Next Article in Journal
A Temporal Second-Order Difference Scheme for Variable-Order-Time Fractional-Sub-Diffusion Equations of the Fourth Order
Next Article in Special Issue
Existence and Stability of Solutions for p-Proportional ω-Weighted κ-Hilfer Fractional Differential Inclusions in the Presence of Non-Instantaneous Impulses in Banach Spaces
Previous Article in Journal
Error-Based Switched Fractional Order Model Reference Adaptive Control for MIMO Linear Time Invariant Systems
Previous Article in Special Issue
The Impulsive Coupled Langevin ψ-Caputo Fractional Problem with Slit-Strip-Generalized-Type Boundary Conditions
 
 
Font Type:
Arial Georgia Verdana
Font Size:
Aa Aa Aa
Line Spacing:
Column Width:
Background:
Article

A Unified Approach to Solvability and Stability of Multipoint BVPs for Langevin and Sturm–Liouville Equations with CH–Fractional Derivatives and Impulses via Coincidence Theory

1
Department of Mathematics, School of Electronics & Information Engineering, Taizhou University, Taizhou 318000, China
2
School of Mathematics and Information Technology, Yuxi Normal University, Yuxi 653100, China
3
Applied Technology College, Soochow University, Suzhou 215325, China
*
Author to whom correspondence should be addressed.
Fractal Fract. 2024, 8(2), 111; https://doi.org/10.3390/fractalfract8020111
Submission received: 3 January 2024 / Revised: 6 February 2024 / Accepted: 9 February 2024 / Published: 13 February 2024

Abstract

:
The Langevin equation is a model for describing Brownian motion, while the Sturm–Liouville equation is an important mechanical model. This paper focuses on the solvability and stability of nonlinear impulsive Langevin and Sturm–Liouville equations with Caputo–Hadamard (CH) fractional derivatives and multipoint boundary value conditions. To unify the two types of equations, we investigate a general nonlinear impulsive coupled implicit system. By cleverly constructing relevant operators involving impulsive terms, we establish the coincidence degree theory and obtain the solvability. We explore the stability of solutions using nonlinear analysis and inequality techniques. As the most direct application, we naturally obtained the solvability and stability of the Langevin and Sturm–Liouville equations mentioned above. Finally, an example is provided to demonstrate the validity and availability of our major findings.

1. Introduction

In this paper, we delve into the following nonlinear impulsive CH–fractional differential Langevin Equation (1) and Sturm–Liouville Equation (2) under the m-point boundary value conditions
CH D x l θ 2 [ CH D x l θ 1 λ ] U ( x ) = F ( x , U ( x ) , CH D x l θ 2 + θ 1 U ( x ) ) , x [ α , β ] { x l } l = 0 m , Δ U ( x l ) = I l ( U ( x l ) ) , Δ [ CH D x l θ 1 U ( · ) ] ( x l ) = J l ( U ( x l ) ) , 1 l m , l = 1 m + 1 a l U ( ξ l ) = c , l = 1 m + 1 b l [ CH D x l θ 1 U ( · ) ] ( ξ l ) = d ,
where 0 < α < β , 0 < θ 1 , θ 2 1 , λ > 0 , α = x 0 = ξ 1 < x 1 < ξ 2 < x 2 < ξ 3 < x 3 < < x m < ξ m = x m + 1 = β , a l , b l , c , d R , l = 1 m a l 0 , l = 1 m b l 0 , CH D x l is the CH–fractional derivative, F C ( [ α , β ] × R 2 , R + ) , I l , J l C ( R , R ) , Δ U ( x l ) = U ( x l + ) U ( x l ) , Δ [ CH D x l θ 1 U ( · ) ] ( x l ) = [ CH D x l θ 1 U ( · ) ] ( x l + ) [ CH D x l θ 1 U ( · ) ] ( x l ) , [ CH D x l θ 1 U ( · ) ] ( x l ) = [ CH D x l θ 1 U ( · ) ] ( x l ) , U ( x l ) = U ( x l ) , 1 l m , i = 1 , 2 .
CH D x l θ 2 [ p ( x ) CH D x l θ 1 ] U ( x ) = F ( x , U ( x ) ) , x [ α , β ] { x l } l = 0 m , Δ U ( x l ) = I l ( U ( x l ) ) , Δ [ p ( · ) CH D x l θ 1 U ( · ) ] ( x l ) = J l ( U ( x l ) ) , 1 l m , l = 1 m + 1 a l U ( ξ l ) = c , l = 1 m + 1 b l [ p ( · ) CH D x l θ 1 U ( · ) ] ( ξ l ) = d ,
where p C ( [ α , β ] , ( 0 , + ) ) , F C ( [ α , β ] × R , R + ) . The other conditions are the same as (1).
Remark 1.
In (1) and (2), the impulse functions I l , J l are only related to U ( x l ) since CH D x l * U ( x l ) 0. In addition, (1) is implicit and (2) is explicit.
As is well known, the Langevin equation is a famous mathematical model that describes the random motion of particles annihilating in a fluid due to collisions between particles and fluid molecules. Compared with the integer-order Langevin equation, the fractional-order Langevin equation is more accurate in describing the random motion of particles in complex viscoelastic fluids. In recent years, many papers dealing with the fractional Langevin equations have been published. For example, Ahmadova and Mahmudov [1] studied the explicit analytical solutions for several families of Langevin differential equations with general fractional orders. Salem et al. [2] applied Darbo’s fixed-point theorem to investigate the existence of solutions for the three-point boundary value problem of a fractional Langevin equation in the noncompact Hausdorff space. Zhao, in [3,4,5], discussed the stability of several types of nonlinear fractional Langevin equations with delays and controls. In [6,7,8], the authors explored the controllability problem of fractional Langevin equations. Other papers are [9,10,11] concerned with the dynamics of stochastic Langevin equations.
Furthermore, the Sturm–Liouville equation, which includes the Helmholtz equation, Bessel equation, and Legendre equation, also represents another important class of mathematical and physical equations. Therefore, study of the fractional Sturm–Liouville equation has also become a hot topic in recent years. Afarideh et al. [12] used the pseudospectral method and Chebyshev cardinal functions to solve the Caputo fractional Sturm–Liouville eigenvalue problems. Sadabad and Akbarfam [13] provided an efficient numerical method to estimate the eigenvalues and eigenfunctions of the fractional Sturm–Liouville equation. Allahverdiev et al. [14] obtained a completeness theorem of singular dissipative conformable fractional Sturm–Liouville operators. Goel et al. [15] probed the numerical calculation of mixed boundary value problems for the generalized fractional Sturm–Liouville system. Kumar and Mehra [16] adopted the wavelet method to solve the Sturm–Liouville fractional optimal control problem. In fact, there are many research achievements on fractional Langevin and Sturm–Liouville equations. We will not elaborate further here. However, previous works in the literature have studied the two types of equations separately, and there is rarely a unified approach. Accordingly, it is novel and fascinating to unify Equations (1) and (2) for research purposes.
To address the solvability and stability of Equations (1) and (2) together, we consider a general system including (1) and (2) as follows:
CH D x l θ 1 U 1 ( x ) = F 1 ( x , U 1 ( x ) , U 2 ( x ) , CH D x l θ 1 U 1 ( x ) , CH D x l θ 2 U 2 ( x ) ) , CH D x l θ 2 U 2 ( x ) = F 2 ( x , U 1 ( x ) , U 2 ( x ) , CH D x l θ 1 U 1 ( x ) , CH D x l θ 2 U 2 ( x ) ) , Δ U 1 ( x l ) = I l ( U 1 ( x l ) , U 2 ( x l ) ) , Δ U 2 ( x l ) = J l ( U 1 ( x l ) , U 2 ( x l ) ) , l = 1 m + 1 a l U 1 ( ξ l ) = c , l = 1 m + 1 b l U 2 ( ξ l ) = d ,
where F i C ( [ α , β ] × R 4 , R + ) , I l , J l C ( R 2 , R ) , Δ U i ( x l ) = U i ( x l + ) U i ( x l ) with U i ( x l ) = U i ( x l ) , 1 l m , i = 1 , 2 . The other conditions are the same as (1).
Remark 2.
When boundary conditions l = 1 m + 1 a l U 1 ( ξ l ) = c , l = 1 m + 1 b l U 2 ( ξ l ) = d degenerate to U 1 ( α ) + U 1 ( β ) = 0 , U 2 ( α ) + U 2 ( β ) = 0 , system (3) becomes an impulsive implicit antiperiodic boundary value problem.
The Hadamard fractional calculus proposed by Hadamard in 1892 [17] is a direct and effective extension of Riemann–Liouville (RL) fractional calculus. Its prominent feature is that the logarithmic kernel H ( x , s ) = ( log x s ) ϑ 1 replaces the polynomial kernel G ( x , s ) = ( x s ) ϑ 1 in the RH–calculus definition. These two kernels maintain certain mathematical commonalities. For example, they have singularity when 0 < ϑ 1 , that is, G ( x , s ) , H ( x , s ) , as s x . Some of their properties are also significantly distinguished. For instance, ζ > 0 , H ( ζ x , ζ s ) = H ( x , s ) , but G ( ζ x , ζ s ) = ζ ϑ 1 G ( x , s ) G ( x , s ) . As an important class of differential equations, the theory and application of Hadamard-type (H-type) fractional differential equations (FDEs) have received extensive and in-depth research, which has achieved fruitful results (see [18,19,20,21,22,23,24,25,26,27,28,29,30,31]). Until now, the exploration of various dynamic properties of H-type fractional differential equations has been a very lively research topic. For example, in [32], the authors discussed the logarithmic decay stability of an H-type fractional equation. Rao et al. [33] considered the problem of multiplicity of solutions for a mixed H–fractional Laplacian system. Zhao [34,35] thought about the approximation and Hyers–Ulam-type stability of two classes of H–fractional boundary value problems. In [36,37], the authors studied the numerical calculation problem of H–fractional equations. Ortigueira et al. [38] explored the unification of H–calculus and RL–calculus. Dhawan et al. [39] applied the upper and lower solution method to analyze a neutral H–fractional equation. Ahmad et al. [40] investigated a coupled system of Hilfer–Hadamard fractional equations. Ben Makhlouf et al. [41] studied the existence, uniqueness, and averaging principle for Hadamard Ito–Doob stochastic delay fractional integral equations. Briefly, the properties, research approaches, and generalization of the concept of H–derivative, as well as the effects of delay, impulse, and random factors on H–fractional differential systems, have always attracted the attention of scholars. We further refer to [42,43,44,45,46,47].
Generally speaking, the study of implicit forms of differential equations is relatively more difficult than explicit forms. Therefore, the research achievements on implicit differential equations are also rarer than those on explicit differential equations. Only a small number of published papers deal with the solvability and stability of implicit Hadamard fractional differential equations (see [48,49,50,51,52,53,54,55,56]). Some academic researchers have applied the theory of coincidence degree to study the solvability of integer-order nonlinear functional differential equations and have achieved fruitful results (see [57,58,59,60,61,62,63,64,65,66,67,68]). In the theory of coincidence degree, the construction of relevant operators is highly skilled, which brings difficulties to the application of this method. Consequently, there are relatively few works [69,70,71,72,73] on the existence of solutions to fractional differential equations via coincidence degree theory.
Owed to the aforementioned, it is fascinating and challenging to investigate the solvability of system (3) by coincidence degree theory. The highlights of this paper mainly comprise the following. (a) Our work enriches and fills the gap in the study of nonlocal boundary value problems for implicit and impulsive fractional coupled systems. (b) In the establishment of coincidence degree theory, we cleverly constructed and proved the complete continuity of the relevant operators for the first time in the study of impulsive fractional differential equations. (c) As an important application of our basic results, we obtained the solvability and stability of the Langevin system and Sturm–Liouville system.
The remaining content of this paper is arranged as follows. Some necessary concepts and lemmas are stated in Section 2. Section 3 studies the existence, uniqueness, and stability of solutions to (3). Section 4 discusses the solvability and stability of the Langevin system (1) and Sturm–Liouville system (2), and gives an example to check the validity and availability of our basic findings. Finally, we provide a simple conclusion of research approaches, results, and significance in Section 5.

2. Preliminaries

This section mainly introduces some basic knowledge required for this article. We first state an important result of the coincidence theory for solving operator equations as follows.
Lemma 1
(Mawhin [74]). Let E , F be Banach spaces, Θ E , a bounded open subset. If L : E F is a 0-index Fredholm operator, and N : E × [ 0 , 1 ] F is L -compact on Θ ¯ × [ 0 , 1 ] , then there has to be at least X * Θ ¯ Dom L s.t. L X * = N ( X * , 1 ) provided that the following is true:
(a1)
If X solves L X = η · N ( X , η ) , then X Θ Dom L , η ( 0 , 1 ) ;
(a2)
Q N ( X , 0 ) X 0 , X Θ Ker L ;
(a3)
deg ( J Q N ( X , 0 ) , Θ Ker L , 0 ) 0 ;
where Q , J : F F are projected and homotopy, respectively.
Next, we need to review the basic concepts and results of Caputo–Hadamard fractional calculus.
Definition 1
([75]). Let θ > 0 , 0 < α < β < , and U : [ α , β ] R ; the definition of θ-order Hadamard fractional integral of U is
H J α + θ U ( x ) = 1 Γ ( θ ) α x log x z θ 1 U ( z ) d z z .
Definition 2
([75]). Let θ > 0 , 0 < α < β < , m = [ θ ] + 1 , and x d d x m 1 U ( x ) A C [ α , β ] ; the definition of θ-order Caputo–Hadamard fractional derivative of U is
CH D α + θ U ( x ) = 1 Γ ( m θ ) α x log x z m θ 1 z d d z m U ( z ) d z z .
Lemma 2
([75]). Let θ > 0 , 0 < α < β < , m = [ θ ] + 1 , and x d d x m 1 U ( x ) A C [ α , β ] , then
H J α + θ ( CH D α + θ U ( x ) ) = U ( x ) + j = 0 m 1 z d d z j U ( z ) | z = α j ! log x α j .
To obtain a prior estimate of the solution to BVP (3), the following lemma is required.
Lemma 3.
Let 0 < α < β , 0 < θ 1 , θ 2 1 , a l , b l , c , d R , l = 1 m a l 0 , l = 1 m b l 0 , F i C ( [ α , β ] × R 4 , R + ) , I l , J l C ( R 2 , R ) , 1 l m , i = 1 , 2 . If U ( x ) = ( U 1 ( x ) , U 2 ( x ) ) T solves the BVP (3), then U ( x ) = ( U 1 ( x ) , U 2 ( x ) ) T also solves the following integral equation
U 1 ( x ) = c 0 U + 1 Γ ( θ 1 ) α x log x z θ 1 1 ϕ U 1 ( z ) d z z , x [ α , x 1 ] , U 2 ( x ) = d 0 U + 1 Γ ( θ 2 ) α x log x z θ 2 1 ψ U 2 ( z ) d z z , x [ α , x 1 ] , U 1 ( x ) = c l U + 1 Γ ( θ 1 ) x l x log x z θ 1 1 ϕ U 1 ( z ) d z z , x ( x l , x l + 1 ] , U 2 ( x ) = d l U + 1 Γ ( θ 2 ) x l x log x z θ 2 1 ψ U 2 ( z ) d z z , x ( x l , x l + 1 ] ,
where 1 l m , a = : l = 1 m a l , b = : l = 1 m b l ,
ϕ U 1 ( x ) = F 1 ( x , U 1 ( x ) , U 2 ( x ) , ϕ U 1 ( x ) , ψ U 2 ( x ) ) , ψ U 2 ( x ) = F 2 ( x , U 1 ( x ) , U 2 ( x ) , ϕ U 1 ( x ) , ψ U 2 ( x ) ) ,
c 0 U = 1 a [ c l = 1 m a l k = 1 l ( 1 Γ ( θ 1 ) x k 1 x k log x k z θ 1 1 ϕ U 1 ( z ) d z z + I k ( U 1 ( x k ) , U 2 ( x k ) ) ) 1 Γ ( θ 1 ) l = 1 m a l x l ξ l log ξ l z θ 1 1 ϕ U 1 ( z ) d z z ] , d 0 U = 1 b [ d l = 1 m b l k = 1 l ( 1 Γ ( θ 2 ) x k 1 x k log x k z θ 2 1 ψ U 2 ( z ) d z z + J k ( U 1 ( x k ) , U 2 ( x k ) ) ) 1 Γ ( θ 2 ) l = 1 m b l x l ξ l log ξ l z θ 2 1 ψ U 2 ( z ) d z z ] , c l U = c 0 U + 1 Γ ( θ 1 ) k = 1 l x k 1 x k log x k z θ 1 1 ϕ U 1 ( z ) d z z + k = 1 l I k ( U 1 ( x k ) , U 2 ( x k ) ) , d l U = d 0 U + 1 Γ ( θ 2 ) k = 1 l x k 1 x k log x k z θ 2 1 ψ U 2 ( z ) d z z + k = 1 l J k ( U 1 ( x k ) , U 2 ( x k ) ) .
Proof. 
For x [ ξ 1 , x 1 ] = [ α , x 1 ] , by Lemma 2 and (3), we have
U 1 ( x ) = c 0 U + 1 Γ ( θ 1 ) α x log x z θ 1 1 ϕ U 1 ( z ) d z z , U 2 ( x ) = d 0 U + 1 Γ ( θ 2 ) α x log x z θ 2 1 ψ U 2 ( z ) d z z .
For x ( x 1 , x 2 ] , similar to (5), we obtain
U 1 ( x ) = c 1 U + 1 Γ ( θ 1 ) x 1 x log x z θ 1 1 ϕ U 1 ( z ) d z z , U 2 ( x ) = d 1 U + 1 Γ ( θ 2 ) x 1 x log x z θ 2 1 ψ U 2 ( z ) d z z .
From (5) and (6) and the impulsive conditions of (3), we yield that
c 1 U c 0 U = 1 Γ ( θ 1 ) α x 1 log x 1 z θ 1 1 ϕ U 1 ( z ) d z z + I 1 ( U 1 ( x 1 ) , U 2 ( x 1 ) ) , d 1 U d 0 U = 1 Γ ( θ 2 ) α x 1 log x 1 z θ 2 1 ψ U 2 ( z ) d z z + J 1 ( U 1 ( x 1 ) , U 2 ( x 1 ) ) .
In the same manner, for x ( x l , x l + 1 ] , 2 l m , we obtain
U 1 ( x ) = c l U + 1 Γ ( θ 1 ) x l x log x z θ 1 1 ϕ U 1 ( z ) d z z , U 2 ( x ) = d l U + 1 Γ ( θ 2 ) x l x log x z θ 2 1 ψ U 2 ( z ) d z z ,
and
c l U c l 1 U = 1 Γ ( θ 1 ) x l 1 x l log x l z θ 1 1 ϕ U 1 ( z ) d z z + I l ( U 1 ( x l ) , U 2 ( x l ) ) , d l U d l 1 U = 1 Γ ( θ 2 ) x l 1 x l log x l z θ 2 1 ψ U 2 ( z ) d z z + J l ( U 1 ( x l ) , U 2 ( x l ) ) .
In view of (7) and (9), we derive that, for 1 l m ,
c l U = c 0 U + 1 Γ ( θ 1 ) k = 1 l x k 1 x k log x k z θ 1 1 ϕ U 1 ( z ) d z z + k = 1 l I k ( U 1 ( x k ) , U 2 ( x k ) ) , d l U = d 0 U + 1 Γ ( θ 2 ) k = 1 l x k 1 x k log x k z θ 2 1 ψ U 2 ( z ) d z z + k = 1 l J k ( U 1 ( x k ) , U 2 ( x k ) ) .
It follows from the boundary value conditions in (3) that
l = 1 m a l c l U + 1 Γ ( θ 1 ) x l ξ l log ξ l z θ 1 1 ϕ U 1 ( z ) d z z = c , l = 1 m b l d l U + 1 Γ ( θ 2 ) x l ξ l log ξ l z θ 2 1 ψ U 2 ( z ) d z z = d .
From (10) and (11), we have
c 0 U = 1 a [ c l = 1 m a l k = 1 l 1 Γ ( θ 1 ) x k 1 x k log x k z θ 1 1 ϕ U 1 ( z ) d z z + I k ( U 1 ( x k ) , U 2 ( x k ) ) 1 Γ ( θ 1 ) l = 1 m a l x l ξ l log ξ l z θ 1 1 ϕ U 1 ( z ) d z z ] , d 0 U = 1 b [ d l = 1 m b l k = 1 l 1 Γ ( θ 2 ) x k 1 x k log x k z θ 2 1 ψ U 2 ( z ) d z z + J k ( U 1 ( x k ) , U 2 ( x k ) ) 1 Γ ( θ 2 ) l = 1 m b l x l ξ l log ξ l z θ 2 1 ψ U 2 ( z ) d z z ] .
From (5), (8), and (12), we gain the integral Equation (4). The proof is completed. □

3. Solvability and Stability of (3)

In this section, we first establish the theory of coincidence degree for BVP (3), and apply Lemma 1 to explore its solvability. Let
P C [ α , β ] = C [ α , x 1 ] W C ( x l , x l + 1 ] : W ( x l ) = W ( x l ) , and W ( x l + ) exists , 1 l m ,
X = U = ( U 1 , U 2 ) T : U i P C [ α , β ] , l = 1 m + 1 a l U 1 ( ξ l ) = c , l = 1 m + 1 b l U 2 ( ξ l ) = d , i = 1 , 2 ,
Y = { V = ( U , h 1 , , h m ) : U X , h l = ( h l 1 , h l 2 ) T R 2 , 1 l m } .
Define some norms as follows:
W P C = max 0 l m sup x l x x l + 1 W ( x ) , W P C [ α , β ] ,
U X = max { U 1 P C , U 2 P C } , U X ,
V Y = max { U X , | h l 1 | , | h 12 | , , | h m 1 | , | h m 2 | } , V Y .
Consequently, ( P C [ α , β ] , · P C ) , ( X , · X ) , ( Y , · Y ) are the Banach spaces. Define two operators L : Dom L = X Y and N : X × [ 0 , 1 ] Y as
L U = CH D x l θ 1 U 1 CH D x l θ 2 U 2 , Δ U 1 ( x 1 ) Δ U 2 ( x 1 ) , , Δ U 1 ( x m ) Δ U 2 ( x m ) ,
N ( U , η ) = F 1 ( x , η ) F 2 ( x , η ) , I 1 ( U ( x 1 ) ) J 1 ( U ( x 1 ) ) , , I m ( U ( x m ) ) J m ( U ( x m ) ) ,
where I l ( U ( x l ) ) = I l ( U 1 ( x l ) , U 2 ( x l ) ) , J l ( U ( x l ) ) = J l ( U 1 ( x l ) , U 2 ( x l ) ) , 1 l m ,
F j ( x , η ) = F j ( x , U 1 ( x ) , U 2 ( x ) , η · CH D x l θ 1 U 1 ( x ) , η · CH D x l θ 2 U 2 ( x ) ) , j = 1 , 2 .
Lemma 4.
L defined by (13) is a 0-index Fredholm operator.
Proof. 
L is obviously linear. The kernel of L , Ker L is defined by
Ker L = { U ¯ = ( U ¯ 1 , U ¯ 2 ) T Dom L : L U ¯ = 0 } .
From (13) and (15), it is similar to Lemma 3 that
U ¯ 1 ( x ) = c ¯ 0 , U ¯ 2 ( x ) = d ¯ 0 , x [ α , x 1 ] , U ¯ 1 ( x ) = c ¯ l , U ¯ 2 ( x ) = d ¯ l , x ( x l , x l + 1 ] , 1 l m ,
and
l = 1 m a l c ¯ l = c , l = 1 m b l d ¯ l = d , c ¯ l = c ¯ 0 , d ¯ l = d ¯ 0 , 1 l m .
We derive from (16) and (17) that Ker L = { U ¯ = ( c a , d b ) T } . Therefore, dim ( Ker L ) = 0 .
The image set of L , Im L is defined by
Im L = { V Y : U Dom L s . t . L U = V } .
Obviously, Im L Y . For all V Y ,
V = V 1 V 2 , h 11 h 12 , , h m 1 h m 2 ,
it follows from L U = V that
CH D x l θ 1 U 1 ( x ) = V 1 ( x ) , CH D x l θ 2 U 2 ( x ) = V 2 ( x ) , x [ α , β ] { x l } , Δ U 1 ( x l ) = h l 1 , Δ U 2 ( x l ) = h l 2 , 1 l m , l = 1 m + 1 a l U 1 ( ξ l ) = c , l = 1 m + 1 b l U 2 ( ξ l ) = d .
Similar to Lemma 3, Equation (19) allows a unique solution U * = ( U 1 * , U 2 * ) T as follows:
U 1 * ( x ) = c 0 * + 1 Γ ( θ 1 ) α x log x z θ 1 1 V 1 ( z ) d z z , x [ α , x 1 ] , U 2 * ( x ) = d 0 * + 1 Γ ( θ 2 ) α x log x z θ 2 1 V 2 ( z ) d z z , x [ α , x 1 ] , U 1 * ( x ) = c l * + 1 Γ ( θ 1 ) x l x log x z θ 1 1 V 1 ( z ) d z z , x ( x l , x l + 1 ] , U 2 * ( x ) = d l * + 1 Γ ( θ 2 ) x l x log x z θ 2 1 V 2 ( z ) d z z , x ( x l , x l + 1 ] ,
where
c 0 * = 1 a [ c l = 1 m a l k = 1 l 1 Γ ( θ 1 ) x k 1 x k log x k z θ 1 1 V 1 ( z ) d z z + h k 1 1 Γ ( θ 1 ) l = 1 m a l x l ξ l log ξ l z θ 1 1 V 1 ( z ) d z z ] , d 0 * = 1 b [ d l = 1 m b l k = 1 l 1 Γ ( θ 2 ) x k 1 x k log x k z θ 2 1 V 2 ( z ) d z z + h k 2 1 Γ ( θ 2 ) l = 1 m b l x l ξ l log ξ l z θ 2 1 V 2 ( z ) d z z ] , c l * = c 0 * + 1 Γ ( θ 1 ) k = 1 l x k 1 x k log x k z θ 1 1 V 1 ( z ) d z z + k = 1 l h k 1 , d l * = d 0 * + 1 Γ ( θ 2 ) k = 1 l x k 1 x k log x k z θ 2 1 V 2 ( z ) d z z + k = 1 l h k 2 .
Clearly, U * = ( U 1 * , U 2 * ) T Dom L . Thereby, Y Im L . Thus, we claim that Y = Im L and Im L is closed, as well as
codim ( Im L ) = dim ( Y / Im L ) = dim ( Y / Y ) = 0 = dim ( Ker L ) .
Based on the definition of 0-index Fredholm operator, we know that Lemma 4 is true. The proof is completed. □
P : X X is defined by
P U = P U 1 ( x ) U 2 ( x ) = c a d b .
Obviously, P 2 = P and Ker P = X . Noticing that Ker L is zero space, we yield that Im P = Ker L and X = Ker L Ker P . Therefore, L | Dom L Ker P : X = Dom L Ker P Im L = Y ; there exists an inverse operator K P . For each V Y , K P V = ( U 1 * ( x ) , U 2 * ( x ) ) T X is defined as (19) and (20). Define Q : Y Y as
Q V 1 V 2 , h l 1 h l 2 l = 1 m = 1 β α α β V 1 ( x ) d x 1 β α α β V 2 ( x ) d x , 0 0 l = 1 m ,
then Q 2 = Q , Ker Q = Im L = Y and Y = Im L Im Q . In addition,
Q N ( U , η ) = 1 β α α β F 1 ( x , η ) d x 1 β α α β F 2 ( x , η ) d x , 0 0 l = 1 m .
By substituting N ( U , η ) Q N ( U , η ) into (20) and (21), we can obtain the expression of K P ( I Q ) N ( U , η ) .
Lemma 5.
N defined by (14) is L -compact.
Proof. 
For all bounded subsets Θ X , it suffices to prove that K P ( I Q ) N ( Θ ¯ ) ¯ is relatively compact. Indeed, it follows from the continuity of F i , I i l and J i l ( i = 1 , 2 ; 1 l m ) that K P ( I Q ) N ( U , η ) is uniformly bounded. Set a sequence of functions, n N ,
H ( U n , η ) ( x ) = K P ( I Q ) N ( U n , η ) ( x ) , x [ α , x 1 ] ( x 1 , x l + 1 ] , 1 l m , K P ( I Q ) N ( U n , η ) ( x + ) , x = x l , 1 l m ,
then { H l U n = : K P ( I Q ) N ( U n , η ) ( x ) | [ x l , x l + 1 ] } n N is uniformly bounded on [ x l , x l + 1 ] ( 0 l m ) . Therefore, { H l U n } has a uniformly convergent subsequence { H l U n 1 ( 1 ) } on [ x 0 , x 1 ] = [ α , x 1 ] . Similarly, { H l U n 1 ( 1 ) } has a uniformly convergent subsequence { H l U n 2 ( 2 ) } on [ x 1 , x 2 ] . By repeating the above, { H l U n m 1 ( m 1 ) } has a uniformly convergent subsequence { H l U n m ( m ) } on [ x m , x m + 1 ] = [ x m , β ] . Thus, { H l U n m ( m ) } is a uniformly convergent subsequence on [ α , β ] , which means that K P ( I Q ) N ( Θ ¯ ) ¯ is relatively compact. The proof is completed. □
Theorem 1.
BVP (3) permits a unique solution U ˜ ( x ) = ( U ˜ 1 ( x ) , U ˜ 2 ( x ) ) T X , provided that the following conditions ( A 1 ) ( A 4 ) are true.
(A1)
Assume that 0 < α < β , 0 < θ 1 , θ 2 1 , α = x 0 = ξ 1 < x 1 < ξ 2 < x 2 < ξ 3 < x 3 < < x m < ξ m = x m + 1 = β , a l , b l , c , d R , a = l = 1 m a l 0 , b = l = 1 m b l 0 , F i C ( [ α , β ] × R 4 , R + ) , I l , J l C ( R 2 , R ) , 1 l m , i = 1 , 2 .
(A2)
u = ( u 1 , u 2 , u 3 , u 4 ) T , u ¯ = ( u ¯ 1 , u ¯ 2 , u ¯ 3 , u ¯ 4 ) T R 4 , L i 1 , L i 2 , L i 3 , L i 4 > 0 s.t.
| F i ( x , u ) F i ( x , u ¯ ) | j = 1 4 L i j | u j u ¯ j | , x [ α , β ] , i = 1 , 2 .
(A3)
v = ( v 1 , v 2 ) T , v ¯ = ( v ¯ 1 , v ¯ 2 ) T R 2 , M l 1 , M l 2 , N l 1 , N l 2 > 0 s.t.
| I l ( v ) I l ( v ¯ ) | i = 1 2 M l i | v i v ¯ i | , | J l ( v ) J l ( v ¯ ) | i = 1 2 N l i | v i v ¯ i | , 1 l m .
(A4)
D > 0 , 0 < L 13 , L 24 , ρ 1 , ρ 2 < 1 , where a + = l = 1 m | a l | , b + = l = 1 m | b l | ,
ρ 1 = 1 | a | [ 2 D 1 a + D Γ ( θ 1 + 1 ) k = 1 m log x k x k 1 θ 1 + D 1 D Γ ( θ 1 + 1 ) k = 1 m | a k | log ξ k x k 1 θ 1 + 2 a + k = 1 m ( M k 1 + M k 2 ) ] + D 1 D Γ ( θ 1 + 1 ) log β α θ 1 ,
ρ 2 = 1 | b | [ 2 D 2 b + D Γ ( θ 2 + 1 ) k = 1 m log x k x k 1 θ 2 + D 2 D Γ ( θ 2 + 1 ) k = 1 m | b k | log ξ k x k 1 θ 2 + 2 b + k = 1 m ( N k 1 + N k 2 ) ] + D 2 D Γ ( θ 2 + 1 ) log β α θ 2 ,
D = 1 L 13 L 14 L 23 1 L 24 , D 1 = L 11 + L 12 L 14 L 21 + L 22 1 L 24 , D 2 = 1 L 13 L 11 + L 12 L 23 L 21 + L 22 .
Proof. 
We will apply Lemma 1 to prove Theorem 1. To this end, define L , N , P , and Q as (13), (14), (22) and (23), respectively. From Lemmas 4 and 5, we claim that L is a 0-index Fredholm operator, and N is L -compact. If U ( x ) = ( U 1 ( x ) , U 2 ( x ) ) T X is a solution of BVP (3), then we know from Lemma 3 that U ( x ) = ( U 1 ( x ) , U 2 ( x ) ) T X satisfies (4). Noticing that F i ( x , 0 , 0 , 0 , 0 ) = 0 ( i = 1 , 2 ) , we derive from ( A 2 ) that
| ϕ U 1 ( x ) | = | F 1 ( x , U 1 ( x ) , U 2 ( x ) , ϕ U 1 ( x ) , ψ U 2 ( x ) ) F 1 ( x , 0 , 0 , 0 , 0 ) | L 11 | U 1 ( x ) |   +   L 12 | U 2 ( x ) |   +   L 13 | ϕ U 1 ( x ) |   +   L 14 | ψ U 2 ( x ) | , ( L 11 + L 12 ) U X + L 13 | ϕ U 1 ( x ) |   + L 14 | ψ U 2 ( x ) | ,
and
| ψ U 2 ( x ) | = | F 2 ( x , U 1 ( x ) , U 2 ( x ) , ϕ U 1 ( x ) , ψ U 2 ( x ) ) F 2 ( x , 0 , 0 , 0 , 0 ) | L 21 | U 1 ( x ) |   +   L 22 | U 2 ( x ) |   +   L 23 | ϕ U 1 ( x ) |   +   L 24 | ψ U 2 ( x ) | ( L 21 + L 22 ) U X + L 23 | ϕ U 1 ( x ) |   + L 24 | ψ U 2 ( x ) | .
Equations (24) and (25) lead to
( 1 L 13 ) | ϕ U 1 ( x ) |     L 14 | ψ U 2 ( x ) |   ( L 11 + L 12 ) U X , L 23 | ϕ U 1 ( x ) |   + ( 1 L 24 ) | ψ U 2 ( x ) |   ( L 21 + L 22 ) U X .
From ( A 4 ) and (26), we have
| ϕ U 1 ( x ) |   D 1 D U X , | ψ U 2 ( x ) |   D 2 D U X , x [ α , β ] .
In addition, by ( A 3 ) , we yield
| I l ( U 1 ( x l ) , U 2 ( x l ) ) |     | I l ( U 1 ( x l ) , U 2 ( x l ) ) I l ( 0 , 0 ) |   +   | I l ( 0 , 0 ) | M l 1 | U 1 ( x l ) |   +   M l 2 | U 2 ( x l ) | + | I l ( 0 , 0 ) |     ( M l 1 + M l 2 ) U X + | I l ( 0 , 0 ) | .
It is similar to obtain
| J l ( U 1 ( x l ) , U 2 ( x l ) ) |     ( N l 1 + N l 2 ) U X + | J l ( 0 , 0 ) | .
In view of (4), (27), and (28), we obtain
| c 0 U |   1 | a | [ | c | + l = 1 m | a l | k = 1 m ( 1 Γ ( θ 1 ) x k 1 x k log x k z θ 1 1 | ϕ U 1 ( z ) | d z z + | I k ( U 1 ( x k ) , U 2 ( x k ) ) | ) + 1 Γ ( θ 1 ) l = 1 m | a l | x l 1 ξ l log ξ l z θ 1 1 | ϕ U 1 ( z ) | d z z ] 1 | a | [ | c | + l = 1 m | a l | k = 1 m ( D 1 D Γ ( θ 1 ) U X x k 1 x k log x k z θ 1 1 d z z + | I k ( 0 , 0 ) | + ( M k 1 + M k 2 ) U X ) + D 1 D Γ ( θ 1 ) U X l = 1 m | a l | x l 1 ξ l log ξ l z θ 1 1 d z z ] = 1 | a | [ | c | + l = 1 m | a l | k = 1 m ( D 1 D Γ ( θ 1 + 1 ) U X log x k x k 1 θ 1 + | I k ( 0 , 0 ) | + ( M k 1 + M k 2 ) U X ) + D 1 D Γ ( θ 1 + 1 ) U X l = 1 m | a l | log ξ l x l 1 θ 1 ] = 1 | a | [ D 1 a + D Γ ( θ 1 + 1 ) k = 1 m log x k x k 1 θ 1 + D 1 D Γ ( θ 1 + 1 ) k = 1 m | a k | log ξ k x k 1 θ 1 + a + k = 1 m ( M k 1 + M k 2 ) ] · U X + 1 | a | | c | + a + k = 1 m | I k ( 0 , 0 ) | ,
and
| c l U |   | c 0 U | + 1 Γ ( θ 1 ) k = 1 m x k 1 x k log x k z θ 1 1 | ϕ U 1 ( z ) | d z z + k = 1 m | I k ( U 1 ( x k ) , U 2 ( x k ) ) | | c 0 U |   +   D 1 D Γ ( θ 1 + 1 ) U X k = 1 m log x k x k 1 θ 1 + k = 1 m [ ( M k 1 + M k 2 ) U X + I k ( 0 , 0 ) ] 1 | a | [ 2 D 1 a + D Γ ( θ 1 + 1 ) k = 1 m log x k x k 1 θ 1 + D 1 D Γ ( θ 1 + 1 ) k = 1 m | a k | log ξ k x k 1 θ 1 + 2 a + k = 1 m ( M k 1 + M k 2 ) ] · U X + 1 | a | | c | + 2 a + k = 1 m | I k ( 0 , 0 ) | .
Similar to (30) and (31), we have
| d 0 U |   1 | b | [ D 2 b + D Γ ( θ 2 + 1 ) k = 1 m log x k x k 1 θ 2 + D 2 D Γ ( θ 2 + 1 ) k = 1 m | b k | log ξ k x k 1 θ 2 + b + k = 1 m ( N k 1 + N k 2 ) ] · U X + 1 | b | | d | + b + k = 1 m | J k ( 0 , 0 ) | ,
and
| d l U |   1 | b | [ 2 D 2 b + D Γ ( θ 2 + 1 ) k = 1 m log x k x k 1 θ 2 + D 2 D Γ ( θ 2 + 1 ) k = 1 m | b k | log ξ k x k 1 θ 2 + 2 b + k = 1 m ( N k 1 + N k 2 ) ] · U X + 1 | b | | d | + 2 b + k = 1 m | J k ( 0 , 0 ) | .
In addition, for 0 l m , we obtain
g 1 ( x ) = 1 Γ ( θ 1 ) x l x log x z θ 1 1 | ϕ U 1 ( z ) | d z z D 1 D Γ ( θ 1 + 1 ) log β α θ 1 · U X ,
and
g 2 ( x ) = 1 Γ ( θ 2 ) x l x log x z θ 2 1 | ψ U 2 ( z ) | d z z D 2 D Γ ( θ 2 + 1 ) log β α θ 2 · U X .
From (4), (30)–(35) and ( A 4 ) , we yield
| U 1 ( x ) |     | c l U |   +   g 1 ( x ) ρ 1 U X + ϖ 1 ρ U X + ϖ , l = 0 , 1 , 2 , , m ,
| U 2 ( x ) |     | d l U |   +   g 2 ( x ) ρ 2 U X + ϖ 2 ρ U X + ϖ , l = 0 , 1 , 2 , , m ,
which implies that
U X ρ U X + ϖ ,
where ρ = max { ρ 1 , ρ 2 } , ϖ = max { ϖ 1 , ϖ 2 } , and
ϖ 1 = 1 | a | | c | + 2 a + k = 1 m | I k ( 0 , 0 ) | , ϖ 2 = 1 | b | | d | + 2 b + k = 1 m | J k ( 0 , 0 ) | .
Equation (36) derives that
U X ϖ 1 ρ .
Based on (37), we choose R > ϖ 1 ρ and Ω = { U X : U X < R } ; then, Ω X is a nonempty bounded open subset. Now, we shall verify that ( a 1 ) in Lemma 1 is true. In fact, if U * * X is a solution of L U = η · N ( U , η ) , then, similar to Lemma 3, it can be expressed in the form of an integral equation. A derivation similar to (24)–(37) yields that U * * X     ϖ 1 ρ , which means that U * * Ω . Noticing that Ω Ker L = { ( c a , d b ) T } and F i C ( [ α , β ] × R 4 , R + ) , we claim that Q N ( ( c a , d b ) T , 0 ) ( c a , d b ) T > 0 , which means that ( a 2 ) in Lemma 1 holds. Let J be an identity; then, we have
deg ( J Q N ( U , 0 ) , Ω Ker L , 0 ) = deg ( J Q N ( U , 0 ) , ( c / a , d / b ) T , 0 ) = 1 .
Thus, ( a 3 ) in Lemma 1 is also true. Therefore, it follows from Lemma 4 that BVP (3) has at least a solution U ˜ ( x ) = ( U ˜ 1 ( x ) , U ˜ 2 ( x ) ) T X .
Next, we will prove the uniqueness of the solution. Let U ^ ( x ) = ( U ^ 1 ( x ) , U ^ 2 ( x ) ) T X be another solution to (3), and denote W ( x ) = U ˜ ( x ) U ^ ( x ) ; then, it follows from Lemma 3 that
W 1 ( x ) = c 0 W + 1 Γ ( θ 1 ) α x log x z θ 1 1 Φ ( z ) d z z , x [ α , x 1 ] , W 2 ( x ) = d 0 W + 1 Γ ( θ 2 ) α x log x z θ 2 1 Ψ ( z ) d z z , x [ α , x 1 ] , W 1 ( x ) = c l W + 1 Γ ( θ 1 ) x l x log x z θ 1 1 Φ ( z ) d z z , x ( x l , x l + 1 ] , W 2 ( x ) = d l W + 1 Γ ( θ 2 ) x l x log x z θ 2 1 Ψ ( z ) d z z , x ( x l , x l + 1 ] ,
c 0 W = 1 a [ l = 1 m a l k = 1 l 1 Γ ( θ 1 ) x k 1 x k log x k z θ 1 1 Φ ( z ) d z z + I k ( x k ) 1 Γ ( θ 1 ) l = 1 m a l x l ξ l log ξ l z θ 1 1 Φ ( z ) d z z ] , d 0 W = 1 b [ l = 1 m b l k = 1 l 1 Γ ( θ 2 ) x k 1 x k log x k z θ 2 1 Ψ ( z ) d z z + J k ( x k ) 1 Γ ( θ 2 ) l = 1 m b l x l ξ l log ξ l z θ 2 1 Ψ ( z ) d z z ] , c l W = c 0 W + 1 Γ ( θ 1 ) k = 1 l x k 1 x k log x k z θ 1 1 Φ ( z ) d z z + k = 1 l I ( x k ) , d l W = d 0 W + 1 Γ ( θ 2 ) k = 1 l x k 1 x k log x k z θ 2 1 Ψ ( z ) d z z + k = 1 l J ( x k ) ,
where
Φ ( z ) = ϕ U ˜ 1 ( z ) ϕ U ^ 1 ( z ) , Ψ ( z ) = ϕ U ˜ 1 ( z ) ϕ U ^ 1 ( z ) ,
I k ( x k ) = I k ( U ˜ 1 ( x k ) , U ˜ 2 ( x k ) ) I k ( U ^ 1 ( x k ) , U ^ 2 ( x k ) ) ,
J k ( x k ) = J k ( U ˜ 1 ( x k ) , U ˜ 2 ( x k ) ) J k ( U ^ 1 ( x k ) , U ^ 2 ( x k ) ) .
It is similar to (24)–(27) that
| Φ ( x ) |   =   | ϕ U ˜ 1 ( x ) ϕ U ^ 1 ( x ) | D 1 D U ˜ U ^ X , x [ α , β ] ,
| Ψ ( x ) |   =   | ψ U ˜ 2 ( x ) ψ U ^ 2 ( x ) | D 2 D U ˜ U ^ X , x [ α , β ] ,
| I k ( x l ) |   =   | I l ( U ˜ 1 ( x l ) , U ˜ 2 ( x l ) ) I l ( U ^ 1 ( x l ) , U ^ 2 ( x l ) ) |     ( M l 1 + M l 2 ) U ˜ U ^ X ,
| J k ( x l ) |   =   | I l ( U ˜ 1 ( x l ) , U ˜ 2 ( x l ) ) I l ( U ^ 1 ( x l ) , U ^ 2 ( x l ) ) |     ( N l 1 + N l 2 ) U ˜ U ^ X .
Similar to (30)–(36), we derive from (38)–(42) that W X ρ W X . Noting that 0 < ρ < 1 , we know that W X = 0 , which implies that U ˜ ( x ) = U ^ ( x ) . The proof is completed. □
Now, we discuss the Hyers–Ulam (HU) stability of BVP (3). The concept of HU–stability of (3) is given as follows.
Definition 3.
BVP (3) is HU–stable if ζ > 0 ; there exists a unique solution U ˜ ( x ) = ( U ˜ 1 ( x ) , U ˜ 2 ( x ) ) X solving (3) s.t.
U ( x ) U ˜ ( x ) X B · ζ ,
where B > 0 is a constant and U = ( U 1 , U 2 ) X is any solution of the following inequality
CH D x l θ 1 U 1 ( x ) F 1 ( x , U 1 ( x ) , U 2 ( x ) , CH D x l θ 1 U 1 ( x ) , CH D x l θ 2 U 2 ( x ) ) ζ , CH D x l θ 2 U 2 ( x ) F 2 ( x , U 1 ( x ) , U 2 ( x ) , CH D x l θ 1 U 1 ( x ) , CH D x l θ 2 U 2 ( x ) ) ζ , Δ U 1 ( x l ) = I l ( U 1 ( x l ) , U 2 ( x l ) ) , Δ U 2 ( x l ) = J l ( U 1 ( x l ) , U 2 ( x l ) ) , l = 1 m + 1 a l U 1 ( ξ l ) = c , l = 1 m + 1 b l U 2 ( ξ l ) = d .
Theorem 2.
BVP (3) is HU–stable provided that the conditions ( A 1 ) ( A 4 ) are true.
Proof. 
U = ( U 1 , U 2 ) X solves the inequality (43) if it also solves the following system
CH D x l θ 1 U 1 ( x ) = F 1 ( x , U 1 ( x ) , U 2 ( x ) , CH D x l θ 1 U 1 ( x ) , CH D x l θ 2 U 2 ( x ) ) + ω 1 ( x ) , CH D x l θ 2 U 2 ( x ) = F 2 ( x , U 1 ( x ) , U 2 ( x ) , CH D x l θ 1 U 1 ( x ) , CH D x l θ 2 U 2 ( x ) ) + ω 2 ( x ) , Δ U 1 ( x l ) = I l ( U 1 ( x l ) , U 2 ( x l ) ) , Δ U 2 ( x l ) = J l ( U 1 ( x l ) , U 2 ( x l ) ) , l = 1 m + 1 a l U 1 ( ξ l ) = c , l = 1 m + 1 b l U 2 ( ξ l ) = d ,
where ω ( x ) = ( ω 1 ( x ) , ω 2 ( x ) ) X and | ω i ( x ) | ζ ( i = 1 , 2 ) , x [ α , β ] .
From Lemma 3 and (44), any solution of inequality (43) is represented by
U 1 ( x ) = c 0 ω + 1 Γ ( θ 1 ) α x log x z θ 1 1 [ ϕ U 1 ( z ) + ω 1 ( z ) ] d z z , x [ α , x 1 ] , U 2 ( x ) = d 0 ω + 1 Γ ( θ 2 ) α x log x z θ 2 1 [ ψ U 2 ( z ) + ω 2 ( z ) ] d z z , x [ α , x 1 ] , U 1 ( x ) = c l ω + 1 Γ ( θ 1 ) x l x log x z θ 1 1 [ ϕ U 1 ( z ) + ω 1 ( z ) ] d z z , x ( x l , x l + 1 ] , U 2 ( x ) = d l ω + 1 Γ ( θ 2 ) x l x log x z θ 2 1 [ ψ U 2 ( z ) + ω 2 ( z ) ] d z z , x ( x l , x l + 1 ] ,
where
c 0 ω = 1 a [ c l = 1 m a l k = 1 l ( 1 Γ ( θ 1 ) x k 1 x k log x k z θ 1 1 [ ϕ U 1 ( z ) + ω 1 ( z ) ] d z z + I k ( U 1 ( x k ) , U 2 ( x k ) ) ) 1 Γ ( θ 1 ) l = 1 m a l x l ξ l log ξ l z θ 1 1 [ ϕ U 1 ( z ) + ω 1 ( z ) ] d z z ] , d 0 ω = 1 b [ d l = 1 m b l k = 1 l ( 1 Γ ( θ 2 ) x k 1 x k log x k z θ 2 1 [ ψ U 2 ( z ) + ω 2 ( z ) ] d z z + J k ( U 1 ( x k ) , U 2 ( x k ) ) ) 1 Γ ( θ 2 ) l = 1 m b l x l ξ l log ξ l z θ 2 1 ψ U 2 ( z ) d z z ] , c l ω = c 0 ω + 1 Γ ( θ 1 ) k = 1 l x k 1 x k log x k z θ 1 1 [ ϕ U 1 ( z ) + ω 1 ( z ) ] d z z + k = 1 l I k ( U 1 ( x k ) , U 2 ( x k ) ) , d l ω = d 0 ω + 1 Γ ( θ 2 ) k = 1 l x k 1 x k log x k z θ 2 1 [ ψ U 2 ( z ) + ω 2 ( z ) ] d z z + k = 1 l J k ( U 1 ( x k ) , U 2 ( x k ) ) .
From Theorem 1, the unique solution U ˜ = ( U ˜ 1 , U ˜ 2 ) X is written as (4). Similar to (39)–(42), we have
| ϕ U 1 ( x ) + ω 1 ( x ) ϕ U ˜ 1 ( x ) |     | ϕ U 1 ( x ) ϕ U ˜ 1 ( x ) |   +   | ω 1 ( x ) |     D 1 D U U ˜ X + ζ ,
| ψ U 2 ( x ) + ω 2 ( x ) ϕ U ˜ 2 ( x ) |     | ψ U 2 ( x ) ψ U ˜ 2 ( x ) |   +   | ω 2 ( x ) | D 2 D U U ˜ X + ζ ,
| I l ( U 1 ( x l ) , U 2 ( x l ) ) I l ( U ˜ 1 ( x l ) , U ˜ 2 ( x l ) ) |     ( M l 1 + M l 2 ) U U ˜ X ,
| J l ( U 1 ( x l ) , U 2 ( x l ) ) J l ( U ˜ 1 ( x l ) , U ˜ 2 ( x l ) ) |     ( N l 1 + N l 2 ) U U ˜ X .
In the same way as (30)–(33), we yield from (45)–(49) that
| c 0 ω c 0 U ˜ |     1 | a | [ D 1 a + D Γ ( θ 1 + 1 ) k = 1 m log x k x k 1 θ 1 + D 1 D Γ ( θ 1 + 1 ) k = 1 m | a k | log ξ k x k 1 θ 1 + a + k = 1 m ( M k 1 + M k 2 ) ] ( U U ˜ X + ζ ) ,
| c l ω c l U ˜ |     1 | a | [ 2 D 1 a + D Γ ( θ 1 + 1 ) k = 1 m log x k x k 1 θ 1 + D 1 D Γ ( θ 1 + 1 ) k = 1 m | a k | log ξ k x k 1 θ 1 + 2 a + k = 1 m ( M k 1 + M k 2 ) ] ( U U ˜ X + ζ ) ,
| d 0 ω d 0 U ˜ |     1 | b | [ D 2 b + D Γ ( θ 2 + 1 ) k = 1 m log x k x k 1 θ 2 + D 2 D Γ ( θ 2 + 1 ) k = 1 m | b k | log ξ k x k 1 θ 2 + b + k = 1 m ( N k 1 + N k 2 ) ] ( U U ˜ X + ζ ) ,
| d l ω d l U ˜ |     1 | b | [ 2 D 2 b + D Γ ( θ 2 + 1 ) k = 1 m log x k x k 1 θ 2 + D 2 D Γ ( θ 2 + 1 ) k = 1 m | b k | log ξ k x k 1 θ 2 + 2 b + k = 1 m ( N k 1 + N k 2 ) ] ( U U ˜ X + ζ ) .
Additionally, we have
h 1 ( x ) = 1 Γ ( θ 1 ) x l x log x z θ 1 1 | ϕ U 1 ( x ) + ω 1 ( x ) ϕ U ˜ 1 ( x ) | d z z D 1 D Γ ( θ 1 + 1 ) log β α θ 1 ( U U ˜ X + ζ ) ,
h 2 ( x ) = 1 Γ ( θ 2 ) x l x log x z θ 2 1 | ψ U 2 ( x ) + ω 2 ( x ) ϕ U ˜ 2 ( x ) | d z z D 2 D Γ ( θ 2 + 1 ) log β α θ 2 ( U U ˜ X + ζ ) .
We derive from (4), (45) and (50)–(55) that
| U 1 ( x ) U ˜ 1 ( x ) |     | c l ω c l U ˜ |   +   h 1 ( x ) ρ 1 ( U U ˜ X + ζ ) , l = 0 , 1 , 2 , , m ,
| U 2 ( x ) U ˜ 2 ( x ) |     | d l ω d l U ˜ |   +   h 2 ( x ) ρ 2 ( U U ˜ X + ζ ) , l = 0 , 1 , 2 , , m .
Equations (56) and (57) lead to
U     U ˜ X     ρ ( U U ˜ X   +   ζ ) ,
which implies that
U     U ˜ X     ρ 1 ρ · ζ ,
where 0 < ρ = max { ρ 1 , ρ 2 } < 1 . From Definition 3, we know that BVP (3) is HU–stable. The proof is completed. □

4. Solvability and Stability of (1) and (2)

In this section, we apply our main methods and results to discuss the existence, uniqueness, and HU–stability of solutions for the Langevin system (1) and Sturm–Liouville system (2).
Theorem 3.
The Langevin Equation (1) has a unique solution in P C [ α , β ] which is HU–stable, provided that the following conditions ( A 1 ) ( A 4 ) are fulfilled.
(A′1)
Assume that 0 < α < β , 0 < θ 1 , θ 2 1 , λ > 0 , α = x 0 = ξ 1 < x 1 < ξ 2 < x 2 < ξ 3 < x 3 < < x m < ξ m = x m + 1 = β , a l , b l , c , d R , a = l = 1 m a l 0 , b = l = 1 m b l 0 , F C ( [ α , β ] × R 2 , R + ) , I l , J l C ( R , R ) , 1 l m .
(A′2)
u = ( u 1 , u 2 ) T , u ¯ = ( u ¯ 1 , u ¯ 2 ) T R 2 , L 1 , L 2 > 0 s.t.
| F ( x , u ) F ( x , u ¯ ) |     j = 1 2 L j | u j u ¯ j | , x [ α , β ] .
(A′3)
v , v ¯ R , M l , N l > 0 s.t.
| I l ( v ) I l ( v ¯ ) |     M l | v     v ¯ | , | J l ( v ) J l ( v ¯ ) |     N l | v v ¯ | , 1 l m .
(A′4)
0 < L 2 , ρ 1 , ρ 2 < 1 , where a + = l = 1 m | a l | , b + = l = 1 m | b l | ,
ρ 1 = 1 | a | [ 2 a + Γ ( θ 1 + 1 ) k = 1 m log x k x k 1 θ 1 + 1 Γ ( θ 1 + 1 ) k = 1 m | a k | log ξ k x k 1 θ 1 + 2 a + k = 1 m M k ] + 1 Γ ( θ 1 + 1 ) log β α θ 1 ,
ρ 2 = 1 | b | [ 2 ( λ + L 2 ) b + ( 1 L 2 ) Γ ( θ 2 + 1 ) k = 1 m log x k x k 1 θ 2 + λ + L 2 ( 1 L 2 ) Γ ( θ 2 + 1 ) k = 1 m | b k | log ξ k x k 1 θ 2 + 2 b + k = 1 m N k ] + λ + L 2 ( 1 L 2 ) Γ ( θ 2 + 1 ) log β α θ 2 ,
Proof. 
Let U ( x ) = U 1 ( x ) , CH D x l θ 1 U ( x ) = U 2 ( x ) ; then, the Langevin system (1) becomes
CH D x l θ 1 U 1 ( x ) = U 2 ( x ) , x [ α , β ] { x l } , CH D x l θ 2 U 2 ( x ) = λ U 1 ( x ) + F ( x , U 1 ( x ) , CH D x l θ 2 U 2 ( x ) ) , x [ α , β ] { x l } , Δ U 1 ( x l ) = I l ( U 1 ( x l ) ) , Δ U 2 ( x l ) = J l ( U 1 ( x l ) ) , 1 l m , l = 1 m + 1 a l U 1 ( ξ l ) = c , l = 1 m + 1 b l U 2 ( ξ l ) = d .
Therefore, the solvability of the Langevin system (1) and BVP (60) is equivalent. It suffices to discuss the existence of solutions for BVP (60). Indeed, let F 1 = U 2 ( x ) , F 2 = λ U 1 ( x ) + F ( x , U 1 ( x ) , CH D x l θ 2 U 2 ( x ) ) ; then, BVP (60) is transformed into the form of (3). Condition ( A 1 ) and Condition ( A 1 ) correspond exactly. From ( A 2 ) and ( A 3 ) , a simple calculation provides that L 11 = L 13 = L 14 = L 22 = L 23 = 0 , L 12 = 1 , L 21 = λ + L 1 , L 24 = L 2 , M l 2 = N l 2 = 0 , M l 1 = M l , N l 2 = N l . Substituting these values into Condition ( A 4 ) yields Condition ( A 4 ) . From Theorems 1 and 2, we declare that BVP (60) has a unique solution in X which is HU–stable. The proof is completed. □
Theorem 4.
The Sturm–Liouville system (2) has a unique solution in P C [ α , β ] which is HU–stable, provided that the following conditions ( A 1 ) ( A 4 ) are fulfilled.
(A″1)
Assume that 0 < α < β , 0 < θ 1 , θ 2 1 , α = x 0 = ξ 1 < x 1 < ξ 2 < x 2 < ξ 3 < x 3 < < x m < ξ m = x m + 1 = β , a l , b l , c , d R , a = l = 1 m a l 0 , b = l = 1 m b l 0 , p C ( [ α , β ] , ( 0 , + ) ) , F C ( [ α , β ] × R , R + ) , I l , J l C ( R , R ) , 1 l m .
(A″2)
u , u ¯ R , L > 0 s.t.
| F ( x , u ) F ( x , u ¯ ) |     L | u u ¯ | , x [ α , β ] .
(A″3)
v , v ¯ R , M l , N l > 0 s.t.
| I l ( v ) I l ( v ¯ ) |     M l | v     v ¯ | , | J l ( v ) J l ( v ¯ ) |     N l | v v ¯ | , 1 l m .
(A″4)
0 < ρ 1 , ρ 2 < 1 , where a + = l = 1 m | a l | , b + = l = 1 m | b l | , p = min { p ( x ) : α x β } ,
ρ 1 = 1 | a | [ 2 a + p Γ ( θ 1 + 1 ) k = 1 m log x k x k 1 θ 1 + 1 p Γ ( θ 1 + 1 ) k = 1 m | a k | log ξ k x k 1 θ 1 + 2 a + k = 1 m M k ] + 1 p Γ ( θ 1 + 1 ) log β α θ 1 ,
ρ 2 = 1 | b | [ 2 L b + Γ ( θ 2 + 1 ) k = 1 m log x k x k 1 θ 2 + L Γ ( θ 2 + 1 ) k = 1 m | b k | log ξ k x k 1 θ 2 + 2 b + k = 1 m N k ] + L Γ ( θ 2 + 1 ) log β α θ 2 ,
Proof. 
Let U ( x ) = U 1 ( x ) , p ( x ) CH D x l θ 1 U ( x ) = U 2 ( x ) ; then, the Sturm–Liouville system (1) changes into
CH D x l θ 1 U 1 ( x ) = 1 p ( x ) U 2 ( x ) , x [ α , β ] { x l } , CH D x l θ 2 U 2 ( x ) = F ( x , U 1 ( x ) ) , x [ α , β ] { x l } , Δ U 1 ( x l ) = I l ( U 1 ( x l ) ) , Δ U 2 ( x l ) = J l ( U 1 ( x l ) ) , 1 l m , l = 1 m + 1 a l U 1 ( ξ l ) = c , l = 1 m + 1 b l U 2 ( ξ l ) = d .
So it suffices to discuss the existence of solutions for BVP (61). In fact, let F 1 = 1 p ( x ) U 2 ( x ) , F 2 = F ( x , U 1 ( x ) ) ; then, BVP (61) is transformed into the form of (3). Condition ( A 1 ) and Condition ( A 1 ) correspond exactly. From ( A 2 ) and ( A 3 ) , a simple computation gives that L 11 = L 13 = L 14 = L 22 = L 23 = L 24 = 0 , L 12 = 1 p , L 21 = L , M l 2 = N l 2 = 0 , M l 1 = M l , N l 2 = N l . Substituting these values into Condition ( A 4 ) yields Condition ( A 4 ) . From Theorems 1 and 2, we declare that BVP (61) has a unique solution in X which is HU–stable. The proof is completed. □
To illustrate the availability and correctness of Theorem 1, we provide an example of the three-point boundary value problem with two impulse points as follows.
Example 1.
Consider the following nonlinear impulsive coupled implicit system
CH D x l 0.9 U 1 ( x ) = F 1 ( x , U 1 ( x ) , U 2 ( x ) , CH D x l 0.9 U 1 ( x ) , CH D x l 0.7 U 2 ( x ) ) , x [ 1 , e ] { x l } , CH D x l 0.7 U 2 ( x ) = F 2 ( x , U 1 ( x ) , U 2 ( x ) , CH D x l 0.9 U 1 ( x ) , CH D x l 0.7 U 2 ( x ) ) , x [ 1 , e ] { x l } , Δ U 1 ( x l ) = I l ( U 1 ( x l ) , U 2 ( x l ) ) , Δ U 2 ( x l ) = J l ( U 1 ( x l ) , U 2 ( x l ) ) , l = 1 , 2 , 1 2 U 1 ( ξ 1 ) + 1 3 U 1 ( ξ 2 ) + 1 6 U 1 ( ξ 3 ) = 2 , 1 4 U 2 ( ξ 1 ) + 1 5 U 2 ( ξ 2 ) + 11 20 U 2 ( ξ 3 ) = 3 ,
where α = ξ 1 = 1 , β = ξ 3 = e , θ 1 = 0.9 , θ 2 = 0.7 , m = 2 , x 1 = e + 3 4 , x 2 = 3 e + 1 4 , ξ 2 = e + 1 2 , a 1 = 1 2 , a 2 = 1 3 , a 3 = 1 6 , b 1 = 1 4 , b 2 = 1 5 , b 3 = 11 20 , c = 2 , d = 3 , F 1 ( x , u 1 , u 2 , u 3 , u 4 ) = x 2 + 2 + e u 1 2 + cos ( u 2 ) + sin ( u 3 + u 4 ) 100 , F 2 ( x , u 1 , u 2 , u 3 , u 4 ) = u 1 + u 2 100 arctan ( u 1 + u 2 ) + 1 200 log [ 1 + ( u 1 + u 2 ) 2 ] + x 100 [ 1 + cos ( u 3 , u 4 ) ] , I 1 ( v 1 , v 2 ) = I 2 ( v 1 , v 2 ) = sin ( v 1 + v 2 ) 100 . J 1 ( v 1 , v 2 ) = J 2 ( v 1 , v 2 ) = ( v 1 + v 2 ) arccos ( v 1 + v 2 ) + 1 ( v 1 + v 2 ) 2 100 . Obviously, a = a 1 + a 2 + a 3 = 1 , b = b 1 + b 2 + b 3 = 1 , F 1 , F 2 C ( [ 1 , e ] × R 4 , R + ) , I 1 , I 2 , J 1 , J 2 C ( R 2 , R ) . Therefore, the condition ( A 1 ) holds. We perform a simple calculation to yield that F 1 u 1 = u 1 50 e u 1 2 , F 1 u 2 = sin ( u 2 ) 100 , F 1 u 3 = F 1 u 4 = cos ( u 3 + u 4 ) 100 , F 2 u 1 = F 2 u 2 = arctan ( u 1 + u 2 ) 100 , F 2 u 3 = F 2 u 4 = x sin ( u 3 + u 4 ) 100 , I i v 1 = I i v 2 = cos ( v 1 + v 2 ) 100 ( i = 1 , 2 ) , J i v 1 = J i v 2 = arccos ( v 1 + v 2 ) 100 ( i = 1 , 2 ) . Hence, we derive that
| F 1 ( x , u 1 , u 2 , u 3 , u 4 ) F 1 ( x , u ¯ 1 , u ¯ 2 , u ¯ 3 , u ¯ 4 ) |     2 e 0.5 100 | u 1 u ¯ 1 |   +   1 100 j = 2 4 | u j u ¯ j | ,
| F 2 ( x , u 1 , u 2 , u 3 , u 4 ) F 2 ( x , u ¯ 1 , u ¯ 2 , u ¯ 3 , u ¯ 4 ) |     π 200 i = 1 2 | u i u ¯ i |   +   e 100 j = 2 4 | u j u ¯ j | ,
| I i ( v 1 , v 2 ) |     I i ( v ¯ 1 , v ¯ 2 ) 1 100 j = 1 2 | v j v ¯ j | , | J i ( v 1 , v 2 ) |     J i ( v ¯ 1 , v ¯ 2 ) π 100 j = 1 2 | v j v ¯ j | ,
L 11 = 2 e 0.5 100 , L 12 = L 13 = L 14 = 1 100 , L 21 = L 22 = π 200 , L 23 = L 24 = e 100 ,
M 11 = M 12 = M 21 = M 22 = 1 100 , N 11 = N 12 = N 21 = N 22 = π 100 ,
a + = l = 1 3 | a l |   =   1 , b + = l = 1 3 | b l | = 1 , D = 1 L 13 L 14 L 23 1 L 24 0.9628 < 1 ,
D 1 = L 11 + L 12 L 14 L 21 + L 22 1 L 24 0.0184 , D 2 = 1 L 13 L 11 + L 12 L 23 L 21 + L 22 0.0316 .
ρ 1 = 1 | a | [ 2 D 1 a + D Γ ( θ 1 + 1 ) k = 1 m log x k x k 1 θ 1 + D 1 D Γ ( θ 1 + 1 ) k = 1 m | a k | log ξ k x k 1 θ 1 + 2 a + k = 1 m ( M k 1 + M k 2 ) ] + D 1 D Γ ( θ 1 + 1 ) log β α θ 1 0.1466 < 1 ,
ρ 2 = 1 | b | [ 2 D 2 b + D Γ ( θ 2 + 1 ) k = 1 m log x k x k 1 θ 2 + D 2 D Γ ( θ 2 + 1 ) k = 1 m | b k | log ξ k x k 1 θ 2 + 2 b + k = 1 m ( N k 1 + N k 2 ) ] + D 2 D Γ ( θ 2 + 1 ) log β α θ 2 0.3950 < 1 .
Thus, the conditions ( A 2 ) ( A 4 ) are also satisfied. From Theorems 1 and 2, we conclude that system (60) admits a unique solution, which is HU–stable.

5. Conclusions

This section first provides further analysis and then discussion of our main results. In Theorems 1–4, the most important condition is that 0 < ρ 1 , ρ 2 , ρ 1 , ρ 2 , ρ 1 , ρ 2 < 1 . This condition is determined by the response functions F 1 , F 2 , pulse functions I l , J l , impulsive points x l , boundary value points α , β , ξ l , and coefficients a l , b l together, l = 1 , 2 , , m . The more complex calculation to verify this condition is that k = 1 m log x k x k 1 θ and k = 1 m log ξ k x k 1 θ , where θ { θ 1 , θ 2 } . Since x k 1 < ξ k < x k ( k = 1 , 2 , , m ) , the positions of adjacent impulsive points x k 1 and x k will result in two different cases for verifying this condition. Case 1: when x k x k 1 < e , then 0 < log x k x k 1 θ < 1 and 0 < log ξ k x k 1 θ < 1 . Thus, the condition 0 < ρ 1 , ρ 2 , ρ 1 , ρ 2 , ρ 1 , ρ 2 < 1 is relatively easy to satisfy. Example 1 belongs to this case. Case 2: when x k x k 1 > e , then log x k x k 1 θ > 1 . So, the condition 0 < ρ 1 , ρ 2 , ρ 1 , ρ 2 , ρ 1 , ρ 2 < 1 is more difficult to satisfy. This requires controlling the values of L i j , M i j , N i j , a l and b l , i , j = 1 , 2 ; l = 1 , 2 , , m . In addition, since this paper considers the CH–fractional differential equations with certain singularities, the ODE toolboxes in MATHLAB cannot be applied in numerical simulations. This requires the design of new numerical algorithms, which is also one of our future research directions.
Next, we make a brief summary. Hadamard fractional calculus in the Caputo sense is an important type of fractional calculus, which is a generalization of RL–fractional calculus in the Caputo sense. The CH–fractional differential equation is used to solve many practical problems and has become a popular object of concern for many academic researchers. There have been some good results in the study of CH–fractional differential explicit systems. However, studies on the solvability and stability of CH–fractional coupled implicit systems under impulsive influence are relatively rare because it is difficult to estimate the existence region of the solution. Additionally, the theory of coincidence degree is an important route to solve the existence of solutions to nonlinear differential equations. In this paper, we creatively establish a framework for the coincidence degree theory for system (3) with impulsive effects and prove the existence of a solution. Simultaneously, our main results are applied to solve the solvability of the Langevin system (1) and Sturm–Liouville system (2). Our research objects and findings enrich the theory of CH–fractional differential equations. Our approach also provides a paradigm for uniformly solving such problems.

Author Contributions

K.Z., conceptualization, methodology, investigation, writing—original draft, and writing—review and editing; J.L., conceptualization, methodology, investigation, and formal analysis; X.L., conceptualization, investigation, validation, and project administration. All authors have read and agreed to the published version of the manuscript.

Funding

The APC was funded by research start-up funds for high-level talents of Taizhou University.

Informed Consent Statement

Not applicable.

Data Availability Statement

No data were used to support this study.

Acknowledgments

The authors would like to express their heartfelt gratitude to the editors and reviewers for their constructive comments.

Conflicts of Interest

The authors declare no conflicts of interest.

References

  1. Ahmadova, A.; Mahmudov, N.I. Langevin differential equations with general fractional orders and their applications to electric circuit theory. J. Computat. Appl. Math. 2021, 388, 113299. [Google Scholar] [CrossRef]
  2. Salem, A.; Alshehri, H.M.; Almaghamsi, L. Measure of noncompactness for an infinite system of fractional Langevin equation in a sequence space. Adv. Differ. Equ. 2021, 2021, 132. [Google Scholar] [CrossRef]
  3. Zhao, K.H. Stability of a nonlinear Langevin system of ML-type fractional derivative affected by time-varying delays and differential feedback control. Fractal Fract. 2022, 6, 725. [Google Scholar] [CrossRef]
  4. Zhao, K.H. Stability of a nonlinear fractional Langevin System with nonsingular exponential kernel and delay control. Discret. Dyn. Nat. Soc. 2022, 2022, 9169185. [Google Scholar] [CrossRef]
  5. Zhao, K.H. Stability of a nonlinear ML-nonsingular kernel fractional Langevin system with distributed lags and integral control. Axioms 2022, 11, 350. [Google Scholar] [CrossRef]
  6. Kumar, V.; Stamov, G.; Stamova, I. Controllability results for a class of piecewise nonlinear impulsive fractional dynamic systems. Appl. Math. Comput. 2022, 439, 127625. [Google Scholar] [CrossRef]
  7. Radhakrishnan, B.; Sathya, T. Controllability of Hilfer fractional Langevin dynamical system with impulse in an abstract weighted space. J. Optimiz. Theory Appl. 2022, 195, 265–281. [Google Scholar] [CrossRef]
  8. Venkatesan, G.; Pitchaikkannu, S.K. Trajectory controllability of nonlinear fractional Langevin systems. Int. J. Nonlin. Sci. Num. 2023, 24, 1079–1093. [Google Scholar] [CrossRef]
  9. Zhou, Z.B.; Yu, W.X. Studying stochastic resonance phenomenon in the fractional-order Lorenz-like chaotic System. Int. J. Bifurcat. Chaos 2022, 32, 2250154. [Google Scholar] [CrossRef]
  10. Yan, Z.; Guirao, J.L.G.; Saeed, T.; Chen, H.; Liu, X. Different stochastic resonances induced by multiplicative polynomial Trichotomous noise in a fractional order oscillator with time delay and fractional Gaussian noise. Fractal Fract. 2022, 6, 191. [Google Scholar] [CrossRef]
  11. Joo, S.; Jeon, J.H. Viscoelastic active diffusion governed by nonequilibrium fractional Langevin equations: Underdamped dynamics and ergodicity breaking. Chaos Soliton. Fract. 2023, 177, 114288. [Google Scholar] [CrossRef]
  12. Afarideh, A.; Saei, F.D.; Lakestani, M.; Saray, B.N. Pseudospectral method for solving fractional Sturm-Liouville problem using Chebyshev cardinal functions. Phys. Scr. 2021, 96, 125267. [Google Scholar] [CrossRef]
  13. Sadabad, M.; Akbarfam, A.J. An efficient numerical method for estimating eigenvalues and eigenfunctions of fractional Sturm-Liouville problems. Math. Comput. Simulat. 2021, 185, 547–569. [Google Scholar] [CrossRef]
  14. Allahverdiev, B.; Tuna, H.; Yalcinkaya, Y. A completeness theorem for dissipative conformable fractional Sturm-Liouville operator in singular case. Filomat 2023, 36, 2461–2474. [Google Scholar] [CrossRef]
  15. Goel, E.; Pandey, R.K.; Yadav, S.; Agrawal, O.P. A numerical approximation for generalized fractional Sturm-Liouville problem with application. Math. Comput. Simul. 2023, 207, 417–436. [Google Scholar] [CrossRef]
  16. Kumar, N.; Mehra, M. Muntz-Legendre wavelet method for solving Sturm-Liouville fractional optimal control problem with error estimates. Math. Methods Appl. Sci. 2023, 46, 12505–12528. [Google Scholar] [CrossRef]
  17. Hadamard, J. Essai sur l’étude des fonctions données par leur développment de Taylor. J. Math. Pures Appl. 1892, 8, 101–186. [Google Scholar]
  18. Kilbas, A.A. Hadamard-type fractional calculus. J. Korean Math. Soc. 2001, 38, 1191–1204. [Google Scholar]
  19. Butzer, P.L.; Kilbas, A.A.; Trujillo, J.J. Compositions of Hadamard-type fractional integration operators and the semigroup property. J. Math. Anal. Appl. 2002, 269, 387–400. [Google Scholar] [CrossRef]
  20. Jarad, F.; Abdeljawad, T.; Baleanu, D. Caputo-type modification of the Hadamard fractional derivatives. Adv. Differ. Equ. 2012, 2012, 142. [Google Scholar] [CrossRef]
  21. Ahmad, B.; Ntouyas, S. A fully Hadamard type integral boundary value problem of a coupled system of fractional differential equations. Fract. Calc. Appl. Anal. 2014, 17, 348–360. [Google Scholar] [CrossRef]
  22. Wang, J.R.; Zhang, Y.R. On the concept and existence of solutions for fractional impulsive systems with Hadamard derivatives. Appl. Math. Lett. 2015, 39, 85–90. [Google Scholar] [CrossRef]
  23. Ahmad, B.; Ntouyas, S.K. On Hadamard fractional integro-differential boundary value problems. J. Appl. Math. Comput. 2015, 47, 119–131. [Google Scholar] [CrossRef]
  24. Aljoudi, S.; Ahmad, B.; Nieto, J.J.; Alsaedi, A. A coupled system of Hadamard type sequential fractional differential equations with coupled strip conditions. Chaos Soliton Fract. 2016, 91, 39–46. [Google Scholar] [CrossRef]
  25. Pei, K.; Wang, G.T.; Sun, Y.Y. Successive iterations and positive extremal solutions for a Hadamard type fractional integro-differential equations on infinite domain. Appl. Math. Comput. 2017, 312, 158–168. [Google Scholar] [CrossRef]
  26. Abbas, S.; Benchohra, M.; Lagreg, J.E.; Alsaedi, A.; Zhou, Y. Existence and Ulam stability for fractional differential equations of Hilfer-Hadamard type. Adv. Differ. Equ. 2017, 2017, 180. [Google Scholar] [CrossRef]
  27. Wang, G.T.; Pei, K.; Agarwal, R.P.; Zhang, L.; Ahmad, B. Nonlocal Hadamard fractional boundary value problem with Hadamard integral and discrete boundary conditions on a half-line. J. Comput. Appl. Math. 2018, 343, 230–239. [Google Scholar] [CrossRef]
  28. Zhao, K.H. Generalized UH-stability of a nonlinear fractional coupling (p1,p2)-Laplacian system concerned with nonsingular Atangana-Baleanu fractional calculus. J. Inequal. Appl. 2023, 2023, 96. [Google Scholar] [CrossRef]
  29. Aljoudi, S.; Ahmad, B.; Alsaedi, A. Existence and uniqueness results for a coupled system of Caputo-Hadamard fractional differential equations with nonlocal Hadamard type integral boundary conditions. Fractal Fract. 2020, 4, 13. [Google Scholar] [CrossRef]
  30. Zhao, K.H. Solvability and GUH-stability of a nonlinear CF-fractional coupled Laplacian equations. AIMS Math. 2023, 8, 13351–13367. [Google Scholar] [CrossRef]
  31. Zhao, K.H. Existence, Stability and simulation of a class of nonlinear fractional Langevin equations involving nonsingular Mittag-Leffler kernel. Fractal Fract. 2022, 6, 469. [Google Scholar] [CrossRef]
  32. Li, C.P.; Li, Z.Q. Stability and logarithmic decay of the solution to Hadamard-type fractional differential equation. J. Nolinear Sci. 2021, 31, 31. [Google Scholar] [CrossRef]
  33. Rao, S.; Ahmadini, A. Multiple positive solutions for system of mixed Hadamard fractional boundary value problems with (p1,p2)-Laplacian operator. AIMS Math. 2023, 8, 14767–14791. [Google Scholar] [CrossRef]
  34. Zhao, K.H. Existence and UH-stability of integral boundary problem for a class of nonlinear higher-order Hadamard fractional Langevin equation via Mittag-Leffler functions. Filomat 2023, 37, 1053–1063. [Google Scholar] [CrossRef]
  35. Zhao, K.H. Solvability, approximation and stability of periodic boundary value problem for a nonlinear Hadamard fractional differential equation with p-Laplacian. Axioms 2023, 12, 733. [Google Scholar] [CrossRef]
  36. Istafa, G.; Rehman, G.M.U. Numerical solutions of Hadamard fractional differential equations by generalized Legendre functions. Math. Methods Appl. Sci. 2022, 46, 6821–6842. [Google Scholar] [CrossRef]
  37. Zaky, M.A.; Hendy, A.S.; Suragan, D. Logarithmic Jacobi collocation method for Caputo-Hadamard fractional differential equations. Appl. Numer. Math. 2022, 181, 326–346. [Google Scholar] [CrossRef]
  38. Ortigueira, M.D.; Bohannan, G.W. Fractional scale calculus: Hadamard vs. Liouville. Fractal Fract. 2023, 7, 296. [Google Scholar] [CrossRef]
  39. Dhawan, K.; Vats, R.K.; Vijayakumar, V. Analysis of neutral fractional differential equation via the method of upper and lower solution. Qual. Theor. Dyn. Syst. 2023, 22, 93. [Google Scholar] [CrossRef]
  40. Ahmad, B.; Aljoudi, S. Investigation of a coupled system of Hilfer-Hadamard fractional differential equations with nonlocal coupled Hadamard fractional integral boundary conditions. Fractal Fract. 2023, 7, 178. [Google Scholar] [CrossRef]
  41. Ben Makhlouf, A.; Mchiri, L.; Mtiri, F. Existence, uniqueness, and averaging principle for Hadamard Ito-Doob stochastic delay fractional integral equations. Math. Methods Appl. Sci. 2023, 46, 14814–14827. [Google Scholar] [CrossRef]
  42. Vivek, D.; Kanagarajan, K.; Elsayed, E.M. Attractivity and Ulam-Hyers stability results for fractional delay differential equations. Filomat 2022, 36, 5707–5724. [Google Scholar] [CrossRef]
  43. Rhaima, M. Ulam-Hyers stability for an impulsive Caputo-Hadamard fractional neutral stochastic differential equations with infinite delay. Math. Comput. Simul. 2023, 210, 281–295. [Google Scholar] [CrossRef]
  44. Arfaoui, H.; Ben Makhlouf, A. Some results for a class of delayed fractional partial differential equations with Caputo-Hadamard derivative. Math. Methods Appl. Sci. 2023, 46, 9954–9965. [Google Scholar] [CrossRef]
  45. Hammou, A.; Hamani, S.; Henderson, J. Impulsive fractional differential equations with state-dependent delay involving the Caputo-Hadamard derivative. Filomat 2023, 37, 1581–1590. [Google Scholar] [CrossRef]
  46. Rafeeq, A.S.; Thabet, S.M.; Mohammed, M.O.; Kedim, I.; Vivas-Cortez, M. On Caputo-Hadamard fractional pantograph problem of two different orders with Dirichlet boundary conditions. Alex. Eng. J. 2024, 86, 386–398. [Google Scholar] [CrossRef]
  47. Bohner, M.; Domoshnitsky, A.; Litsyn, E.; Padhi, S.; Narayan Srivastava, S. Vallée-Poussin theorem for Hadamard fractional functional differential equations. Appl. Math. Sci. Eng. 2023, 31, 2259057. [Google Scholar] [CrossRef]
  48. Benchohra, M.; Bouriah, S.; Graef, J.R. Boundary value problems for nonlinear implicit Caputo-Hadamard-type fractional differential equations with impulses. Mediterr. J. Math. 2017, 14, 206. [Google Scholar] [CrossRef]
  49. Benchohra, M.; Bouriah, S.; Nieto, J.J. Existence of periodic solutions for nonlinear implicit Hadamard’s fractional differential equations. Rev. Real Acad. Cienc. Exactas Físicas Nat. Ser. Mat. 2018, 112, 25–35. [Google Scholar] [CrossRef]
  50. Riaz, U.; Zada, A.; Ali, Z.; Cui, Y.; Xu, J. Analysis of coupled systems of implicit impulsive fractional differential equations involving Hadamard derivatives. Adv. Differ. Equ. 2019, 2019, 226. [Google Scholar] [CrossRef]
  51. Van, H.N.; Ho, V. A survey on the initial value problems of fuzzy implicit fractional differential equations. Fuzzy Set. Syst. 2020, 400, 90–133. [Google Scholar]
  52. Zhao, K.H. Study on the stability and its simulation algorithm of a nonlinear impulsive ABC-fractional coupled system with a Laplacian operator via F-contractive mapping. Adv. Cont. Discr. Mod. 2024, 2024, 5. [Google Scholar] [CrossRef]
  53. Salim, A.; Benchohra, M.; Graef, J.R. Boundary value problem for fractional order generalized Hilfer-type fractional derivative with non-instantaneous impulses. Fractal Fract. 2021, 5, 1. [Google Scholar] [CrossRef]
  54. Abbas, M.; Feckan, M. Investigation of an implicit Hadamard fractional differential equation with Riemann-Stieltjes integral boundary condition. Math. Slovaca 2022, 72, 925–934. [Google Scholar] [CrossRef]
  55. Hammad, H.A.; Aydi, H.; Isik, H.; De la Sen, M. Existence and stability results for a coupled system of impulsive fractional differential equations with Hadamard fractional derivatives. AIMS Math. 2023, 8, 6913–6941. [Google Scholar] [CrossRef]
  56. Guo, L.M.; Riaz, U.; Zada, A.; Alam, M. On implicit coupled Hadamard fractional differential equations with generalized Hadamard fractional integro-differential boundary conditions. Fractal Fract. 2023, 7, 13. [Google Scholar] [CrossRef]
  57. Wu, Z.Q.; Muhammadhaji, A. Dynamics in a competitive Nicholson’s blowflies model with continuous time delays. Symmetry 2023, 15, 1495. [Google Scholar] [CrossRef]
  58. Zhao, K.H. Asymptotic stability of a periodic GA-predation system with infinite distributed lags on time scales. Int. J. Control 2023, in press. [Google Scholar] [CrossRef]
  59. Guo, K.; Ma, W.B. Existence of positive periodic solutions for a periodic predator-prey model with fear effect and general functional responses. Adv. Contin. Discret. Models 2023, 2023, 22. [Google Scholar] [CrossRef]
  60. Zhao, K.H. Existence and stability of a nonlinear distributed delayed periodic AG-ecosystem with competition on time scales. Axioms 2023, 12, 315. [Google Scholar] [CrossRef]
  61. Feltrin, G.; Sampedro, J.C.; Zanolin, F. Periodic solutions to superlinear indefinite planar systems: A topological degree approach. J. Differ. Equ. 2023, 363, 546–581. [Google Scholar] [CrossRef]
  62. Zhao, K.H. Global asymptotic stability for a classical controlled nonlinear periodic commensalism AG-ecosystem with distributed lags on time scales. Filomat 2023, 37, 9899–9911. [Google Scholar]
  63. Lyu, Y.; Tan, X.L.; Yang, X.; Li, Y. Controllability of nonlinear discrete systems with degeneracy. Acta Math. Appl. Sin. Engl. Ser. 2023, 39, 293–305. [Google Scholar] [CrossRef]
  64. Zhao, K.H. Local exponential stability of four almost-periodic positive solutions for a classic Ayala-Gilpin competitive ecosystem provided with varying-lags and control terms. Int. J. Control 2023, 96, 1922–1934. [Google Scholar] [CrossRef]
  65. Taie, R.O.A.; Bakhit, D.A.M. Existence and uniqueness criterion of a periodic solution for a third-order neutral differential equation with multiple delay. Bound. Value Probl. 2023, 2023, 25. [Google Scholar] [CrossRef]
  66. Zhao, K.H. Local exponential stability of several almost periodic positive solutions for a classical controlled GA-predation ecosystem possessed distributed delays. Appl. Math. Comput. 2023, 437, 127540. [Google Scholar] [CrossRef]
  67. Wang, J.R.; Wang, B.; Miao, Y.W.; Yu, X.C. Existence and multiplicity of positive periodic solutions to a class of Lienard equations with repulsive singularities. J. Fix. Point Theory A 2022, 24, 64. [Google Scholar] [CrossRef]
  68. Zhao, K.H. Coincidence theory of a nonlinear periodic Sturm-Liouville system and its applications. Axioms 2022, 11, 726. [Google Scholar] [CrossRef]
  69. Salim, A.; Benchohra, M.; Lazreg, J.E. On implicit k-generalized ψ-Hilfer fractional differential coupled systems with periodic conditions. Qual. Theor. Dyn. Syst. 2023, 22, 75. [Google Scholar] [CrossRef]
  70. Iatime, K.; Guedda, L.; Djebali, S. System of fractional boundary value problems at resonance. Fract. Calc. Appl. Anal. 2023, 26, 1359–1383. [Google Scholar] [CrossRef]
  71. Salem, A.; Almaghamsi, L. Solvability of sequential fractional differential equation at resonance. Mathematics 2023, 11, 1044. [Google Scholar] [CrossRef]
  72. Domoshnitsky, A.; Srivastava, S.N.; Padhi, S. Existence of solutions for a higher order Riemann-Liouville fractional differential equation by Mawhin’s coincidence degree theory. Math. Methods Appl. Sci. 2023, 46, 12018–12034. [Google Scholar] [CrossRef]
  73. Imaga, O.S.; Iyase, S.A.; Ogunniyi, P.O. Existence results for an m-point mixed fractional-order problem at resonance on the half-Line. Axioms 2022, 11, 630. [Google Scholar] [CrossRef]
  74. Gaines, R.E.; Mawhin, J.L. Coincidence Degree and Nonlinear Differetial Equitions; Springer: Berlin/Heidelberg, Germany, 1977. [Google Scholar]
  75. Kilbas, A.A.; Srivastava, H.M.; Trujillo, J.J. Theory and Applications of Fractional Differential Equations; Elsevier: Amsterdam, The Netherlands, 2006. [Google Scholar]
Disclaimer/Publisher’s Note: The statements, opinions and data contained in all publications are solely those of the individual author(s) and contributor(s) and not of MDPI and/or the editor(s). MDPI and/or the editor(s) disclaim responsibility for any injury to people or property resulting from any ideas, methods, instructions or products referred to in the content.

Share and Cite

MDPI and ACS Style

Zhao, K.; Liu, J.; Lv, X. A Unified Approach to Solvability and Stability of Multipoint BVPs for Langevin and Sturm–Liouville Equations with CH–Fractional Derivatives and Impulses via Coincidence Theory. Fractal Fract. 2024, 8, 111. https://doi.org/10.3390/fractalfract8020111

AMA Style

Zhao K, Liu J, Lv X. A Unified Approach to Solvability and Stability of Multipoint BVPs for Langevin and Sturm–Liouville Equations with CH–Fractional Derivatives and Impulses via Coincidence Theory. Fractal and Fractional. 2024; 8(2):111. https://doi.org/10.3390/fractalfract8020111

Chicago/Turabian Style

Zhao, Kaihong, Juqing Liu, and Xiaojun Lv. 2024. "A Unified Approach to Solvability and Stability of Multipoint BVPs for Langevin and Sturm–Liouville Equations with CH–Fractional Derivatives and Impulses via Coincidence Theory" Fractal and Fractional 8, no. 2: 111. https://doi.org/10.3390/fractalfract8020111

Article Metrics

Back to TopTop