) = υ(s) = 4, s ∈ [0, 3].
2  3, s ∈]1, 3],

Thus,
s α
(s − α)ϱ(s)−1 (α − σ)ϖ(α)−1
Z Z
ϱ(s) ϖ(s)
I0+ I0+ υ(s) =

υ(σ) dσ dα,
0 Γ ϱ(s) 0 Γ ϖ(α)
s Z 1 Z α
(s − α)ϱ(s)−1 (α − σ)0 (α − σ)2
Z 
=
4 dσ + 4 dσ dα,
0 Γ ϱ(s) 0 Γ(1) 1 Γ(3)
s
(s − α)ϱ(s)−1
Z  
2
=
4 + (α − 1)3 dα,
0 Γ ϱ(s) 3

307
Bouazza et al. Malaysian J. Math. Sci. 17(3): 305–332(2023) 305 - 332

and
s
(s − α)ϱ(s)+ϖ(s)−1
Z
ϱ(s)+ϖ(s)
I0+ υ(s) =
υ(α) dα.
0 Γ ϱ(s) + ϖ(s)

Consequently, we get
2
(2 − α)0
Z  
ϱ(s) ϖ(s) 2
I0+ I0+ υ(s)|s=2 = 4 + (α − 1)3 dα,
0 Γ(1) 3
Z 2 
2 3 10
= α − 2α2 + 2α + dα,
0 3 3
= 8.

2
(2 − α)ϱ(s)+ϖ(s)−1
Z
ϱ(s)+ϖ(s)
I0+ υ(s)|s=2 =
υ(α) dα,
0 Γ ϱ(s) + ϖ(s)
Z 1 Z 2
(2 − α)1 (2 − α)3
= 4 dα + 4 dα,
0 Γ(2) 1 Γ(4)
Z 1
2 2
Z
=4 (−α + 2) dα + (−α3 + 6α2 − 12α + 8) dα,
0 3 1
1 37
=6+ = .
6 6
Therefore, we obtain
ϱ(s) ϖ(s) ϱ(s)+ϖ(s)
I0+ I0+ υ(s)|s=2 ̸= I0+ υ(s)|s=2 .
Lemma 2.2. ([31]) Let φ ∈ C(Υ, (1, 2]) and 0 ≤ δ ≤ mins∈Υ |φ(s)|, then for any υ ∈ Cδ (Υ, Λ) where
φ(s)
Cδ (Υ, Λ) = {υ(s) ∈ C(Υ, Λ), sδ υ(s) ∈ C(Υ, Λ)}, the integral I0+ υ(s) exists for any s ∈ Υ.
φ(s)
Lemma 2.3. ([31]) If φ ∈ C(Υ, (1, 2]), then I0+ υ(s) ∈ C(Υ, Λ) for any υ ∈ C(Υ, Λ).
Definition 2.3. ([30, 13, 28] ) Let I ⊂ R;

1. The generalised interval I is either an interval or {a} or ∅.

2. A partition of I is a finite set P, such that, for any x in I lies in exactly one of the generalised intervals
E in P.
3. Let P be a partition of I, we say that the function h : I → R is piecewise constant with respect to P
if for each E in P, h admits constant values on E.

2.1 The technique of measures of non-compactness

Definition 2.4. ([3]) Let ΩΛ the bounded subsets of a Banach space Λ. The (MNK) is a mapping
ϑ : ΩΛ → [0, ∞] that follows the format below,
ϑ(∆) = inf{ϵ > 0 : ∆(∈ ΩΛ ) ⊆ ∪nι=1 ∆ι , diam(∆ι ) ≤ ϵ},
where
diam(∆ι ) = sup{||x − y|| : x, y ∈ ∆ι }.

308
Bouazza et al. Malaysian J. Math. Sci. 17(3): 305–332(2023) 305 - 332

Proposition 2.1. ([3, 4]) Let ∆, ∆1 , ∆2 be a bounded subsets of Λ, then:

1. ∆ is relatively compact⇐⇒ ϑ(∆) = 0.

2. ϑ(∅) = 0.
3. ϑ(∆) = ϑ(∆) = ϑ(conv∆).
4. ∆1 ⊂ ∆2 =⇒ ϑ(∆1 ) ≤ ϑ(∆2 ).
5. ϑ(∆1 + ∆2 ) ≤ ϑ(∆1 ) + ϑ(∆2 ).
6. ϑ(Π∆) = |Π|ϑ(∆), Π ∈ R.
7. ϑ(∆1 ∪ ∆2 ) = M ax{ϑ(∆i ), i = 1, 2}.
8. ϑ(∆1 ∩ ∆2 ) = M in{ϑ(∆i ), i = 1, 2}.
9. ϑ(∆ + x0 ) = ϑ(∆) for any x0 ∈ Λ.
Lemma 2.4. ([11]) Let Λ be a Banach space. If χ is an bounded and equicontinuous subset of the space
C(Υ, Λ) of continuous functions, then;

(i) ϑ(χ(s)) is a continuous function for s ∈ Υ, and

ϑ(χ)
b = sup ϑ(χ(s)).
s∈Υ

!
Z K Z K

(ii) ϑ x(θ) dθ : x ∈ χ ≤ ϑ χ(θ) dθ,
0 0

where
χ(α) = {x(α) : x ∈ χ}, α ∈ Υ.
Theorem 2.1. (DFPTh) [3] Let Υ be a nonempty, closed, bounded, and convex subset of a Banach space
Λ, and assume that 𭟋 : Υ −→ Υ is a continuous operator fulfilling:
ϑ(𭟋(S)) ≤ lϑ(S), for any S(̸= ∅) ⊂ Υ, l ∈ [0, 1),
i.e., 𭟋 is l-set contractions. Therefore, 𭟋 has at least one fixed point in Υ.
Definition 2.5. ([19]) The equation (1) is Ulam-Hyers stable if ch > 0 exists, such that for any ϵ > 0
and for every solution z ∈ C(Υ, Λ), the following inequality holds.
 
φ(s) φ(s)
D0+ z(s) + h s, z(s), D0+ z(s) ≤ ϵ, s ∈ Υ.

There is a solution ξ ∈ C(Υ, Λ) of equation (1) with:

∥z(s) − ξ(s)∥ ≤ ch ϵ, s ∈ Υ.
Definition 2.6. ([19]) The equation (1) is (HUR) stable with respect to ψ if ch > 0 exists such that for
each ϵ > 0 and every solution z ∈ C(Υ, Λ) of the following inequality:
 
φ(s) φ(s)
D0+ z(s) + h s, z(s), D0+ z(s) ≤ ϵψ(s), s ∈ Υ,

(1) has a solution ξ ∈ C(Υ, Λ) satisfying

∥z(s) − ξ(s)∥ ≤ ch ϵψ(s), s ∈ Υ.

309
Bouazza et al. Malaysian J. Math. Sci. 17(3): 305–332(2023) 305 - 332

3 The Existence of Solutions for Boundary Value Problem

Let’s introduce these presumptions:

(H1) Let n ∈ N and β = {Υ1 := [0, K1 ], Υ2 := (K1 , K2 ], Υ3 := (K2 , K3 ], ...Υn := (Kn−1 , K]}
represent a partition of the interval Υ, and let φ(t) : Υ → (1, 2] be a piecewise constant
function with respect to β, i.e.,


 φ1 , if s ∈ Υ1 ,



φ2 , if s ∈ Υ2 ,


n


X 
φ(s) = φι Iι (s) = .

ι=1
 .



.




φn , if s ∈ Υn ,


where 1 < φι ≤ 2 are constants and Iι indicates the interval for

Υι := (Kι−1 , Kι ], ι = 1, 2, ..., n, (with K0 = 0, Kn = K) such that,

1, for s ∈ Υι ,
Iι (s) =
0, for elsewhere.

(H2) Let sδ h : Υ × Λ × Λ → Λ be a continuous function (0 ≤ δ ≤ mins∈Υ |(φ(s))|), there exist a

constants γ1 , γ2 > 0 such that

sδ h(s, π1 , ς1 ) − h(s, π2 , ς2 ) ≤ γ1 ∥π1 − π2 ∥ + γ2 ∥ς1 − ς2 ∥,

for any π1 , π2 , ς1 , ς2 ∈ Λ.

Remark 3.1. According to the previous observation of [5], we can easily show that the condition (H2) and
the following inequality
 
ϑ tδ h(s, D1 , D2 ) ≤ γ1 ϑ(D1 ) + γ2 ϑ(D2 ),

are equivalent for any bounded sets D1 , D2 ⊂ Λ and every s ∈ Υ.

Furthermore, let’s define φ : Υ → Λ for a given set of functions χ indicated by,

n o
χ(s) = x(s), x ∈ χ , s ∈ Υ,

and
n o
χ(Υ) = χ(s) : x ∈ χ, s∈Υ .

Now, using the concepts of (MNK) and (DFPTh), we can show that there is a solution to (BVP)(1).

The notation Ξι = C(Υι , Λ) signifies the Banach space of continuous functions ξ : Υι → Λ for
each ι ∈ {1, 2, ..., n}, equipped with,

∥ξ∥Ξι = sup ∥ξ(s)∥.

s∈Υι

310
Bouazza et al. Malaysian J. Math. Sci. 17(3): 305–332(2023) 305 - 332

We begin by analysing BVP as defined in (1).

Because of (3), we can express the equation of the BVP(1) as follows:

Z s
d2 (s − α)1−φ(s) 
φ(s)


ξ(α) dα + h s, ξ(s), D0+ ξ(s) = 0, s ∈ Υ. (4)
ds2 0 Γ 2 − φ(s)
The equation (4) in Υι , where ι = 1, 2, ..., n may be presented as follows while accounting for
(H1),
Z K1 Z s !
d2 (s − α)1−φ1 (s − α)1−φι 
φι

ξ(α) dα+...+ ξ(α) dα +h s, ξ(s), D0 + ξ(s) = 0, s ∈ Υι .
ds2 0 Γ(2 − φ1 ) Kι−1 Γ(2 − φι )
(5)
Definition 3.1. BVP (1) has a solution, if there exists functions ξι , ι = 1, 2, ..., n, such that,
ξι ∈ C([0, Kι ], Λ) satisfying equation (5) and ξι (0) = 0 = ξι (Kι ).

If we consider ξ(s) ≡ 0 for 0 ≤ s ≤ Kι−1 , then (5) is expressed as,

 
φι φι
DK + ξ(s) + h s, ξ(s), DK+
ξ(s) = 0, s ∈ Υι .
ι−1 ι−1

We will address the following BVPs:

  
 Dφι+ ξ(s) + h s, ξ(s), Dφι+ ξ(s) = 0,

s ∈ Υι ,
Kι−1 Kι−1 (6)

 ξ(K ) = 0,
ι−1 ξ(Kι ) = 0.

Lemma 3.1. The solution ξ of (6) is written as the following integral equation
!
Z Kι Z Kι
ξ(s) = Gι (s, α)h α, Gι (s, σ)ξ(σ)dσ, ξ(α) dα, (7)
Kι−1 Kι−1

φι
where DK + ξ(s) = ξ(s), Gι (s, α) is the Green’s function, defined as follows:
ι−1

 " #
1
(Kι − Kι−1 )1−φι (s − Kι−1 )φι −1 (Kι − α)φι −1 − (s − α)φι −1 ,



Γ(φι )






Kι−1 ≤ α ≤ s ≤ Kι ,


Gι (s, α) =
 1
(Kι − Kι−1 )1−φι (s − Kι−1 )φι −1 (Kι − α)φι −1 ,



Γ(φ )



 ι


Kι−1 ≤ s ≤ α ≤ Kι ,


where ι = 1, 2, ..., n.

φι
Proof. We assume that ξ ∈ Ξι is the solution to the problem (6), and we take DK + ξ(s) = ξ(s).
ι−1
φι
Applying the operator IK + to both sides of (6), we find (see Lemma(2.1)),
ι−1

φι
ξ(s) = ω1 (s − Kι−1 )φι −1 + ω2 (s − Kι−1 )φι −2 − IK + ξ(s), s ∈ Υι .
ι−1

Due to the assumption of function h together with ξ(Kι−1 ) = 0, we conclude that ω2 = 0.

311
Bouazza et al. Malaysian J. Math. Sci. 17(3): 305–332(2023) 305 - 332

Let ξ(s) satisfies ξ(Kι ) = 0. We can observe that

φι
ω1 = (Kι − Kι−1 )1−φι IK + ξ(Kι ).
ι−1

Then, we observe
φι φι
ξ(s) = (Kι − Kι−1 )1−φι (s − Kι−1 )φι −1 IK + ξ(Kι ) − I + ξ(s),
K
s ∈ Υι .
ι−1 ι−1

Then the solution to problem (6) is given by:

Z Kι
1
ξ(s) = (Kι − Kι−1 )1−φι (s − Kι−1 )φι −1 (Kι − α)φι −1 ξ(α) dα
Γ(φι ) Kι−1
Z s
1
− (s − α)φι −1 ξ(α) dα,
Γ(φι ) Kι−1
"Z
s
1  
= (Kι − Kι−1 )1−φι (s − Kι−1 )φι −1 (Kι − α)φι −1 − (s − α)φι −1 ξ(α) dα
Γ(φι ) Kι−1
Z Kι #
1−φι φι −1 φι −1
+ (Kι − Kι−1 ) (s − Kι−1 ) (Kι − α) ξ(α) dα .
s

The continuation of the function of the Green suggests that:

!
Z Kι Z Kι
ξ(s) = Gι (s, α)h α, Gι (s, σ)ξ(σ)dσ, ξ(α) dα.
Kι−1 Kι−1

Let ξ ∈ Ξι be the solution of the integral equation (7). Based on the continuity of the function
sδ h and the Lemma(2.1), we can conclude that ξ is the solution to the problem (6).

The next proposition will be necessary.

Proposition 3.1. ([30]) Assume that sδ h : Υ × Λ × Λ → Λ, (0 ≤ δ ≤ mins∈Υ |(φ(s))|) is a function
that is continuous, φ(s) : Υ → (1, 2] satisfies (H1). Then, the Green functions of the boundary value
problem (6) satisfy the following characteristics:

(1) Gι (s, α) ≥ 0 ∀Kι−1 ≤ s, α ≤ Kι .

(2) max Gι (s, α) = Gι (α, α), α ∈ Υι .
s∈Υι

(3) Gι (α, α) has one unique maximum given by:

!φι −1
1 Kι − Kι−1
max Gι (α, α) = , for ι = 1, 2, ..., n.
α∈Υι Γ(φι ) 4

Theorem (2.1) is the basis for our first existence result.

Theorem 3.1. Let the conditions (H1)-(H2) to be satisfied, and if
 
1−δ
(Kι − Kι−1 )φι −1 Kι1−δ − Kι−1
!
γ1 (Kι − Kι−1 )φι
+ γ2 < 1, (8)
(1 − δ)4φι −1 Γ(φι ) 4φι −1 Γ(φι )

the problem (6) has thus at least one solution on Υ.

312
Bouazza et al. Malaysian J. Math. Sci. 17(3): 305–332(2023) 305 - 332

Proof. We establish the operator

S : Ξι → Ξι ,

following:
Z Kι
Sξ(s) = Gι (s, α)ξ(α) dα, s ∈ Υι , (9)
Kι−1

where
!
Z Kι
ξ(α) = h α, Gι (s, σ)ξ(σ) dσ, ξ(α) .
Kι−1

The operator S defined in (9) is well defined from the continuity of function sδ h and the properties
of fractional integrals.

Let,
h⋆ (Kι − Kι−1 )φι
4φι −1 Γ(φι )
Rι ≥   !,
1−δ
(Kι − Kι−1 )φι −1 Kι1−δ − Kι−1 γ1 (Kι − Kι−1 ) φι
1− + γ2
(1 − δ)4φι −1 Γ(φι ) 4φι −1 Γ(φι )

with

h⋆ = sup ∥h(s, 0, 0)∥.

s∈Υι

We evaluate the set

n o
BRι = ξ ∈ Ξι , ∥ξ∥Ξι ≤ Rι ,

where BRι is bounded, convex, closed and nonempty.

Now, we will demonstrate the satisfaction of Theorem (2.1) in four steps as outlined below.

STEP 1: S(BRι ) ⊆ (BRι ).

By Proposition (3.1) and (H2), we have

Z Kι
∥Sξ(s)∥ = Gι (s, α)ξ(α) dα ,
Kι−1
Z Kι
≤ Gι (s, α)∥ξ(α)∥ dα,
Kι−1
!φι −1 Z
Kι
1 Kι − Kι−1
≤ ∥ξ(α)∥ dα,
Γ(φι ) 4 Kι−1
!φι −1 Z !
Kι Kι
Kι − Kι−1
Z
1
≤ h α, Gι (s, σ)ξ(σ) dσ, ξ(α) dα,
Γ(φι ) 4 Kι−1 Kι−1

313
Bouazza et al. Malaysian J. Math. Sci. 17(3): 305–332(2023) 305 - 332

!φι −1 Z !
Kι Kι
Kι − Kι−1
Z
1
≤ h α, Gι (s, σ)ξ(σ) dσ, y(α) − h(α, 0, 0) dα
Γ(φι ) 4 Kι−1 Kι−1
!φι −1 Z
Kι
1 Kι − Kι−1
+ ∥h(α, 0, 0)∥ dα,
Γ(φι ) 4 Kι−1
!φι −1 Z !
Kι Z Kι
1 Kι − Kι−1 −δ
≤ α γ1 Gι (s, σ)ξ(σ) dσ + γ2 ∥ξ(α)∥ dα
Γ(φι ) 4 Kι−1 Kι−1

h⋆ (Kι − Kι−1 )φι

+ ,
4φι −1 Γ(φι )
!φι −1 Z !
Kι φι
1 Kι − Kι−1 γ 1 (K ι − K ι−1 )
≤ α−δ ∥ξ∥Ξι + γ2 ∥ξ∥Ξι dα
Γ(φι ) 4 Kι−1 4φι −1 Γ(φι )
h⋆ (Kι − Kι−1 )φι
+ ,
4φι −1 Γ(φι )
 
1−δ
(Kι − Kι−1 )φι −1 Kι1−δ − Kι−1
!
γ1 (Kι − Kι−1 )φι h⋆ (Kι − Kι−1 )φι
≤ + γ2 Rι + ,
(1 − δ)4φι −1 Γ(φι ) 4 φι −1 Γ(φι ) 4φι −1 Γ(φι )
≤ Rι .

It indicates S(BRι ) ⊆ BRι .

STEP 2: S is continuous.

The sequence (ξn ) is assumed to converge to ξ in Ξι and s ∈ Υι . Then,

Z Kι
∥(Sξn )(s) − (Sξ)(s)∥ ≤ Gι (s, α) ξn (α) − ξ(α) dα,
Kι−1
!φι −1 Z
Kι
1 Kι − Kι−1
≤ ξn (α) − ξ(α) dα,
Γ(φι ) 4 Kι−1
!φι −1 Z !
Kι Z Kι
1 Kι − Kι−1
≤ h α, Gι (s, σ)ξn (σ) dσ, ξn (α)
Γ(φι ) 4 Kι−1 Kι−1
Z Kι !
− h α, Gι (s, σ)y(σ) dσ, ξ(α) dα,
Kι−1
!φι −1 Z
Kι Z Kι
1 Kι − Kι−1
≤ α−δ γ1 Gι (s, σ)ξn (σ) dσ
Γ(φι ) 4 Kι−1 Kι−1
Z Kι !
− Gι (s, σ)ξ(σ) dσ + γ2 ∥ξn (α) − ξ(α)∥ dα,
Kι−1
!φι −1 Z
Kι Kι
Kι − Kι−1
Z
1 −δ
≤ α γ1 Gι (s, σ) ξn (σ) − ξ(σ) dσ
Γ(φι ) 4 Kι−1 Kι−1
!
+ γ2 ∥ξn (α) − ξ(α)∥ dα,

314
Bouazza et al. Malaysian J. Math. Sci. 17(3): 305–332(2023) 305 - 332

!φι −1 Z
Kι
1 Kι − Kι−1 γ1 (Kι − Kι−1 )φι
≤ α−δ ∥ξn − ξ∥Ξι
Γ(φι ) 4 Kι−1 4φι −1 Γ(φι )
!
+ γ2 ∥ξn − ξ∥Ξι dα,
 
1−δ
(Kι − Kι−1 )φι −1 Kι1−δ − Kι−1
!
γ1 (Kι − Kι−1 )φι
≤ + γ2 ∥ξn − ξ∥Ξι ,
(1 − δ)4φι −1 Γ(φι ) 4φι −1 Γ(φι )
i.e., we get
∥(Sξn ) − (Sξ)∥Ξι → 0 as n → ∞.
Therefore, S is a continuous operator.

STEP 3: S is bounded and equicontinuous.

Step 2 yields S(BRι ) = {S(ξ) : ξ ∈ BRι } ⊂ BRι . Consequently, for each ξ ∈ BRι , ∥S(ξ)∥Ξι ≤ Rι
holds, showing that S(BRι ) is bounded. It remains essential to establish that S(BRι ) is equicon-
tinuous.

For s1 , s2 ∈ Υι , s1 < s2 and ξ ∈ BRι , we obtain

Z Kι Z Kι
∥(Sξ)(s2 ) − (Sξ)(s1 )∥ = Gι (s2 , α)ξ(α) dα − Gι (s1 , α)ξ(α) dα ,
Kι−1 Kι−1
Z Kι  
≤ Gι (s2 , α) − Gι (t1 , α) ξ(α) dα,
Kι−1
!
Z Kι Z Kι
≤ ∥Gι (s2 , α) − Gι (s1 , α)∥ h α, Gι (s, σ)ξ(σ) dσ, ξ(α) dα,
Kι−1 Kι−1
!
Z Kι Z Kι
≤ Gι (s2 , α) − Gι (s1 , α) h α, Gι (s, σ)ξ(σ) dσ, ξ(α)
Kι−1 Kι−1
!
− h(α, 0, 0) + ∥h(α, 0, 0)∥ dα,
"
Z Kι Z Kι
≤ Gι (s2 , α) − Gι (s1 , α) α−δ γ1 Gι (s, σ)ξ(σ) dσ
Kι−1 Kι−1
! #
⋆
+ γ2 ∥ξ(α)∥ +h dα,
" !
Kι
γ1 (Kι − Kι−1 )φι
Z
−δ
≤ Gι (s2 , α) − Gι (s1 , α) α ∥ξ∥Ξι + γ2 ∥ξ∥Ξι
Kι−1 4φι −1 Γ(φι )
#
⋆
+h dα,
! Z Kι
γ1 (Kι − Kι−1 )φι (Kι−1 )−δ −δ
≤ + γ2 (Kι−1 ) ∥ξ∥Ξι Gι (s2 , α)
4φι −1 Γ(φι ) Kι−1
Z Kι
⋆
− Gι (s1 , α) dα + h Gι (s2 , α) − Gι (s1 , α) dα,
Kι−1

315
Bouazza et al. Malaysian J. Math. Sci. 17(3): 305–332(2023) 305 - 332

using the continuity of Green’s function. Hence ∥(Sξ)(s2 ) − (Sξ)(s1 )∥Eι → 0 as |s2 − s1 | → 0. It
implies that S(BRι ) is equicontinuous.

STEP 4: S is l-set contractions.

If χ ∈ BRι , s ∈ Υι , we receive,

ϑ(S(χ)(s)) = ϑ((Sξ)(s), ξ ∈ χ),

(Z ! )
Kι Z Kι
≤ Gι (s, α)ϑh α, Gι (s, σ)ξ(σ) dσ, ξ(α) dα, ξ∈χ .
Kι−1 Kι−1

Then, according to Remark 3.1, we have

(Z " ! # )
Kι Z Kι
ϑ(S(χ)(s)) ≤ Gι (s, α)α−δ γ1 ϑ Gι (s, σ)ξ(σ)dσ + γ2 ϑ(ξ(α)) dα, y∈χ ,
Kι−1 Kι−1
!φι −1 "  φι
γ1 Kι − Kι−1
(
Kι
Kι − Kι−1
Z
1
≤ ϑ(χ)
b α−δ dα
Γ(φι ) 4 4φι −1 Γ(φι ) Kι−1
Z Kι # )
−δ
+ γ2 ϑ(χ)
b α dα , ξ∈χ ,
Kι−1
  φι −1  φι
Kι 1−δ − Kι−1 1−δ Kι − Kι−1 γ1 Kι − Kι−1
!
≤ + γ2 ϑ(χ).
b
4φι −1 (1 − δ)Γ(φι ) 4φι −1 Γ(φι )

Therefore,
  φι −1  φι
Kι 1−δ − Kι−1 1−δ Kι − Kι−1 γ1 Kι − Kι−1
!
ϑ(Sχ)
b ≤ + γ2 ϑ(χ).
b
4φι −1 (1 − δ)Γ(φι ) 4φι −1 Γ(φι )

As a result of (8), we conclude that S is a ι-set contraction. According to Theorem (2.1), problem
(6) has at least one solution yeι in BRι .
Let 
 0, s ∈ [0, Kι−1 ],
ξι = (10)
ξι , s ∈ Υι .
 e

ξι ∈ C([0, Kι ], Λ) is defined by (10) and known to satisfy equation:

Z K1 Z s !
d2 (s − α)1−φ1 (s − α)1−φι 
φι

ξ ι (α)dα + ... + ξ ι (α)dα + h α, ξ ι (α), D0 + ξ ι (α) = 0,
ds2 0 Γ(2 − φ1 ) Kι−1 Γ(2 − φι )

for s ∈ Υι , implying that ξι is a solution of (5) with ξι (0) = 0, ξι (Kι ) = ξeι (Kι ) = 0.

316
Bouazza et al. Malaysian J. Math. Sci. 17(3): 305–332(2023) 305 - 332

Then,
ξ1 (s), s ∈ Υ1 ,




 




  0, s ∈ Υ1 ,



 ξ2 (s) =
ξe2 , s ∈ Υ2 ,

 

ξ(s) = .



 .



 . 



  0, s ∈ [0, Kι−1 ],



 ξn (s) =
ξeι , s ∈ Υι .


It constitutes a solution of BVP(1).

4 Hyers-Ulam-Rassias Stability

Theorem 4.1. Consider (H1), (H2), (8) and

(H3) Let ψ ∈ C(Υι , Λ) is an increasing function, and λψ > 0 exists such that
φι
IK + ψ(s) ≤ λψ(s) ψ(s), for any s ∈ Υι .
ι−1

The equation of (1) is hence (HUR) stable with respect to ψ.

Proof. Consider z ∈ C(Υι , Λ) to be a solution to the following inequality:

 
φι φι
DKι−1
+ z(s) + h s, z(s), DK
ι−1
+ z(s) ≤ ϵψ(s), s ∈ Υι . (11)

ξ ∈ C(Υι , Λ) is a solution to the given problem:

  
 Dφι + ξ(s) + h s, ξ(s), Dφι + ξ(s) = 0, s ∈ Υι ,
Kι−1 Kι−1

ξ(Kι−1 ) = 0, ξ(Kι ) = 0.


By using Lemma (3.1), we have

!
Z Kι Z Kι
ξ(s) = Gι (s, α)h α, Gι (s, σ)ξ(σ) dσ, ξ(α) dα.
Kι−1 Kι−1

From the integration of (11) and (H3), we get

!
Kι Kι s
(s − α)φ(ι)−1
Z Z Z
z(s) + Gι (s, α)h α, Gι (s, σ)z(σ) dσ, z(α) dα ≤ ϵ ψ(α) dα,
Kι−1 Kι−1 Kι−1 Γ(φ(ι))
≤ ϵλψ(s) ψ(s).

317
Bouazza et al. Malaysian J. Math. Sci. 17(3): 305–332(2023) 305 - 332

On the other hand, by Proposition (3.1), we have, for each s ∈ Υι

!
Z Kι Z Kι
∥z(s) − ξ(s)∥ ≤ z(s) − Gι (s, α)h α, Gι (s, σ)z(σ) dσ, z(α) dα
Kι−1 Kι−1
!
Z Kι Z Kι
+ Gι (s, α) h α, Gι (s, σ)z(σ) dσ, z(α)
Kι−1 Kι−1
!
Z Kι
− h α, Gι (s, σ)y(σ) dσ, ξ(α) dα,
Kι−1
!
Z Kι Z Kι
≤ z(s) + Gι (s, α)h α, Gι (s, σ)z(σ) dσ, z(α) dα
Kι−1 Kι−1
!
Z Kι Z Kι
+ Gι (s, α) h α, Gι (s, σ)z(σ) dσ, z(α)
Kι−1 Kι−1
!
Z Kι
− h α, Gι (s, σ)y(σ) dσ, ξ(α) dα,
Kι−1

≤ λψ(s) ϵψ(s)
!φι −1 Z
Kι Kι
Kι − Kι−1
Z
1 −δ
+ α γ1 Gι (s, σ)∥z(σ) − ξ(σ)∥ dσ
Γ(φι ) 4 Kι−1 Kι−1
!
+ γ2 ∥z(α) − ξ(α)∥ dα,

!φι −1 Z  φι
1 Kι − Kι−1 Kι Kι − Kι−1
≤ λψ(s) ϵψ(s) + α−δ γ1 ∥z − ξ∥Ξι
Γ(φι ) 4 Kι−1 4φι −1 Γ(φι )
!
+ γ2 ∥z − ξ∥Ξι dα,
 φι −1    φι
1−δ
Kι − Kι−1 Kι1−δ − Kι−1 γ1 Kι − Kι−1
!
≤ λψ(s) ϵψ(s) + + γ2 ∥z − ξ∥Ξι .
(1 − δ)4φι −1 Γ(φι ) 4φι −1 Γ(φι )

Hence,
 φι −1    φι
1−δ
Kι − Kι−1 Kι1−δ − Kι−1 γ1 Kι − Kι−1
" !#
∥z − ξ∥Eι 1 − + γ2 ≤ λψ(s) ϵψ(s).
(1 − δ)4φι −1 Γ(φι ) 4φι −1 Γ(φι )

We get, for any s ∈ Υι

 φι −1    φι !#−1
1−δ
Kι − Kι−1 Kι1−δ − Kι−1 γ1 Kι − Kι−1
"
∥z − ξ∥Ξι ≤ 1 − + γ2 λψ(s) ϵψ(s),
(1 − δ)4φι −1 Γ(φι ) 4φι −1 Γ(φι )
:= cg ϵψ(s).

Then, BVP (1) is (HUR) stable.

318
Bouazza et al. Malaysian J. Math. Sci. 17(3): 305–332(2023) 305 - 332

5 Approximate Numerical Examples

5.1 Example 5.1

Consider the following BVP:

 φ(s)
 Dφ(s) ξ(s) + |ξ(s)| + |D0+ ξ(s)|
s ∈ Υ := [0, 2],

0+ 1 = 0,
(s + 1) 2 (9 + es )(1 + ξ 2 (s)) (12)


ξ(0) = 0, ξ(2) = 0.
Let,
π+z
h(s, π, z) = 1 , (s, π, z) ∈ [0, 2] × [0, +∞) × [0, +∞).
(s + 1) (9 + es )(1 + π 2 )
2

9

 8,

 s ∈ Υ1 := [0, 1],
φ(s) = (13)
 7,

s ∈ Υ2 :=]1, 2].

4
Then, we have
1 1 π1 + ς 1 π2 + ς 2
s 2 h(s, π1 , ς1 ) − h(s, π2 , ς2 ) ≤ − ,
9 + et 1 + π12 1 + π22
1
≤ (π1 + ς1 ) − (π2 + ς2 ) ,
10
1 1
≤ |π1 − π2 | + |ς1 − ς2 |.
10 10
1 1
Consequently, (H2) holds if δ = , γ1 = γ2 = .
2 10

The problem’s equation (12) is decomposed into two expressions by (13).

 9
 98
 |ξ(s)| + |D08+ ξ(s)|
 D
 0+
 ξ(s) + 1 = 0, s ∈ Υ1 ,
(s + 1) 2 (9 + es )(1 + ξ 2 (s))


 7
|ξ(s)| + |D14+ ξ(s)|

 7

 D1+ ξ(s) +

 4
1 = 0, s ∈ Υ2 .
(s + 1) 2 (9 + es )(1 + ξ 2 (s))
For s ∈ Υ1 , (12) corresponds to the next problem:
 9

 9 |ξ(s)| + |D08+ ξ(s)|
 D0+ ξ(s) +
 8
1 = 0, s ∈ Υ1 ,
(s + 1) 2 (9 + es )(1 + ξ 2 (s)) (14)



ξ(0) = 0, ξ(1) = 0.


Next, we demonstrate that the condition (8) is satisfied

 φ1 −1    φ1
K1 − K0 K11−δ − K01−δ γ1 K1 − K0
! 1 1 9
!
1
(1) 8 (1 2 ) 10 (1) 1
8
+ γ2 = 1 18 1 + ,
(1 − δ)4φ1 −1 Γ(φ1 ) 4φ1 −1 Γ(φ1 ) 9
2 4 Γ( 8 ) 4 8 Γ( 98 ) 10
≃ 0.3380 < 1.

319
Bouazza et al. Malaysian J. Math. Sci. 17(3): 305–332(2023) 305 - 332

1
Let ψ(s) = s 2 ,
Z s
1 1 1
I0φ+1 ψ(t) = (s − α) 8 α 2 dα,
Γ( 98 ) 0
Z s
1 1
≤ (s − α) 8 dα,
Γ( 98 ) 0
8
≤ ψ(s).
9Γ( 98 )

1 8
Consequently, if ψ(s) = s 2 and λψ(s) = , then (H3) is satisfied.
9Γ( 98 )

The problem (14) has a solution ξ1 ∈ Ξ1 according to the Theorem (3.1).

The problem (12) may be formulated as follows for s ∈ Υ2 :

 7

 7 |ξ(s)| + |D14+ ξ(s)|
 D1+ ξ(s) +
 4
1 = 0, s ∈ Υ2 ,
(s + 1) 2 (9 + es )(1 + ξ 2 (s)) (15)



ξ(1) = 0, ξ(2) = 0.


We see that,
 φ2 −1    φ2
K2 − K1 K21−δ − K11−δ γ1 K 2 − K 1
! 3 1 1 7
!
1
(1) 4 (2 2 − 1 2 ) 10 (1) 1
4
+ γ2 = 1 34 3 + ,
(1 − δ)4φ2 −1 Γ(φ2 ) 4φ2 −1 Γ(φ2 ) 7
2 4 Γ( 4 ) 4 4 Γ( 47 ) 10
≃ 0.0441 < 1.

Therefore, (8) is satisfied, and

Z s
1 3 1
I1φ+2 ψ(s) = 7 (s − α) 4 α 2 dα,
Γ( 4 ) 1
√ Z s
2 3
≤ (s − α) 4 dα,
Γ( 74 ) 1
√
4 2
≤ ψ(s) := λψ(s) ψ(s).
7Γ( 74 )
√
1 4 2
Consequently, (H3) is satisfied by ψ(s) = s and λψ(s)
2 = .
7Γ( 74 )

The problem (15) has to have a solution ξe2 ∈ Ξ2 according to Theorem (3.1).

It’s well known,


 0, s ∈ Υ1 ,
ξ2 (s) =
ξ2 (s), s ∈ Υ2 .
 e

320
Bouazza et al. Malaysian J. Math. Sci. 17(3): 305–332(2023) 305 - 332

As a result, by Definition (3.1), the boundary value problem (12) has a solution:


 ξ1 (s), s ∈ Υ1 ,


 
ξ(s) =  0, s ∈ Υ1 ,
ξ (s) =

2



ξ2 (s), s ∈ Υ2 .
  e

Furthermore, according to Theorem (4.1), the equation of (12) is (HUR) stable.

In those figures, we present the plot of the function h(s, π, z) in the two interval [0, 1] and [1, 2].

Figure 1: A plot of h(s, π, z) for different values of s ∈ [0, 1].

321
Bouazza et al. Malaysian J. Math. Sci. 17(3): 305–332(2023) 305 - 332

Figure 2: A plot of h(s, π, z) for different values of s ∈]1, 2].

322
Bouazza et al. Malaysian J. Math. Sci. 17(3): 305–332(2023) 305 - 332

Figure 3: The 3D plot of h(s, π, z) for s ∈ [0, 1] and ]1, 2].

In this table, we present the value of maxπ,z h(s, π, z) for different values of s ∈ [0, 1] and ]1, 2].

s ∈ [0, 1] 0.0010 0.0100 0.0200 0.0800 0.1000 0.5000 1.0000

maxπ,z h(s, π, z) 0.2117 0.2105 0.2093 0.2021 0.1998 0.1624 0.1278

s ∈]1, 2] 1.0010 1.0100 1.0200 1.0800 1.1000 1.5000 2.0000

maxπ,z h(s, π, z) 0.1277 0.1272 0.1266 0.1229 0.1217 0.0994 0.0746

We observe that the value of maxπ,z h(s, π, z) is decreasing with respect to the value of s in the
two intervals.

5.2 Example 5.2

Consider the following BVP:

 −1

 Dφ(s)
 s 4 e−s

 0+ ξ(s) + s
φ(s)
= 0, s ∈ Υ := [0, 3],
4e 1+s2 + 5es + 2 1 + |ξ(s)| + |D0+ ξ(s)| (16)



ξ(0) = 0, ξ(3) = 0.


Let,
−1
s e−s
4
h(s, π, z) = s

, (s, π, z) ∈ [0, 3] × [0, +∞) × [0, +∞).
4e 1+s2 + 5es + 2 1 + π + z

323
Bouazza et al. Malaysian J. Math. Sci. 17(3): 305–332(2023) 305 - 332

3


 , s ∈ Υ1 := [0, 1],



 2

8

φ(s) = , s ∈ Υ2 :=]1, 2], (17)


 5

 9,


s ∈ Υ3 :=]2, 3].

5
Then, we have

1 1 1 1
s 4 h(s, π1 , ς1 ) − h(s, π2 , ς2 ) ≤ s − ,
1 + π1 + ς 1 1 + π2 + ς2


es4e +
1+s2 +2 5es
1
≤ (π2 + ς2 ) − (π1 + ς1 ) ,
11
1 1
≤ |π1 − π2 | + |ς1 − ς2 |.
11 11
1 1
Consequently, (H2) holds if δ = , γ1 = γ 2 = .
4 11

The problem’s equation (16) is decomposed into two expressions by (17),

−1
s 4 e−s

3
 D 2+ ξ(s) +
= 0, s ∈ Υ1 ,

0 s

φ(s)
4e 1+s2 + 5es + 2 1 + |ξ(s)| + |D0+ ξ(s)|







 −1
s 4 e−s

 8
D1+ ξ(s) +
5
s
φ(s)
= 0, s ∈ Υ2 ,


 4e 1+s2 + 5es + 2 1 + |ξ(s)| + |D1+ ξ(s)|



−1

s 4 e−s


 9
 D25+ ξ(s) +

 s
φ(s)
= 0, s ∈ Υ3 .
4e 1+s2 + 5es + 2 1 + |ξ(s)| + |D2+ ξ(s)|

For s ∈ Υ1 , (16) corresponds to the next problem:

 −1
 3 s 4 e−s
 D0+ ξ(s) +

 2
s
φ(s)
= 0, s ∈ Υ1 ,
4e 1+s2 + 5es + 2 1 + |ξ(s)| + |D0+ ξ(s)| (18)



ξ(0) = 0, ξ(1) = 0.


Next, we demonstrate that the condition (8) is satisfied.

 φ1 −1    φ1
K1 − K0 K11−δ − K01−δ γ1 K1 − K0
! 1 3 3
!
1
(1) 2 (1 4 ) 11 (1) 1
2
+ γ2 = 3 12 1 + ,
(1 − δ)4φ1 −1 Γ(φ1 ) 4φ1 −1 Γ(φ1 ) 3
4 4 Γ( 2 )
3
4 2 Γ( 2 ) 11
≃ 0.1070 < 1.

The problem (18) has a solution ξ1 ∈ Ξ1 according to the Theorem (3.1).

The problem (16) may be formulated as follows for s ∈ Υ2 :

 −1
 8 s 4 e−s
 D1+ ξ(s) +

 5
s
φ(s)
= 0, s ∈ Υ2 ,
4e 1+s2 + 5es + 2 1 + |ξ(s)| + |D1+ ξ(s)| (19)



ξ(1) = 0, ξ(2) = 0.


324
Bouazza et al. Malaysian J. Math. Sci. 17(3): 305–332(2023) 305 - 332

We see that
 φ2 −1    φ2
K2 − K1 K21−δ − K11−δ γ1 K2 − K1
! 3 3 3 8
!
1
(1) 5 (2 4 − 1 4 ) 11 (1) 1
5
+ γ2 = 3 35 3 + ,
(1 − δ)4φ2 −1 Γ(φ2 ) 4φ2 −1 Γ(φ2 ) 8
4 4 Γ( 5 )
8
4 5 Γ( 5 ) 11
≃ 0.0599 < 1.

Therefore, (8) is satisfied.

The problem (20) has to have a solution ξe2 ∈ Ξ2 according to Theorem (3.1).

The problem (16) may be formulated as follows for s ∈ Υ3 :

 −1
 D 59 ξ(s) +
 s 4 e−s

 2+ s
φ(s)
= 0, s ∈ Υ3 ,
4e 1+s2 + 5es + 2 1 + |ξ(s)| + |D2+ ξ(s)| (20)



ξ(2) = 0, ξ(3) = 0.


We see that
 φ3 −1    φ3
K3 − K2 K31−δ − K21−δ γ1 K3 − K2
! 4 3 3 9
!
1
(1) 5 (3 4 − 2 4 ) 11 (1) 1
5
+ γ2 = 3 45 4 + ,
(1 − δ)4φ3 −1 Γ(φ3 ) 4φ3 −1 Γ(φ3 ) 9
4 4 Γ( 5 )
9
4 5 Γ( 5 ) 11
≃ 0.0347 < 1.

Therefore, (8) is satisfied.

The problem (20) has to have a solution ξe3 ∈ Ξ2 according to Theorem (3.1).

As a result, by definition (3.1), the boundary value problem (16) has a solution:

s ∈ Υ1 ,

 ξ1 (s),



 



  0, s ∈ Υ1 ,



 ξ 2 (s) =
ξ2 (s), s ∈ Υ2 ,

  e
ξ(s) =
s ∈ Υ1 ,



  0,

 


 



 ξ3 (s) = 0, s ∈ Υ2 ,

 


 
 e
ξ3 (s), s ∈ Υ3 .

In those figures, we present the plot of the function h(s, π, z) in the two interval [0, 1], [1, 2] and
[2, 3],

325
Bouazza et al. Malaysian J. Math. Sci. 17(3): 305–332(2023) 305 - 332

Figure 4: A plot of h(s, π, z) for different values of s ∈ [0, 1].

326
Bouazza et al. Malaysian J. Math. Sci. 17(3): 305–332(2023) 305 - 332

Figure 5: A plot of h(s, π, z) for different values of s ∈]1, 2].

327
Bouazza et al. Malaysian J. Math. Sci. 17(3): 305–332(2023) 305 - 332

Figure 6: A plot of h(s, π, z) for different values of s ∈]2, 3].

328
Bouazza et al. Malaysian J. Math. Sci. 17(3): 305–332(2023) 305 - 332

Figure 7: The 3D plot of h(s, π, z) for s ∈ [0, 1], ]1, 2] and ]2, 3].

In this table, we present the value of maxπ,z h(s, π, z) for different values of s ∈ [0, 1], [1, 2] and
[2, 3]. We have made the same observation as in Example 5.1.

s ∈ [0, 1] 0.1000 0.2000 0.3000 0.4000 0.5000 0.7000 1.0000

maxπ,z h(s, π, z) 0.1347 0.0945 0.0714 0.0558 0.0445 0.0294 0.0166

s ∈]1, 2] 1.1000 1.2000 1.3000 1.4000 1.5000 1.7000 2.0000

maxπ,z h(s, π, z) 0.0138 0.0114 0.0095 0.0079 0.0066 0.0045 0.0025

s ∈]2, 3] 2.1000 2.2000 2.3000 2.4000 2.5000 2.7000 3.0000

maxπ,z h(s, π, z) 0.0021 0.0017 0.0014 0.0012 0.0010 0.0006 0.0004

6 Conclusions

This research introduced a boundary value problem of Riemann-Liouville fractional differen-

tial equations of variable order φ(s), in which φ(s) is a piecewise constant function. The analytical

329
Bouazza et al. Malaysian J. Math. Sci. 17(3): 305–332(2023) 305 - 332

solutions have been successfully investigated via three strategies: the DFPTH, Kuratowski’s MNK
and HUR stability concept. Finally, we illustrated the theoretical results by some numerical ex-
amples. Therefore, the results showcased in this paper exhibit immense promise for utilization in
diverse applications of multidisciplinary sciences.

Acknowledgement This research is funded by Universiti Malaysia Sabah under the Research
Grant SBk0508-2021. The authors would like to express their gratitude to all anonymous review-
ers for their valuable feedback and suggestions, which greatly contributed to the improvement of
this paper.

Conflicts of Interest We hereby to declare that there is no conflict of interest among us in conduct-
ing this research.

References

[1] D. Baleanu, A. Jajarmi, H. Mohammadi & S. Rezapour (2020). A new study on the mathe-
matical modelling of human liver with Caputo-Fabrizio fractional derivative. Chaos, Solitons
& Fractals, 134, 109705. https://doi.org/10.1016/j.chaos.2020.109705.
[2] D. Baleanu, H. Mohammadi & S. Rezapour (2020). Analysis of the model of HIV-1 infection
of CD4+ T-cell with a new approach of fractional derivative. Advances in Difference Equations,
2020(1), 1–17. https://doi.org/10.1186/s13662-020-02544-w.
[3] J. Banas̀ & K. Goebel (1980). Measures of noncompactness in Banach spaces volume 13. Marcel
Dekker, New York, NY.
[4] J. Banas̀ & L. Olszowy (2001). Measures of non-compactness related to monotonicity. Annales
Societatis Mathematicae Polonae. Seria 1: Commentationes Mathematicae, [Z] 41, 13–23.
[5] M. Benchohra, S. Bouriah, J. E. Lazreg & J. J. Nieto (2016). Nonlinear implicit Hadamard’s
fractional differential equations with delay in Banach space. Acta Universitatis Palackianae
Olomucensis. Facultas Rerum Naturalium. Mathematica, 55(1), 15–26. https://eudml.org/doc/
286701.
[6] A. Benkerrouche, M. S. Souid, K. Sitthithakerngkiet & A. Hakem (2021). Implicit nonlinear
fractional differential equations of variable order. Boundary Value Problems, 2021(64), 1–16.
https://doi.org/10.1186/s13661-021-01540-7.
[7] A. Djumaklych (2013). Boundary value problem for higher order equation with fractional
derivative. Malaysian Journal of Mathemathical Sciences, 7(2), 203–217.
[8] S. Dob, H. Lakhal & M. Maouni (2021). Existance and uniqueness of solutions for a nonlinear
fractional elliptic system. Malaysian Journal of Mathemathical Sciences, 15(3), 347–356.
[9] S. Etemad, I. Avci, P. Kumar, D. Baleanu & S. Rezapour (2022). Some novel mathematical
analysis on the fractal-fractional model of the AH1N1/09 virus and its generalized Caputo-
type version. Chaos, Solitons & Fractals, 162, 112511. https://doi.org/10.1016/j.chaos.2022.
112511.
[10] S. Gul, R. A. Khan, K. Shah & T. Abdeljawad (2023). On a general class of nth order sequential
hybrid fractional differential equations with boundary conditions. AIMS Mathematics, 8(4),
9740–9760. https://doi.org/10.3934/math.2023491.

330
Bouazza et al. Malaysian J. Math. Sci. 17(3): 305–332(2023) 305 - 332

[11] D. Guo, V. Lakshmikantham & X. Liu (1996). Nonlinear integral equations in abstract spaces.
Kluwer Academic Publishers, Dordrecht.
[12] S. Hussain, E. N. Madi, H. Khan, H. Gulzar, S. Etemad, S. Rezapour & M. K. Kaabar (2022).
On the stochastic modeling of COVID-19 under the environmental white noise. Journal of
Function Spaces, 2022, Article ID 4320865. https://doi.org/10.1155/2022/4320865.
[13] A. Jiahui & C. Pengyu (2019). Uniqueness of solutions to initial value problem of fractional
differential equations of variable-order. Dynamic Systems and Applications, 28(3), 607–623.
[14] H. Khan, K. Alam, H. Gulzar, S. Etemad & S. Rezapour (2022). A case study of fractal-
fractional tuberculosis model in China: Existence and stability theories along with numerical
simulations. Mathematics and Computers in Simulation, 198, 455–473. https://doi.org/10.1016/
j.matcom.2022.03.009.
[15] A. A. Kilbas, H. M. Srivastava & J. J. Trujillo (2006). Theory and applications of fractional differ-
ential equations volume 204. Elsevier, New York, NY.
[16] M. M. Matar, M. I. Abbas, J. Alzabut, M. K. A. Kaabar, S. Etemad & S. Rezapour (2021).
Investigation of the p-Laplacian nonperiodic nonlinear boundary value problem via gen-
eralized Caputo fractional derivatives. Advances in Difference Equations, 2021(68), 1–18.
https://doi.org/10.1186/s13662-021-03228-9.
[17] H. Mohammadi, S. Kumar, S. Rezapour & S. Etemad (2021). A theoretical study of the
Caputo-Fabrizio fractional modeling for hearing loss due to Mumps virus with optimal con-
trol. Chaos, Solitons & Fractals, 144, 110668. https://doi.org/10.1016/j.chaos.2021.110668.
[18] S. Rezapour, M. S. Souid, S. Etemad, Z. Bouazza, S. K. Ntouyas, S. Asawasamrit & J. Tariboon
(2021). Mawhin’s continuation technique for a nonlinear BVP of variable order at resonance
via piecewise constant functions. Fractal and Fractional, 5(4), 216. https://doi.org/10.3390/
fractalfract5040216.
[19] I. A. Rus (2010). Ulam stabilities of ordinary differential equations in a Banach space.
Carpathian Journal of Mathematics, 26(1), 103–107. https://www.jstor.org/stable/43999438.
[20] S. G. Samko (1995). Fractional integration and differentiation of variable order. Analysis
Mathematica, 21, 213–236. https://doi.org/10.1007/BF01911126.
[21] S. G. Samko & B. Boss (1993). Integration and differentiation to a variable fractional
order. Integral Transforms and Special Functions, 1(4), 277–300. https://doi.org/10.1080/
10652469308819027.
[22] K. Shah & W. Hussain (2019). Investigating a class of nonlinear fractional differential equa-
tions and its Hyers-Ulam stability by means of topological degree theory. Numerical Func-
tional Analysis and Optimization, 40(12), 1355–1372. https://doi.org/10.1080/01630563.2019.
1604545.
[23] K. Shah & R. A. Khan (2016). Existence and uniqueness results to a coupled system of frac-
tional order boundary value problems by topological degree theory. Numerical Functional
Analysis and Optimization, 37(7), 887–899. https://doi.org/10.1080/01630563.2016.1177547.
[24] K. Shah, M. Sher, A. Ali & T. Abdeljawad (2022). On degree theory for non-monotone type
fractional order delay differential equations. AIMS Mathematics, 7(5), 9479–9492. https://
doi.org/10.3934/math.2022526.
[25] N. H. Tuan, H. Mohammadi & S. Rezapour (2020). A mathematical model for COVID-
19 transmission by using the Caputo fractional derivative. Chaos, Solitons & Fractals, 140,
1101079. https://doi.org/10.1016/j.chaos.2020.110107.

331
Bouazza et al. Malaysian J. Math. Sci. 17(3): 305–332(2023) 305 - 332

[26] D. Valerio & J. Costa (2011). Variable-order fractional derivatives and their numerical ap-
proximations. Signal Processing, 91(3), 470–483. https://doi.org/10.1016/j.sigpro.2010.04.
006.

[27] S. Zhang (2013). Existence of solutions for two point boundary value problems with singular
differential equations of variable order. Electronic Journal of Differential Equations, 2013(245),
1–16.
[28] S. Zhang (2018). The uniqueness result of solutions to initial value problems of differential
equations of variable-order. Revista de la Real Academia de Ciencias Exactas, Físicas y Naturales.
Serie A. Matemáticas, (112), 407–423. https://doi.org/10.1007/s13398-017-0389-4.
[29] S. Zhang & L. Hu (2019). Unique existence result of approximate solution to initial value
problem for fractional differential equation of variable-order involving the derivative argu-
ments on the half-axis. Mathematics, 7(3), 286. https://doi.org/10.3390/math7030286.
[30] S. Zhang & L. Hu (2020). The existence of solutions and generalized Lyapunov-type inequal-
ities to boundary value problems of differential equations of variable order. AIMS Mathemat-
ics, 5(4), 2923–2943. https://doi.org/10.3934/math.2020189.
[31] S. Zhang, S. Sun & L. Hu (2018). Approximate solutions to initial value problem for differ-
ential equation of variable order. Journal of Fractional Calculus and Applications, 9(2), 93–112.

[32] S. Zhang, S. Li & L. Hu (2018). The existeness and uniqueness result of solutions to initial
value problems of nonlinear diffusion equations involving with the conformable variable
derivative. Revista de la Real Academia de Ciencias Exactas, Físicas y Naturales. Serie A. Matemáti-
cas, 113(1), 1601–1623. https://doi.org/10.1007/s13398-018-0572-2.

332