# Solvability for a coupled system of fractional differential equations with impulses at resonance

Xiaozhi Zhang, Chuanxi Zhu* and Zhaoqi Wu

Author Affiliations

Department of Mathematics, Nanchang University, Nanchang, 330031, P.R. China

For all author emails, please log on.

Boundary Value Problems 2013, 2013:80  doi:10.1186/1687-2770-2013-80

 Received: 9 November 2012 Accepted: 18 March 2013 Published: 8 April 2013

This is an Open Access article distributed under the terms of the Creative Commons Attribution License (http://creativecommons.org/licenses/by/2.0), which permits unrestricted use, distribution, and reproduction in any medium, provided the original work is properly cited.

### Abstract

In this paper, some Banach spaces are introduced. Based on these spaces and the coincidence degree theory, a 2m-point boundary value problem for a coupled system of impulsive fractional differential equations at resonance is considered, and the new criterion on existence is obtained. Finally, an example is also given to illustrate the availability of our main results.

MSC: 34A08, 34B10, 34B37.

##### Keywords:
coupled system; impulsive fractional differential equations; at resonance; coincidence degree

### 1 Introduction

Recently, Wang et al.[1] presented a counterexample to show an error formula of solutions to the traditional boundary value problem for impulsive differential equations with fractional derivative in [2-5]. Meanwhile, they introduced the correct formula of solutions for an impulsive Cauchy problem with the Caputo fractional derivative. Shortly afterwards, many works on the better formula of solutions to the Cauchy problem for impulsive fractional differential equations have been reported by Li et al.[6], Wang et al.[7], Fečkan [8], etc.

Fractional differential equations have been paid much attention to in recent years due to their wide applications such as nonlinear oscillations of earthquakes, Nutting’s law, charge transport in amorphous semiconductors, fluid dynamic traffic model, non-Markovian diffusion process with memory etc.[9-11]. For more details, see the monographs of Hilfer [12], Miller and Ross [13], Podlubny [14], Lakshmikantham et al.[15], Samko et al.[16], and the papers of [2,17-19] and the references therein.

In recent years, many researchers paid much attention to the coupled system of fractional differential equations due to its applications in different fields [20-25]. Zhang et al.[25] investigated a three-point boundary value problem at resonance for a coupled system of nonlinear fractional differential equations given by

{ D 0 + α u ( t ) = f ( t , v ( t ) , D 0 + β 1 v ( t ) ) , 0 < t < 1 ; D 0 + β v ( t ) = g ( t , u ( t ) , D 0 + α 1 u ( t ) ) , 0 < t < 1 ; u ( 0 ) = v ( 0 ) = 0 , u ( 1 ) = σ 1 u ( η 1 ) , v ( 1 ) = σ 2 v ( η 2 ) ,

where 1 < α , β 2 , 0 < η 1 , η 2 < 1 , σ 1 , σ 2 > 0 , σ 1 η 1 α 1 = σ 2 η 2 β 1 = 1 , D 0 + α is the standard Riemann-Liouville fractional derivative and f , g : [ 0 , 1 ] × R 2 R are continuous. And Wang et al.[23] considered a 2m-point boundary value problem (BVP) at resonance for a coupled system as follows:

{ D 0 + α u ( t ) = f ( t , v ( t ) , D 0 + β 1 v ( t ) , D 0 + β 2 v ( t ) ) , 0 < t < 1 ; D 0 + β v ( t ) = g ( t , u ( t ) , D 0 + α 1 u ( t ) , D 0 + α 2 u ( t ) ) , 0 < t < 1 ; I 0 + 3 α u ( 0 ) = 0 , D 0 + α 2 u ( 1 ) = i = 1 m a i D 0 + α 2 u ( ξ i ) , u ( 1 ) = i = 1 m b i u ( η i ) ; I 0 + 3 β v ( 0 ) = 0 , D 0 + β 2 v ( 1 ) = i = 1 m c i D 0 + β 2 v ( γ i ) , v ( 1 ) = i = 1 m d i v ( δ i ) ,

where 1 < α , β 2 . With the help of the coincidence degree theory, many existence results have been given in the above literatures. It is worth mentioning that the orders of derivative in the nonlinear function on the right-hand of equal signs are all fixed in the above works, but the opposite case is more difficult and complicated, then this work attempts to deal exactly with this case. What is more, this case of arbitrary order derivative included in the nonlinear functions is very important in many aspects [20,22].

There are significant developments in the theory of impulses especially in the area of impulsive differential equations with fixed moments, which provided a natural description of observed evolution processes, regarding as important tools for better understanding several real word phenomena in applied sciences [1,7,26-29]. In addition, motivated by the better formula of solutions cited by the work of Zhou et al.[1,7,8], the aim of this work is to discuss a boundary value problem for a coupled system of impulsive fractional differential equation. Exactly, this paper deals with the 2m-point boundary value problem of the following coupled system of impulsive fractional differential equations at resonance:

{ D 0 + α u ( t ) = f ( t , v ( t ) , D 0 + p v ( t ) ) , D 0 + β v ( t ) = g ( t , u ( t ) , D 0 + q u ( t ) ) , 0 < t < 1 ; Δ u ( t i ) = A i ( v ( t i ) , D 0 + p v ( t i ) ) , Δ D 0 + q u ( t i ) = B i ( v ( t i ) , D 0 + p v ( t i ) ) , i = 1 , 2 , , k ; Δ v ( t i ) = C i ( u ( t i ) , D 0 + q u ( t i ) ) , Δ D 0 + p v ( t i ) = D i ( u ( t i ) , D 0 + q u ( t i ) ) , i = 1 , 2 , , k ; D 0 + α 1 u ( 0 ) = i = 1 m a i D 0 + α 1 u ( ξ i ) , u ( 1 ) = i = 1 m b i η i 2 α u ( η i ) ; D 0 + β 1 v ( 0 ) = i = 1 m c i D 0 + β 1 v ( ζ i ) , v ( 1 ) = i = 1 m d i θ i 2 β v ( θ i ) , (1.1)

where 1 < α , β < 2 , α q 1 , β p 1 and 0 < ξ 1 < ξ 2 < < ξ m < 1 , 0 < η 1 < η 2 < < η m < 1 , 0 < ζ 1 < ζ 2 < < ζ m < 1 , 0 < θ 1 < θ 2 < < θ m < 1 . f , g : [ 0 , 1 ] × R 2 R satisfy Carathéodory conditions, A i , B i , C i , D i : R × R R . Δ w ( t i ) = w ( t i + ) w ( t i ) , Δ D 0 + r w ( t i ) = D 0 + r w ( t i + ) D 0 + r w ( t i ) , here w { u , v } , r { p , q } , w ( t i + ) and w ( t i ) denote the right and left limits of w ( t ) at t = t i , respectively, and the fractional derivative is understood in the Riemann-Liouville sense. k, m, a i , b i , c i , d i ( i = 1 , 2 , , m ) are fixed constant satisfying i = 1 m a i = i = 1 m b i = i = 1 m c i = i = 1 m d i = 1 and i = 1 m b i η i = i = 1 m d i θ i = 1 .

The coupled system (1.1) happens to be at resonance in the sense that the associated linear homogeneous coupled system

{ D 0 + α u ( t ) = 0 , D 0 + β v ( t ) = 0 , 0 < t < 1 ; D 0 + α 1 u ( 0 ) = i = 1 m a i D 0 + α 1 u ( ξ i ) , u ( 1 ) = i = 1 m b i η i 2 α u ( η i ) ; D 0 + β 1 v ( 0 ) = i = 1 m c i D 0 + β 1 v ( ζ i ) , v ( 1 ) = i = 1 m d i θ i 2 β v ( θ i )

has ( u ( t ) , v ( t ) ) = ( h 1 t α 1 + h 2 t α 2 , h 3 t β 1 + h 4 t β 2 ) , c i R , i = 1 , 2 , 3 , 4 as a nontrivial solution. To solve this interesting and important problem and to overcome the difficulties caused by the impulses, we will construct some Banach spaces, then we shall obtain the new solvability results for the coupled system (1.1) with the help of a coincidence degree continuation theorem. The main contributions of this work are Lemma 2.1 and Lemma 3.1 in Section 3 since the calculations are disposed well.

The plan of this work is organized as follows. Section 2 contains some necessary notations, definitions and lemmas that will be used in the sequel. In Section 3, we establish a theorem on the existence of solutions for the coupled system (1.1) based on the coincidence degree theory due to Mawhin [30,31].

### 2 Background materials and preliminaries

For the convenience of the readers, we recall some notations and an abstract existence theorem [30,31].

Let Y, Z be real Banach spaces, L : dom ( L ) Y Z be a Fredholm map of index zero and P : Y Y , Q : Z Z be continuous projectors such that Im ( P ) = Ker ( L ) , Ker ( Q ) = Im ( L ) and Y = Ker ( L ) Ker ( P ) , Z = Im ( L ) Im ( Q ) . It follows that L | dom ( L ) Ker ( P ) : dom ( L ) Ker ( P ) Im ( L ) is invertible. We denote the inverse of the map by K P . If Ω is an open bounded subset of Y such that dom ( L ) Ω , the map N : Y Z will be called L-compact on Ω ¯ if Q N ( Ω ¯ ) is bounded and K P ( I Q ) N : Ω ¯ Y is compact.

The main tool we used is Theorem 2.4 of [30].

Theorem 2.1LetLbe a Fredholm operator of index zero, and letNbeL-compact on Ω ¯ . Assume that the following conditions are satisfied:

(i) L x λ N x for every ( x , λ ) [ ( dom ( L ) Ker ( L ) ) Ω ] × ( 0 , 1 ) ;

(ii) N x Im ( L ) for every x Ker ( L ) Ω ;

(iii) deg ( Q N | Ker ( L ) , Ω Ker ( L ) , 0 ) 0 , where Q : Z Z is a projection as above with Im ( L ) = Ker ( Q ) .

Then the equation L x = N x has at least one solution in dom ( L ) Ω ¯ .

Now, we present some basic knowledge and definitions about fractional calculus theory, which can be found in the recent works [13,16,32].

Definition 2.1 The fractional integral of order α > 0 of a function y : ( 0 , ) R is defined by

I 0 + α y ( t ) = 0 t ( t s ) α 1 Γ ( α ) y ( s ) d s ,

provided the right-hand side is pointwise defined on ( 0 , ) .

Definition 2.2 The fractional derivative of order α > 0 of a function y : ( 0 , ) R is defined by

D 0 + α y ( t ) = 1 Γ ( n α ) ( d d t ) n 0 t ( t s ) n α 1 y ( s ) d s ,

where n = [ α ] + 1 , provided the right-hand side is pointwise defined on ( 0 , ) .

Remark 2.1 It can be directly verified that the Riemann-Liouville fractional integration and fractional differentiation operators of the power functions t μ yield power functions of the same form. For α 0 , μ 1 , we have

I 0 + α t μ = Γ ( μ + 1 ) Γ ( μ + α + 1 ) t μ + α , D 0 + α t μ = Γ ( μ + 1 ) Γ ( μ α + 1 ) t μ α ( μ α ) . (2.1)

Proposition 2.1[17]

Assume that y C ( 0 , 1 ) L [ 0 , 1 ] with a fractional derivative of order α > 0 that belongs to C ( 0 , 1 ) L [ 0 , 1 ] . Then

I 0 + α D 0 + α y ( t ) = y ( t ) + c 1 t α 1 + c 2 t α 2 + + c N t α N (2.2)

for some c i R , i = 1 , 2 , , N , whereNis the smallest integer grater than or equal toα.

Proposition 2.2[32]

If α > 0 , β > 0 , then the equation

( I 0 + α I 0 + β y ) ( t ) = ( I 0 + α + β y ) ( t )

is satisfied for a continuous functiony.

If α > 0 , m N and D = d / d t , the fractional derivatives ( D 0 + α y ) ( t ) and ( D 0 + α + m y ) ( t ) exist, then

( D m D 0 + α y ) ( t ) = ( D 0 + α + m y ) ( t ) .

If α > 0 , then the equation

( D 0 + α I 0 + α y ) ( t ) = y ( t )

is satisfied for a continuous functiony.

If α > β > 0 , then the relation

( D 0 + β I 0 + α y ) ( t ) = ( I 0 + α β y ) ( t )

holds for a continuous functiony.

Let C [ 0 , 1 ] = { u | u  is continuous in  [ 0 , 1 ] } with the norm u = max t [ 0 , 1 ] | u ( t ) | and

P C [ 0 , 1 ] = { x : x C ( t i , t i + 1 ] ,  there exist  x ( t i )  and x ( t i + )  with  x ( t i ) = x ( t i ) , i = 1 , 2 , , k 1 }

with the norm x P C = sup t [ 0 , 1 ] | x ( t ) | . Denote

where u α ( t ) = t 2 α u ( t ) , v β ( t ) = t 2 β v ( t ) with the norm

Thus, Y = Y 1 × Y 2 is a Banach space with the norm defined by ( u , v ) Y = max { u Y 1 , v Y 2 } .

Set Z 1 = Z 2 = P C [ 0 , 1 ] × R 2 k equipped with the norm

x Z 1 = max { y P C , | c | } , x = ( y , c ) Z 1 ,

thus Z = Z 1 × Z 2 is a Banach space with the norm defined by ( x , y ) Z = max { x Z 1 , y Z 2 } .

Define the operator L : Y Z , L ( u , v ) = ( L 1 u , L 2 v ) , dom ( L ) = dom ( L 1 ) × dom ( L 2 ) , where

with

Let N : Y Z be defined as N ( u , v ) = ( N 1 v , N 2 u ) , where

Then the coupled system of boundary value problem (1.1) can be written as

L ( u , v ) = N ( u , v ) .

For the sake of simplicity, we define the operators T 1 , T 2 : Z 1 Z 1 for X = ( x , δ 1 , , δ k , ω 1 , , ω k ) as follows:

(2.3)

(2.4)

By the same way, we define the operators T 3 , T 4 : Z 2 Z 2 for Y = ( y , ρ 1 , , ρ k , τ 1 , , τ k ) as follows:

(2.5)

(2.6)

In what follows, we present the following lemmas which will be used to prove our main results.

Lemma 2.1If the following condition is satisfied:

(H1) σ 1 = | σ 11 σ 12 σ 13 σ 14 | 0 , σ 2 = | σ 21 σ 22 σ 23 σ 24 | 0 , where

then L : dom ( L ) Y Z is a Fredholm operator of index zero. Moreover, Ker ( L ) = Ker ( L 1 ) × Ker ( L 2 ) , where

Ker ( L 1 ) = { h 1 t α 1 + h 2 t α 2 , h 1 , h 2 R } , Ker ( L 2 ) = { h 3 t β 1 + h 4 t β 2 , h 3 , h 4 R } (2.7)

and Im ( L ) = Im ( L 1 ) × Im ( L 2 ) , here

(2.8)

(2.9)

Proof It is clear that (2.7) holds. For ( u , v ) Ker ( L ) , we have L ( u , v ) = ( L 1 u , L 2 v ) = ( 0 , 0 ) , i.e., L 1 u = 0 , L 2 v = 0 , then u Ker ( L 1 ) , v Ker ( L 2 ) , so Ker ( L ) = Ker ( L 1 ) × Ker ( L 2 ) . Similarly, it is not difficult to see that Im ( L ) = Im ( L 1 ) × Im ( L 2 ) . Next, we will show that (2.8) and (2.9) hold.

If Z 1 = ( z 1 , δ 1 , , δ k , ω 1 , , ω k ) Im ( L 1 ) , Z 2 = ( z 2 , ρ 1 , , ρ k , τ 1 , , τ k ) Im ( L 2 ) , then there exist u dom ( L 1 ) and v dom ( L 2 ) such that

{ D 0 + α u ( t ) = z 1 ( t ) , Δ u ( t i ) = δ i , Δ D 0 + q u ( t i ) = ω i , { D 0 + β v ( t ) = z 2 ( t ) , Δ v ( t i ) = ρ i , Δ D 0 + p v ( t i ) = τ i (2.10)

and

(2.11)

(2.12)

Proposition 2.1 together with (2.10)-(2.12) gives that

(2.13)

(2.14)

Substituting the boundary condition D 0 + α 1 u ( 0 ) = i = 1 m a i D 0 + α 1 u ( ξ i ) into (2.13), one has

i = 1 m a i ( 0 ξ i z 1 ( s ) d s + Γ ( α q ) t i < ξ i ω i t i q + 1 α ) = 0 , (2.15)

and substituting the boundary condition u ( 1 ) = i = 1 m b i η i 2 α u ( η i ) into (2.13), one has

(2.16)

By the same way, if we substitute the condition (2.12) into (2.14), then we can obtain that

i = 1 m c i ( 0 ζ i z 2 ( s ) d s + Γ ( β p ) t i < ζ i τ i t i p + 1 β ) = 0 , (2.17)

and

(2.18)

Conversely, if (2.15)-(2.18) hold, set

It is easy to check that the above u, v satisfy equation (2.10)-(2.12). Thus, (2.8) and (2.9) hold.

Define the operator Q : Z Z , Q ( x , y ) = ( Q 1 x , Q 2 y ) with Q 1 X = Q 11 X + Q 12 X t , Q 2 Y = Q 21 Y + Q 22 Y t , here

In what follows, we will show that Q 1 and Q 2 are linear projectors. By some direct computations, we have

As a result,

Q 1 ( Q 1 X ) = Q 1 ( Q 11 X + Q 12 X t ) = Q 11 ( Q 11 X + Q 12 X t ) + Q 12 ( Q 11 X + Q 12 X t ) t = Q 11 2 X + Q 11 ( Q 12 X t ) + [ Q 12 ( Q 11 X ) + Q 12 ( Q 12 X t ) ] t = Q 11 X + Q 12 X t = Q 1 X .

Similarly, we can see that Q 2 ( Q 2 Y ) = Q 2 Y . Then for ( X , Y ) Z , we have Q 2 ( X , Y ) = Q ( Q 1 X , Q 2 Y ) = ( Q 1 2 X , Q 2 2 Y ) = ( Q 1 X , Q 2 Y ) = Q ( X , Y ) . It means that the operator Q : Z Z is a projector.

Now, we show that Ker ( Q ) = Im ( L ) . Obviously, Im ( L ) Ker ( Q ) . On the other hand, for ( X , Y ) Ker ( Q ) , then Q ( X , Y ) = ( 0 , 0 ) implies that

{ σ 11 T 1 X σ 12 T 2 X = ( 0 , 0 , , 0 ) , σ 13 T 1 X σ 14 T 2 X = ( 0 , 0 , , 0 ) , { σ 21 T 1 Y σ 22 T 2 Y = ( 0 , 0 , , 0 ) , σ 23 T 1 Y σ 24 T 2 Y = ( 0 , 0 , , 0 ) .

The condition (H1) guarantees that T 1 X = T 2 X = ( 0 , 0 , , 0 ) , T 3 Y = T 4 Y = ( 0 , 0 , , 0 ) , then ( X , Y ) Im ( L ) . Hence, Ker ( Q ) = Im ( L ) .

For W Z , let W = ( W Q W ) + Q W . Then W Q W Ker ( Q ) = Im ( L ) , Q W Im ( Q ) , it means that Z = Im ( L ) + Im ( Q ) . Moreover, Ker ( Q ) = Im ( L ) gives that Im ( L ) Im ( Q ) = ( 0 , 0 ) . Thus, Z = Im ( L ) Im ( Q ) . Then dim Ker ( L ) = dim Im ( Q ) = codim Im ( L ) = 4 , L is a Fredholm map of index zero. □

Define the operator P : Y Y with P ( u , v ) = ( P 1 u , P 2 v ) , here P 1 : Y 1 Y 1 , P 2 : Y 2 Y 2 are defined as follows:

Moreover, we define K P : Im ( L ) dom ( L ) Ker ( P ) as K P ( X , Y ) = ( K P 1 X , K P 2 Y ) , where K P i : Im ( L i ) dom ( L i ) Ker ( P i ) , i = 1 , 2 is defined as follows:

Lemma 2.2Assume that Ω Y is an open bounded subset with dom ( L ) Ω ¯ , thenNisL-compact on Ω ¯ .

Proof Obviously, Im ( P ) = Ker ( L ) . By a direct computation, we have that

P 1 2 u = 1 Γ ( α ) D 0 + α 1 P 1 u ( 0 ) t α 1 + lim t 0 t 2 α P 1 u ( t ) t α 2 = 1 Γ ( α ) D 0 + α 1 u ( 0 ) t α 1 + lim t 0 t 2 α u ( t ) t α 2 = P 1 u .

Similarly, P 2 2 v = P 2 v . This gives that P 2 ( u , v ) = P ( P 1 u , P 2 v ) = ( P 1 2 u , P 2 2 v ) = ( P 1 u , P 2 v ) = P ( u , v ) , that is to say, the operator P is a linear projector. It is easy to check from w = ( w P w ) + P w that Y = Ker ( P ) + Ker ( L ) . Moreover, we can see that Ker ( P ) Ker ( L ) = ( 0 , 0 ) . Thus, Y = Ker ( P ) Ker ( L ) .

In what follows, we will show that K P defined above is the inverse of L | dom ( L ) Ker ( P ) .

If ( X , Y ) Im ( L ) , then L 1 K P 1 X = X , L 2 K P 2 Y = Y , which gives that

L K P ( X , Y ) = ( L 1 K P 1 X , L 2 K P 2 Y ) = ( X , Y ) .

On the other hand, for ( u , v ) dom ( L ) Ker ( P ) , we have

( K P 1 L 1 ) u ( t ) = K P 1 ( D 0 + α u ( t ) , δ 1 , , δ k , ω 1 , , ω k ) = u ( t ) + ( h 1 + Γ ( α q ) Γ ( α ) t i < t ω i t i q + 1 α ) t α 1 + ( h 2 + t i < t δ i t i 2 α Γ ( α q ) Γ ( α ) t i < t ω i t i q + 2 α ) t α 2 .

Since u K P 1 and K P 1 L 1 u Ker ( P 1 ) , then

(2.19)

(2.20)

By some calculations, (2.19) and (2.20) imply that

It means that K P 1 L 1 u = u . Analogously, K P 2 L 2 v = v . Thus, K P L ( u , v ) = ( K P 1 L 1 u , K P 2 L 2 v ) = ( u , v ) . So, K P is the inverse of L | dom ( L ) Ker ( P ) .

Finally, we show that N is L-compact on Ω ¯ . Denote Q 1 N 1 v = ( v , 0 , , 0 ) , Q 2 N 2 u = ( u , 0 , , 0 ) , where

Then we can see that

K P ( I Q ) N ( u , v ) = K P ( I Q ) ( N 1 v , N 2 u ) = ( K P 1 ( I Q 1 ) N 1 v , K P 2 ( I Q 2 ) N 2 u ) ,

where

So, we can see that Q 1 N 1 is bounded and K P 1 ( I Q 1 ) N 1 is uniformly bounded.

For 0 t 1 < t 2 1 , we have

(2.21)

(2.22)

The equicontinuity of t α , t α + 1 together with (2.21) and (2.22) gives that | K P 1 ( I Q 1 ) N 1 v ( t 2 ) K P 1 ( I Q 1 ) N 1 v ( t 1 ) | 0 as t 2 t 1 , which yields that K P 1 ( I Q 1 ) N 1 is equicontinuous. By the Ascoli-Arzela theorem, we can see that K P 1 ( I Q 1 ) N 1 is compact. By the same way, Q 2 N 2 is bounded and K P 2 ( I Q 2 ) N 2 is compact. Since Q N ( u , v ) = Q ( N 1 v , N 2 u ) = ( Q 1 N 1 v , Q 2 N 2 u ) and K P ( I Q ) N ( u , v ) = ( K P 1 ( I Q 1 ) N 1 v , K P 2 ( I Q 2 ) N 2 u ) , then QN is bounded and K P ( I Q ) N is compact. This means that N is L-compact on Ω ¯ . □

### 3 Main results

In this section, we present the existence results of the coupled system (1.1). To do this, we need the following hypotheses.

(H2) There exist functions φ i , ψ i , γ i C [ 0 , 1 ] , i = 1 , 2 , such that

where ψ i , γ i ( i = 1 , 2 ) satisfy

here

(H3) For ( u , v ) dom ( L ) , there exist constants e i ( 0 , 1 ) ( i = 0 , 1 , 2 ), M i > 0 ( i = 1 , 2 ) such that

(1) if either | u ( t ) | > M 1 or | v ( t ) | > M 1 for t [ e 0 , e 1 ] , then either T 2 N 1 v ( t ) 0 or T 4 N 2 u ( t ) 0 ;

(2) if either | D 0 + q u ( t ) | > M 2 or | D 0 + p v ( t ) | > M 2 , t [ e 2 , 1 ] , then either T 1 N 1 v ( t ) 0 or T 3 N 2 u ( t ) 0 .

(H4) For ( u , v ) Ker ( L ) , there exist constants g i > 0 ( i = 1 , 2 ) such that if either | h 1 | g 1 or | h 2 | g 1 , either | h 3 | g 2 or | h 4 | g 2 , then either (1) or (2) holds, where

(1)

here s 1 , s 2 are positive constants;

(2)

here s 3 , s 4 are negative constants.

Lemma 3.1Suppose that (H2)-(H3) hold. Then the set

Ω 1 = { ( u , v ) dom ( L ) Ker ( L ) | L ( u , v ) = λ N ( u , v ) , λ ( 0 , 1 ) }

is bounded inY.

Proof For ( u , v ) Ω 1 , by L ( u , v ) = ( L 1 u , L 2 v ) = λ N ( u , v ) = ( λ N 1 v , λ N 2 u ) and ( u , v ) dom ( L ) , we have

(3.1)

(3.2)

(3.3)

(3.4)

Since N 1 v Im ( L 1 ) , N 2 u Im ( L 2 ) , then T 1 ( N 1 v ) = T 2 ( N 1 v ) = 0 , T 3 ( N 2 u ) = T 4 ( N 2 u ) = 0 . Then we can see, from the condition (H3), that there exist constants e 0 , e , e ( 0 , 1 ) such that | u ( t ) | M 1 , | v ( t ) | M 1 for t [ e 0 , e ] and | D 0 + q u ( t ) | M 2 , | D 0 + p v ( t ) | M 2 for t [ e , 1 ] . So, we can see from (3.1) and (3.2) that

| h 1 | Γ ( α q ) Γ ( α ) t α q 1 M 2 + 1 Γ ( α ) sup t [ 0 , 1 ] | f ( t , v ( t ) , D 0 + p v ( t ) ) | + Γ ( α q ) Γ ( α ) t i < t | B i | t i q + 1 α Γ ( α q ) Γ ( α ) e α q 1 M 2 + 1 Γ ( α ) sup t [ 0 , 1 ] | f ( t , v ( t ) , D 0 + p v ( t ) ) | + Γ ( α q ) Γ ( α ) i = 1 k | B i | t i q + 1 α (3.5)

and

| h 2 | M 1 + 1 Γ ( α ) sup t [ 0 , 1 ] | f ( t , v ( t ) , D 0 + p v ( t ) ) | + | h 1 | + Γ ( α q ) Γ ( α ) ( t i < t | B i | t i q + 1 α ) t + t i < t | A i | t i 2 α + Γ ( α q ) Γ ( α ) t i < t | B i | t i q + 2 α M 1 + 1 Γ ( α ) sup t [ 0 , 1 ] | f ( t , v ( t ) , D 0 + p v ( t ) ) | + | h 1 | + i = 1 k | A i | t i 2 α + Γ ( α q ) Γ ( α ) i = 1 k | B i | t i q + 1 α ( 1 + t i ) . (3.6)

Then for t [ 0 , 1 ] and u dom ( L 2 ) , we have

(3.7)

(3.8)

(3.9)

Similarly, for u dom ( L 2 ) , we have that

(3.10)

(3.11)

(3.12)

Substitute (3.11) and (3.12) into (3.8), then we have

u α P C 4 Γ ( α ) φ 1 + 4 Γ ( α ) ψ 1 × ( 4 Γ ( β ) [ φ 2 + ψ 2 u α P C + γ 2 D 0 + q u P C ] + R 2 ) + 4 Γ ( α ) γ 1 ( 2 Γ ( β p ) [ φ 2 + ψ 2 u α P C + γ 2 D 0 + q u P C ] + R 3 ) = ( 16 ψ 1 Γ ( α ) Γ ( β ) + 8 γ 1 Γ ( α ) Γ ( β p ) ) [ φ 2 + ψ 2 u α P C + γ 2 D 0 + q u P C ] + 4 Γ ( α ) ( φ 1 + ψ 1 R 2 + γ 1 R 3 ) . (3.13)

It means that

u α P C A D 0 + q u P C + B ,

similarly,

v β P C A D 0 + p v P C + B .

Substituting the above two into (3.9) and (3.12), we can see that

D 0 + q u P C 2 Γ ( α q ) ( ψ 1 A + γ 1 ) D 0 + p v P C + 2 Γ ( α q ) ( φ 1 + ψ 1 B ) R 3 (3.14)

and

D 0 + p v P C 2 Γ ( β p ) ( ψ 2 A + γ 2 ) D 0 + q u P C + 2 Γ ( β p ) ( φ 2 + ψ 2 B ) R 3 . (3.15)

From the condition (H2), (3.14) and (3.15) give that D 0 + q u P C and D 0 + p v P C are bounded, then u α P C and v β P C are also bounded. Thus, by the definition of the norm on Y, u Y 1 and v Y 2 are bounded. That is, Ω 1 is bounded in Y. □

Lemma 3.2Suppose that the condition (H3) holds. Then the set

Ω 2 = { ( u , v ) | ( u , v ) Ker ( L ) , N ( u , v ) Im ( L ) }

is bounded inY.

Proof For ( u , v ) Ker ( L ) , we have that ( u , v ) = ( h 1 t α 1 + h 2 t α 2 , h 3 t β 1 + h 4 t β 2 ) , where h i , i { 1 , 2 , 3 , 4 } . Since N ( u , v ) Im ( L ) , so we have

T 1 N 1 ( h 3 t β 1 + h 4 t β 2 ) = T 2 N 1 ( h 3 t β 1 + h 4 t β 2 ) = 0

and

T 3 N 2 ( h 1 t α 1 + h 2 t α 2 ) = T 4 N 2 ( h 1 t α 1 + h 2 t α 2 ) = 0 .

From (H3), there exist positive constants M , M , e 0 , e , e such that for t [ e , 1 ] ,

| D 0 + q u ( t ) | = Γ ( α ) Γ ( α q ) | h 1 | t α q 1 M ,

which means that | h 1 | M Γ ( α q ) Γ ( α ) e α q 1 . And for t [ e 0 , e ] ,

| u ( t ) | = | h 1 t α 1 + h 2 t α 2 | M ,

which means that | h 2 | = | u ( t ) t 2 α + h 1 t | | u ( t ) | + | h 1 | M + | h 1 | . So, we can see that for t [ 0 , 1 ] ,

The above two arguments imply that | u | Y 1 is bounded. In the same way, | v | Y 2 is bounded. Thus, Ω 2 is bounded in Y. □

Lemma 3.3The set

Ω 3 = { ( u , v ) Ker ( L ) | λ J ( u , v ) + ( 1 λ ) θ Q N ( u , v ) = ( 0 , 0 , , 0 ) , λ [ 0 , 1 ] }

is bounded inY, where J : Ker ( L ) Im ( Q ) is the linear isomorphism given by

and

θ = { 1 , if  (H ) (1)  hold 4 , 1 , if  (H ) (2)  hold 4 .

Proof For ( u , v ) Ker ( L ) , set u = h 1 t α 1 + h 2 t α 2 , v = h 3 t β 1 + h 4 t β 2 , then λ J ( u , v ) + ( 1 λ ) θ Q N ( u , v ) = ( 0 , 0 , , 0 ) implies that

(3.16)

(3.17)

(3.18)

(3.19)

From (3.16) and (3.17), we have

λ ( h 1 2 + h 2 2 , 0 , , 0 ) + ( 1 λ ) θ [ h 1 T 1 N 1 ( v ) + h 2 T 2 N 1 ( v ) ] = ( 0 , 0 , , 0 ) ,

the condition (H4) gives that

λ ( h 1 2 + h 2 2 ) = ( 1 λ ) θ s < 0 ,

where

s = { s 1 , if (H ) (1) hold 4 , s 3 , if (H ) (2) hold 4 ,

which is a contradiction. As a result, there exist positive constants g 1 , g 2 such that | h 1 | g 1 , | h 2 | g 2 . Similarly, from (3.18)-(3.19) and the second part of (1) or (2) of (H4), there exist two positive constants g 3 , g 4 such that | h 3 | g 3 , | h 4 | g 4 . It follows that u Y 1 , v Y 1 are bounded, that is, Ω 3 is bounded in Y. □

Theorem 3.1Suppose that (H1)-(H4) hold. Then the problem (1.1) has at least one solution inY.

Proof Let Ω be a bounded open set of Y such that i = 1 3 Ω ¯ i Ω . It follows from Lemma 2.2 that N is L-compact on Ω ¯ . By means of above Lemmas 3.1-3.3, one obtains that

(i) L ( u , v ) λ N ( u , v ) for every ( ( u , v ) , λ ) [ ( dom ( L ) Ker ( L ) ) Ω ] × ( 0 , 1 ) ;

(ii) N ( u , v ) Im ( L ) for every ( u , v ) Ker ( L ) Ω .

Then we need only to prove

(iii) deg ( Q N | Ker ( L ) , Ω Ker ( L ) , ( 0 , 0 , , 0 ) ) 0 .

Take

H ( u , v , λ ) = ± λ J + ( 1 λ ) N ( u , v ) .

According to Lemma 3.3, we know H ( ( u , v ) , λ ) ( 0 , 0 , , 0 ) for all ( u , v ) Ω Ker ( L ) . Thus, the homotopy invariance property of degree theory gives that

deg ( Q N | Ker ( L ) , Ω Ker ( L ) , ( 0 , 0 , , 0 ) ) = deg ( H ( , 0 ) , Ω Ker ( L ) , ( 0 , 0 , , 0 ) ) = deg ( H ( , 1 ) , Ω Ker ( L ) , ( 0 , 0 , , 0 ) ) = deg ( ± J , Ω Ker ( L ) , ( 0 , 0 , , 0 ) ) 0 .

Then, by Theorem 2.1, L ( u , v ) = N ( u , v ) has at least one solution in dom ( L ) Ω ¯ , i.e., the problem (1.1) has at least one solution in Y, which completes the proof. □

### 4 An example

Example 4.1

Consider the following boundary value problem for coupled systems of impulsive fractional differential equations:

{ D 0 + 3 2 u ( t ) = f ( t , v ( t ) , D 0 + 1 6 v ( t ) ) , D 0 + 4 3 v ( t ) = g ( t , u ( t ) , D 0 + 1 4 u ( t ) ) , 0 < t < 1 ; Δ u ( 3 4 ) = A 1 ( v ( 3 4 ) , D 0 + 1 6 v ( 3 4 ) ) , Δ D 0 + 1 4 u ( 3 4 ) = B 1 ( v ( 3 4 ) , D 0 + 1 6 v ( 3 4 ) ) ; Δ v ( 2 3 ) = C 1 ( u ( 2 3 ) , D 0 + 1 4 u ( 2 3 ) ) , Δ D 0 + 1 6 v ( 2 3 ) = D 1 ( u ( 2 3 ) , D 0 + 1 4 u ( 2 3 ) ) ; D 0 + 1 2 u ( 0 ) = 3 D 0 + 1 2 u ( 1 6 ) 2 D 0 + 1 2 u ( 1 4 ) , u ( 1 ) = 5 u ( 1 5 ) + 2 3 u ( 1 3 ) ; D 0 + 1 3 v ( 0 ) = 4 D 0 + 1 3 v ( 1 5 ) 3 D 0 + 1 3 v ( 1 4 ) , v ( 1 ) = 3 ( 1 3 ) 2 3 v ( 1 3 ) + 4 ( 1 2 ) 2 3 v ( 1 2 ) , (4.1)

where

and

Due to the coupled problem (1.1), we have that α = 3 2 , β = 4 3 , p = 1 6 , q = 1 4 , a 1 = 3 , a 2 = 2 , b 1 = 5 , b 2 = 6 , c 1 = 4 , c 2 = 3 , d 1 = 3 , d 2 = 4 . ξ 1 = 1 6 , ξ 2 = 1 4 ; η 1 = 1 5 , η 2 = 1 3 ; ζ 1 = 1 5 , ζ 2 = 1 4 ; θ 1 = 1 3 , θ 2 = 1 2 . Obviously, a 1 + a 2 = b 1 + b 2 = c 1 + c 2 = d 1 + d 2 = 1 and b 1 η 1 + b 2 η 2 = d 1 θ 1 + d 2 θ 2 = 1 . By direct calculation, we obtain that

It is easy to see that

where

So, ψ 1 = 1 24 , γ 1 = 1 4 π , ψ 2 = 1 20 , γ 2 = 3 100 . And

where A = 0.0527672 , A = 0.0110034 . Thus, the condition (H2) holds.

Taking M 1 = 1 , for any v dom ( L 2 ) , assume that | D 0 + 1 6 v ( t ) | > 1 holds for any t [ 1 12 , 1 6 ] . Thus either D 0 + 1 6 v ( t ) > 1 or D 0 + 1 6 v ( t ) < 1 for any t [ 1 12 , 1 6 ] . If D 0 + 1 6 v ( t ) > 1 , t [ 1 12 , 1 6 ] , then

If D 0 + v ( t ) < 1 , t [ 1 12 , 1 6 ] , then

Similarly, assume that | D 0 + 1 4 v ( t ) | > 1 holds for any t [ 1 10 , 1 5 ] . Thus either D 0 + 1 4 v ( t ) > 1 or D 0 + 1 4 v ( t ) < 1 for any t [ 1 10 , 1 5 ] . If D 0 + 1 6 v ( t ) > 1 , t [ 1 10 , 1 5 ] , then

If D 0 + 1 6 v ( t ) < 1 , t [ 1 10 , 1 5 ] , then

So, from the above arguments, the first part of the condition (H3) is true for M 1 = 1 , t [ 1 12 , 1 6 ] .

Taking M 21 = 16 , assume that | v | > 16 holds for any t [ 1 3 , 1 ] . Then either v > 16 or v < 16 for t [ 1 3 , 1 ] . If v > 16 for t [ 1 3 , 1 ] , then

If v < 16 for t [ 1 3 , 1 ] , then

By the same way, taking M 22 = 10 / 3 3 , assume that | u ( t ) | > M 22 holds for any t [ 1 2 , 1 ] . Then either v > 10 / 3 3 or v < 10 / 3 3 for t [ 1 2 , 1 ] . If v > 10 / 3 3 for t [ 1 2 , 1 ] , then

If v > 10 / 3 3 for t [ 1 2 , 1 ] , then

So, from the above arguments, the second part of the condition (H3) holds for M 2 = max { M 21 , M 22 } = 16 , t [ 1 3 , 1 ] .

On the other hand, for ( u , v ) = ( h 1 t α 1 + h 2 t α 2 , h 3 t β 1 + h 4 t β 2 ) Ker ( L ) , taking g = 8 , assume that h i < 8 , i = 1 , 2 , 3 , 4 , then h 3 t β 1 + h 4 t β 2 < 16 for t [ 1 3 , 1 ] , h 1 t α 1 + h 2 t α 2 < 10 3 3 for t [ 1 2 , 1 ] . And D 0 + 1 4 u < 1 for t [ 1 10 , 1 5 ] , D 0 + 1 6 u < 1 for t [ 1 12 , 1 6 ] . Then we can see, from the above arguments, that T 1 N 1 v = ( r 1 , 0 , , 0 ) , T 2 N 1 v = ( r 2 , 0 , , 0 ) , T 1 N 2 u = ( r 3 , 0 , , 0 ) , T 2 N 2 u = ( r 4 , 0 , , 0 ) , where r i < 0 , i = 1 , 2 , 3 , 4 . Thus,

where s 1 > 0 , s 2 > 0 . So, the condition (H4) holds. Hence, from Theorem 3.1, the coupled problem (4.1) has at least one solution in { u 3 2 , D 0 + 1 2 u P C [ 0 , 1 ] } × { v 4 3 , D 0 + 1 3 v P C [ 0 , 1 ] } .

### Competing interests

The authors declare that they have no competing interests.

### Authors’ contributions

CZ conceived the main idea of the study, XZ carried out the main parts of the draft. CZ gave many valuable suggestions and corrected the main theorems in the discussion. All authors read and approved the final manuscript.

### Acknowledgements

The authors would like to thank the editor and referee for their valuable comments and remarks which lead to a great improvement of the article. This research is supported by the National Natural Science Foundation of China (11071108), the Provincial Natural Science Foundation of Jiangxi, China (2010GZS0147, 20114BAB201007).

### References

1. Wang, JR, Fečkan, M, Zhou, Y: On the new concept of solutions existence results for impulsive fractional evolution equations. Dyn. Partial Differ. Equ.. 8, 345–361 (2011)

2. Agarwal, RP, Benchohra, M, Hamani, S: A survey on existence results for boundary value problems of nonlinear fractional differential equations and inclusions. Acta Appl. Math.. 109, 973–1033 (2010). Publisher Full Text

3. Ahmad, B, Sivasundaram, S: Existence results for nonlinear impulsive hybrid boundary value problems involving fractional differential equations. Nonlinear Anal. Hybrid Syst.. 3, 251–258 (2009). Publisher Full Text

4. Balachandran, K, Kiruthika, S: Existence of solutions of abstract fractional impulsive semilinear evolution equations. Electron. J. Qual. Theory Differ. Equ.. 4, 1–12 (2010)

5. Wang, G, Zhang, L, Song, G: Systems of first order impulsive functional differential equations with deviating arguments and nonlinear boundary conditions. Nonlinear Anal.. 74, 974–982 (2011). Publisher Full Text

6. Li, XP, Chen, FL, Li, XZ: Generalized anti-periodic boundary value problems of impulsive fractional differential equations. Commun. Nonlinear Sci. Numer. Simul.. 18, 28–41 (2013). Publisher Full Text

7. Wang, JR, Zhou, Y, Fečkan, M: On recent developments in the theory of boundary value problems for impulsive fractional differential equations. Comput. Math. Appl.. 64(10), 3008–3020 (2012). Publisher Full Text

8. Fečkan, M, Zhou, Y, Wang, JR: On the concept and existence of solution for impulsive fractional differential equations. Commun. Nonlinear Sci. Numer. Simul.. 17, 3050–3060 (2012). Publisher Full Text

9. Metzler, R, Klafter, J: Boundary value problems for fractional diffusion equations. Physica A. 278, 107–125 (2000). Publisher Full Text

10. Scher, H, Montroll, E: Anomalous transit-time dispersion in amorphous solids. Phys. Rev. B. 12, 2455–2477 (1975). Publisher Full Text

11. Mainardi, F: Fractional diffusive waves in viscoelastic solids. In: Wegner JL, Norwood FR (eds.) Nonlinear Waves in Solids, pp. 93–97. ASME/AMR, Fairfield (1995)

12. Hilfer, R: Applications of Fractional Calculus in Physics, World Scientific, Singapore (2000)

13. Miller, KS, Ross, B: An Introduction to the Fractional Calculus and Differential Equations, Wiley, New York (1993)

14. Podlubny, I: Fractional Differential Equation, Academic Press, San Diego (1999)

15. Lakshmikantham, V, Leela, S, Devi, JV: Theory of Fractional Dynamic Systems, Cambridge Academic, Cambridge (2009)

16. Samko, SG, Kilbas, AA, Marichev, OI: Fractional Integrals and Derivatives, Theory and Applications, Gordon & Breach, Switzerland (1993)

17. Bai, CZ, Lü, H: Positive solutions of boundary value problems of nonlinear fractional differential equation. J. Math. Anal. Appl.. 311, 495–505 (2005). Publisher Full Text

18. Lakshmikantham, V, Vatsala, AS: General uniqueness and monotone iterative technique for fractional differential equations. Appl. Math. Lett.. 21, 828–834 (2008). Publisher Full Text

19. Zhang, XZ, Zhu, CX, Wu, ZQ: The Cauchy problem for a class of fractional impulsive differential equations with delay. Electron. J. Qual. Theory Differ. Equ.. 37, 1–13 (2012)

20. Ahmada, B, Nieto, JJ: Existence results for a coupled system of nonlinear fractional differential equations with three-point boundary conditions. Comput. Math. Appl.. 58, 1838–1843 (2009). Publisher Full Text

21. Hu, ZG, Liu, WB, Chen, TY: Existence of solutions for a coupled system of fractional differential equations at resonance. Bound. Value Probl.. 2012, (2012) Article ID 98

22. Su, XW: Boundary value problem for a coupled system of nonlinear fractional differential equations. Appl. Math. Lett.. 22, 64–69 (2009). Publisher Full Text

23. Wang, G, Liu, WB, Zhu, SN, Zheng, T: Existence results for a coupled system of nonlinear fractional 2m-point boundary value problems at resonance. Adv. Differ. Equ.. 2011, (2011) Article ID 44

24. Yang, WG: Positive solutions for a coupled system of nonlinear fractional differential equations with integral boundary conditions. Comput. Math. Appl.. 63, 288–297 (2012). Publisher Full Text

25. Zhang, YH, Bai, ZB, Feng, TT: Existence results for a coupled system of nonlinear fractional three-point boundary value problems at resonance. Comput. Math. Appl.. 61, 1032–1047 (2011). Publisher Full Text

26. Agarwal, RP, Benchohra, M, Slimani, BA: Existence results for differential equations with fractional order and impulses. Mem. Differ. Equ. Math. Phys.. 44, 1–21 (2008). Publisher Full Text

27. Bai, CZ: Solvability of multi-point boundary value problem of nonlinear impulsive fractional differential equation at resonance. Electron. J. Qual. Theory Differ. Equ.. 89, 1–19 (2011)

28. Tian, YS, Bai, ZB: Existence results for the three-point impulsive boundary value problem involving fractional differential equations. Comput. Math. Appl.. 59, 2601–2609 (2010). Publisher Full Text

29. Wang, GT, Ahmad, B, Zhang, LH: Some existence results for impulsive nonlinear fractional differential equations with mixed boundary conditions. Comput. Math. Appl.. 62, 1389–1397 (2011). Publisher Full Text

30. Mawhin, J: Topological degree methods in nonlinear boundary value problems. NSFCBMS Regional Conference Series in Mathematics, Am. Math. Soc., Providence (1979)

31. Mawhin, J: Topological degree and boundary value problems for nonlinear differential equations. In: Fitzpatrick PM, Martelli M, Mawhin J, Nussbaum R (eds.) Topological Methods for Ordinary Differential Equations, pp. 74–142. Springer, Berlin (1991)

32. Kilbas, AA, Srivastava, HM, Trujillo, JJ: Theory and Applications of Fractional Differential Equations, Elsevier, Amsterdam (2006)