We investigate the existence of solutions for a Caputo fractional difference equation boundary value problem. We use Schauder fixed point theorem to deduce the existence of solutions. The proofs are based upon the theory of discrete fractional calculus. We also provide some examples to illustrate our main results.
1. Introduction
The theory of fractional difference equations and their applications have been receiving intensive attention. In the last ten years, new research achievements kept emerging (see [1–16] and the references therein). In [1–3], the authors introduced fractional sum and difference operators, studied their behavior, and developed a complete theory governing their compositions. Abdeljawad [3] considered the initial value problem to Caputo fractional difference equation. The authors explored BVP of fractional difference equation, and they deduced the existence of one or more positive solutions in [4–7]. Abdeljawad and Baleanu [8] introduced the fractional differences and integration by parts. In [9] the authors studied the stability of discrete nonautonomous systems within the frame of the Caputo fractional difference by using Lyapunov’s direct method. They discussed the conditions for uniform stability, uniform asymptotic stability, and uniform global stability.
In [10, 11], the authors studied multiple solutions to fractional difference boundary value problems by means of Krasnosel’skii theorem and Schauder fixed point theorem. They obtained sufficient conditions of the existence of two positive solutions for the boundary value problem of fractional difference equations depending on parameters in [12]. Chen et al. [13] presented the existence of at least one positive solution for Caputo fractional boundary value problems. In [14], Kang et al. discussed existence of positive solutions for a system of Caputo fractional difference equations depending on parameters on the basis of [13].
Recently, Wu and Baleanu introduced some applications of the Caputo fractional difference to discrete chaotic maps in [15, 16]. Thus, the fractional difference equation has recently attracted increasing attention from a growing number of researchers.
Following this trend, we investigate the following boundary value problem of fractional difference equation (FBVP): (1)ΔCνyt=-ft+ν-1,yt+ν-1,yν-3=Δyb+ν=Δ2yν-3=0,where t∈{0,1,…,b+1}≔[0,b+1]N0, b⩾5, is an integer. f:[ν-2,b+ν]Nν-2×R→R is continuous and f is not identically zero, 2<ν⩽3, and ΔCνy(t) is the standard Caputo difference.
The reason for studying (1) is that, in [13], Chen et al. studied similar FBVP (1) by using the cone fixed point theorem. In this note, we consider existence of solutions to FBVP (1) by using Schauder theorem. Here, the nonnegative function f is not necessary, but f is nonnegative in [13]. We extend the domain of f.
This paper is organized as follows. Section 1 introduces the developing of fractional difference equations in a simple way. We present some necessary definitions and lemmas in Section 2. In Section 3, we prove existence of solution to FBVP (1). Finally, we provide some examples to illustrate our main results.
2. Preliminaries
We first introduce some theory about fractional sums and differences. The definitions and some basic results about fractional sums and differences can be seen in [1–7], so we omit their proof.
Definition 1 (see [3]).
For any t and ν, one defines (2)tν_=Γt+1Γt+1-νfor which the right-hand side is defined. One appeals to the convention that if t+1-ν is a pole of the Gamma function and t+1 is not a pole, then tν_=0.
Definition 2 (see [3]).
The νth fractional sum of a function f is (3)Δ-νft=1Γν∑s=at-νt-s-1ν-1_fs,for ν>0 and t∈a+ν,a+ν+1,…≔Na+ν. One also defines the νth Caputo fractional difference for ν>0 by (4)ΔCνft=Δ-n-νΔnft=1Γn-ν∑s=at-n-νt-s-1n-ν-1_Δanfs,where n-1<ν⩽n.
Lemma 3 (see [3]).
Assume that ν>0 and f is defined on domains Na, then (5)Δa+n-ν-νΔCνft=ft-∑k=0n-1ckt-ak_,where ck∈R, k=0,1,…,n-1, and n-1<ν⩽n.
Lemma 4 (see [13]).
Let 2<ν⩽3 and g:[ν-2,b+ν]Nν-2→R be given. Then the solution of FBVP,(6)ΔCνyt=-gt+ν-1,yν-3=Δyb+ν=Δ2yν-3=0,is given by (7)yt=∑s=0b+1Gt,sgs+ν-1,where Green’s function G:[ν-2,b+ν]Nν-2×[0,b+1]N0→R is defined by (8)Gt,s=1Γνν-1t-ν+3b+ν-s-1ν-2_-t-s-1ν-1_,0⩽s<t-ν+1⩽b+1,ν-1t-ν+3b+ν-s-1ν-2_,0⩽t-ν+1⩽s⩽b+1.
Remark 5.
Notice that G(ν-3,s)=0, G(t,b+2)=0. G could be extended to [ν-3,b+ν]Nν-3×[0,b+2]N0, so we only discuss (t,s)∈[ν-2,b+ν]Nν-2×[0,b+1]N0.
Lemma 6 (see [13]).
The Green function G has the following properties:
G(t,s)>0, (t,s)∈[ν-2,b+ν]Nν-2×[0,b+1]N0.
maxt∈[ν-2,b+ν]Nν-2G(t,s)=G(b+ν,s), s∈[0,b+1]N0.
Let (9)K=y∣y:ν-3,ν+bNν-3⟶R,yν-3=Δyν+b=Δ2yν-3=0.
It is clear that K is a Banach space with the norm y=maxt∈[ν-2,ν+b]Nν-2yt. Now consider the operator T defined by(10)Tyt=∑s=0b+1Gt,sfs+ν-1,ys+ν-1,where t∈[ν-2,b+ν]Nν-2, f:[ν-2,b+ν]Nν-2×R→R is continuous, and f is not identically zero. It is easy to see that y=y(t) is a solution of FBVP (1) if and only if y=y(t) is a fixed point of T.
Lemma 7 (see [17] (Schauder fixed point theorem)).
Suppose that K is a Banach space. Let B be a bounded closed convex set of K, and let T:B→B be a complete continuous operator. Then T has at least one fixed point in B.
3. Main Results
In this section, we give the main result of this paper. We will prove this result by using Schauder fixed point theorem and provide some examples to illustrate our main results.
For the sake of convenience, we write out the conditions as follows:
There exists a nonnegative function a∈C([ν-2,b+ν]Nν-2) and a constant c such that |f(t,y)|⩽a(t)+c|y|ρ, where c⩾0, 0<ρ<1.
|f(t,y)|⩽c|y|ρ, where c>0, ρ>1.
Theorem 8.
Let f:[ν-2,b+ν]Nν-2×R→R be a continuous function. Suppose that one of conditions (H1) and (H2) is satisfied. Then FBVP (1) has at least one solution.
Proof.
First, suppose that condition (H1) is satisfied. Let (11)B=yt∣yt∈K,y⩽R,t∈ν-2,b+νNν-2,where (12)R⩾max2∑s=0b+1Gb+ν,sas+ν-1,2cΓb+ν+1Γb+1Γν1ν+b+3b+11/1-ρ.Obviously B is the ball in the Banach space.
Now we prove that T:B→B. For any y∈B, then (13)Tyt⩽∑s=0b+1Gt,sas+ν-1+cys+ν-1ρ⩽∑s=0b+1Gt,sas+ν-1+∑s=0b+1cyρ·Gt,s⩽∑s=0b+1Gb+ν,sas+ν-1+cRρ∑s=0t-νGt,s+∑s=t-ν+1b+1Gt,s⩽∑s=0b+1Gb+ν,sas+ν-1+cRρΓν∑s=0t-νt-s-1ν-1_+cRρΓν∑s=0b+1ν-1·t-ν+3b+ν-s-1ν-2_⩽∑s=0b+1Gb+ν,s·as+ν-1+cRρmaxt∈ν-2,ν+bNν-21Γν·∑s=0t-νt-s-1ν-1_+cRρmaxt∈ν-2,ν+bNν-2t-ν+3Γν-1·∑s=0b+1b+ν-s-1ν-2_.As (14)1Γν∑s=0t-νt-s-1ν-1_=1Γν-1νt-sν_s=0t-ν+1=tν_Γν+1≤Γb+ν+1Γb+1Γν+1,t-ν+3Γν-1∑s=0b+1b+ν-s-1ν-2_⩽b+3Γν-1-1ν-1b+ν-sν-1_s=0b+2=b+3Γb+ν+1Γb+2Γν,it follows that (15)Tyt⩽∑s=0b+1Gb+ν,sas+ν-1+cRρΓb+ν+1Γb+1Γν+1+cRρb+3Γb+ν+1Γb+2Γν=∑s=0b+1Gb+ν,sas+ν-1+cRρΓb+ν+1Γb+1Γν1ν+b+3b+1⩽R2+R2=R.Hence, Ty⩽R. Namely, T:B→B.
Second, let condition (H2) be valid. Choose (16)0<R⩽ΓνΓb+1cΓb+ν+11ν+b+3b+1-11/ρ-1.
Repeating the course of the above, we obtain Ty⩽R. Consequently, we get T:B→B. By means of the continuity of G and f, it is easy to see that operator T is continuous. Next, we show that T is a completely continuous operator. For this, we take (17)M=maxt∈ν-2,b+νNν-2ft,yt.
For any y∈B, let t,τ∈[ν-2,b+ν]Nν-2 such that t<τ; then (18)Tyt-Tyτ=∑s=0b+1Gt,s-Gτ,s·fs+ν-1,ys+ν-1⩽M∑s=0t-νGt,s-Gτ,s+∑s=t-ν+1τ-νGt,s-Gτ,s+∑s=τ-ν+1b+1Gt,s-Gτ,s=MΓν∑s=0t-νν-1·t-ν+3b+ν-s-1ν-2_-t-s-1ν-1_-ν-1τ-ν+3b+ν-s-1ν-2_+τ-s-1ν-1_+MΓν∑s=t-ν+1τ-νν-1t-ν+3·b+ν-s-1ν-2_-ν-1τ-ν+3·b+ν-s-1ν-2_+MΓν∑s=τ-ν+1b+1ν-1·t-ν+3b+ν-s-1ν-2_-ν-1τ-ν+3·b+ν-s-1ν-2_⩽MΓν∑s=0t-νν-1·b+ν-s-1ν-2_t-τ+τ-s-1ν-1_-t-s-1ν-1_+MΓν∑s=t-ν+1τ-νν-1·b+ν-s-1ν-2_t-τ+τ-s-1ν-1_+MΓν·∑s=τ-ν+1b+1ν-1b+ν-s-1ν-2_t-τ=MΓν∑s=0b+1ν-1b+ν-s-1ν-2_t-τ+∑s=0τ-ντ-s-1ν-1_-∑s=0t-νt-s-1ν-1_=MΓν-b+ν-sν-1_s=0b+2t-τ+-1ντ-sν_s=0τ-ν+1--1νt-sν_s=0t-ν+1=MΓν-ν-2ν-1_+b+νν-1_t-τ+1ντν_-tν_=MΓνb+νν-1_t-τ+1ντν_-tν_.
Since functions tν_, t are uniformly continuous on interval [ν-2,b+ν]Nν-2, we conclude that T(B) is an equicontinuous set. Obviously, it is uniformly bounded since T(B)⊂B. Thus, we know T is completely continuous.
Consequently, it follows at once by Schauder fixed point theorem that T has a fixed point y; namely, y is a solution of (1). The theorem is proved.
Remark 9.
In this paper, f is only continuous function, without nonnegative assumptions on function f.
Remark 10.
If ρ=1 in (H1), we need condition c(Γν+b+1/Γb+1Γν)1/ν+(b+3)/(b+1)⩽1/2. At this moment, choose (19)R⩾2∑s=0b+1Gs+ν-1,sas+ν-1.
If ρ=1 in (H2), we only need condition cΓ(ν+b+1)/Γ(b+1)Γ(ν)1/ν+(b+3)/(b+1)⩽1. Then the conclusion of Theorem 8 remains true.
Example 11.
Consider the following Caputo fractional difference boundary value problem: (20)ΔC9/4yt=-sinπt+541+y1/2t+5/4y2t+5/4+50,y-34=Δy454=Δ2y-34=0,where ν=9/4, b=9, f(t,y)=sinπt·1+y1/2/y2+50. Since (21)ft,y⩽1+150y1/2,then a(t)=1, c=1/50, and ρ=1/2. At this moment, we take R=270.918. The condition of (H1) in Theorem 8 is satisfied. Applying Theorem 8, FBVP (20) has at least one solution y, and y⩽R.
Example 12.
Consider the following Caputo fractional difference boundary value problem: (22)ΔC5/2yt=-2sinπt+3/22t+323y2t+32,y-12=Δy232=Δ2y-12=0,where t∈[0,10]N0 and ν=5/2, b=9, and f(t,y)=sinπt/t+160y2. As (23)ft,y⩽1160y2,then a(t)=0, c=1/160, and ρ=2>1. At this moment, we take R=0.3525. The condition of (H2) in Theorem 8 is satisfied. Applying Theorem 8, FBVP (22) has at least one solution y, and |y|⩽R.
Conflict of Interests
The authors declare that there is no conflict of interests regarding the publication of this paper.
Acknowledgments
The authors are very grateful to the referee for her/his valuable suggestions. This paper is supported by the National Natural Science Foundation of China (Grant no. 11271235) and Shanxi Datong University Institute (2013K5).
HolmM.Sum and difference compositions in discrete fractional calculus201113315318410.4067/s0719-06462011000300009MR2895482AticiF. M.EloeP. W.Discrete fractional calculus with the nabla operator20093112AbdeljawadT.On Riemann and Caputo fractional differences20116231602161110.1016/j.camwa.2011.03.036MR28247472-s2.0-79960975255GoodrichC. S.Some new existence results for fractional difference equations201131-214516210.1504/ijdsde.2011.038499MR27970462-s2.0-79951536621GoodrichC. S.On positive solutions to nonlocal fractional and integer-order difference equations20115112213210.2298/AADM110111001GMR2809040ZBL1289.390082-s2.0-79953326334GoodrichC. S.Existence of a positive solution to a system of discrete fractional boundary value problems201121794740475310.1016/j.amc.2010.11.029MR2745153ZBL1215.390032-s2.0-78650170131GoodrichC. S.On a discrete fractional three-point boundary value problem201218339741510.1080/10236198.2010.503240MR29018292-s2.0-79953294826AbdeljawadT.BaleanuD.Fractional differences and integration by parts2011133574582MR27524282-s2.0-79960258288JaradF.AbdeljawadT.BaleanuD.On the stability of some discrete fractional nonautonomous systems20122012947658110.1155/2012/476581ChenH. Q.CuiY. Q.ZhaoX. L.Multiple solutions to fractional difference boundary value problems20142014687938010.1155/2014/879380MR3206824KangS. G.LiY.ChenH. Q.Positive solutions to boundary value problems of fractional difference equation with nonlocal conditions20142014, article 71210.1186/1687-1847-2014-7MR33504302-s2.0-84899505475KangS.ZhaoX.ChenH.Positive solutions for boundary value problems of fractional difference equations depending on parameters20132013, article 376MR3337228ChenH. Q.JinZ.KangS. G.Existence of positive solutions for Caputo fractional difference equation20154411210.1186/s13662-015-0367-0MR33119562-s2.0-84922486261KangS. G.ChenH. Q.GuoJ. M.Existence of positive solutions for a system of Caputo fractional difference equations depending on parameters20152015, article 13810.1186/s13662-015-0466-yMR3343136WuG.-C.BaleanuD.Discrete fractional logistic map and its chaos2014751-228328710.1007/s11071-013-1065-7MR31448522-s2.0-84891517157WuG.-C.BaleanuD.Chaos synchronization of the discrete fractional logistic map2014102969910.1016/j.sigpro.2014.02.0222-s2.0-84897525673AgarwalR. P.MeehanM.O'ReganD.2001Cambridge, UKCambridge University Press10.1017/cbo9780511543005MR1825411