Professional Documents
Culture Documents
Existence and Stability of Implicit Fractional Differential Equations With Stieltjes Boundary Conditions Involving Hadamard Derivatives
Existence and Stability of Implicit Fractional Differential Equations With Stieltjes Boundary Conditions Involving Hadamard Derivatives
Complexity
Volume 2021, Article ID 8824935, 36 pages
https://doi.org/10.1155/2021/8824935
Research Article
Existence and Stability of Implicit Fractional Differential
Equations with Stieltjes Boundary Conditions Involving
Hadamard Derivatives
Danfeng Luo ,1 Mehboob Alam,2 Akbar Zada ,2 Usman Riaz,2 and Zhiguo Luo 3
1
Department of Mathematics, Guizhou University, Guiyang 550025, China
2
Department of Mathematics, University of Peshawar, Peshawar, Khyber Pakhtunkhwa, Pakistan
3
Key Laboratory of Computing and Stochastic Mathematics (Ministry of Education), School of Mathematics and Statistics,
Hunan Normal University, Changsha 410081, China
Received 4 September 2020; Revised 7 November 2020; Accepted 22 February 2021; Published 23 March 2021
Copyright © 2021 Danfeng Luo et al. This is an open access article distributed under the Creative Commons Attribution License,
which permits unrestricted use, distribution, and reproduction in any medium, provided the original work is properly cited.
In this article, we make analysis of the implicit fractional differential equations involving integral boundary conditions associated
with Stieltjes integral and its corresponding coupled system. We use some sufficient conditions to achieve the existence and
uniqueness results for the given problems by applying the Banach contraction principle, Schaefer’s fixed point theorem, and
Leray–Schauder result of the cone type. Moreover, we present different kinds of stability such as Hyers–Ulam stability, generalized
Hyers–Ulam stability, Hyers–Ulam–Rassias stability, and generalized Hyers–Ulam–Rassias stability by using the classical
technique of functional analysis. At the end, the results are verified with the help of examples.
differential equations is also very significant because these where J � [0, 1] and Dαq denote the standard Rie-
types of systems appear naturally in many problems of mann–Liouville fractional q-derivatives of order α, with
applied nature. For details and examples, the reader may see α ∈ (2, 3], φ ∈ L1 [0, 1] being nonnegative.
[15–18] and the references cited therein. Benchohra and Lazreg [8] studied the existence and
Another area of research which has received great at- Hyers–Ulam stability of the following implicit fractional
tention from researchers in the field of fractional differential differential equations (FDE′ s) involving Hadamard
equations is the notion of stability in the sense of Ulam. derivatives:
There are different kinds of Ulam’s stabilities, i.e., α
HD u(σ) � ϕ1 σ, u(σ), Dα u(σ), α ∈ (0, 1),
Hyers–Ulam stability (HUS), generalized Hyers–Ulam (2)
stability (GHUS), Hyers–Ulam–Rassias stability u(1) � u1 , u1 ∈ R,
(HURS), and generalized Hyers–Ulam–Rassias stability
(GHURS). The Ulam-type stability was first introduced where σ ∈ [1, T], T > 1 and H Dα denotes the Hadamard
by Ulam [19] in 1940 and then was studied and generalized fractional derivative of order α.
by many mathematicians with different approaches Riaz et al. [10] studied the existence, uniqueness, and
[10, 20–29]. stability of the solution of a nonlinear coupled system of
Ren and Zhai [30] discussed the existence of a unique impulsive FDEs using Hadamard derivatives of the fol-
solution and multiple positive solutions with nonlocal lowing form:
boundary conditions for the system involving standard
Riemann–Liouville fractional q-derivatives:
⎧
⎪
⎪ Dαq u(σ) + ϕ(σ, u(σ)) � 0, σ ∈ J,
⎪
⎪
⎪
⎨ u(σ)| � Dα− 2 u(σ)| � 0,
σ�0 q σ�0 (1)
⎪
⎪
⎪
⎪ ζ
⎪
⎩ Dα− 1 β
q u(σ)|σ�1 � α[u] + φ(σ)Dq u(σ)dq σ,
0
α
⎧
⎪
⎪
⎪ H D u(σ) − f(σ, u(σ), v(σ)) � 0, σ ∈ J, σ ≠ σ i , i � 1, 2, . . . , m,
⎪
⎪
⎪
⎪
⎪
⎪ β
⎪
⎪ H D v(σ) − g(σ, u(σ), v(σ)) � 0, σ ∈ J, σ ≠ σ j , j � 1, 2, . . . , m,
⎪
⎪
⎪
⎪
⎪
⎪
⎪
⎪ i u σ i , i � 1, 2, . . . , m,
⎪
⎪ Δu σ i � Ii u σ i , Δu′ σ i � I
⎪
⎪
⎪
⎪
⎪
⎪
⎪
⎪
⎨ Δvσ j � Ij vσ j , Δv′ σ j � Ij vσ j , j � 1, 2, . . . , m,
⎪
⎪ (3)
⎪
⎪
⎪
⎪ μ ln 2u(2) + ] ln 2u′ (2) � ϕ(u),
⎪
⎪
⎪
⎪
⎪
⎪
⎪
⎪
⎪
⎪ μu(T) + ]u′ (T) � φ(u),
⎪
⎪
⎪
⎪
⎪
⎪
⎪
⎪ μ ln 2v(2) + ] ln 2v′ (2) � ϕ(v),
⎪
⎪
⎪
⎪
⎪
⎪
⎪
⎩
μv(T) + ]v′ (T) � φ(v),
Complexity 3
where 1 <α, β ≤ 2,f,g: J × R2 ⟶ R and ϕ, φ: C(J, R) ⟶ R Riaz et al. [31] studied the existence, uniqueness, and
are continuous functions. stability of solution of coupled implicit impulsive fractional
differential equations using Hadamard derivatives:
α β
⎪
⎧
⎪ H D u(σ) − fσ, u(σ), H D v(σ) � 0, σ ∈ J, σ ≠ σ i , i � 1, 2, . . . , m,
⎪
⎪
⎪
⎪
⎪
⎪
⎪
⎪
⎪
⎪
⎪
⎪
⎪
⎪ H Dβ v(σ) − g σ, v(σ), H Dα u(σ) � 0, σ ∈ J, σ ≠ σ j , j � 1, 2, . . . , m,
⎪
⎪
⎪
⎪
⎪
⎪
⎪
⎪
⎪
⎪
⎪
⎪
⎪
⎪
⎪
⎪ i u σ i , i � 1, 2, . . . , m,
Δu σ i � Ii u σ i , Δu′ σ i � I
⎪
⎪
⎪
⎪
⎪
⎪
⎪
⎨
⎪ (4)
⎪ Δvσ � I vσ , Δv′ σ � I
⎪ j vσ j , j � 1, 2, . . . , m,
⎪
⎪ j j j j
⎪
⎪
⎪
⎪
⎪
⎪
⎪
⎪
⎪
⎪
⎪
⎪ T (ln(T/π))α− 1
⎪
⎪ dπ
⎪
⎪ u(T) � ϕ(π, u(π)) , u′ (T) � φ(u),
⎪
⎪ 1 Γ(α) π
⎪
⎪
⎪
⎪
⎪
⎪
⎪
⎪
⎪
⎪
⎪
⎪
⎪
⎪ T (ln(T/π))β− 1 dπ
⎪
⎪
⎩ v(T) � ϕ(π, v(π)) , v′ (T) � φ(v),
1 Γ(β) π
where 1 < α, β ≤ 2, f, g: J × R2 ⟶ R and Stieltjes integral condition instead of the Riemann–Liouville
ϕ, φ: C(J, R) ⟶ R are continuous functions. fractional q-derivative and present its existence, uniqueness,
In this article, we extend the system of Ren and Zhai [30] at least one solution, and different kinds of Ulam’s stabilities.
to implicit Hadamard fractional derivatives having the We study the following system:
α α
⎪
⎧
⎪ H D u(σ) + ϕ σ, u(σ), H D u(σ) � 0, σ ∈ J,
⎪
⎪
⎪
⎪
⎪
⎨ u(σ)| � Dα− 2 u(σ)| � 0,
σ�1 H σ�1 (5)
⎪
⎪
⎪
⎪
⎪
⎪ ζ
⎪
⎩ H Dα− 1 u(σ)|σ�2 � α[u] + φ(σ)H Dβ u(σ)dσ,
1
where α ∈ (2, 3], ϕ: J � [1, 2] × R × R ⟶ R is the con- involving the Stieltjes integral with respect to the function
tinuous function and H Dα , H Dβ are the Hadamard frac- μ: [1, 2] ⟶ R, μ(σ) is right continuous on [1, 2), left con-
tional derivatives of order α, with α − 1 − β > 0, φ ∈ L1 [1, 2] tinuous at σ � 2. Particularly, μ is a nondecreasing function
being non-negative, and α[u] is a linear function given by with μ(1) � 0; then, dμ is a positive Stieltjes measure.
We also extend system (5) to the coupled system and
2
discuss its existence, uniqueness, at least one solution, and
α[u] � u(σ)dμ(σ), (6)
1 different kinds of Ulam’s stabilities of the problem:
4 Complexity
⎪
⎧
α c
H D u(σ) + ϕ1 σ, u(σ), H D v(σ) � 0, σ ∈ J, (M6 ): let Ψ ∈ C(J, R) be an increasing function; then,
⎪
⎪ there is ΩΨ > 0 such that, for each σ ∈ J, the inequality
⎪
⎪ c α
⎪
⎪
⎪ H D v(σ) + ϕ2 σ, v(σ), H D u(σ) � 0, σ ∈ J,
α
⎪
⎪ HI Ψ(σ) ≤ ΩΨ Ψ(σ), (12)
⎪ α− 2
⎪ u(σ)|σ�1 � H D u(σ)|σ�1 � 0,
⎪
⎪
⎪
⎨ ζ
(7) holds.
⎪
⎪ H Dα− 1 u(σ)|σ�2 � α[u] + φα (σ)H Dβ u(σ)dσ, ζ
⎪
⎪
⎪ 1 (M7 ): Δα � Γ(α − β) − 1 φα (σ)(lnσ)α− 1− β
dσ − Γ (α −
⎪
⎪ v(σ)| � Dc− 2 v(σ)| � 0, α− 1
⎪
⎪ β)/Γ(α) 21 (lnσ) dμ(σ), and
⎪
⎪ σ�1 H σ�1
⎪
⎪ ξ
⎪
⎪ ξ
⎩ H Dc− 1 v(σ)|σ�2 � c[v] + φc (σ)H Dδ v(σ)dσ, 1− ξ
1 Δc � Γ(c − ξ) − φc (σ)(ln σ)c− dσ
1
where α, c ∈ (2, 3], ϕ1 , ϕ2 : J � [1, 2] × R × R ⟶ R are the (13)
continuous functions and H Dα , H Dβ , H Dc , and H Dδ are the 2
Hadamard fractional derivatives of orders α, c, respectively Γ(c − ξ)
− (ln σ)c− 1 dμ(σ).
with α − 1 − β > 0, c − 1 − δ > 0, φα , φc ∈ L1 [1, 2] being Γ(c)
1
nonnegative, and α[u], c[v] are linear functions given by
2 (M8 ): ϕ1 , ϕ2 : J × R × R ⟶ R are continuous.
α[u] � u(σ)dμ(σ),
1 (M9 ): for σ ∈ J and u, z ∈ X, there are
2
(8) π4 , π5 , π6 ∈ C(J, R+ ) such that
c[v] � v(σ)dμ(σ),
1 ϕ (σ, u(σ), z(σ)) ≤ π4 (σ) + π5 (σ)|u(σ)| + π6 (σ)|z(σ)|,
1
1 ≤ σ ≤ 2,
(20) Theorem
� � [34]). Let X be a Banach space containing a
4 (see
cone C. B � ⊂ C with 0 ∈ B � is a relatively open set, and the
where Γ(·) is the gamma function. � �
operator T: B ⟶ B is completely continuous. Then, one of
the following conditions holds:
Lemma 1 (see [31]). Let α > 0 and u be any function; then, � and δ ∈ (0, 1) such that
the homogeneous differential equation along with Hadamard (1) There exist (u, v) ∈zB
fractional order H Dα u(σ) � 0 has a solution (u, v) � δT(u, v)
(2) T has a fixed point in B �
n
u(σ) � ci (ln σ)α− i , (21)
i�1
3. Existence and Uniqueness
and the following formula holds: In this section, we give the existence and uniqueness of
n
solutions of (5) and its coupled system (7).
α α
H I H D u(σ) � u(σ) + ci (ln σ)α− i , (22)
i�1
3.1. Existence and Uniqueness Solution for System (5).
where ci ∈ R, i � 1, 2, . . . , n, and n − 1 < α < n. Our first result is stated as follows.
Theorem 1 (see [32]) (Arzela-Ascoli’s Theorem). Let Lemma 3. Let us assume that f ∈ C(J, R) and Δ > 0; then,
B ⊂ C(J, R) be relatively compact, and the fractional differential equation
α
(A) B is a uniformly bounded set such that there exists ⎧ H D u(σ) + f(σ) � 0, α ∈ (2, 3], σ ∈ J,
⎪
⎪
⎪
⎪
ϖ > 0 with ⎪
⎨ u(σ)| � Dα− 2 u(σ)| � 0,
σ�1 H σ�1
‖ϕ‖ � sup |ϕ(y)| < ϖ, for every ϕ ∈ B. ⎪
⎪
(23) ⎪
⎪ 2 ζ
y∈J ⎪
⎩ H Dα− 1 u(σ)|σ�2 � u(σ)dμ(σ) + φ(σ)H Dβ u(σ)dσ,
1 1
(B) B is an equicontinuous set, i.e., for every ϵ > 0, there (27)
exists δ > 0 such that, for any
has a solution
z, z ∈ J, 2 dη
for every ϕ ∈ B. u(σ) � G(σ, η)f(η) , σ ∈ J, (28)
|z − z| ≤ δ⇒|ϕ(z) − ϕ(z)| ≤ ε, 1 η
(24) where
6 Complexity
G(σ, η) � G1 (σ, η) + G2 (σ, η) + G3 (σ, η), Now, considering the condition H Dα− 2 u(σ)|σ�1 � 0, we
get
α− 1 α− 1
⎪
⎧ (ln σ) − (ln σ/η) α− 2
� c1H Dα− 2 (ln σ)α− 1 + c2H Dα− 2 (ln σ)α− 2
⎪
⎪ , 1 ≤ η ≤ σ ≤ 2, HD u(σ)
⎪
⎪ Γ(α)
⎪
⎨ α− 2 α
G1 (σ, η) � ⎪ − HD HI f(σ),
⎪
⎪
⎪
⎪ (ln σ)α− 1
⎪
⎩ , 1 ≤ σ ≤ η ≤ 2, Γ(α) Γ(α − 1) 2
Γ(α) � c1 (ln σ) + c2 − HI f(σ),
Γ(2) Γ(1)
(ln σ)α− 1 ζ (33)
G2 (σ, η) � φ(σ)Q1 (σ, η)dσ,
Δ 1 in view of H I2 f(σ) ⟶ 0 as σ ⟶ 0, so we must set c2 � 0.
Then, we have
Γ(α − β)(ln σ)α− 1 2
G3 (σ, η) � G1 (σ, η)dμ(σ), σ α− 1
ΔΓ(α) 1 1 σ dη
u(σ) � c1 (ln σ)α− 1 − ln f(η) . (34)
Γ(α) 1 η η
⎪
⎧
⎪
⎪ (ln σ)α− 1− β − (ln(σ/η))α− 1− β
⎪
⎪ , 1 ≤ η ≤ σ ≤ 2, Furthermore, we obtain
⎪
⎪ Γ(α)
⎨ α− 1
Q1 (σ, η) � ⎪ HD u(σ) � c1H Dα− 1 (ln σ)α− 1 − HD
α− 1 α
H I f(σ),
⎪
⎪
⎪
⎪ α− 1− β
⎪ (ln σ)
⎪
⎩ , 1 ≤ σ ≤ η ≤ 2. σ dη
Γ(α) � c1 Γ(α) − f(η) ,
1 η
(29)
β
HD u(σ) � c1H Dβ (ln σ)α− 1 − β α
H D H I f(σ),
Proof. Consider
Γ(α) 1
α � c1 (ln σ)α− 1− β
−
HD u(σ) + f(σ) � 0, Γ(α − β) Γ(α − β)
(30)
α ∈ (2, 3], σ ∈ J.
σ α− 1− β
σ dη
For σ ∈ J, Lemma 1 gives · ln f(η) .
1 η η
u(σ) � c1 (ln σ)α− 1 + c2 (ln σ)α− 2 + c3 (ln σ)α− 3
(35)
σ α− 1 (31) Then, we obtain
1 σ dη
− ln f(η) ,
Γ(α) 1 η η
σ α− 1 (32)
1 σ dη
− ln f(η) .
Γ(α) 1 η η
2 ζ
u(σ)dμ(σ) + φ(σ)H Dβ u(σ)dσ,
1 1
2 Γ(α) ζ
� c1 (ln σ)α− 1 dμ(σ) + c1 φ(σ)(ln σ)α− 1− β
dσ (36)
1 Γ(α − β) 1
2 σ α− 1 ζ σ α− 1− β
1 σ dη 1 σ dη
− ln f(η) dμ(σ) − φ(σ) ln f(η) dσ.
Γ(α) 1 1 η η Γ(α − β) 1 1 η η
Complexity 7
2
1 f(η)(dη/η)
c1 � ζ 2
Γ(α) − (Γ(α)/Γ(α − β)) 1 φ(σ)(ln σ)α− 1− β dσ − 1 (ln σ)α− 1 dμ(σ)
ζ σ
1 φ(σ) 1 (ln(σ/η))α− 1− β
f(η)(dη/η)dσ
− ζ 2
(37)
α− 1− β
Γ(α)Γ(α − β) − Γ(α) 1 φ(σ)(ln σ) dσ − Γ(α − β) 1 (ln σ)α− 1 dμ(σ)
2 σ
1 1 (ln(σ/η))α− 1 f(η)(dη/η)dμ(σ)
− ζ 2
.
Γ2 (α) − Γ2 (α)/Γ(α − β) 1 φ(σ)(ln σ)α− 1− β dσ − Γ(α) 1 (ln σ)α− 1 dμ(σ)
2
(ln σ)α− 1 1 f(η)(dη/η)
u(σ) � ζ 2
Γ(α)Γ(α − β) − Γ(α) 1 φ(σ)(ln σ)α− 1− β dσ − Γ(α − β) 1 (ln σ)α− 1 dμ(σ)
ζ Γ(α − β) 2
× Γ(α − β) − φ(σ)(ln σ)α− 1− β
dσ − (ln σ)α− 1 dμ(σ)
1 Γ(α) 1
ζ α− 1
Γ(α − β) 2 1 σ σ dη
+ φ(σ)(ln σ)α− 1− β
dσ + (ln σ)α− 1 dμ(σ) − ln f(η)
1 Γ(α) 1 Γ(α) 1 η η
ζ σ
(ln σ)α− 1 1 φ(σ) 1 (ln(σ/η))α− 1− β
f(η)(dη/η)dσ
− ζ 2
α− 1− β
Γ(α)Γ(α − β) − Γ(α) 1 φ(σ)(ln σ) dσ − Γ(α − β) 1 (ln σ)α− 1 dμ(σ)
2 σ
(ln σ)α− 1 1 1 (ln(σ/η))α− 1 f(η)(dη/η)dμ(σ)
− ζ 2
,
Γ2 (α) − Γ2 (α)/Γ(α − β) 1 φ(σ)(ln σ)α− 1− β dσ − Γ(α) 1 (ln σ)α− 1 dμ(σ)
(ln σ)α− 1
� ζ 2
Γ(α − β) − 1 φ(σ)(ln σ)α− 1− β dσ − (Γ(α − β)/Γ(α)) 1 (ln σ)α− 1 dμ(σ)
ζ 2 ζ σ α− 1− β
1 ⎛ ⎝ φ(σ)(ln σ)α− 1− β dη σ dη ⎞
× f(η) dσ − φ(σ)ln f(η) dσ ⎠
Γ(α) 1 1 η 1 1 η η
1 2 2 dη
× (ln σ)α− 1 f(η) dμ(σ)
Γ(α) 1 1 η
(ln σ)α− 1
− ζ 2
Γ2 (α)Γ(α − β) − Γ2 (α) 1 φ(σ)(ln σ)α− 1− β dσ − Γ(α)Γ(α − β) 1 (ln σ)α− 1 dμ(σ)
2 σ α− 1 2
σ dη dη
× ln f(η) dμ(σ) + G1 (σ, η)f(η) ,
1 1 η η 1 η
8 Complexity
(ln σ)α− 1
� ζ 2
Γ(α − β) − 1 φ(σ)(ln σ)α− 1− β dσ − (Γ(α − β)/Γ(α)) 1 (ln σ)α− 1 dμ(σ)
ζ σ α− 1− β ζ 2
1 ⎡⎢ ⎝(ln σ)α− 1− β σ ⎠f(η) dη dσ + φ(σ)(ln σ)α− 1− β dη ⎤⎥
× ⎣ φ(σ)⎛ − ln ⎞ f(η) dσ ⎦
Γ(α) 1 1 η η 1 σ η
2 dη (ln σ)α− 1 ζ 2 dη
+ G1 (σ, η)f(η) � φ(σ) Q1 (σ, η)f(η) dσ
1 η Δ 1 1 η
2 dη Γ(α − β)(ln σ)α− 1 2 2 dη
+ G1 (σ, η)f(η) + G1 (σ, η)f(η) dμ(σ),
1 η ΔΓ(α) 1 1 η
2 dη 2 dη
� G1 (σ, η) + G2 (σ, η) + G3 (σ, η)f(η) � G(σ, η)f(η) .
1 η 1 η
By replacing f(σ) � ϕ(σ, u(σ),H Dα u(σ)) in Lemma 3, (1) G(σ, η)⩾0 for all (σ, η) ∈ J
we get the integral equation of problem (5) as (2) G(σ, η) is continuous over J × J
2 2
dη (3) maxσ∈J 1 |G(σ, η)|dη/η ⩽ Y,
u(σ) � G(σ, η)ϕ η, u(η), H Dα u(η) . (39)
1 η where
□
Lemma 4. Green’s function G(σ, η), which is obtained in
(39), has the following properties:
α− 1− β
(ln 2)α− 1 ⎝
⎛αEζ(ln 2) (− 1)α− 1− β− k (α − 1 − β)!
Y� (ln ζ)k
∇Γ(α + 1) k�0
k!
α− β
αEζ β− k (α − β)!
+ (− 1)α− (ln ζ)k
α − β k�0 k!
(40)
α− 1− β
1− β− k (α − 1 − β)! (ln ρ)α− 1 PΓ(α − β)
∇ � Γ(α − β) + Eζ (− 1)α− (ln ζ)k +
k�0
k! Γ(α)
Complexity 9
Proof. It is very easy to prove (1) and (2), so we leave it. (3) Since we know that Green’s function of the con-
sidered problem is in the form
α− 1− β
2 dη (ln σ)α− 1 ⎣ ζ 2
⎡ φ(σ)(ln σ)α− 1− β dη
ζ σ σ dη ⎦⎤
|G(σ, η)| � dσ − φ(σ)ln dσ
1 η ΔΓ(α) 1 1 η 1 1 η η
α− 1
(ln σ)α− 1 Γ(α − β) ⎣⎡ 2 2 dη 2 σ σ dη
+ 2 (ln σ)α− 1 dμ(σ) − ln dμ(σ)⎦⎤
ΔΓ (α) 1 1 η 1 1 η η
α− 1
(ln σ)α− 1 2 dη 1 σ σ dη
+ − ln
Γ(α) 1 η Γ(α) 1 η η
(41)
α− 1− β
(ln σ)α− 1 ⎡ ζ 2
⎣ |φ(σ)|(ln σ) α− 1− β dη ζ σ
σ dη ⎤⎦
≤ dσ + |φ(σ)|ln dσ
η
|Δ|Γ(α) 1 1 η 1 1 η
α− 1
(ln σ)α− 1 Γ(α − β) ⎡⎣ 2 2 2 σ dη
α− 1 dη σ
+ 2 (ln σ) dμ(σ) + ln dμ(σ)⎤⎦
|Δ|Γ (α) 1 1 η 1 1 η η
(ln σ)α− 1 2 dη 1 σ σ
α− 1
dη
+ + ln ,
Γ(α) 1 η Γ(α) 1 η η
α− 1− β
2 dη (ln 2)α− 1 ⎛
⎝αEζ(ln 2) (− 1)α− 1− β− k (α − 1 − β)!
max |G(σ, η)| ≤ (ln ζ)k
σ∈J 1 η ∇Γ(α + 1) k�0
k!
α− β
αEζ β− k (α − β)!
+ (− 1)α− (ln ζ)k (42)
α − β k�0 k!
⎠ � Y.
+ P(ln ρ)α− 1 Γ(α − β)(α(ln 2) +(ln ρ)) + ∇(α(ln 2) +(ln 2))⎞
Hence, the proof of (3) is complete. If u is the solution of the given system (5) and σ ∈ J, then
(ln σ)α− 1 ζ 2 dη
u(σ) � φ(σ)(ln σ)α− 1− β
ϕ(η, u(η), z(η)) dσ
ΔΓ(α) 1 1 η
α− 1− β
(ln σ)α− 1 ζ σ σ dη
− φ(σ)ln ϕ(η, u(η), z(η)) dσ
ΔΓ(α) 1 1 η η
(ln σ)α− 1 Γ(α − β) 2 2 dη
+ 2 (ln σ)α− 1 ϕ(η, u(η), z(η)) dμ(σ) (43)
ΔΓ (α) 1 1 η
α− 1
(ln σ)α− 1 Γ(α − β) 2 σ σ dη
− 2 ln ϕ(η, u(η), z(η)) dμ(σ)
ΔΓ (α) 1 1 η η
α− 1
(ln σ)α− 1 2 dη 1 σ σ dη
+ ϕ(η, u(η), z(η)) − ln ϕ(η, u(η), z(η)) ,
Γ(α) 1 η Γ(α) 1 η η
10 Complexity
α− 1− β
(ln σ)α− 1 ⎡ ζ 2
⎣ φ(σ)(ln σ)α− 1− β dη ζ σ σ dη ⎤⎦
Tu(σ) � − z(η) dσ − φ(σ)ln z(η) dσ
ΔΓ(α) 1 1 η 1 1 η η
α− 1
(ln σ)α− 1 Γ(α − β) ⎣ 2 2 2 σ
⎡ (ln σ)α− 1 z(η) dη dμ(σ) − ln σ dη
− 2 z(η) dμ(σ)⎤⎦ (45)
ΔΓ (α) 1 1 η 1 1 η η
α− 1
(ln σ)α− 1 2 dη 1 σ σ dη
− z(η) + ln z(η) ,
Γ(α) 1 η Γ(α) 1 η η
where z ∈ C(J, R) such that Proof. To show that the operator T is compact, we follow
several steps.
z(σ) � − ϕ(σ, u(σ), z(σ)). (46)
□ Step 1: we consider a sequence un such that un ⟶ u
in X; then, for each σ ∈ J, we have
Theorem 5. Let (M2 ) − (M5 ) hold. Then, the operator T,
defined in (45), is compact.
α− 1 ζ 2
T un (σ) − T(u)(σ) � − (ln σ) φ(σ)(ln σ)α− 1− β
zn (η) − z(η)
dη
dσ
ΔΓ(α) 1 1 η
α− 1− β
(ln σ)α− 1 ζ σ σ dη
+ φ(σ)ln zn (η) − z(η) dσ
ΔΓ(α) 1 1 η η
From (46), we can write T un (σ) − T(u)(σ) ⟶ 0, as n ⟶ ∞. (51)
zn (σ) � − ϕ σ, un (σ), zn (σ),
(48) This implies
z(σ) � − ϕ(σ, u(σ), z(σ)). �� �
��T un − T(u)��� ⟶ 0, as n ⟶ ∞. (52)
X
Now, by (M4 ), we have
Hence, T is continuous.
zn (σ) − z(σ) � − ϕ σ, un (σ), zn (σ) − ϕ(σ, u(σ), z(σ))
Step 2: now, we are going to prove that the operator T is
≤ Lϕ un (σ) − u(σ) + L∗ zn (σ) − z(σ),
ϕ bounded in set X. For this, we show that, for any �p > 0,
(49) there exists �
q > 0 such that, for each
which implies
Lϕ u ∈ Q � u ∈ X : ‖u‖X ≤ �
p, (53)
zn (σ) − z(σ) ≤
∗ un (σ) − u(σ). (50)
1 − Lϕ we have
Since we supposed that un ⟶ u, zn ⟶ z as n ⟶ ∞ ‖T(u)‖X ≤�
q. (54)
for each σ ∈ J. So, by Lebesgue dominated convergence
theorem [36], (47) gives From (45), for each σ ∈ J, we have
α− 1− β
(ln σ)α− 1 ⎡ ζ 2
⎣ φ(σ)(ln σ)α− 1− β dη ζ σ σ dη ⎤⎦
Tu(σ) � − z(η) dσ − φ(σ)ln z(η) dσ
ΔΓ(α) 1 1 η 1 1 η η
α− 1
(ln σ)α− 1 Γ(α − β) ⎣ 2 2 2 σ
⎡ (ln σ)α− 1 z(η) dη dμ(σ) − ln σ dη
− z(η) dμ(σ)⎤⎦ (55)
ΔΓ2 (α) 1 1 η 1 1 η η
α− 1
(ln σ)α− 1 2 dη 1 σ σ dη
− z(η) + ln z(η) .
Γ(α) 1 η Γ(α) 1 η η
α− 1− β
π∗1 + π∗2 �
p ⎡⎢ (ln 2)α− 1 ⎛
⎝αEζ(ln 2) (− 1)α− 1− β− k (α − 1 − β)!
‖Tu‖X ≤ ⎣ (ln ζ)k
1 − π∗3 ∇Γ(α + 1) k�0
k!
α− β
αEζ β− k (α − β)!
+ (− 1)α− (ln ζ)k
α − β k�0 k!
(58)
⎠⎤⎥⎦
+ P(ln ρ)α− 1 Γ(α − β)(α(ln 2) +(ln ρ)) + ∇(α(ln 2) +(ln 2))⎞
π∗1 + π∗2 �
p
≤ Y ��
q.
1 − π∗3
12 Complexity
ζ 2
T(u) σ − T(u) σ � − 1 ln σ α− 1 φ σ ln σ α− 1− β
z(η)
dη
dσ
1 2
ΔΓ(α) 1
1 1
1 1
η 1
ζ 2 dη
α− 1 α− 1− β
− ln σ 2 φ σ 2 ln σ 2 z(η) dσ
1 1 η 2
ζ σ1 α− 1− β
α− 1 σ1 dη
− ln σ 1 φ σ 1 ln z(η) dσ
1 1 η η 1
ζ σ2 α− 1− β
α− 1 σ2 dη
+ ln σ 2 φ σ 2 ln z(η) dσ ⎤⎦
1 1 η η 2
Γ(α − β) 2 2 dη
α− 1 α− 1
− 2 ln σ 1 ln σ 1 z(η) dμ σ 1
ΔΓ (α) 1 1 η
2 2 dη
α− 1 α− 1
− ln σ 2 ln σ 2 z(η) dμ σ 2
1 1 η
2 σ1 α− 1
α− 1 σ dη
− ln σ 1 ln 1 z(η) dμ σ 1
1 1 η η
2 σ2 α− 1 (59)
α− 1 σ2 dη
+ ln σ 2 ln z(η) dμ σ 2 ⎤⎦
1 1 η η
α− 1 α− 1 σ1 2 α− 1
ln σ 1 − ln σ 2 dη 1 σ dη
− + ln 1
z(η) z(η)
Γ(α) 1 η Γ(α) σ 2 η η
dη
σ2 α− 1 α− 1
1 ⎝
⎛ σ2 σ1 ⎠
⎞
− ln − ln z(η)
Γ(α) 1 η η η
α− 1 α− 1 α− 1− β
⎢ ln σ 1 − ln σ 2 1− β− k (α − 1 − β)!
≤⎡
⎣ ⎝αEζ(ln 2) (− 1)α−
⎛ (ln ζ)k
∇Γ(α + 1) k�0
k!
α− β
αEζ β− k (α − β)!
+ (− 1)α− (ln ζ)k
α − β k�0 k!
⎠
+ P(ln ρ)α− 1 Γ(α − β)(α(ln 2) +(ln ρ)) + ∇α(ln 2)⎞
α α
ln σ 1 − ln σ 2 π∗1 + π∗2 �
p
+ .
Γ(α + 1) 1 − π∗3
The right-hand side of (59) approaches to zero as σ 1 ⟶ σ 2 . Theorem 6. Let the hypotheses (M3 ) and (M5 ) hold, and if
Hence, T(Q) is equicontinuous. As a consequence of Step 1 N < 1, the given problem (5) has at least one solution in X.
to 3, the operator T is completely continuous. Therefore, in
view of the Arzelà-Ascoli theorem, the operator T is Proof. For the proof of this theorem, we are considering a set
compact. □ B ⊂ X which is defined in the following form:
Complexity 13
B � {u ∈ X | u � δTu, 0 < δ < 1}. (60) u(σ) � δTu(σ), where 0 < δ < 1. (61)
We have to show that the set B is bounded. Let u ∈ X Then, for each σ ∈ J, we have
such that
(ln σ)α− 1 ζ 2 dη
|u(σ)| � − δ φ(σ)(ln σ)α− 1− β
z(η) dσ
ΔΓ(α) 1 1 η
α− 1− β
(ln σ)α− 1 ζ σ σ dη
− φ(σ)ln z(η) dσ
ΔΓ(α) 1 1 η η
α− 1
(ln σ)α− 1 Γ(α − β) 2 σ σ dη
− 2 ln z(η) dμ(σ)
ΔΓ (α) 1 1 η η
(62)
(ln σ)α− 1 2 dη 1 σ σ
α− 1
dη⎞
⎠
+ z(η) − ln z(η)
Γ(α) 1 η Γ(α) 1 η η
α− 1− β
(ln 2)α− 1 ⎝αEζ(ln 2) (− 1)α− 1− β− k (α − 1 − β)!
≤ |z(η)|⎛ (ln ζ)k
∇Γ(α + 1) k�0
k!
α− β
αEζ β− k (α − β)!
+ (− 1)α− (ln ζ)k
α − β k�0 k!
π1 (σ) + π2 (σ)|u(σ)|
Now, by (M3 ), we have |z(σ)| ≤ . (64)
1 − π3 (σ)
|z(σ)| � | − ϕ(σ, u(σ), z(σ))|
(63) Putting (64) in (62) and taking supσ>1 , we get
≤ π1 (σ) + π2 (σ)|u(σ)| + π3 (σ)|z(σ)|.
So, we get
α− 1− β
(ln 2)α− 1 π∗1 + π∗2 ‖u‖X ⎛ α− 1− β− k (α − 1 − β)!
‖u‖X ≤ ⎝αEζ(ln 2) (− 1) (ln ζ)k
∇Γ(α + 1) 1 − π∗3 k�0
k!
α− β
αEζ β− k (α − β)!
+ (− 1)α− (ln ζ)k (65)
α − β k�0 k!
α− β
αEζ β− k (α − β)!
+ (− 1)α− (ln ζ)k
α − β k�0 k!
⎠,
+ P(ln ρ)α− 1 Γ(α − β)(α(ln 2) +(ln ρ)) + ∇(α(ln 2) +(ln 2))⎞
(66)
α− 1− β
π∗2 (ln 2)α− 1 ⎝αEζ(ln 2) (− 1)α− 1− β− k (α − 1 − β)!
N� ⎛ (ln ζ)k
∇ 1 − π∗3 Γ(α + 1) k�0
k!
α− β
αEζ β− k (α − β)!
+ (− 1)α− (ln ζ)k
α − β k�0 k!
⎠.
+ P(ln ρ)α− 1 Γ(α − β)(α(ln 2) +(ln ρ)) + ∇(α(ln 2) +(ln 2))⎞
So, (65) becomes Theorem 7. Suppose that the hypothesis (M2 ), (M4 ), and
(M5 ) hold. Then, the given problem (5) has a unique solution
‖u‖X ≤ M + N‖u‖X , (67)
in X if
which implies Lϕ
Y < 1. (69)
M 1 − L∗ϕ
‖u‖X ≤ . (68)
1− N
This shows that the set B is bounded. So, by Theorems 3 Proof. We shall use the Banach contraction principle to
and 5, we get that the operator T has at least one fixed point. prove that the operator T has a unique fixed point, which
Therefore, the given problem (5) has at least one solution in will be the unique solution of the given system (5), by
X. □ considering the operator T: X ⟶ X defined in (45).
Let u, u be the solution of (5), and for σ ∈ J, we have
(ln σ)α− 1 ζ 2 dη
|T(u)(σ) − T(u)(σ)| ≤ |φ(σ)|(ln σ)α− 1− β
|z(η) − z(η)| dσ
|Δ|Γ(α) 1 1 η
α− 1− β
(ln σ)α− 1 ζ σ σ dη
+ |φ(σ)|ln |z(η) − z(η)| dσ
|Δ|Γ(α) 1 1 η η
α− 1
(ln σ)α− 1 Γ(α − β) 2 σ σ dη
+
ln |z(η) − z(η)| dμ(σ)
|Δ|Γ2 (α) 1 1 η η
1
(ln σ)α− 1 2 dη 1 σ σ
α−
dη
+ |z(η) − z(η)| + ln |z(η) − z(η)| ,
Γ(α) 1 η Γ(α) 1 η η
Complexity 15
Lϕ (ln σ)α− 1 ζ 2 dη
|T(u)(σ) − T(u)(σ)| ≤ |φ(σ)|(ln σ)α− 1− β
|u(η) − u(η)| dσ
1− L∗ϕ |Δ|Γ(α) 1 1 η
α− 1− β
(ln σ)α− 1 ζ σ σ dη
+ |φ(σ)|ln |u(η) − u(η)| dσ
|Δ|Γ(α) 1 1 η η
α− 1
(ln σ)α− 1 Γ(α − β) 2 σ σ dη
+
ln |u(η) − u(η)| dμ(σ)
|Δ|Γ2 (α) 1 1 η η
1
(ln σ)α− 1 2 dη 1 σ σ
α−
dη
+ |u(η) − u(η)| + ln |u(η) − u(η)| ⎤⎦.
Γ(α) 1 η Γ(α) 1 η η
α− 1− β
Lϕ ⎡ ⎢ (ln 2)α− 1 ⎛
⎝αEζ(ln 2) (− 1)α− 1− β− k (α − 1 − β)!
‖T(u) − T(u)‖X ≤ ∗
⎣ (ln ζ)k
1 − Lϕ ∇Γ(α + 1) k�0
k!
α− β
αEζ β− k (α − β)!
+ (− 1)α− (ln ζ)k (75)
α − β k�0 k!
⎠⎤⎥⎦‖u − u‖ .
+ P(ln ρ)α− 1 Γ(α − β)(α(ln 2) +(ln ρ)) + ∇(α(ln 2) +(ln 2))⎞ X
This implies that fixed point, which is a unique solution of the given problem
(5). □
Lϕ
‖T(u) − T(u)‖X ≤ Y‖u − u‖X . (76)
1 − L∗ϕ
3.2. Existence and Uniqueness Solution for System (7). In this
Hence, the operator T is a contraction. Thus, by the section, we show the existence and uniqueness of the so-
Banach contraction principle, we get that T has a unique lution of the system (7). First, we have the following:
16 Complexity
⎪
⎧
⎪ (ln σ)α− 1 − (ln(σ/η))α− 1
⎪
⎪ , 1 ≤ η ≤ σ ≤ 2,
⎪
⎨ Γ(α)
G1 (σ, η) � ⎪
⎪
⎪
⎪
⎪ (ln σ)α− 1
⎩ , 1 ≤ σ ≤ η ≤ 2,
Γ(α)
(ln σ)α− 1 ζ
G2 (σ, η) � φα (σ)Q1 (σ, η)dσ,
Δα 1
⎪
⎧
⎪
⎪ (ln σ) − (ln(σ/η))c− 1
c− 1
⎪
⎪ , 1 ≤ η ≤ σ ≤ 2,
⎪
⎨ Γ(c)
G4 (σ, η) � ⎪
⎪
⎪
⎪
⎪ (ln σ)c− 1
⎪
⎩ , 1 ≤ σ ≤ η ≤ 2, (79)
Γ(c)
1
(ln σ)c− ξ
G5 (σ, η) � φc (σ)Q1 (σ, η)dσ,
Δc 1
1
Γ(c − δ)(ln σ)c− 2
G6 (σ, η) � G4 (σ, η)dμ(σ),
Δc Γ(c) 1
⎧
⎪
⎪
⎪ (ln σ)α− 1− β − (ln(σ/η))α− 1− β
⎪
⎪ , 1 ≤ η ≤ σ ≤ 2,
⎪
⎨ Γ(α)
Q1 (σ, η) � ⎪
⎪
⎪ α− 1− β
⎪
⎪ (ln σ)
⎪
⎩ , 1 ≤ σ ≤ η ≤ 2,
Γ(α)
⎪
⎧
⎪
⎪ (ln σ)c− 1− δ − (ln(σ/η))c− 1− δ
⎪
⎪ , 1 ≤ η ≤ σ ≤ 2,
⎪
⎨ Γ(c)
Q2 (σ, η) � ⎪
⎪
⎪
⎪
⎪
⎪ (ln σ)c− 1− δ
⎩ , 1 ≤ σ ≤ η ≤ 2.
Γ(c)
Complexity 17
Proof. The proof is similar to that given in Lemma 3 and, where y, z ∈ X satisfying the functional equation. □
hence, is not included here.We use the following notations
for convenience:
Lemma 6. Green’s function Gα,c (σ, η) � (Gα (σ, η), Gc
y(σ) ≔ − ϕ1 σ, u(σ),H Dc v(σ) � − ϕ1 (σ, u(σ), z(σ)), (σ, η)) of the system (7) have the following properties:
z(σ) ≔ − ϕ2 σ, v(σ),H Dα u(σ) � − ϕ2 (σ, v(σ), y(σ)). (1 )Gα,c (σ, η) is continuous over J × J
2 2
(80) (2) maxσ∈J 1 |Gα (σ, η)|dη/ η ⩽ Yα and maxσ∈J 1 |Gc
(σ, η)|dη/η ⩽ Yc ,
Hence, for σ ∈ J, (78) becomes
2
where
⎪
⎧
⎪ dη
⎪
⎪ u(σ) � Gα (σ, η)y(η) ,
⎪
⎪ 1 η
⎨
⎪ (81)
⎪
⎪
⎪
⎪ 2 dη
⎪
⎩ v(σ) � Gc (σ, η)z(η) ,
1 η
α− 1− β
(ln 2)α− 1 ⎛
⎝αE ζ(ln 2) (− 1)α− 1− β− k (α − 1 − β)!
Yα � α (ln ζ)k
∇α Γ(α + 1) k�0
k!
α− β
αEα ζ β− k (α − β)!
+ (− 1)α− (ln ζ)k
α − β k�0 k!
c− 1− δ
(ln 2)c− 1 ⎝
⎛cEc ξ(ln 2) (− 1)c− 1− δ− k (c − 1 − δ)!
Yc � (ln ξ)k
∇c Γ(c + 1) k�0
k!
(82)
c− δ
cEc ξ δ− k (c − δ)!
+ (− 1)c− (ln ξ)k
c − δ k�0 k!
α− 1− β
1− β− k (α − 1 − β)! (ln ρ)α− 1 PΓ(α − β)
∇α � Γ(α − β) + Eα ζ (− 1)α− (ln ζ)k + ,
k�0
k! Γ(α)
c− 1− δ
1− δ− k (c − 1 − δ)! (ln ρ)c− 1 PΓ(c − δ)
∇c � Γ(c − δ) + Ec ξ (− 1)c− (ln ξ)k + .
k�0
k! Γ(c)
Proof. (1) It is easy to prove that Gα,c (σ, η) is continuous, so (2) Since we know that Green’s functions of the con-
we leave it. sidered problem (7) is in the form
18 Complexity
2
dη (ln σ)α− 1 ζ 2 1− β dη
ζ σ σ
α− 1− β
dη ⎤⎦
Gα (σ, η) � ⎣ φα (σ)(ln σ)α−
⎡ dσ − φα (σ)ln dσ
1 η Δα Γ(α) 1 1 η 1 1 η η
α− 1
(ln σ)α− 1 Γ(α − β) ⎡⎣ 2 2 dη 2 σ σ dη
+ 2 (ln σ)α− 1 dμ(σ) − ln dμ(σ)⎤⎦
Δα Γ (α) 1 1 η 1 1 η η
α− 1
(ln σ)α− 1
dη 21 σ σ dη
+ − ln
Γ(α) 1 η Γ(α) 1 η η
(83)
1− β
(ln σ) α− 1 ζ 2
β dη
ζ
σ σ
α−
dη ⎤⎦
≤ ⎣ φα (σ)(ln σ)α−
⎡ 1−
dσ + φα (σ)ln dσ
Δα Γ(α) 1 1 η 1 1 η η
1
(ln σ)α− 1 Γ(α − β) ⎣⎡ 2 2 2 σ α−
dη
α− 1 dη σ
+ 2 (ln σ) dμ(σ) + ln dμ(σ)⎤⎦
Δα Γ (α) 1 1 η 1 1 η η
1
(ln σ)α− 1
dη 21 σ σ
α−
dη
+ + ln ,
η
Γ(α) 1 η Γ(α) 1 η
α− 1− β
2
dη (ln 2)α− 1 1− β− k (α − 1 − β)!
max Gα (σ, η) ≤ ⎝αEζ(ln 2) (− 1)α−
⎛ (ln ζ)k
σ∈J 1 η ∇α Γ(α + 1) k�0
k!
α− β (84)
αEζ β− k (α − β)!
+ (− 1)α− (ln ζ)k
α − β k�0 k!
2 dη (ln 2)c− 1 c− 1− δ
1− δ− k (c − 1 − δ)!
maxσ∈J Gc (σ, η) ≤ ⎝cEξ(ln 2) (− 1)c−
⎛ (ln ξ)k
1 η ∇c Γ(c + 1) k�0
k!
c− δ
cEξ δ− k (c − δ)!
+ (− 1)c− (ln ξ)k (85)
c − δ k�0 k!
Hence, the proof of 2 is complete. If u, v are the solutions of the given system (7) and σ ∈ J;
then,
(lnσ)α− 1 ζ 2 dη
u(σ) � φ (σ)(lnσ)α− 1− β
ϕ1 (η, u(η), z(η)) dσ
Δα Γ(α) 1 1 α η
α− 1− β
(lnσ)α− 1 ζ σ σ dη
− φ (σ)ln ϕ1 (η, u(η), z(η)) dσ
Δα Γ(α) 1 1 α η η
(lnσ)α− 1 Γ(α − β) 2 2 dη
+ 2 (lnσ)α− 1 ϕ1 (η, u(η), z(η)) dμ(σ)
Δα Γ (α) 1 1 η
α− 1
(lnσ)α− 1 Γ(α − β) 2 σ σ dη
− 2 ln ϕ1 (η, u(η), z(η)) dμ(σ)
Δα Γ (α) 1 1 η η
α− 1
(lnσ)α− 1 2 dη 1 σ σ dη
+ ϕ1 (η, u(η), z(η)) − ln ϕ1 (η, u(η), z(η)) ,
Γ(α) 1 η Γ(α) 1 η η
(86)
(lnσ)c− 1 ξ 2 1− δ dη
v(σ) � φ (σ)(lnσ)c− ϕ2 (η, v(η), y(η)) dσ
Δc Γ(c) 1 1 c η
c− 1− δ
(lnσ)c− 1 ξ σ σ dη
− φc (σ)ln ϕ2 (η, v(η), y(η)) dσ
Δc Γ(c) 1 1 η η
(lnσ)c− 1 Γ(c − δ) 2 2 dη
+ 2 (lnσ)c− 1 ϕ2 (η, v(η), y(η)) dμ(σ)
Δc Γ (c) 1 1 η
c− 1
(lnσ)c− 1 Γ(c − δ) 2 σ σ dη
− 2 ln ϕ2 (η, v(η), y(η)) dμ(σ)
Δc Γ (c) 1 1 η η
1 c− 1
(lnσ)c− 2 dη 1 σ σ dη
+ ϕ2 (η, v(η), y(η)) − ln ϕ(η, v(η), y(η)) .
Γ(c) 1 η Γ(c) 1 η η
2 dη
Gα (σ, η)ϕ1 (η, u(η), z(η)) ⎞
⎜
⎛
⎜
⎜
⎜
⎜ 1 η ⎟
⎟
⎟
⎟
⎟ Tα (u, z)(σ)
⎜
⎜ ⎟
⎟
⎟ ⎜ ⎟
⎜
T(u, v)(σ) � ⎜
⎜ ⎟
⎟ � ⎛
⎝ ⎞
⎠. (87)
⎜
⎜ ⎟
⎟
⎟
⎜
⎜
⎜ ⎟
⎟ T (v, y)(σ)
⎝ 2 dη ⎠ c
Gc (σ, η)ϕ2 (η, v(η), y(η))
1 η
20 Complexity
α− 1− β
(lnσ)α− 1 ⎣ ζ 2
⎡ φα (σ)(lnσ)α− 1− β dη ζ σ σ dη ⎦⎤
Tα (u, z)(σ) � − y(η) dσ − φα (σ)ln y(η) dσ
Δα Γ(α) 1 1 η 1 1 η η
α− 1
(lnσ)α− 1 Γ(α − β) ⎣ 2 2 2 σ
⎡ (lnσ)α− 1 y(η) dη dμ(σ) − ln σ dη
− 2 y(η) dμ(σ)⎤⎦
Δα Γ (α) 1 1 η 1 1 η η
α− 1
(lnσ)α− 1 2 dη 1 σ σ dη
− y(η) + ln y(η) ,
Γ(α) 1 η Γ(α) 1 η η
(88)
c− 1 ξ 2 ξ σ c− 1− δ
(lnσ) ⎣ ⎡ φc (σ)(lnσ)c− 1− δ dη σ dη ⎤⎦
Tc (v, y)(σ) � − z(η) dσ − φc (σ)ln z(η) dσ
Δc Γ(c) 1 1 η 1 1 η η
c− 1
(lnσ)c− 1 Γ(c − δ) ⎣ 2 2 2 σ
⎡ (lnσ)c− 1 z(η) dη dμ(σ) − ln σ dη
− 2 z(η) dμ(σ)⎦⎤
Δc Γ (c) 1 1 η 1 1 η η
1 c− 1
(lnσ)c− 2 dη 1 σ σ dη
− z(η) + ln z(η) .
Γ(c) 1 η Γ(c) 1 η η
□
Theorem 8. Let ϕ1 , ϕ2 : J × R ×� R ⟶� R and (M9 ), (M11 ) Proof. In view of continuity of ϕ1�, ϕ2 and Gα,c (σ, η),
� T is
hold. Then, the operator T: C ⟶ C defined in (87) is also continuous for all (u, z) ∈ C. Suppose B⊆C � is a
completely continuous. � we have
bounded set. So, for every u ∈ B,
(lnσ)α− 1 ζ 2 α− 1− β dη ζ σ σ
α− 1− β
dη ⎤⎦
Tα (u, v)(σ) � − ⎣
⎡ y(η) y(η)
Δα Γ(α) 1 1 φα (σ)(lnσ) η
dσ − φα (σ)ln
1 1 η η
dσ
α− 1
(lnσ)α− 1 Γ(α − β) ⎡ 2 2 2 σ
⎣ (lnσ)α− 1 y(η) dη dμ(σ) − ln σ dη
− y(η) dμ(σ)⎤⎦
Δα Γ2 (α) 1 1 η 1 1 η η
α− 1
(lnσ)α− 1
dη2 1 σ σ dη
− y(η) + ln y(η)
Γ(α) 1 η Γ(α) 1 η η
(89)
1− β
(lnσ) ⎡ α− 1 ζ 2 dη ζ σσ
α−
dη
⩽ ⎣ φα (σ)(lnσ)α− 1− β
|y(η)| dσ + φα (σ)ln y(η) dσ ⎤⎦
Δα Γ(α) 1 1 η 1 1 η η
1
(lnσ)α− 1 Γ(α − β) ⎡ 2 2 2 σ α−
+ 2 ⎣ (lnσ) |y(η)| dμ(σ) + ln σ
α− 1 dη dη
|y(η)| dμ(σ)⎤⎦
Δα Γ (α) 1 1 η 1 1 η η
1
(lnσ)α− 1 2 dη 1 σ σ
α−
dη
+ |y(η)| + ln |y(η)| .
Γ(α) 1 η Γ(α) 1 η η
Complexity 21
|y(σ)| � − ϕ1 (σ, u(σ), z(σ)) ⩽ π4 (σ) + π5 (σ)|u(σ)| + π6 (σ)|z(σ)|,
� π4 (σ) + π5 (σ)u(σ) + π6 (σ)ϕ2 (σ, v(σ), y(σ)) (90)
So, we obtain Now, using 2 of Lemma 6, (M11 ), and (91) in (89), we get
π4 (σ) + π6 (σ)π7 (σ) π5 (σ)|u(σ)| + π 6 (σ)π8 (σ)|v(σ)|
|y(σ)| ⩽ + .
1 − π6 (σ)π9 (σ) 1 − π6 (σ)π9 (σ)
(91)
�� �
��Tα (u, v)��� π∗4 + π∗6 π∗7 π∗5 ‖u‖X + π∗6 π∗8 ‖v‖X
⩽ +
X×X 1 − π∗6 π∗9 1 − π∗6 π∗9
α− 1− β
⎢ (ln 2)α− 1 ⎛ 1− β− k (α − 1 − β)!
×⎡
⎣ ⎝αE ζ(ln 2) (− 1)α−
α (lnζ)k ⎞
⎠
∇α Γ(α + 1) k�0
k!
α− β (92)
αEα ζ β− k (α − β)!
+ (− 1)α− (lnζ)k
α − β k�0 k!
π∗ + π∗ π∗ π∗ ‖u‖ + π∗ π∗ ‖v‖
⩽ 4 ∗6 ∗7 + 5 X ∗6 ∗8 X Yα .
1 − π6 π9 1 − π6 π9
In the same way, we obtain Thus, from (92) and (93), we get
�� � π∗7 + π∗4 π∗9 π∗5 π∗9 ‖u‖X + π∗8 ‖v‖X
��T (u, v)��� ⩽ + Yc .
� c �X×X 1 − π∗ π∗ 1 − π∗ π∗
6 9 6 9
(93)
�� �� �� ��
‖T(u, v)‖X×X ⩽ ��Tα (u, v)��X×X + ���Tc (u, v)���
X×X
(94)
π∗4 + π∗6 π∗7 π∗5 ‖u‖X + π∗6 π∗8 ‖v‖X π∗7 + π∗4 π∗9 π∗5 π∗9 ‖u‖X + π∗8 ‖v‖X
⩽ + Y + + Yc � M.
1 − π∗6 π∗9 1 − π∗6 π∗9 α
1 − π∗6 π∗9 1 − π∗6 π∗9
22 Complexity
Tα (u, v) σ 1 − Tα (u, v) σ 2 ,
ln σ α− 1 ζ 2 dη
� − 1 α− 1− β
φα σ 1 ln σ 1 y(η) dσ
Δα Γ(α) 1 1 η 1
α− 1
ζ 2
ln σ 2 α− 1− β dη
+ φα σ 2 ln σ 2 y(η) dσ
Δα Γ(α) 1 1 η 2
α− 1
ζ σ2 α− 1− β
ln σ 2 σ dη
− φα σ 2 ln 2 y(η) dσ
Δα Γ(α) 1 1 η η 2
α− 1
ζ σ1 α− 1− β
ln σ 1 σ dη
+ φα σ 1 ln 1 y(η) dσ
Δα Γ(α) 1 1 η η 1
α− 1
ln σ 1 Γ(α − β) 2 2 α− 1 dη
− 2 ln σ 1 y(η) dμ σ 1
Δα Γ (α) 1 1 η
α− 1
ln σ 2 Γ(α − β) 2 2 α− 1 dη
+ ln σ 2 y(η) dμ σ 2
Δα Γ2 (α) 1 1 η
α− 1 α− 1 (95)
ln σ 2 Γ(α − β) 2 σ 2 σ dη
− 2 ln 2 y(η) dμ σ 2
Δα Γ (α) 1 1 η η
α− 1 α− 1
ln σ 1 Γ(α − β) 2 σ 1 σ dη
+ 2 ln 1 y(η) dμ σ 1
Δα Γ (α) 1 1 η η
α− 1 α− 1 2 σ1 α− 1
ln σ 1 − ln σ 2 dη 1 σ dη
− y(η) + ln 1 y(η)
Γ(α) 1 η Γ(α) σ 2 η η
1 σ2 α− 1 α− 1
− ⎛ ⎝ln σ 2 σ
− ln 1 ⎞ dη
⎠z(η)
Γ(α) 1 η η η
α− 1 α− 1
π∗ + π∗ π∗ π∗ ‖u‖ + π∗ π∗ ‖v‖ ln σ 1 − ln σ 2
≤ 4 ∗6 ∗7 + 5 X ∗6 ∗8 X ⎡⎢⎣
1 − π6 π9 1 − π6 π9 ∇Γ(α + 1)
α− 1− β α− β
⎝αEζ(ln 2) (− 1)α− 1− β− k (α − 1 − β)! αEζ (α − β)!
×⎛ (ln ζ)k + (− 1)α− β− k
(ln ζ)k
k�0
k! α − β k�0 k!
α α
ln σ 1 − ln σ 2
+ P(ln ρ)α− 1 Γ(α − β)(α(ln 2) +(ln ρ)) + ∇α(ln 2) + .
Γ(α + 1)
Complexity 23
c− 1 c− 1
π∗ + π∗ π∗ π∗ π∗ ‖u‖ + π∗ ‖v‖ lnσ 1 − lnσ 2
⩽ 7 ∗4 ∗9 + 5 9 X ∗ ∗8 X ⎡⎢⎣
1 − π6 π9 1 − π6 π9 ∇Γ(c + 1)
c− 1− δ c− δ (96)
⎝cEξ(ln 2) (− 1)c− 1− δ− k (c − 1 − δ)! cEξ δ− k (c − δ)!
×⎛ (lnξ)k + (− 1)c− (lnξ)k
k�0
k! c − δ k�0 k!
c c
lnσ 1 − lnσ 2
+ P(lnρ)c− 1 Γ(c − δ)(c(ln 2) +(lnρ)) + ∇c(ln 2) + .
Γ(c + 1)
α− 1 ζ 2
Tα (u, v)(σ) − Tα (u, v)(σ) ⩽ (lnσ) α− 1− β dη
Δα Γ(α) 1 1 φα (σ)(lnσ) η
dσ
α− 1− β
(lnσ)α− 1 ζ σ σ dη
+ φ (σ)ln dσ
Δα Γ(α) 1 1 α η η
(lnσ)α− 1 Γ(α − β) 2 2 dη
+ (lnσ)α− 1 dμ(σ) (98)
Δα Γ2 (α) 1 1 η
α− 1
(lnσ)α− 1 Γ(α − β) 2 σ σ dη
+ 2 ln dμ(σ)
Δα Γ (α) 1 1 η η
α− 1
(lnσ)α− 1 2 dη 1 σ σ dη
+ + ln |y(η) − y
(η)|,
Γ(α) 1 η Γ(α) 1 η η
|y(σ) − y(σ)| � − ϕ1 (σ, u(σ), z(σ) − ϕ1 (σ, u(σ), z(σ))
Substituting (100) in (98) and taking supσ∈J , we get �� � Yc Lϕ1 L∗ϕ2 + Lϕ2
��T (u, v) − T (u, v)��� ⩽ ⎢
⎡
⎣ ⎤⎥⎦
� c c �X×X 1 − L∗ L ∗
�� � Y L + L∗ L
��Tα (u, v) − Tα (u, v)��� ⩽ ⎡⎢⎣ α ϕ1 ∗ ϕ∗1 ϕ2 ⎤⎥⎦ ϕ1 ϕ2 (102)
X×X 1 − Lϕ1 Lϕ2 (101) · ‖(u, v) − (u, v
)‖X×X .
· ‖(u, v) − (u, v)‖X×X .
So, from (101) and (102), we get
In the same way, we can obtain
Thus, T is a contraction. Therefore, by the Banach Theorem 10. In view of the continuity of the functions ϕ1 , ϕ2
contraction principle, T has a fixed point. So, we infer that and supposing (M9 ) and (M13 ) with
the given coupled system (7) has a unique solution. □
hold, the coupled system (7) has at least one solution. where max2N∗11 /1 − 2N∗12 , 2N∗21 /1 − 2N∗22 < RB
� . Further-
�
� be defined as more, the operator defined by T: B � ⟶ C in (87) is
Proof. Let a set B
� then, by defi-
completely continuous. Suppose (u, v) ∈ B;
� � (u, v) ∈ X × X: ‖(u, v)‖X×X < R � ,
B (105)
B � we have ‖(u, v)‖X×X < R � ;
nition of B, B
�� � 2 dη
��Tα (u, z)���
Gα (σ, η)ϕ1 (η, u(η), z(η))
X×X ⩽ max
σ∈J η
1
2
π∗ + π∗ π∗ dη 2
π∗5 + π∗6 π∗8 dη
⩽ Gα (2, η) 4 ∗6 ∗7 + RB
� Gα (2, η) ,
1 1 − π6 π9 η 1 1 − π∗6 π∗9 η
RB �
� N∗11 + RB ∗
� N12 ⩽ .
2
Similarly, Therefore,
�� � RB
��T (v, y)��� �
(107) ‖T(u, v)‖X×X ⩽ RB
�, (108)
� c �X×X ⩽ 2 .
Complexity 25
�� �� 2
dη
‖u‖X � ��δTα (u, z)��X ⩽ maxσ∈J Gα (σ, η)ϕ1 (η, u(η), z(η))
1 η
2
π (η)|u(η)| + π6 (η)π8 (η)|v(η)| dη
+ maxσ∈J Gα (σ, η) 5 (110)
1 1 − π6 (η)π9 (η) η
2
π∗ + π∗ π∗ dη 2
∗ ∗ ∗
π5 + π6 π8 dη
⩽ Gα (2, η) 4 ∗6 ∗7 + RB� Gα (2, η) ∗ ∗ ,
1 1 − π6 π9 η 1 1 − π6 π9 η
RB �
� N∗11 + RB ∗
� N12 ⩽ .
2
Lemma 7. Let 2 < α ⩽ 3; if u ∈ C1 (J, R) is the solution of the So, for σ ∈ J, the solution of (120) will be in the form
inequality (113), then u will be the solution of the following 2 dη
integral inequality: u(σ) � G(σ, η)ϕ η, u(η),H Dα u(η)
1 η
u(σ) − ω1 (σ) ⩽ Yϵ. (119) (121)
2 dη
+ G(σ, η)Uϕ (η) .
1 η
Proof. Let u be the solution of the inequality (113). So, in
view 2 of Remark 2, we have From equation (121), we have
α α
⎪
⎧
⎪ H D u(σ) + ϕ σ, u(σ), H D u(σ) � Uϕ (σ), σ ∈ J,
⎪
⎪
⎪
⎨ u(σ)| � Dα− 2 u(σ)| � 0,
σ�1 H σ�1
⎪
⎪
⎪
⎪ ζ
⎪
⎩ H Dα− 1 u(σ)|σ�2 � α[u] + φ(σ)H Dβ u(σ)dσ.
1
(120)
α− 1− β
(lnσ)α− 1 ⎡ ζ 2
⎣ φ(σ)(lnσ)α− 1− β dη ζ σ σ dη ⎤⎦
u(σ) � − z(η) dσ − φ(σ)ln z(η) dσ
ΔΓ(α) 1 1 η 1 1 η η
α− 1
(lnσ)α− 1 Γ(α − β) ⎡ 2 2 2 σ
⎣ (lnσ)α− 1 z(η) dη dμ(σ) − ln σ dη
− 2 z(η) dμ(σ)⎤⎦
ΔΓ (α) 1 1 η 1 1 η η
α− 1
(lnσ)α− 1 2 dη 1 σ σ dη
− z(η) + ln z(η)
Γ(α) 1 η Γ(α) 1 η η
(122)
α− 1 ζ 2 ζ σ α− 1− β
(lnσ) ⎣⎡ φ(σ)(lnσ)α− 1− β dη σ dη ⎤⎦
+ Uϕ (η) dσ − φ(σ)ln Uϕ (η) dσ
ΔΓ(α) 1 1 η 1 1 η η
α− 1
(lnσ)α− 1 Γ(α − β) ⎣ 2 2 2 σ
⎡ (lnσ)α− 1 Uϕ (η) dη dμ(σ) − ln σ dη
+ 2 Uϕ (η) dμ(σ)⎤⎦
ΔΓ (α) 1 1 η 1 1 η η
α− 1
(lnσ)α− 1 2 dη 1 σ σ dη
+ Uϕ (η) − ln Uϕ (η) .
Γ(α) 1 η Γ(α) 1 η η
α− 1− β
(lnσ)α− 1 ⎡ ζ 2
⎣ φ(σ)(lnσ)α− 1− β dη ζ σ σ dη ⎤⎦
ω1 (σ) � − z(η) dσ − φ(σ)ln z(η) dσ
ΔΓ(α) 1 1 η 1 1 η η
α− 1
(lnσ)α− 1 Γ(α − β) ⎡ 2 2 2 σ
⎣ (lnσ)α− 1 z(η) dη dμ(σ) − ln σ dη
− 2 z(η) dμ(σ)⎤⎦ (123)
ΔΓ (α) 1 1 η 1 1 η η
α− 1
(lnσ)α− 1 2 dη 1 σ σ dη
− z(η) + ln z(η) .
Γ(α) 1 η Γ(α) 1 η η
Complexity 27
(lnσ)α− 1 Γ(α − β) 2 2 dη
+ 2 (lnσ)α− 1 Uϕ (η) dμ(σ)
|Δ|Γ (α) 1 1 η
α− 1
(lnσ)α− 1 Γ(α − β) 2 σ σ dη
+
ln Uϕ (η) dμ(σ)
|Δ|Γ2 (α) 1 1 η η
(124)
1
(lnσ)α− 1 2 dη 1 σ σ
α−
dη
+ Uϕ (η) + ln Uϕ (η)
Γ(α) 1 η Γ(α) 1 η η
α− 1− β
⎢ (ln 2)α− 1 ⎝ 1− β− k (α − 1 − β)!
⩽⎡
⎣ ⎛αEζ(ln 2) (− 1)α− (lnζ)k
∇Γ(α + 1) k�0
k!
α− β
αEζ β− k (α − β)!
+ (− 1)α− (lnζ)k
α − β k�0 k!
Consider
holds, then the given system (5) is HUS and, consequently,
GHUS. |u(σ) − u(σ)| ⩽ u(σ) − ω1 (σ) + ω1 (σ) − u(σ). (129)
Proof. Let u ∈ C(J, R) be the solution of (113) and u be the Using Lemma 7 in (129), we have
unique solution of the system given by
|u(σ) − u(σ)|
α− 1− β
(lnσ)α− 1 ζ 2 (lnσ)α− 1 ζ σ
1− β dη σ dη
⩽ Yϵ + φ(σ)(lnσ)α− dσ + φ(σ)ln dσ
ΔΓ(α) 1 1 η ΔΓ(α) 1 1 η η
α− 1 (130)
(lnσ)α− 1 Γ(α − β) 2 2 α− 1 dη (lnσ)α− 1 Γ(α − β) 2 σ σ dη
+ 2 (lnσ) dμ(σ) + 2 ln dμ(σ)
ΔΓ (α) 1 1 η ΔΓ (α) 1 1 η η
α− 1
(lnσ)α− 1 2dη 1 σ σ dη
+ + ln zu (η) − zu (η),
Γ(α) 1 η Γ(α) 1 η η
28 Complexity
α− 1 ζ 2
u(σ) − ω1 (σ) ⩽ (lnσ) dη
|φ(σ)|(lnσ)α− 1− β Uϕ (η) dσ
|Δ|Γ(α) 1 1 η
α− 1− β
(lnσ)α− 1 ζ σ σ dη
+
|φ(σ)|ln Uϕ (η) dσ
|Δ|Γ(α) 1 1 η
η
(lnσ)α− 1 Γ(α − β) 2 2 dη
+ (lnσ)α− 1 Uϕ (η) dμ(σ) (138)
|Δ|Γ2 (α) 1 1 η
α− 1
(lnσ)α− 1 Γ(α − β) 2 σ σ dη
+ 2 ln Uϕ (η) dμ(σ)
|Δ|Γ (α) 1 1 η η
α− 1
(lnσ)α− 1 2 dη 1 σ σ dη
+ Uϕ (η) + ln Uϕ (η) .
Γ(α) 1 η Γ(α) 1 η η
α− 1− β
⎢ (ln 2)α− 1 (α − 1 − β)!
u(σ) − ω1 (σ) ⩽ ⎡
⎣ ⎝αEζ(ln 2) (− 1)α−
⎛ 1− β− k
(lnζ)k
∇Γ(α + 1) k�0
k!
α− β
αEζ β− k (α − β)!
+ (− 1)α− (lnζ)k (139)
α − β k�0 k!
⩽ YΩΨ Ψ(σ)ϵ.
□
Complexity 29
Theorem 12. Under the hypothesis (M2 ) and (M4 )–(M6 ) 4.2. Hyers–Ulam Stability Concepts for System (7)
and if the inequality
Definition 7 (see [38]). The given system (7) has HUS if
1 − L∗ϕ − YLϕ
>0 (140) there exists Cα,c � max(Cα , Cc ) > 0 such that there exist
1 − L∗ϕ some εα,c � max(εα , εc ) > 0, and for every solution
(u, v) ∈ X × X of the inequality
holds, then the given system (5) is stable in the sense of HUR.
α c
⎨ H D u(σ) + ϕ1 σ, u(σ), H D v(σ) ⩽ ϵα , σ ∈ J,
⎪
⎧
⎪
Proof. Let u ∈ C(J, R) be the solution of (113) and u be the ⎩ Dc v(σ) + ϕ σ, v(σ), Dα u(σ) ⩽ ϵ , σ ∈ J,
H 2 H c
unique solution of the system given by
(149)
α
⎪
⎧
⎪ HD u (σ) + ϕ σ, u(σ), H Dα u
(σ) � 0, σ ∈ J,
⎪
⎪ there exists a solution (u, v) ∈ X × X with
⎪
⎪
⎪
⎨ u(σ)| � Dα− 2 u(σ)| � 0,
σ�1 H σ�1 (141) ‖(u, v)(σ) − (u, v)(σ)‖ ⩽ Cα,c ϵα,c , σ ∈ J. (150)
⎪
⎪
⎪
⎪
⎪
⎪ ζ
⎪
⎩ H Dα− 1 u(σ)|σ�2 � α[u] + φ(σ)H Dβ u(σ)dσ.
1 Definition 8 (see [38]). The given system (7) has GHUS if
there exists C′ ∈ C(R+ , R+ ) with C′ (0) � 0 such that, for
Then, for σ ∈ J, the solution of (141) is
any solution (u, v) ∈ X × X of the inequality (149), there
2 exists a solution (u, v) ∈ X × X of (7) satisfying
dη
u(σ) � G(σ, η)ϕ η, u(η),H Dα u(η) . (142)
1 η ‖(u, v)(σ) − (u, v)(σ)‖ ⩽ C′ ϵα,c , σ ∈ J. (151)
Consider
Definition 9 (see [38]). The given system (7) has HURS
|u(σ) − u(σ)| ⩽ u(σ) − ω1 (σ) + ω1 (σ) − u(σ). (143) with respect to Ψα,c � max(Ψα , Ψc ) with Ψα,c ∈ C(J, R) if
there exists some constant CΨα,c � max(CΨα , CΨc ) > 0 such
Using (M4 ) in a similar way as used in Theorem 11, we that, for any εα,c � max(εα , εc ) > 0 and for any solution
get (u, v) ∈ X × X of the inequalities
YLϕ
ω1 (σ) − u(σ) ⩽ |u(η) − u
(η)|. (144) ⎧
α c
⎨ H D u(σ) + ϕ1 σ, u(σ), H D v(σ) ⩽ Ψα (σ)ϵα , σ ∈ J,
1 − L∗ϕ ⎩ Dc v(σ) + ϕ σ, v(σ), Dα u(σ) ⩽ Ψ (σ)ϵ , σ ∈ J,
H 2 H c c
Now, by Lemma 8 and by (144), (143) becomes (152)
YLϕ
‖u − u‖X ⩽ YΩΨ Ψ(σ)ϵ + ‖u − u‖X , (145) there exists a solution (u, v) ∈ X × X with
1 − L∗ϕ
‖(u, v)(σ) − (u, v)(σ)‖ ⩽ CΨα,c Ψα,c (σ)ϵα,c , σ ∈ J. (153)
which implies
1 − L∗ϕ
‖u − u‖X ⩽ YΩΨ Ψ(σ)ϵ. (146) Definition 10 (see [38]). The given system (7) has GHURS
1 − L∗ϕ − YLϕ with respect to Ψα,c � max(Ψα , Ψc ) with Ψα,c ∈ C(J, R) if
there exists some constant CΨα,c � max(CΨα , CΨc ) > 0 such
Thus, we have that, for any solution (u, v) ∈ X × X of the inequality (152),
there exists a solution (u, v) ∈ X × X of (7) satisfying
‖X ⩽ KΨ Ψ(σ)ϵ,
‖u − u (147)
‖(u, v)(σ) − (u, v)(σ)‖ ⩽ CΨα,c Ψα,c (σ), σ ∈ J. (154)
where
1 − L∗ϕ
KΨ � YΩΨ , Remark 5. Let (u, v) ∈ X × X be a solution of the inequality
1 − L∗ϕ − YLϕ
(149), if there exist functions Uϕ1 , Uϕ2 ∈ C(J, R) depending
(148) on u, v, respectively, such that
1 − L∗ϕ − YLϕ
> 0. (1 )|Uϕ1 (σ)| ⩽ ϵα , |Uϕ2 (σ)| ⩽ ϵc , σ ∈ J
1 − L∗ϕ
Dα u(σ) + ϕ1 (σ, u(σ), H Dc v(σ)) � Uϕ1 (σ), σ ∈ J
Hence, the given system (5) is HUR stable. □ (2) H c α
H D v(σ) + ϕ2 (σ, v(σ), H D u(σ)) � Uϕ2 (σ), σ ∈ J
Remark 4. If ϵ � 1 in (147), then by Definition 6, the given Lemma 9. Let (u, v) ∈ X × X be the solution of (149); then,
system (5) is GHURS. for σ ∈ J, we have
30 Complexity
⎨ u(σ) − ω1 (σ) ⩽ Yα ϵα ,
⎧ Proof. By 2 of Remark 5 and for σ ∈ J, we have
⎩ v(σ) − ω (σ) ⩽ Y ϵ . (155)
2 c c
α c
⎪
⎧
⎪ H D u(σ) + ϕ1 σ, u(σ), H D v(σ) � Uϕ1 (σ), σ ∈ J,
⎪
⎪
⎪ c α
⎪ D v(σ) + ϕ σ, v(σ), D u(σ) � Uϕ (σ), σ ∈ J,
⎪
⎪ H 2 H
⎪
⎪ 2
⎨ ζ
⎪ u(σ)| � D α− 2
u(σ)| � 0, Dα− 1
u(σ)| � α[u] + φα (σ)H Dβ u(σ)dσ, (156)
⎪
⎪ σ�1 H σ�1 H σ�2
⎪
⎪ 1
⎪
⎪
⎪
⎪ ξ
⎪
⎩ v(σ)|σ�1 � H Dc− 2 v(σ)|σ�1 � 0, H Dc− 1 v(σ)|σ�2 � c[v] + φc (σ)H Dβ v(σ)dσ.
1
2 dη 2 dη
⎪
⎧
⎪ c
⎪
⎪ u(σ) � Gα (σ, η)ϕ 1 η, u(η), H D v(η) + Gα (σ, η)Uϕ1 (η) ,
⎪
⎪ 1 η 1 η
⎨
⎪ (157)
⎪
⎪
⎪ 2 2
⎪ v(σ) � G (σ, η)ϕ η, v(η), Dα u(η) dη + G (σ, η)U (η) dη.
⎪
⎩ c 2 H c ϕ2
1 η 1 η
α− 1− β
(lnσ)α− 1 ⎡ ζ 2
⎣ φα (σ)(lnσ)α− 1− β dη ζ σ σ dη ⎤⎦
u(σ) � − y(η) dσ − φα (σ)ln y(η) dσ
Δα Γ(α) 1 1 η 1 1 η η
α− 1
(lnσ)α− 1 Γ(α − β) ⎡ 2 2 2 σ
⎣ (lnσ)α− 1 y(η) dη dμ(σ) − ln σ dη
− 2 y(η) dμ(σ)⎤⎦
Δα Γ (α) 1 1 η 1 1 η η
α− 1
(lnσ)α− 1 2 dη 1 σ σ dη
− y(η) + ln y(η)
Γ(α) 1 η Γ(α) 1 η η
(158)
α− 1 ζ 2 ζ σ α− 1− β
(lnσ) ⎣ ⎡ φα (σ)(lnσ)α− 1− β dη σ dη ⎤⎦
+ Uϕ1 (η) dσ − φα (σ)ln Uϕ1 (η) dσ
Δα Γ(α) 1 1 η 1 1 η η
α− 1
(lnσ)α− 1 Γ(α − β) ⎡ 2 2 2 σ
⎣ (lnσ)α− 1 Uϕ (η) dη dμ(σ) − ln σ dη
+ 2 Uϕ1 (η) dμ(σ)⎤⎦
Δα Γ (α) 1 1
1
η 1 1 η η
α− 1
(lnσ)α− 1 2 dη 1 σ σ dη
+ Uϕ1 (η) − ln Uϕ1 (η) .
Γ(α) 1 η Γ(α) 1 η η
Complexity 31
α− 1− β
(lnσ)α− 1 ⎣ ζ 2
⎡ φα (σ)(lnσ)α− 1− β dη ζ σ σ dη ⎦⎤
ω1 (σ) � − y(η) dσ − φα (σ)ln y(η) dσ
Δα Γ(α) 1 1 η 1 1 η η
α− 1
(lnσ)α− 1 Γ(α − β) ⎣ 2 2 2 σ
⎡ (lnσ)α− 1 y(η) dη dμ(σ) − ln σ dη
− y(η) dμ(σ)⎤⎦ (159)
Δα Γ2 (α) 1 1 η 1 1 η η
α− 1
(lnσ)α− 1 2 dη 1 σ σ dη
− y(η) + ln y(η) .
Γ(α) 1 η Γ(α) 1 η η
So, from the above and taking the absolute value, (158)
becomes
(lnσ)α− 1 ζ 2 dη
u(σ) − ω1 (σ) ⩽ 1− β
α−
Uϕ (η) dσ
Δα Γ(α) 1 1 φα (σ)(lnσ) 1
η
1− β
(lnσ)α− 1 ζ σ σ
α−
dη
+ φα (σ)ln Uϕ (η) dσ
Δα Γ(α) 1 1 η 1 η
(lnσ)α− 1 Γ(α − β) 2 2 dη
1
+ 2 (lnσ)α− Uϕ (η) dμ(σ) (160)
Δα Γ (α) 1 1
1
η
α− 1
(lnσ)α− 1 Γ(α − β) 2 σ σ dη
+ 2 ln Uϕ (η) dμ(σ)
Δα Γ (α) 1 1 η 1 η
1
(lnσ)α− 1 2 dη 1 σ σ
α−
dη
+
Uϕ1 (η) + ln Uϕ (η) .
Γ(α) 1 η Γ(α) 1 η 1 η
1 − K1 K3 > 0 (163)
Using 2 of Lemma 6 and 1 of Remark 5, we get
holds, then the given system (7) is stable in the sense of HU.
u(σ) − ω1 (σ) ⩽ Yα ϵα . (161)
In the same way, we have Proof. Let (u, v) ∈ X × X be the solution of (149) and
v(σ) − ω2 (σ) ⩽ Yc ϵc . (162) (u, v) ∈ X × X be the solution to the system
□
Theorem 13. Under the hypothesis (M10 ) and if
α
⎪
⎧ HD u (σ) + ϕ1 σ, u(σ), H Dc v(σ) � 0, σ ∈ J,
⎪
⎪
⎪
⎪
⎪
⎪ HD v
c
(σ) + ϕ2 σ, v(σ), H Dα u(σ) � 0, σ ∈ J,
⎪
⎪
⎨ ζ
⎪ u(σ)|σ�1 � H Dα− 2 u (σ)|σ�1 � 0, H Dα− 1 u(σ)|σ�2 � α[u] + φα (σ)H Dβ u(σ)dσ, (164)
⎪
⎪
⎪
⎪ 1
⎪
⎪
⎪
⎪ ξ
⎩ v(σ)|σ�1 � H Dc− 2 v(σ)|σ�1 � 0, H Dc− 1 v(σ)|σ�2 � c[v] + φc (σ)H Dβ v(σ)dσ.
1
32 Complexity
|u(σ) − u(σ)| ⩽ u(σ) − ω1 (σ) + ω1 (σ) − u(σ)
(lnσ)α− 1 ζ 2
1− β
dη
⩽ Yα ϵα + φα (σ)(lnσ)α− y(η) − yu (η) dσ
Δ
α Γ(α) 1 1 η
1− β
(lnσ)α− 1 ζ σ σ
α−
dη
+ φα (σ)ln y(η) − yu (η) dσ
Δα Γ(α) 1 1 η η
Yα L∗ϕ1 Lϕ2 Yα ϵ α
‖u − u‖X − ‖v − v‖X ⩽ ∗ ∗ ,
1 − L∗ϕ1 L∗ϕ2 − Yα Lϕ1 1 − Yα Lϕ1 /1 − Lϕ1 Lϕ2
(173)
Yc Lϕ1 L∗ϕ2 Yc ϵc
‖v − v‖X − ∗ ∗ ‖u − u‖X ⩽ ∗ ∗ ,
1 − Lϕ1 Lϕ2 − Yc Lϕ2 1 − Yc Lϕ2 /1 − Lϕ1 Lϕ2
Complexity 33
respectively. Let
Yα L∗ϕ1 Lϕ2 Yα
K1 � , K2 � ∗ ∗ ,
1 − L∗ϕ1 L∗ϕ2 − Yα Lϕ1 1 − Yα Lϕ1 /1 − Lϕ1 Lϕ2
(174)
Yc Lϕ1 L∗ϕ2 Yc
K3 � ∗ ∗ and K4 � ∗ ∗ .
1 − Lϕ1 Lϕ2 − Yc Lϕ2 1 − Yc Lϕ2 /1 − Lϕ1 Lϕ2
‖(u, v) − (u, v)‖X×X ⩽ Cα,c ϵα,c , (179) Hence, (M4 ) is satisfied with Lϕ � L∗ϕ � 1/10e2 .
Also, for any y, z ∈ R, we have
where 1
|ϕ(σ, y, z)| ⩽ (2 +|y| +|z|). (184)
K2 K4 K1 K4 K2 K3 10eσ+2
Cα,c � + + + .
1 − K1 K3 1 − K1 K3 1 − K1 K3 1 − K1 K3
Hence, (M3 ) is satisfied with
(180)
1 1
π1 (ϱ) � σ+2 , π 2 (ϱ) � π 3 (ϱ) � , (185)
□ 5e 10eσ+2
Remark 6. By setting C′ (ϵα,c ) � Cα,c ϵα,c , C′ (0) � 0 in (179), where π∗1 � 1/5e2 and π∗2 � π∗3 � 1/10e2 .
by Definition 8, the given system (7) is GHUS. From Theorem 7, we use the inequality which are found
as
Remark 7. Under the hypothesis (M12 ) and (163) and by Lϕ
Y ≈ 0.0134252 < 1. (186)
using Definitions 9 and 10, one can repeat the process of 1 − L∗ϕ
Lemma 9 and the Theorem 13, and the system (7) will be
HUR and GHURS. Hence, (181) has a unique solution.
34 Complexity
Furthermore, K > 0 with condition (126) holds, and Hence, (M4 ) is satisfied with Lϕ � L∗ϕ � 1/90.
Lϕ Also, for any y, z ∈ R, we have
Y ≈ 0.0134252 < 1. (187) 1 1
1 − L∗ϕ |ϕ(σ, y, z)| ≤ (1 +|y|) + |z|. (191)
90(σ + 4) 90 + σ 2
Hence, with the help of Theorem 11, the given system
(181) is HUS and, hence, GHUS. Also, by checking the Hence, (M3 ) is satisfied with
conditions of Theorem 12, it can be easily verified that the 1
π1 (σ) � π2 (σ) � ,
considered problem (181) is HURS and GHURS. 90(σ + 4)
(192)
Example 2. 1
π3 (σ) � ,
D5/2 u(σ) 90 + σ 2
⎪
⎧
⎪ 5/2 1 +|u(σ)| sin H
⎪
⎪
⎪ H D u(σ) + + 2 � 0, where π∗1 � π∗2 � π∗3 � 1/90.
⎪
⎪ 90(σ + 4)(1 +|u(σ)|) 90 + σ
⎪
⎪ From Theorem 7, we use the inequality which are found
⎪
⎪
⎨ as
⎪ u(σ)|σ�1 � H D1/2 u(σ)|σ�1 � 0,
⎪
⎪
⎪ Lϕ
⎪
⎪ Y ≈ 0.0160445 < 1. (193)
⎪
⎪ 1 − L∗ϕ
⎪
⎪ 50
⎪
⎩ D3/2 u(σ)|σ�2 � α[u] + σ 5/2 D1/2 u(σ)dσ,
H H
1 Hence, (188) has a unique solution.
(188) Furthermore, K > 0 with condition (126) holds, and
where α � 5/2, β � 1/2, ζ � 50, P � 50, E � 50, ρ � 3/2, Lϕ
Y ≈ 0.0160445 < 1. (194)
φ(σ) � σ 5/ 2 ⩾ 0, for σ ∈ [1, 2]. Moreover, α[u] � u(1/5)/15. 1 − L∗ϕ
Set
Hence, with the help of Theorem 11, the given system
1 +|u(σ)| sinH D5/2 u(σ) (188) is HUS and, hence, GHUS. Also, by checking the
ϕ(σ, u(σ), z(σ)) � + .
90(σ + 4)(1 +|u(σ)|) 90 + σ 2 conditions of Theorem 11, we can find that the considered
(189) problem (188) is HURS and GHURS.
⎪
⎧ 5 +|u(σ)| + D(7/3) v(σ)
⎪
⎪ (8/3) H
⎪
⎪ HD u(σ) + � 0,
⎪
⎪ 2 +|u(σ)| + H D v(σ)
σ+45 (7/3)
⎪
⎪ 50e
⎪
⎪
⎪
⎪
⎪
⎪
⎪
⎪
⎪
⎪ D(8/3) v(σ)
⎪
⎪ σ cos(u(σ)) − u(σ)sin(σ) H
⎪
⎪ (7/3)
� 0,
⎪
⎪ D v(σ) + + v(σ)
⎨ H (8/3)
80 45 + H D
⎪ (195)
⎪
⎪
⎪
⎪
⎪
⎪
⎪
⎪ 2
⎪
⎪ (2/3)
u(σ)|σ�1 � 0, H D(5/3) u(σ)|σ�2 � α[u] + σ (8/3) H D(2/3) u(σ)dσ,
⎪ u(σ)|σ�1 � H D
⎪
⎪
⎪ 1
⎪
⎪
⎪
⎪
⎪
⎪
⎪
⎪ 2
⎪
⎪
⎩ v(σ)|σ�1 � H D(1/3) v(σ)|σ�1 � 0, H D(4/3) v(σ)|σ�2 � α[v] + σ (7/3) H D(2/3) v(σ)dσ,
1
Complexity 35
c
5 +|u(σ)| + H D(7/3) v(σ)
ϕ1 σ, u(σ), H D v(σ) � ,
50eσ+45 2 +|u(σ)| + H D(7/3) v(σ)
(196)
D(8/3) v(σ)
α σ cos(u(σ)) − u(σ)sin(σ) H
ϕ2 σ, v(σ), H D u(σ) � + .
80 45 + H D(8/3) v(σ)
[11] G. Wang, K. Pei, and Y. Chen, “Stability analysis of nonlinear Mathematical Methods in the Applied Sciences, vol. 43, no. 8,
Hadamard fractional differential system,” Journal of the pp. 5097–5113, 2020.
Franklin Institute, vol. 356, no. 12, pp. 6538–6546, 2019. [30] J. Ren and C. Zhai, “Nonlocal q-fractional boundary value
[12] Z. Ali, A. Zada, and K. Shah, “Ulam stability results for the problem with Stieltjes integral conditions,” Nonlinear Anal-
solutions of nonlinear implicit fractional order differential ysis: Modelling and Control, vol. 24, no. 4, pp. 582–602, 2019.
equations,” Hacettepe Journal of Mathematics and Statistics, [31] U. Riaz, A. Zada, Z. Ali, Y. Cui, and J. Xu, “Analysis of coupled
vol. 48, no. 4, pp. 1092–1109, 2018. systems of implicit impulsive fractional differential equations
[13] R. Rizwan and A. Zada, “Nonlinear impulsive Langevin involving Hadamard derivatives,” Advances in Difference
equation with mixed derivatives,” Mathematical Methods in Equations, vol. 2019, no. 226, pp. 1–27, 2019.
the Applied Sciences, vol. 43, no. 1, pp. 427–442, 2020. [32] J. K. Hale and S. M. V. Lunel, Introduction to Functional
[14] A. Zada, M. Alam, and U. Riaz, “Analysis of q-fractional Differential Equations, p. 99, Springer-Verlag, New York, NY,
implicit boundary value problems having stieltjes integral USA, 1993.
conditions,” Mathematical Methods in the Applied Sciences, [33] A. Granas and J. Dugundji, Fixed Point Theory, Springer-
vol. 44, no. 6, pp. 4381–4413, 2020. Verlag, New York, NY, USA, 2003.
[15] Z. Ali, A. Zada, and K. Shah, “On ulam’s stability for a coupled [34] V. Gafiychuk, B. Datsko, V. Meleshko, and D. Blackmore,
systems of nonlinear implicit fractional differential equa- “Analysis of the solutions of coupled nonlinear fractional
tions,” Bulletin of the Malaysian Mathematical Sciences So- reaction-diffusion equations,” Chaos, Solitons & Fractals,
vol. 41, no. 3, pp. 1095–1104, 2009.
ciety, vol. 42, no. 5, pp. 2681–2699, 2019.
[35] R. L. Wheeden, Measure and Integral: An Introduction to Real
[16] Z. Ali, A. Zada, and K. Shah, “Ulam stability to a toppled
Analysis, p. 308, CRC Press, Boco Raton, FL, USA, 2015.
systems of nonlinear implicit fractional order boundary value
[36] A. Browder, Mathematical Analysis: An Introduction,
problem,” Boundary Value Problems, vol. 175, 2018.
Springer-Verlag, New York, NY, USA, 1996.
[17] V. D. Gejji, “Positive solutions of a system of non-autono-
[37] J. Wang, A. Zada, and W. Ali, “Ulam’s-type stability of first-
mous fractional differential equations,” Journal of Mathe-
order impulsive differential equations with variable delay in
matical Analysis and Applications, vol. 302, pp. 56–64, 2005. quasi-banach spaces,” International Journal of Nonlinear
[18] M. P. Lazarević, “Finite time stability analysis of PD α frac- Sciences and Numerical Simulation, vol. 19, no. 5, pp. 553–560,
tional control of robotic time-delay systems,” Mechanics 2018.
Research Communications, vol. 33, pp. 269–279, 2006. [38] I. A. Rus, “Ulam stabilities of ordinary differential equations
[19] S. M. Ulam, A Collection of the Mathematical Problems, in a Banach space,” Carpathian Journal of Mathematics,
Interscience, New York, NY, USA, 1960. vol. 26, pp. 103–107, 2010.
[20] D. H. Hyers, G. Isac, and T. M. Rassias, Stability of Functional
Equations in Several Variables, Birkh¨auser, Boston, MA,
USA, 1998.
[21] S.-M. Jung, “Hyers-Ulam stability of linear differential
equations of first order, II,” Applied Mathematics Letters,
vol. 19, no. 9, pp. 854–858, 2006.
[22] T. Li and A. Zada, “Connections between Hyers–Ulam sta-
bility and uniform exponential stability of discrete evolution
families of bounded linear operators over Banach spaces,”
Advances in Difference Equations, vol. 2016, no. 1, p. 153, 2016.
[23] M. Obloza, “Hyers stability of the linear differential equation,”
Rocznik NaukDydakt, Prace Mat.vol. 13, pp. 259–270, 1993.
[24] T. M. Rassias, “On the stability of functional equations and a
problem of Ulam,” Acta Applicandae Mathematicae, vol. 62,
no. 1, pp. 23–130, 2000.
[25] M. Ahmad, A. Zada, and J. Alzabut, “Hyers-Ulam stability of a
coupled system of fractional differential equations of Hilfer-
Hadamard type,” Demonstratio Mathematica, vol. 52, no. 1,
pp. 283–295, 2019.
[26] M. Ahmad, A. Zada, and X. Wang, “Existence, uniqueness and
stability of implicit switched coupled fractional differential
equations of ψ$\boldsymbol {\psi} $-hilfer type Ψ –hilfer
type,” International Journal of Nonlinear Sciences and Nu-
merical Simulation, vol. 21, no. 3-4, p. 327, 2020.
[27] J. Wang, A. Zada, and H. Waheed, “Stability analysis of a
coupled system of nonlinear implicit fractional anti-periodic
boundary value problem,” Mathematical Methods in the
Applied Sciences, vol. 42, no. 18, pp. 6706–6732, 2019.
[28] A. Zada, S. Ali, and T. Li, “Analysis of a new class of impulsive
implicit sequential fractional differential equations,” Inter-
national Journal of Nonlinear Sciences and Numerical Sim-
ulation, vol. 21, no. 6, p. 571, 2020.
[29] A. Zada, B. Pervaiz, S. O. Shah, and J. Xu, “Stability analysis of
first-order impulsive nonautonomous system on timescales,”