See discussions, stats, and author profiles for this publication at: https://www.researchgate.

net/publication/342330058

Existence and Ulam stability results of a coupled system for terminal value
problems involving ψ-Hilfer fractional operator

Article  in  Advances in Difference Equations · June 2020

DOI: 10.1186/s13662-020-02775-x

CITATIONS READS

0 58

4 authors:

Mohammed S. Abdo Kamal Shah

Dr. Babasaheb Ambedkar Marathwada University University of Malakand
50 PUBLICATIONS   155 CITATIONS    176 PUBLICATIONS   1,073 CITATIONS   

SEE PROFILE SEE PROFILE

Satish K Panchal Hanan A. Wahash

Dr. Babasaheb Ambedkar Marathwada University Dr. Babasaheb Ambedkar Marathwada University
70 PUBLICATIONS   172 CITATIONS    12 PUBLICATIONS   18 CITATIONS   

SEE PROFILE SEE PROFILE

Some of the authors of this publication are also working on these related projects:

Fuzzy Fractional Differential Equation View project

Fractional Calculus View project

All content following this page was uploaded by Kamal Shah on 25 June 2020.

The user has requested enhancement of the downloaded file.

Abdo et al. Advances in Difference Equations (2020) 2020:316
https://doi.org/10.1186/s13662-020-02775-x

RESEARCH Open Access

Existence and Ulam stability results of a

coupled system for terminal value problems
involving ψ-Hilfer fractional operator
Mohammed S. Abdo1,2* , Kamal Shah3 , Satish K. Panchal1 and Hanan A. Wahash1
*
Correspondence:
msabdo1977@gmail.com Abstract
1
Department of Mathematics, Dr.
Babasaheb Ambedkar Marathwada The work reported in this paper deals with the study of a coupled system for
University, Aurangabad, India fractional terminal value problems involving ψ -Hilfer fractional derivative. The
2
Department of Mathematics, existence and uniqueness theorems to the problem at hand are investigated. Besides,
Hodeidah University, AL-Hodeidah,
Yemen the stability analysis in the Ulam–Hyers sense of a given system is studied. Our
Full list of author information is discussion is based upon known fixed point theorems of Banach and Krasnoselskii.
available at the end of the article Examples are also provided to demonstrate the applicability of our results.
MSC: Primary 26A33; secondary 34A07; 93A30; 35R11
Keywords: Hilfer fractional differential equation; Terminal value problem; Existence
and stability theory; Fixed point theorem

1 Introduction
Over the past few years, mathematicians have realized that fractional calculus has many
applications in various scopes of applied science and engineering. Several researchers have
employed the fractional calculus as an experiential style of describing the properties of
natural phenomena such as chemistry, biology, physics, bioengineering, electrochemistry,
finance, economic, etc., for more details, see [12, 20, 21, 23] and many other references.
The interesting issue about this theme is that it is completely unlike classical derivatives,
because it deals with arbitrary and noninteger order, e.g., the fractional derivative of non-
integer order depends not just on the diagram of the function very near to the point but
also on some chronicle.
Recently, there has been considerable growth in fractional differential equations (FDEs)
involving several different fractional derivative operators, we indicate here the more fa-
mous operators like Riemann–Liouville (RL), Caputo, Hilfer, Hadamard, Katugampola,
and several other generalized operators.
So, this implies that different categories of FDEs involving several fractional opera-
tors have been considered. Researchers who are concerned with this topic have pre-
sented many generalizations of fractional derivatives like Hilfer–Katugampola, Hilfer–
Hadamard, ψ-Caputo, and ψ-Hilfer, see [10, 16–18, 24, 29]. Some existence, uniqueness,

© The Author(s) 2020. This article is licensed under a Creative Commons Attribution 4.0 International License, which permits use,
sharing, adaptation, distribution and reproduction in any medium or format, as long as you give appropriate credit to the original
author(s) and the source, provide a link to the Creative Commons licence, and indicate if changes were made. The images or other
third party material in this article are included in the article’s Creative Commons licence, unless indicated otherwise in a credit line
to the material. If material is not included in the article’s Creative Commons licence and your intended use is not permitted by
statutory regulation or exceeds the permitted use, you will need to obtain permission directly from the copyright holder. To view a
copy of this licence, visit http://creativecommons.org/licenses/by/4.0/.
Abdo et al. Advances in Difference Equations (2020) 2020:316 Page 2 of 21

and Ulam–Hyers stability results of generalized FDEs have been studied, see the recent
references [1–4, 9, 11, 22, 25, 30–34].
On the other hand, many interesting and recent results on the existence and stability of
a coupled system for different categories of FDEs have been investigated, see the following
studies [6–8, 26] and the references therein.
Terminal value problems (TVPs) for FDEs for the time being play a major role in the
modeling of numerous phenomena in engineering, science, and simulation. In short, the
existence results for ordinary and fractional TVPs have been studied by many investiga-
tors, see [5, 13–15, 27, 28, 35]. For example, Benchohra et al. in [13] obtained the existence
and uniqueness of solution to the fractional implicit TVP

θ,η  θ,η 
ρ
Da+ y(t) = f t, y(t), ρ Da+ y(t) , a < t ≤ T, a > 0, (1)

under the terminal condition

y(T) = w ∈ R, (2)

θ,η
where ρ Da+ is the fractional derivative of order (θ , η) in the Hilfer–Katugampola sense
(0 < θ < 1, 0 ≤ η ≤ 1), ρ > 0, and f : (a, T] × R → R is a certain function.
Motivated by the aforementioned works, the target of this work is to investigate the exis-
tence, uniqueness, and Ulam–Hyers stability of solutions of a coupled system for fractional
TVPs involving generalized Hilfer fractional derivative of the type
⎧
⎨Dθ1+,η1 ;ψ y(t) = f (t, x(t)), a < t ≤ T, a > 0,
a 1
(3)
⎩Dθ2+,η2 ;ψ x(t) = f2 (t, y(t)), a < t ≤ T, a > 0,
a

under the terminal conditions

⎧
⎨y(T) = w ∈ R,
1
(4)
⎩x(T) = w2 ∈ R,

θ ,η ;ψ
where 0 < θi < 1, 0 ≤ ηi ≤ 1, Dai+ i (i = 1, 2) is the Hilfer fractional derivative of order
θi and type ηi with respect to ψ and f : (a, T] × R → R is a certain function under the
conditions listed later.
As far as we know, no papers about a coupled system for fractional TVPs exist in the
literature, specifically for those encompassing the generalized fractional derivative in the
ψ-Hilfer sense. Moreover, the results of the problem at hand are obtained under minimal
assumptions on nonlinear functions f1 , f2 .
The rest of the structure of this paper is as follows. In Sect. 2, we briefly state some es-
sential definitions and the results that are applied throughout the paper. Section 3 studies
the existence and uniqueness results on ψ-Hilfer FDEs with the terminal conditions via
fixed point techniques of Banach and Krasnoselskii. The stability analysis in the concept
Ulam–Hyers of the proposed system is investigated in Sect. 4. At the end, some examples
are included to illustrate the applicability of the obtained results in Sect. 5.
Abdo et al. Advances in Difference Equations (2020) 2020:316 Page 3 of 21

2 Auxiliary results
Let [a, T] ⊂ R+ with (0 < a < T < ∞), we also consider C[a, T] the Banach space of real-
valued continuous functions defined on [a, T] with the norm

 
σ C = max σ (t), σ ∈ C[a, T].
t∈[a,T]

Let ς = θ + η(1 – θ ) where 0 < θ < 1, 0 ≤ η ≤ 1, and let ψ ∈ C 1 [a, T] be an increasing

function with ψ  (t) = 0 for each t ∈ [a, T]. For 0 < ς ≤ 1, the weighted spaces C1–ς ;ψ [a, T],
n
C1–ς ;ψ [a, T] of continuous functions are defined by [29]


1–ς 
C1–ς ;ψ [a, T] = σ : (a, T] → R; ψ(t) – ψ(a) σ (t) ∈ C[a, T] ,
 
;ψ [a, T] = σ ∈ C [a, T], σ (n) ∈ C1–ς ;ψ [a, T] .
n n–1
C1–ς

n
Obviously, C1–ς ;ψ [a, T] and C1–ς ;ψ [a, T] are Banach spaces endowed with the norms


1–ς 
σ C1–ς ;ψ = max  ψ(t) – ψ(a) σ (t),
t∈[a,T]


n–1
(j)
σ C1–ς
n = σ + σ (n) ,
;ψ C C 1–ς ;ψ
j=0

0
respectively. For n = 0, C1–ς ;ψ [a, T] = C1–ς ;ψ [a, T]. Let us introduce the following space:

 
E = σ (t) : σ (t) ∈ C1–ς ;ψ [a, T]

endowed with the norm defined by

σ E = σ c1–ς ;ψ .

It is easy to perceive that, for σ ∈ E, (E, σ E ) is a Banach space. Then, for (σ , ρ) ∈ E × E,

the product space (E × E, (σ , ρ)E×E ) is a Banach space too, where

 
(σ , ρ) = max σ E , ρE , σ , ρ ∈ E.
E×E

Definition 1 ([19, Sect. 2.5, Eq. (2.5.1)]) Let θ > 0, σ ∈ L1 [a, T]. Then the left-sided ψ-RL
fractional integral of order θ of a function σ w.r.t. ψ is described by


1 t  θ–1
ψ  (s) ψ(t) – ψ(s)
θ,ψ
Ia+ σ (t) = σ (s) ds.
Γ (θ ) a

Definition 2 ([19, Eq. (2.5.17)], [10, Definition 1]) Let n – 1 < θ < n, and σ ∈ C n [a, T].
Then the left-sided ψ-fractional derivatives in the concepts ψ-RL and ψ-Caputo of order
θ of a function σ w.r.t. ψ are described by

θ,ψ n–θ,ψ
Da+ σ (t) = Dn,ψ Ia+ σ (t)
Abdo et al. Advances in Difference Equations (2020) 2020:316 Page 4 of 21

and

θ,ψ n–θ,ψ
C
Da+ σ (t) = Ia+ σψ[n] (t),

respectively, where σψ[n] (t) = Dn,ψ σ (t), Dn,ψ = [ ψ 1(t) dtd ]n , and n = [θ ] + 1.

Definition 3 ([29, Definition 7]) Let ς = θ + η(n – θ ) where n – 1 < θ < n ∈ N, 0 ≤ η ≤ 1,

and σ ∈ C n [a, T]. Then the left-sided ψ-fractional derivative in the concept ψ-Hilfer of
order θ and type η of a function σ w.r.t. ψ is given by

θ,η,ψ η(n–θ);ψ (1–η)(n–θ);ψ

Da+ σ (t) = Ia+ Dn,ψ Ia+ σ (t)
η(n–θ);ψ ς ;ψ
= Ia+ Da+ σ (t), (5)

where

ς ;ψ (1–η)(n–θ);ψ
Da+ σ (t) = Dn,ψ Ia+ σ (t).

θ,ψ
Lemma 1 ([32, Lemma 1]) Let θ > 0 and 0 < ς ≤ 1. Then (Ia+ is bounded) : C1–ς ;ψ [a, T] →
C1–ς ;ψ [a, T].

Lemma 2 ([32, Lemma 1], [29, Lemma 4]) Let θ > 0, 0 < ς ≤ 1, and σ ∈ C1–ς ,ψ [a, T]. If
θ,ψ
ς < θ , then (Ia+ is bounded) : C1–ς ;ψ [a, T] → C[a, T]. Besides,

θ;ψ θ;ψ
Ia+ σ (a) = lim+ Ia+ σ (t) = 0.
t→a

θ,η
Definition 4 [32, Definition 3] The weighted continuous spaces C1–ς ;ψ [a, T] and
ς
C1–ς ;ψ [a, T] are described by

θ,η  θ,η;ψ 
C1–ς ;ψ [a, T] = σ ∈ C1–ς ;ψ [a, T], Da+ σ ∈ C1–ς ;ψ [a, T]

and

ς  ς ;ψ 
C1–ς ;ψ [a, T] = σ ∈ C1–ς ;ψ [a, T], Da+ σ ∈ C1–ς ;ψ [a, T] , (6)

where ς = θ + η(1 – θ ), 0 < θ < 1, and 0 ≤ η ≤ 1. Observe that

ς θ,η
C1–ς ;ψ [a, T] ⊂ C1–ς ;ψ [a, T] ⊂ C1–ς ;ψ [a, T] ⊂ C[a, T].

Lemma 3 ([2, Lemma 2.3]) Let ς = θ + η(1 – θ ) where 0 < θ < 1, 0 ≤ η ≤ 1, and σ ∈
ς
C1–ς ;ψ [a, T]. Then

ς ;ψ ς ;ψ θ;ψ θ,η;ψ ς ;ψ θ;ψ η(1–θ);ψ

Ia+ Da+ σ = Ia+ Da+ σ and Da+ Ia+ σ = Da+ σ.

Lemma 4 ([29, Theorem 7]) Let θ > 0, 0 < ς ≤ 1, and σ ∈ C1–ς [a, T], η ∈ [0, 1]. Then

θ,ψ θ,ψ
Da+ Ia+ σ (t) = σ (t).
Abdo et al. Advances in Difference Equations (2020) 2020:316 Page 5 of 21

1–ς ;ψ
Lemma 5 ([29, Theorem 1]) Let θ > 0, 0 < ς ≤ 1, and σ ∈ C1–ς ;ψ [a, T] and Ia+ σ ∈
1
C1–ς ,ψ [a, T]. Then

σ (a)  ς –1
1–ς ;ψ
ς ;ψ ς ,ψ Ia+
Ia+ Da+ σ (t) = σ (t) – ψ(t) – ψ(a) .
Γ (ς)

Lemma 6 ([19, Property 2.18]) Let t > a, and consider χ ς (t) := [ψ(t) – ψ(a)]ς –1 . Then, for
θ > 0 and ς > 0,

θ,ψ Γ (ς)  θ+ς –1

Ia+ χ ς (t) = ψ(t) – ψ(a) .
Γ (θ + ς)
Besides, for 0 < θ < 1,

θ,ψ
Da+ χ θ (t) = 0.

Theorem 1 ([36, Theorem 1.45]) Let Ξ be a Banach space, and let ∅ ⊂ Ξ be a nonempty,
closed, convex, and bounded set, and Π1 , Π2 be two operators satisfying
(i) Π1 u + Π2 v ∈ ∅ for all u, v ∈ ∅;
(ii) Π1 is continuous and compact;
(iii) Π2 is a contraction operator.
Then there exists u∗ ∈ ∅ such that u∗ = Π1 u∗ + Π2 u∗ .

Theorem 2 ([36, Theorem 1.41]) Let Ξ be a Banach space and ∅ be a nonempty closed
subset of Ξ , then any contraction operator Π : ∅ → ∅ has a unique fixed point.

The results we study in this paper will only concern 0 < θ < 1 (n = 1).

3 Existence and uniqueness results

ς
To shorten the length of equations, we set Kψ (t, a) := [ψ(t) – ψ(a)]ς –1 and Hψθ (t, s) :=
ψ  (s)[ψ(t) – ψ(s)]θ–1 .

Theorem 3 Let ς = θ + η(1 – θ ) where 0 < θ < 1 and 0 ≤ η ≤ 1. If σ : (a, T] → R is a

ς
function such that σ (·) ∈ C1–ς ,ψ [a, T], then y ∈ C1–ς ,ψ (a, T] satisfies the TVP for ψ-Hilfer
FDEs

θ,η;ψ
Da+ y(t) = σ (t), t ∈ (a, T], a > 0, (7)

y(T) = w ∈ R, (8)

if and only if y fulfills the following fractional integral equation:

   
Kψς (t, a) 1 T
1 t
y(t) = w– Hψθ (T, s)σ (s) ds + Hψθ (t, s)σ (s) ds. (9)
Kψς (T, a) Γ (θ ) a Γ (θ ) a

ς
Proof Assume that y ∈ C1–ς ,ψ [a, T] is the solution of the TVP for ψ-Hilfer FDEs (7)–(8).
ς ς ;ψ
According to the definitions of C1–ς ,ψ (a, T] and Da+ , applying Lemma 1, we have

1–ς ,ψ ς ;ψ 1–ς ,ψ
Ia+ y(t) ∈ C1–ς ,ψ [a, T] and Da+ y(t) = D1,ψ Ia+ y(t) ∈ C1–ς ,ψ [a, T]. (10)
Abdo et al. Advances in Difference Equations (2020) 2020:316 Page 6 of 21

n
Thanks to the definition of C1–ς ,ψ [a, T], we have

1–ς ,ψ
Ia+ y(t) ∈ C1–ς
1
,ψ [a, T]. (11)

Take advantage of Lemma 5 to get

1–ς ;ψ
ς ;ψ ς ,ψ Ia+ y(a) ς
Ia+ Da+ y(t) = y(t) – Kψ (t, a), t ∈ (a, T]. (12)
Γ (ς)

ς
It follows from the assumption y ∈ C1–ς ,ψ [a, T], Lemma 3, and equation (7) that

ς ;ψ ς ;ψ θ;ψ
Ia+ Da+ y(t) = Ia+ σ (t). (13)

Equating both sides of equations (12) and (13), we find that

1–ς ;ψ
Ia+ y(a) ς
y(t) = Kψ (t, a) + Iaθ;ψ
+ σ (t). (14)
Γ (ς)

Using the terminal condition y(T) = w, we get

1–ς ;ψ Γ (ς) θ;ψ

Ia + y(a) = ς w1 – Ia+ σ (T) . (15)

Kψ (T, a)

Now, from equations (14) and (15), we conclude that

   
Kψς (t, a) 1 T
1 t
y(t) = w– Hψθ (T, s)σ (s) ds + Hψθ (t, s)σ (s) ds.
Kψς (T, a) Γ (θ ) a Γ (θ ) a

Hence y(t) satisfies the TVP for ψ-Hilfer FDEs (7)–(8).

ς
Conversely, suppose that y ∈ C1–ς ,ψ [a, T] satisfying integral equation (9). Applying frac-
ς ;ψ
tional derivative Da+ on both sides of integral equation (9) and employing Lemmas 6 and
3, we will surely find

ς ;ψ 1 θ;ψ
ς ;ψ ς ς ;ψ θ;ψ
Da+ y(t) = w – Ia+ σ (T) Da+ Kψ (t, a) + Da+ Ia+ σ (t)
Kψς (T, a)
η(1–θ);ψ
= D a+ σ (t), (16)

ς ;ψ ς ς ;ψ
where Da+ Kψ (t, a) = 0. From (10), we have Da+ y ∈ C1–ς ;ψ [a, T], then (16) implies

ς ;ψ 1–η(1–θ),ψ η(1–θ);ψ
Da+ y(t) = D1,ψ Ia+ σ (t) = Da+ σ (t) ∈ C1–ς ;ψ [a, T]. (17)

As σ (t) ∈ C1–ς ;ψ [a, T], it follows from Lemma 1 that

1–η(1–θ);ψ
Ia+ σ ∈ C1–ς ;ψ [a, T]. (18)

n
From the definition of C1–ς ;ψ (a, T] with the aid of equations (17), (18), we get

1–η(1–θ);ψ
Ia+ σ ∈ C1–ς
1
;ψ [a, T].
Abdo et al. Advances in Difference Equations (2020) 2020:316 Page 7 of 21

1–η(1–θ);ψ
Thus σ and Ia+ σ satisfy assumptions of Lemma 5. Now, we can apply the fractional
η(1–θ);ψ
integral Ia+ on both sides of (17) and utilize Lemmas 2, 5 to get

1–η(1–θ);ψ
η(1–θ);ψ ς ;ψ Ia + σ (a) η(1–θ)
Ia+ Da+ y(t) = σ (t) – K (t, a)
Γ (η(1 – θ )) ψ
= σ (t). (19)

From (5) with n = 1, equation (19) reduces to

θ,η;ψ
Da+ y(t) = σ (t), t ∈ (a, T].

ς
This proves that y also satisfies FDE (7). Undoubtedly, if y ∈ C1–ς ;ψ [a, T] satisfies integral
equation (9), then it also fulfills terminal condition (8). 

Before we present our main results, we consider that the following assumptions are sat-
isfied:
(H1 ) f1 , f2 : (a, T] × R → R are such that
  η1 (1–θ1 )
f1 ·, y(·) ∈ C1–ς 1 ;ψ
[a, T], y ∈ C1–ς1 ;ψ [a, T],
  η2 (1–θ2 )
f2 ·, x(·) ∈ C1–ς 2 ;ψ
[a, T], y ∈ C1–ς2 ;ψ [a, T].

(H2 ) There exist L1 (> 0) and L2 (> 0) such that, for p, p∗ , q, q∗ ∈ R, t ∈ (a, T], we have
    
f1 (t, p) – f1 t, p∗  ≤ L1 p – p∗ ,
    
f2 (t, q) – f2 t, q∗  ≤ L2 q – q∗ .

In the forthcoming theorem, by using Theorem 2, we prove the unique solution of a

coupled system for ψ-Hilfer terminal FDEs (3)–(4). In view of Theorem 3, we get the
following lemma.

Lemma 7 Let i = 1, 2, ςi = θi + ηi (1 – θi ), where 0 < θi < 1 and 0 ≤ ηi ≤ 1. If f1 , f2 : (a, T] ×

R → R two functions such that f1 (·, y) ∈ C1–ς1 ,ψ [a, T] and f2 (·, x) ∈ C1–ς2 ,ψ [a, T], then y ∈
ς1 ς2
C1–ς 1 ,ψ
(a, T] and x ∈ C1–ς 2 ,ψ
(a, T] satisfy the following coupled system of fractional integral
equations:
⎧ ς
T
⎪ K 1 (t,a)
⎪
⎪ y(t) = ςψ1 [w1 – Γ (θ1 1 ) a Hψθ1 (T, s)f1 (s, x(s)) ds]
⎪
⎪ K (T,a)
⎪
⎪
ψ
t
⎨ + Γ (θ1 1 ) a Hψθ1 (t, s)f1 (s, x(s)) ds,
ς
T (20)
⎪
⎪
Kψ2 (t,a)
[w2 – Γ (θ1 2 ) a Hψθ2 (T, s)f2 (s, y(s)) ds]
⎪
⎪ x(t) = ς2
⎪
⎪
Kψ (T,a)
⎪
⎩ t
+ Γ (θ1 2 ) a Hψθ2 (t, s)f2 (s, y(s)) ds.

According to Lemma (7), we consider the operators N1 : E → E and N2 : E → E defined

by
⎧
⎨N x(t) = y(t),
1
⎩N2 y(t) = x(t).
Abdo et al. Advances in Difference Equations (2020) 2020:316 Page 8 of 21

That is,
⎧ ς
 T θ1
⎪ Kψ1 (t,a)
⎪
⎪ N x(t) = 1 (T,a) [w1 – Γ (θ1 ) a Hψ (T, s)f1 (s, x(s)) ds]
1
⎪
⎪
1
K
ς
⎪
⎪
ψ
t
⎨ + Γ (θ1 1 ) a Hψθ1 (t, s)f1 (s, x(s)) ds,
(21)
⎪N y(t) = Kψ2 (t,a) [w – 1  T Hθ2 (T, s)f (s, y(s)) ds]
ς
⎪
⎪
⎪
⎪
2 ς2
Kψ (T,a)
2 Γ (θ2 ) a ψ 2
⎪
⎪ 
⎩ t
+ Γ (θ1 2 ) a Hψθ2 (t, s)f2 (s, y(s)) ds.

Therefore, we define N : E × E → E × E by

 
N (y, x)(t) = N1 (x)(t), N2 (y)(t) .

For the sake of brevity, we set

 
2L1 B (ς1 , θ1 ) θ1 +1
Λf1 = Kψ (T, a),
Γ (θ1 )
w1 2A1
f1 = + K2 (T, a),
Kψς1 (T, a) Γ (θ1 + 1) ψ
 
2L2 B (ς2 , θ2 ) θ2 +1
Λf2 = Kψ (T, a),
Γ (θ2 )
w2 2A2
f2 = + K2 (T, a),
Kψς2 (T, a) Γ (θ2 + 1) ψ

where Ai = maxt∈[a,T] |fi (t, 0)|, i = 1, 2.

Now, via Theorems 2, 1, we obtain the existence and uniqueness results of a coupled
system for ψ-Hilfer FDEs (3)–(4).

Theorem 4 Assume that (H1 ) and (H2 ) hold. If Λf1 < 1 and Λf2 < 1, then ψ-Hilfer coupled
ς
system (3)–(4) has a unique solution in Eς × Eς ⊂ Eθ,η × Eθ,η , where Eς := C1–ς ;ψ [a, T] and
θ,η
Eθ,η := C1–ς ;ψ [a, T].

Proof Define the closed, bounded, convex, and nonempty set

 
SR = (y, x) ∈ E × E : (y, x) E×E ≤ R ⊂ E × E,

with
 
f1 f2
R ≥ max , .
1 – Λf1 1 – Λf2

The analysis of proof will be presented in three steps.

Step(1): N S R ⊂ SR .
Let (y, x) ∈ SR and t ∈ (a, T]. Then


 ψ(t) – ψ(a) 1–ς1 (N1 x)(t)
  T 
1 1   
≤ ς1 w1 + θ1  
Hψ (T, s) f1 s, x(s) ds
Kψ (T, a) Γ (θ1 ) a
Abdo et al. Advances in Difference Equations (2020) 2020:316 Page 9 of 21


1 1 t   
+ Hψθ1 (t, s)f1 s, x(s)  ds
Kψς1 (t, a) Γ (θ1 ) a
  T
1 1
≤ ς1 w1 + Hψθ1 (T, s)
Kψ (T, a) Γ (θ1 ) a

     

× f1 s, x(s) – f1 (s, 0) + f1 (s, 0) ds

 t
1 1
+ Hθ1 (t, s)
Kψς1 (t, a)
Γ (θ1 ) a ψ
     

× f1 s, x(s) – f1 (s, 0) + f1 (s, 0) ds. (22)

Thanks to hypothesis (H2 ), for (y, x) ∈ SR , t ∈ (a, T], we have



 ψ(t) – ψ(a) 1–ς1 (N1 x)(t)
  T 
1 1   1–ς1 
≤ w1 + Hψθ1 (T, s) L1 ψ(s) – ψ(a) xE + A1 ds
Kψς1 (T, a) Γ (θ1 ) a
 t
1 1   ς –1 
+ ς1 Hψθ1 (t, s) L1 ψ(s) – ψ(a) 1 xE + A1 ds
Kψ (t, a) Γ (θ1 ) a
 
1 Γ (ς1 ) A1
≤ ς1 w1 + L1 R Kψθ1 +ς1 (T, a) + Kψθ1 +1 (T, a)
Kψ (T, a) Γ (ς1 + θ1 ) Γ (θ1 + 1)
 
1 Γ (ς1 ) A1
+ ς1 L1 R Kψθ1 +1 (t, a) + Kψθ1 +1 (t, a)
Kψ (t, a) Γ (ς1 + θ1 ) Γ (θ1 + 1)
w1 2A1 B (ς1 , θ1 ) θ1 +1
≤ ς1 + Kψ2 (T, a) + 2L1 R K (T, a), (23)
Kψ (T, a) Γ (θ1 + 1) Γ (θ1 ) ψ

where we used the fact that

1 1
≤ for ς1 ≥ θ1
Kψς1 (t, a) Kψθ1 (t, a)

and

Kψς1 (t, a) ≤ 1 for 0 < ς1 ≤ 1.

Hence, inequality (23) becomes



 ψ(t) – ψ(a) 1–ς1 (N1 x)(t) ≤ Λf + f R,
1 1

which leads to

N1 xE ≤ Λf1 + f1 R ≤ R. (24)

Similarly, we can get that

N2 yE ≤ Λf2 + f2 R ≤ R. (25)

Abdo et al. Advances in Difference Equations (2020) 2020:316 Page 10 of 21

It follows from (24) and (25) that

 
N (y, x) = max N1 xE , N2 yE ≤ max(R, R) = R.
E×E

This proves N S R ⊂ SR .
Step(2): The operator N is a contraction.
Let (y, x), (y∗ , x∗ ) ∈ SR and t ∈ (a, T]. Applying (H2 ) we have


 

 ψ(t) – ψ(a) 1–ς1 (N1 x)(t) – N1 x∗ (t) 

1 1 T     
≤ Hψθ1 (T, s)f1 s, x(s) – f1 s, x∗ (s)  ds
Kψς1 (T, a) Γ (θ1 ) a

1 1 t     
+ Hψθ1 (t, s)f1 s, x(s) – f1 s, x∗ (s)  ds
Kψς1 (t, a) Γ (θ1 ) a

1  θ ,ψ ς 
≤ L1 x – x∗ E Ia1+ Kψ1 (s, a) (T)
Kψς1 (T, a)
1  
+ L1 x – x∗ I θ1+,ψ Kς1 (s, a) (t)
Kψς1 (t, a) E a ψ

B (ς1 , θ1 ) θ1 +1
≤ 2L1 Kψ (T, a) x – x∗ E ,
Γ (θ1 )

which implies

 
(N1 x) – N1 x∗ ≤ Λf x – x∗ .
E 1 E

By the same technique, we can also get

 
(N2 y) – N2 y∗ ≤ Λf x – x∗ .
E 2 E

In view of the conditions Λf1 < 1 and Λf2 < 1, we get

   
N (y, x) – N y∗ , x∗ < (y, x) – y∗ , x∗ E×E .
E×E

Thus N is a contraction mapping. In accordance with Theorem 2, N has a unique fixed

point (y, x) ∈ E × E.
Step(3): We show that such a fixed point ( x) ∈ C1–ς1 ;ψ [a, T] × C1–ς2 ;ψ [a, T] is actually
y,
ς1 ς2
in C1–ς 1 ;ψ
[a, T] × C 1–ς2 ;ψ [a, T]. Since
y and 
x are the unique fixed points of N1 and N2 in
C1–ς1 ;ψ [a, T] × C1–ς2 ;ψ [a, T] respectively, then for t ∈ (a, T], we have

⎧ ς
T
⎪ Kψ1 (t,a)
⎪
⎪y(t) = [w1 – Γ (θ1 1 ) a Hψθ1 (T, s)f1 (s,x(s)) ds]
⎪
⎪
ς
Kψ1 (T,a)
⎪
⎪ t
⎨ + Γ (θ1 1 ) a Hψθ1 (t, s)f1 (s,x(s)) ds,
⎪ K
ς2
(t,a) 
⎪
⎪x(t) = ςψ2
T
[w2 – Γ (θ1 2 ) a Hψθ2 (T, s)f2 (s,
y(s)) ds]
⎪
⎪ Kψ (T,a)
⎪
⎪ 
⎩ t
+ Γ (θ1 2 ) a Hψθ2 (t, s)f2 (s,y(s)) ds.
Abdo et al. Advances in Difference Equations (2020) 2020:316 Page 11 of 21

ς ,ψ ς ,ψ
Multiplying both sides of the last system by Da1+ , Da2+ respectively, it follows from Lem-
mas 6 and 3 that
⎧
⎨Dς1+,ψ ς ,ψ θ ;ψ
y(t) = Da1+ Ia1+ f1 (s,
η (1–θ );ψ
x(s))(t) = Da1+ 1 f1 (s,
x(s)),
a
⎩D 2+ 
ς ,ψ ς ,ψ θ ;ψ
x(t) = D 2+ I 2+ f2 (s,
η (1–θ );ψ
y(s))(t) = D 2+ 2 f2 (s,
y(s)).
a a a a

Since ςi ≥ θi (i = 1, 2) and by (H1 ), we get

⎧
⎨Dη1+(1–θ1 );ψ f (s,x(s)) ∈ C1–ς1 ;ψ [a, T],
a 1
⎩D +
η2 (1–θ2 );ψ
y(s)) ∈ C1–ς ;ψ [a, T].
f2 (s,
a 2

ς ,ψ ς ,ψ
Hence, Da1+  y ∈ C1–ς1 ;ψ [a, T] and Da2+  x ∈ C1–ς2 ;ψ [a, T], it follows from the definition of
ςi ς1 ς2
C1–ςi ;ψ [a, T] (i = 1, 2) that  y ∈ C1–ς1 ;ψ [a, T] and 
x ∈ C1–ς 2 ;ψ
[a, T]. As a sequel to the steps
outlined above, we infer that the ψ-Hilfer coupled system (3)–(4) has a unique solution in
ς1 ς2
C1–ς 1 ;ψ
[a, T] × C1–ς 2 ;ψ
[a, T]. 

We exhibit now the next result, which relies on Theorem 1.

Theorem 5 Assume that (H1 )–(H2 ) hold. Then the ψ-Hilfer coupled system (3)–(4) has at
least one solution.

Proof Let K ⊂ SR ⊆ E × E be a bounded set, and we define the operators F1 , G1 : E → E

and F2 , G2 : E → E by

1 t  
(F1 x)(t) = Hψθ1 (t, s)f1 s, x(s) ds,
Γ (θ1 ) a
ς1   
Kψ (t, a) 1 T  
(G1 x)(t) = w1 – Hψθ1 (T, s)f1 s, x(s) ds ,
Kψς1 (T, a) Γ (θ1 ) a

1 t  
(F2 y)(t) = Hψθ2 (t, s)f2 s, y(s) ds,
Γ (θ2 ) a
ς2   
Kψ (t, a) 1 T  
(G2 y)(t) = w2 – Hψθ2 (T, s)f2 s, y(s) ds .
Kψς2 (T, a) Γ (θ2 ) a

From the above-mentioned operators, we are able to write N1 = F1 + G1 and N2 = F2 + G2 .

Thus, the operator N can be expressed as

N =F +G such that F (y, x) = (F1 x, F2 y) and G (y, x) = (G1 x, G2 y).

The proof will be divided into several stages as follows:

Stage(1): N is continuous.
The continuity of f1 and f2 implies the continuity of N .
Stage(2): F (K) is uniformly bounded.
Let (y, x) ∈ K, t ∈ (a, T]. Then, by using (H2 ), we have


1–ς1 
F1 xE = max  ψ(t) – ψ(a) (F1 x)(t)
t∈[a,T]
Abdo et al. Advances in Difference Equations (2020) 2020:316 Page 12 of 21


1 1 t  
≤ max ς Hψθ1 (t, s)f1 s, x(s) ds
t∈[a,T] Kψ1 (t, a) Γ (θ1 ) a

1 1 t      
≤ max ς1 Hψθ1 (t, s) f1 s, x(s) – f1 (s, 0) + f1 (s, 0) ds
t∈[a,T] Kψ (t, a) Γ (θ1 ) a

1 1 t 
ς –1 
≤ max ς Hψθ1 (t, s) A1 + L1 ψ(s) – ψ(a) 1 xE ds
t∈[a,T] Kψ1 (t, a) Γ (θ1 ) a
 
A1 B (ς1 , θ1 ) θ1 +1
≤ max Kψ2 (t, a) + L1 K (t, a)xE
t∈[a,T] Γ (θ1 + 1) Γ (θ1 ) ψ
 
A1 B (ς1 , θ1 ) 2
≤ + L1 Kψ (T, a)xE := R1 , (26)
Γ (θ1 + 1) Γ (θ1 )

where Kψθ1 +1 (t, a) < Kψ2 (t, a)for 0 < θ1 < 1. In the same fashion, we get

 
A2 B (ς2 , θ2 ) 2
F2 yE ≤ + L2 Kψ (T, a)yE := R2 ,
Γ (θ2 + 1) Γ (θ2 )

which implies

F (y, x) ≤ max(R1 , R2 ).
E×E

This proves F (K) is uniformly bounded.

Stage(3): F (K) is equicontinuous in K.
Let (y, x) ∈ K and t1 , t2 ∈ (a, T] with t1 < t2 . Then we have




 ψ(t2 ) – ψ(a) 1–ς1 (F1 x)(t2 ) – ψ(t1 ) – ψ(a) 1–ς1 (F1 x)(t1 )
 t2
1 1   
≤ ς1 Hψθ1 (t2 , s)f1 s, x(s)  ds
Kψ (t2 , a) Γ (θ1 ) t1
 
1  t1 1 1   
+  ς1
θ1
Hψ (t2 , s) – ς1 Hψ (t1 , s)f1 s, x(s)  ds
θ1
Γ (θ1 ) a Kψ (t2 , a) Kψ (t1 , a)
1  θ1 ,ψ
ς –1   
≤ I ψ(s) – ψ(a) 1 (t2 ) f1 ·, x(·) E
Kψς1 (t2 , a) t1

 1  θ ,ψ
ς –1 
+  ς1 I 1 ψ(s) – ψ(a) 1 (t2 )
Kψ (t2 , a) a

1  θ ,ψ
ς1 –1    
– ς1 I 1 ψ(s) – ψ(a) (t1 ) f1 ·, x(·) E
Kψ (t1 , a) a
1 B (ς1 , θ1 ) 
θ +ς –1   
≤ ψ(t2 ) – ψ(t1 ) 1 1 f1 ·, x(·) E
Kψς1 (t2 , a) Γ (θ1 )

 1 B (ς1 , θ1 ) 
θ +ς –1 
+  ς1 ψ(t2 ) – ψ(a) 1 1
Kψ (t2 , a) Γ (θ 1 )

1 B (ς1 , θ1 ) 
θ1 +ς1 –1   
– ς1 ψ(t1 ) – ψ(a)  f1 ·, x(·)
Kψ (t1 , a) Γ (θ1 )  E
Abdo et al. Advances in Difference Equations (2020) 2020:316 Page 13 of 21


B (ς1 , θ1 )  
f1 ·, x(·) [ψ(t2 ) –ςψ(t1 )]
θ1 +ς1 –1
≤
Γ (θ1 ) E Kψ1 (t2 , a)


θ
θ 
+  ψ(t2 ) – ψ(a) 1 – ψ(t1 ) – ψ(a) 1  ,

which leads to

(F1 x)(t2 ) – (F1 x)(t1 ) → 0, as t2 → t1 .
E

Again applying the same reasoning, we have

F2 y(t2 ) – F2 y(t1 ) → 0, as t2 → t1 .
E

This exhibits that F (K) is equicontinuous. Stages 1–3 show that F is relatively compact
on K. By E(= C1–ς ;ψ ) type Arzelá–Ascoli theorem, F is compact on K.
Stage(4): Q is a contraction operator.
Let (y, x), (y∗ , x∗ ) ∈ E × E and t ∈ (a, T]. Then, by applying (H2 ), we easily get
 
(G1 x) – G1 x∗ ≤ Λf x – x∗
E 1 E

and
 
(G2 y) – G2 y∗ ≤ Λf y – y∗ .
E 2 E

Since Λf1 , Λf2 < 1, G is a contraction mapping. Using Theorem 1, we see that N has
at least one fixed point, which is the corresponding solution of ψ-Hilfer coupled system
(3)–(4). 

4 Ulam–Hyers stability
This part is devoted to proving the Ulam–Hyers (U-H) and generalized Ulam–Hyers (G-
U-H) stability of solution to the ψ-Hilfer coupled system (3)–(4).

Definition 5 The ψ-Hilfer coupled system (3)–(4) is U-H stable if there exists λ =
(λf1 , λf2 ) > 0 with the following property: For some  = (1 , 2 ) > 0 and each ( x) ∈ E × E,
y,
if
 θ1 ,η1 ,ψ  
D +  x(t)  ≤ 1 ,
y(t) – f1 t, (27)
a
 θ2 ,η2 ,ψ  
D +  y(t)  ≤ 2 ,
x(t) – f2 t, (28)
a

then there exists (y, x) ∈ E × E satisfying coupled system (3) with the following coupled
boundary conditions:
⎧
⎨y(T) =
y(T),
(29)
⎩x(T) =
x(T),

complying with

( x) – (y, x) E×E ≤ λ.
y,
Abdo et al. Advances in Difference Equations (2020) 2020:316 Page 14 of 21

Definition 6 The ψ-Hilfer coupled system (3)–(4) is G-U-H stable if there exists ϕ =
(ϕf1 , ϕf2 ) ∈ C(R+ , R+ ) with ϕ(0) = (ϕf1 (0), ϕf2 (0)) = (0, 0) such that for some  = (1 , 2 ) > 0
and for each solution ( x) ∈ E × E of inequalities (27), (28) there exists a solution (y, x) ∈
y,
E × E of coupled system (3)–(29) complying with

( x) – (y, x) E×E ≤ ϕ().
y,

Remark 1 A function ( x) ∈ E × E is a solution of inequalities (27)–(28) if and only if

y,
there exists a function (h1 , h2 ) ∈ E × E (where h1 depends on solution 
y and h2 depends
on solution 
x) such that
(i) |h1 (t)| ≤ 1 and |h2 (t)| ≤ 2 for t ∈ (a, T],
(ii) For t ∈ (a, T],

⎧
⎨Dθ1+,η1 ,ψy(t) = f1 (t,
x(t)) + h1 (t),
a
⎩D 2+ 2 
θ ,η ,ψ
x(t) = f2 (t,
y(t)) + h2 (t).
a

Lemma 8 Let ( x) ∈ E × E be the solution of inequalities (27)–(28). Then (y, x) ∈ E × E is

y,
the solution of the following fractional integral inequalities:

  
 1 t    21 Kψθ1 +1 (T, a)
 x(s) ds ≤
Hψθ1 (t, s)f1 s,
y(t) – Zx – Γ (θ1 ) Γ (θ1 + 1)
a

and

  
 1 t    22 Kψθ2 +1 (T, a)
 y(s) ds ≤
Hψθ2 (t, s)f2 s,
x(t) – Zu – Γ (θ2 ) Γ (θ2 + 1)
,
a

where

  
Kψς1 (t, a) 1 T  
Zx := w1 – Hψθ1 (T, s)f1 s,
x(s) ds
Kψς1 (T, a) Γ (θ1 ) a

and

  
Kψς2 (t, a) 1 T  
Zy := ς2 w2 – Hψθ2 (T, s)f2 s,
y(s) ds .
Kψ (T, a) Γ (θ2 ) a

Proof Thanks to Remark 1, we have

⎧
⎪ θ1 ,η1 ,ψ
⎨ D a+ 
⎪ y(t) = f1 (t,
x(t)) + h1 (t),
θ ,η ,ψ
Da2+ 2 
x(t) = f2 (t,
y(t)) + h2 (t), (30)
⎪
⎪
⎩
y(T) = w1 , x(T) = w2 .
Abdo et al. Advances in Difference Equations (2020) 2020:316 Page 15 of 21

Thanks to Theorem 15, the solution of (30) is defined by

⎧ ς1 T
⎪ Kψ (t,a)
⎪
⎪ [w1 – Γ (θ1 1 ) a Hψθ1 (T, s)f1 (s,
x(s)) ds]
⎪ Kψ (T,a)
⎪
ς1
⎪
⎪ ς
T
⎨ K 1 (t,a)
+ ςψ1 [w1 – Γ (θ1 1 ) a Hψθ1 (T, s)h1 (s) ds]

y(t) = Kψ (T,a) (31)
⎪
⎪ t
⎪
⎪ + Γ (θ1 1 ) a Hψθ1 (t, s)h1 (s) ds
⎪
⎪
⎪
⎩ t
+ Γ (θ1 1 ) a Hψθ1 (t, s)f1 (s,
x(s)) ds

and
⎧ ς2  T θ2
⎪ Kψ (t,a)
⎪
⎪ Kς2 (T,a) [w2 – Γ (θ2 ) a Hψ (T, s)f2 (s,
1
y(s)) ds]
⎪
⎪
⎪
⎪
ψ
ς
T
⎨ 2
K (t,a)
+ ςψ2 [w2 – Γ (θ1 1 ) a Hψθ2 (T, s)h1 (s) ds]

x(t) = Kψ (T,a) (32)
⎪
⎪ t
⎪
⎪ + Γ (θ1 2 ) a Hψθ2 (t, s)h1 (s) ds
⎪
⎪
⎪
⎩ t
+ Γ (θ1 2 ) a Hψθ2 (t, s)f2 (s,
y(s)) ds.

It follows from (31) and (32) with using Lemma 8 that

  
 1 t   
 x(s) ds
Hψθ1 (t, s)f1 s,
y(t) – Zx – Γ (θ1 ) a
 
Kψς1 (t, a) 1 T   1 t  
≤ Hψθ1 (T, s)h1 (s) ds + Hψθ1 (t, s)h1 (s) ds
Kψς1 (T, a) Γ (θ1 ) a Γ (θ1 ) a

Kψς1 (t, a) 1 1
≤ 1 ς1 Kθ1 +1 (T, a) + Kθ1 +1 (t, a)
Kψ (T, a) Γ (θ1 + 1) ψ Γ (θ1 + 1) ψ

21 Kψθ1 +1 (T, a)

≤ ,
Γ (θ1 + 1)

where Kψμ (t, a) ≤ Kψμ (T, a). Along the same lines, we can also get

   θ +1
 1 t    22 Kψ2 (T, a)
 y(s) ds ≤
Hψθ2 (t, s)f2 s,
x(t) – Zu – Γ (θ2 ) Γ (θ2 + 1)
. 
a

Theorem 6 Under the assumptions of Theorem 4, if 1 – Cθ1 Cθ2 = 0, then the ψ-Hilfer cou-
pled system (3)–(4) will be U-H and G-U-H stable in E × E, where Cθ1 := ΓΓ(ς(ς11+θ
)L1
1)
Kψθ1 +1 (T, a)
θ2 +1
and Cθ2 := Γ (ς2 +θ2 ) Kψ (T, a).
Γ (ς2 )L2

Proof Let ( x) ∈ E × E be the solution of coupled system (30) and (y, x) ∈ E × E be a

y,
unique solution of the ψ-Hilfer coupled system (3)–(4) with the conditions
⎧
⎨y(T) =
y(T) = w1 ,
(33)
⎩x(T) =
x(T) = w2 .

That is,

1 t  
y(t) = Zx + Hψθ1 (t, s)f1 s, x(s) ds (34)
Γ (θ1 ) a
Abdo et al. Advances in Difference Equations (2020) 2020:316 Page 16 of 21

and

1 t  
x(t) = Zy + Hψθ2 (t, s)f2 s, y(s) ds. (35)
Γ (θ2 ) a

Due to (33), Zx = Zx and Zy = Zy . Hence relations (34) and (35) become


1 t  
y(t) = Zx + Hψθ1 (t, s)f1 s, x(s) ds
Γ (θ1 ) a

and

1 t  
x(t) = Zy + Hψθ2 (t, s)f2 s, y(s) ds.
Γ (θ2 ) a

Through Lemma 8, we arrive at

  
 1 t   
 x(s) ds ≤ A1 1
Hψθ1 (t, s)f1 s,
y(t) – Zx – Γ (θ1 )
(36)
a

and
  
 1 t   
 y(s) ds ≤ A2 2 ,
Hψθ2 (t, s)f2 s,
x(t) – Zu – Γ (θ2 )
(37)
a

θ +1 θ +1
2K 1 (T,a) 2K 2 (T,a)
where A1 = Γψ(θ1 +1) and A2 = Γψ(θ2 +1) .
Thus, by (H2 ) and inequalities (36), (37), we reach

 

y(t) – y(t)
  
 1 t   
= y(t) – Zx – x(s) ds
Hψθ1 (t, s)f1 s,
Γ (θ1 ) a

1 t     
+ x(s) – f1 s, x(s)  ds
Hψθ1 (t, s)f1 s,
Γ (θ1 ) a

1 t  
≤ A1 1 + Hψθ1 (t, s)L1 
x(s) – x(s) ds
Γ (θ1 ) a

Γ (ς1 )L1 θ1 +ς1

≤ A1 1 + K (t, a)
x – xE ,
Γ (ς1 + θ1 ) ψ

which implies

y – yE ≤ A1 1 + Cθ1 
 x – xE , (38)

A1
where A1 := ς .
Kψ1 (T,a)
Similarly, we have

x – xE ≤ A2 2 + Cθ2 
 y – yE , (39)
Abdo et al. Advances in Difference Equations (2020) 2020:316 Page 17 of 21

A2
where A2 := ς . Inequalities (38) and (39) can be rewritten again as follows:
Kψ2 (T,a)

⎧
⎨
y – yE – Cθ1 
x – xE ≤ A1 1 ,
(40)
⎩
x – xE – Cθ y – yE ≤ A2 2 .
2

Now, we will represent the relations in (40) as matrices as follows:

    
1 –Cθ1 
y – yE A1 1
≤ .
–Cθ2 1 
x – xE A2 2

After simple computations of the above inequality, we can write

   Cθ1
  

y – yE 1
A1 1
≤ C θ × ,

x – xE
2 1

A2 2

where = 1 – Cθ1 Cθ2 = 0. This leads to

A1 Cθ A2

y – yE ≤ 1 + 1 2 ,

Cθ2 A1 A2
x – xE ≤
 1 + 2 .

By compiling the above two inequalities, we get

   
A1 + Cθ2 A1 A2 + Cθ1 A2

y – yE + 
x – xE ≤ 1 + 2 .

For  = max{1 , 2 } and

 
A1 + Cθ2 A1 + A2 + Cθ1 A2
λ= ,

we obtain

( x) – (y, x) E×E ≤ λ.
y, (41)

This proves that the ψ-Hilfer coupled system (3)–(4) is U-H stable.
Moreover, we could put into writing inequality (41) as

( x) – (y, x) E×E ≤ ϕ(),
y,

where ϕ() = λ with ϕ(0) = 0. This shows that the ψ-Hilfer coupled system (3)–(4) is
G-U-H stable. 

5 Examples
Here, we provide some illustrative examples to validate the obtained results.
Abdo et al. Advances in Difference Equations (2020) 2020:316 Page 18 of 21

Example 1 Consider the following coupled system of TVP for ψ-Hilfer FDE:
⎧ t
⎪
⎪
1 ,0;e 3
|x(t)|
⎪
⎪ D02+ y(t) = 1
+ √1 , t ∈ (0, 1],
⎨ t
10et+1 1+|x(t)| t+1
1 ,0;e 3
2 1 |y(t)|
√1 , t ∈ (0, 1], (42)
⎪
⎪ D 0+ x(t) = (t+6) 2 1+|y(t)| +
⎪
⎪
2 t+1
⎩y(1) = 1, x(1) = 2.

Set

1 u 1
f1 (t, u) = t+1
+√ , u, v ∈ R, t ∈ (0, 1]
10e 1 + u t+1

and

1 v 1
f2 (t, v) = + √ u, v ∈ R, t ∈ (0, 1].
(t + 6)2 1 + v 2 t + 1

Then

η (1–θ )   t 1 
1
C1–ς 1 ;ψ
1
[0, 1] = C 01 t [0, 1] = f1 : (0, 1] × R → R; e 3 – 1 2 f1 ∈ C[0, 1] ,
2 ;e 3

η (1–θ )   t 1 
2
C1–ς 2 ;ψ
2
[0, 1] = C 01 t [0, 1] = f2 : (0, 1] × R → R; e 3 – 1 2 f2 ∈ C[0, 1] ,
2 ;e 3

t
with θ1 = θ2 = 12 , η1 = η2 = 0, ς1 = ς2 = 12 , ψ(t) = e 3 , and (a, T] = (0, 1]. Evidently, the func-
tions f1 , f2 ∈ C 1 3t [0, 1]. So, hypothesis (H1 ) holds. For t ∈ (0, 1] and u, u∗ , v, v∗ ∈ R, we
2 ;e
have
    
f1 (t, u) – f1 t, u∗  ≤ 1 u – u∗ ,
10e
  ∗   
f2 (t, v) – f2 t, v  ≤ 1 v – v∗ .
36
1 1
Hence, hypothesis (H2 ) is satisfied with L1 = 10e and L2 = 36 . By some straightforward
calculations, we find that A1 = 1, A2 = 2 , Λf1 ≈ 0.08 < 1, Λf2 ≈ 0.06 < 1, f1 ≈ 1.5, and
1

f2 ≈ 1.7. Thus all the assumptions in Theorem 4 are satisfied. An application of The-
1
orem 4 shows that problem (42) has a unique solution in C 12 t [0, 1]. However, =
3
√ √ √ √2 ,e √
π ( 3 e–1) π ( 3 e–1) π ( 3 e–1)
1 – Cθ1 Cθ1 = 1 – 360e = 0, where Cθ1 = 10e
and Cθ2 = 36
. Therefore, from
Theorem 6, coupled system (42) is U-H and G-U-H stable.

Example 2 A particular case, for θ1 = θ2 = 12 , η1 = η2 = 1, and ψ(t) = t, the coupled system

for ψ-Hilfer FDE (42) reduces to the following coupled system for Caputo FDE:
⎧ 1
⎪
⎪ |x(t)|
⎪ D0+ y(t) = 10et+1 1+|x(t)| + √t+1 , t ∈ [0, 1],
C 2 1 1
⎨ 1
|y(t)|
⎪ D0+ x(t) = (t+2)2 1+|y(t)| + 2 t+1 , t ∈ [0, 1],
C 2 1 √1 (43)
⎪
⎪
⎩y(1) = 1, x(1) = 2.

Clearly, the function f1 , f2 ∈ C[0, 1]. Hence (H1 ) holds. Also, hypothesis (H2 ) is satisfied
1 1
with L1 = 10e and L2 = 36 . Via some straightforward computations, we see that A1 = 1,
Abdo et al. Advances in Difference Equations (2020) 2020:316 Page 19 of 21

A2 = 12 , Λf1 = 5e√
2
π
< 1, Λf2 = 9√1 π < 1, f1 = 1 + √4π , and f2 = 2 + √2π . Thus all the sup-
positions in Theorem 4 are satisfied. An application of Theorem 4 shows that problem
(43) has a unique solution in C[0, 1]. Moreover, since = 1 – Cθ1 Cθ1 = 1 – 90eπ 1
= 0, where
Cθ1 = 5e π and Cθ2 = 18 π , Theorem 6 shows that coupled system (43) is U-H and G-U-H
1
√ 1
√

stable.

Example 3 Consider the special case when ψ(t) = log t of the coupled system of TVP for
Hilfer–Hadamard FDE
⎧ 1 1 √
⎪
⎪ 4 , 3 ;log t 1 x2 (t)+|x(t)|
2, t ∈ (1, e],
⎪
⎨D11+ 1 y(t) = 5+2 t 1+|x(t)| +

4 , 3 ;log t e– log t |y(t)|

⎪ D1+ x(t) = elog t +8 1+|y(t)| + 2 , t ∈ (1, e],
3 (44)
⎪
⎪
⎩y(e) = 1, x(e) = 2.

Through comparing coupled system (3)–(4) with (44), we have θ1 = θ2 = 14 , η1 = η2 = 13 ,

1 u2 +u
√
ς1 = ς2 = 12 , ψ(t) = log t, a = 1, T = e, w1 = 1, w2 = 2 and f1 (t, u) = 5+2 t 1+u + 2 and f2 (t, v) =
e– log t v
elog t +8 1+v
+ 2 for u, v ∈ R, t ∈ (1, e]. Evidently, the functions f1 , f2 ∈ C 1 ;log [1, e]. So, condition
3
2
(H1 ) holds. It is simple to verify that

    
f1 (t, u) – f1 t, u∗  ≤ 1 u – u∗ , t ∈ (1, e], u, u∗ ∈ R,
7
  ∗  1  
f2 (t, v) – f2 t, v  ≤ v – v∗ , t ∈ (1, e], v, v∗ ∈ R.
9

Thus, hypothesis (H2 ) is satisfied with L1 = 17 and L2 = 19 . Through some easy calcula-
√
tions, we infer that A1 = 2, A2 = 32 , Λf1 ≈ 0.4 < 1, Λf2 ≈ 0.3 < 1, f1 ≈ 4.1, and f2 ≈ 5.3.
Since every supposition in Theorem 4 is satisfied, problem (44) has a unique solution in
1
C 12 ,log [1, e]. Further, as shown in Theorem 6, for every  = max(1 , 2 ) > 0, if ( x) ∈ E × E
y,
2
satisfies
⎧ √
⎨|Dθ1+,η1 ,log
y(t) – 1  x2 (t)+|
x(t)|
+ 2| ≤ 1 ,
a 5+2t 1+| x(t)|
⎩|D +
θ2 ,η2 ,log
 e– log t | y(t)|
x(t) – elog t +8 1+|y(t)| + 32 | ≤ 2 ,
a

there exists a unique solution (y, x) ∈ E × E such that

( x) – (y, x) E×E ≤ λ,
y,

where

 
A1 + Cθ2 A1 + A2 + Cθ1 A2
λ= ≈ 5.4 > 0,

√ √
A1 = A2 = 2
Γ ( 54 )
, Cθ1 = π
7Γ ( 34 )
, Cθ2 = π
9Γ ( 34 )
, and = 1 – Cθ1 Cθ1 = 1 – π
63Γ ( 34 )2
= 0.
Hence coupled system (44) is U-H and G-U-H stable. Some graphical presentation of
Example 3 is shown in Fig. 1.
Abdo et al. Advances in Difference Equations (2020) 2020:316 Page 20 of 21

Figure 1 Graphical representation of approximate solution for Example 3 at various fractional order of
coupled system (44)

Acknowledgements
We are thankful to the reviewers for their careful reading and constructive suggestions which improved this paper very
much.

Funding
Funding source is not available.

Abbreviations
FDEs, Fractional differential equations; TVPs, Terminal value problems; RL, Riemann–Liouville; U-H, Ulam–Hyers; G-U-H,
Generalized Ulam–Hyers.

Availability of data and materials

This is not applicable in our paper.

Competing interests
None of the authors has conflict of interests regarding this work.

Authors’ contributions
All authors contributed to this manuscript and approved the final version.

Author details
1
Department of Mathematics, Dr. Babasaheb Ambedkar Marathwada University, Aurangabad, India. 2 Department of
Mathematics, Hodeidah University, AL-Hodeidah, Yemen. 3 Department of Mathematics, University of Malakand,
Chakdara, 18000, Dir(L), Khyber Pakhtunkhwa, Pakistan.

Publisher’s Note
Springer Nature remains neutral with regard to jurisdictional claims in published maps and institutional affiliations.

Received: 14 April 2020 Accepted: 15 June 2020

References
1. Abdo, M.S., Ibrahim, A.G., Panchal, S.K.: Nonlinear implicit fractional differential equation involving ψ -Caputo
fractional derivative. Proc. Jangjeon Math. Soc. 22(3), 387–400 (2019)
2. Abdo, M.S., Panchal, S.K.: Fractional integro-differential equations involving ψ -Hilfer fractional derivative. Adv. Appl.
Math. Mech. 11(2), 338–359 (2019)
3. Abdo, M.S., Panchal, S.K., Hussien, H.S.: Fractional integro-differential equations with nonlocal conditions and ψ -Hilfer
fractional derivative. Math. Model. Anal. 24(4), 564–584 (2019)
4. Abdo, M.S., Panchal, S.K., Saeed, A.M.: Fractional boundary value problem with ψ -Caputo fractional derivative. Proc.
Indian Acad. Sci. Math. Sci. 129(5), 65 (2019)
5. Aftabizadeh, A.R., Lakshmikantham, V.: On the theory of terminal value problems for ordinary differential equations.
Nonlinear Anal., Theory Methods Appl. 5(11), 1173–1180 (1981)
6. Ali, Z., Kumam, P., Shah, K., Zada, A.: Investigation of Ulam stability results of a coupled system of nonlinear implicit
fractional differential equations. Mathematics 7(4), 341 (2019)
7. Ali, Z., Zada, A., Shah, K.: Ulam stability to a toppled systems of nonlinear implicit fractional order boundary value
problem. Bound. Value Probl. 2018, 175, 1–16 (2018)
8. Ali, Z., Zada, A., Shah, K.: On Ulam’s stability for a coupled systems of nonlinear implicit fractional differential
equations. Bull. Malays. Math. Sci. Soc. 42(5), 2681–2699 (2019)
9. Almalahi, M.A., Abdo, M.S., Panchal, S.K.: Existence and Ulam–Hyers–Mittag-Leffler stability results of ψ -Hilfer
nonlocal Cauchy problem. Rend. Circ. Mat. Palermo, II. Ser (2020). https://doi.org/10.1007/s12215-020-00484-8
10. Almeida, R.: A Caputo fractional derivative of a function with respect to another function. Commun. Nonlinear Sci.
Numer. Simul. 44, 460–481 (2017)
 Abdo et al. Advances in Difference Equations (2020) 2020:316 Page 21 of 21

11. Almeida, R.: Fractional differential equations with mixed boundary conditions. Bull. Malays. Math. Sci. Soc. 42(4),
1687–1697 (2019)
12. Atanackovic, T.M., Pilipovic, S., Stankovic, B., Zorica, D.: Fractional Calculus with Applications in Mechanics: Wave
Propagation, Impact and Variational Principles. Wiley, New York (2014)
13. Benchohra, M., Bouriah, S., Nieto, J.J.: Terminal value problem for differential equations with Hilfer–Katugampola
fractional derivative. Symmetry 11(5), 672 (2019)
14. Ford, N.J., Morgado, M.L., Rebelo, M.: A nonpolynomial collocation method for fractional terminal value problems. J.
Comput. Appl. Math. 275, 392–402 (2015)
15. Hallam, T.G.: A comparison principle for terminal value problems in ordinary differential equations. Trans. Am. Math.
Soc. 169, 49–57 (1972)
16. Jarad, F., Abdeljawad, T., Baleanu, D.: Caputo-type modification of the Hadamard fractional derivatives. Adv. Differ.
Equ. 2012(1), 142 (2012)
17. Kassim, M.D., Tatar, N.E.: Well-posedness and stability for a differential problem with Hilfer–Hadamard fractional
derivative. Abstr. Appl. Anal. 2013, 605029 (2013)
18. Katugampola, U.N.: A new approach to generalized fractional derivatives. Bull. Math. Anal. Appl. 6(4), 1–15 (2014)
19. Kilbas, A.A., Srivastava, H.M., Trujillo, J.J.: Theory and Applications of Fractional Differential Equations, vol. 204. Elsevier,
Amsterdam (2006)
20. Magin, R.L.: Fractional Calculus in Bioengineering. Begell House Inc. Publisher (2006)
21. Magin, R.L.: Fractional calculus models of complex dynamics in biological tissues. Comput. Math. Appl. 59,
1586–1593 (2010)
22. Mali, A.D., Kucche, K.D.: Nonlocal boundary value problem for generalized hilfer implicit fractional differential
equations (2020) arXiv preprint. arXiv:2001.08479
23. Manam, S.R.: Multiple integral equations arising in the theory of water waves. Appl. Math. Lett. 24, 1369–1373 (2011)
24. Oliveira, D.S., de Oliveira, E.C.: Hilfer–Katugampola fractional derivatives. Comput. Appl. Math. 37(3), 3672–3690 (2018)
25. Seemab, A., Rehman, M.U., Alzabut, J., Hamdi, A.: On the existence of positive solutions for generalized fractional
boundary value problems. Bound. Value Probl. 2019(1), 186 (2019)
26. Shah, K., Wang, J., Khalil, H., Khan, R.A.: Existence and numerical solutions of a coupled system of integral BVP for
fractional differential equations. Adv. Differ. Equ. 2018(1), 149 (2018)
27. Shah, S.H., Rehman, M.: A note on terminal value problems for fractional differential equations on infinite interval.
Appl. Math. Lett. 52, 118–125 (2016)
28. Shreve, W.E.: Terminal value problems for second order nonlinear differential equations. SIAM J. Appl. Math. 18,
783–791 (1970)
29. Sousa, J.V.D.C., de Oliveira, E.C.: On the ψ -Hilfer fractional derivative. Commun. Nonlinear Sci. Numer. Simul. 60, 72–91
(2018)
30. Sousa, J.V.D.C., de Oliveira, E.C.: Ulam–Hyers stability of a nonlinear fractional Volterra integro-differential equation.
Appl. Math. Lett. 81, 50–56 (2018)
31. Sousa, J.V.D.C., De Oliveira, E.C.: On the Ulam–Hyers–Rassias stability for nonlinear fractional differential equations
using the ψ -Hilfer operator. J. Fixed Point Theory Appl. 20(3), 96 (2018)
32. Sousa, J.V.D.C., de Oliveira, E.C.: A Gronwall inequality and the Cauchy-type problem by means of ψ -Hilfer operator.
Differ. Equ. Appl. 11(1), 87–106 (2019)
33. Sousa, J.V.D.C., De Oliveira, E.C.: Leibniz type rule: ψ -Hilfer fractional operator. Commun. Nonlinear Sci. Numer. Simul.
77, 305–311 (2019)
34. Sousa, J.V.D.C., Rodrigues, F.G., De Oliveira, E.C.: Stability of the fractional Volterra integro-differential equation by
means of ψ -Hilfer operator. Math. Methods Appl. Sci. 42(9), 3033–3043 (2019)
35. Zaky, M.A.: Recovery of high order accuracy in Jacobi spectral collocation methods for fractional terminal value
problems with non-smooth solutions. J. Comput. Appl. Math. 357, 103–122 (2019)
36. Zhou, Y.: Basic Theory of Fractional Differential Equations, vol. 6. World Scientific, Singapore (2014)

View publication stats