We study the existence of positive solutions for the system of nonlinear semipositone boundary value problems with Riemann-Liouville fractional derivatives D0+αD0+αu=f1t,u,u′,v,v′,0<t<1, D0+αD0+αv=f2(t,u,u′,v,v′),0<t<1, u0=u′0=u′(1)=D0+αu(0)=D0+α+1u(0)=D0+α+1u(1)=0, and v(0)=v′(0)=v′(1)=D0+αv(0)=D0+α+1v(0)=D0+α+1v(1)=0, where α∈(2,3] is a real number and D0+α is the standard Riemann-Liouville fractional derivative of order α. Under some appropriate conditions for semipositone nonlinearities, we use the fixed point index to establish two existence theorems. Moreover, nonnegative concave and convex functions are used to depict the coupling behavior of our nonlinearities.
National Natural Science Foundation of China11601048Natural Science Foundation of Chongqingcstc2016jcyjA0181Science and Technology Research Program of Chongqing Municipal Education CommissionKJ1703050Natural Science Foundation of Chongqing Normal University16XYY2415XLB0111. Introduction
In this paper, we investigate the existence of positive solutions for the system of nonlinear semipositone boundary value problems with Riemann-Liouville fractional derivatives(1)D0+αD0+αu=f1t,u,u′,v,v′,0<t<1,D0+αD0+αv=f2t,u,u′,v,v′,0<t<1,u0=u′0=u′1=D0+αu0=D0+α+1u0=D0+α+1u1=0,v0=v′0=v′1=D0+αv0=D0+α+1v0=D0+α+1v1=0,where α∈(2,3] is a real number and D0+α is the standard Riemann-Liouville fractional derivative of order α. The nonlinear terms fi∈C([0,1]×R+4,R)(R+=[0,+∞), R=-∞,+∞) are bounded below; that is, fi(i=1,2) satisfy the following.
(H1) there exists a real number M≥0, such that fi(t,x1,x2,x3,x4)+M≥0,∀t∈[0,1], xj∈R+,i=1,2,j=1,2,3,4.
Existence and multiplicity of solutions for fractional differential equations are widely studied in the literature; see [1–14] and the references therein. For example, in [1], the authors used the Guo-Krasnosel’skii fixed point theorem to investigate the existence of positive solutions for the singular fractional differential system(2)-D0+αut=λf1t,u,v,t∈0,1,-D0+αvt=λf2t,u,v,t∈0,1,u0=u′0=0,v0=v′0=0,u1=avξ,v1=buη,where fi(i=1,2) satisfy(3)fit,u,vu+v=0, or ∞,as u+v⟶+∞, uniformly for t∈0,1 or a subinterval.
Condition (3) is used to study various types of fractional systems (see [1–12] and the references therein).
In this paper we use the fixed point index to study the existence of positive solutions for the system of nonlinear semipositone fractional boundary value problem (1). Under some appropriate conditions for fi(i=1,2), we use the fixed point index to obtain our results. Moreover, nonnegative concave and convex functions are used to depict the coupling behavior of our nonlinearities (see [13–15]), which depend on the unknown functions u,v and their derivatives u′, v′.
2. PreliminaryDefinition 1 (see [16, 17]).
The Riemann-Liouville fractional derivative of order α>0 of a continuous function f:(0,+∞)→R is given by(4)D0+αft=1Γn-αddtn∫0tt-sn-α-1fsds,where n=[α]+1 with [α] denoting the integer part of a number α, provided that the right hand side is pointwise defined on (0,+∞).
We first study the Green functions of problem (1). Let (5)G1t,s≔1Γαtα-11-sα-2-t-sα-1,0≤s≤t≤1,tα-11-sα-2,0≤t≤s≤1.
Then we have (6)G2t,s≔∂∂tG1t,s=α-1Γαtα-21-sα-2-t-sα-2,0≤s≤t≤1,tα-21-sα-2,0≤t≤s≤1.
Lemma 2.
Let fi(i=1,2) be as in (1). Then (1) is equivalent to(7)D0+αx=-f1t,∫01G1t,sxsds,∫01G2t,sxsds,∫01G1t,sysds,∫01G2t,sysds,D0+αy=-f2t,∫01G1t,sxsds,∫01G2t,sxsds,∫01G1t,sysds,∫01G2t,sysds,x0=x′0=x′1=0,y0=y′0=y′1=0,which takes the form(8)xt=∫01G1t,sf1s,∫01G1s,τxτdτ,∫01G2s,τxτdτ,∫01G1s,τyτdτ,∫01G2s,τyτdτds,yt=∫01G1t,sf2s,∫01G1s,τxτdτ,∫01G2s,τxτdτ,∫01G1s,τyτdτ,∫01G2s,τyτdτds.Let D0+αu=-x, D0+αv=-y. Then an argument similar to that in [18, Lemma 2.7] and [19, Lemma 3] establishes the result (we omit the standard details).
Lemma 3 ([19, Lemma 4]).
The functions Gi(t,s)∈C([0,1]×[0,1],R+)(i=1,2). Moreover, the following inequalities are satisfied:(9)tα-1s1-sα-2≤ΓαG1t,s≤s1-sα-2∀t,s∈0,1,(10)α-1α-2tα-21-ts1-sα-2≤ΓαG2t,s≤α-1tα-3s1-sα-2∀t,s∈0,1.
Lemma 4 ([19, Lemma 5]).
Let φ(t)=t(1-t)α-2 for all t∈[0,1]. Let(11)k1≔αΓα-1Γ2α≤k2≔1αα-1Γα,k3≔α-1α-2ΓαΓ2α≤k4≔Γα-1Γ2α-2.Then(12)k2i-1φs≤∫01Git,sφtdt≤k2iφs,i=1,2,∀s∈0,1.
Lemma 5.
(i) If (x∗(t),y∗(t)) is a positive solution of (7), then (x∗(t)+w(t),y∗(t)+w(t)) is a positive solution of the following differential equation:(13)D0+αx=-F1t,∫01G1t,sxs-wsds,∫01G2t,sxs-wsds,∫01G1t,sys-wsds,∫01G2t,sys-wsds,D0+αy=-F2t,∫01G1t,sxs-wsds,∫01G2t,sxs-wsds,∫01G1t,sys-wsds,∫01G2t,sys-wsds,x0=x′0=x′1=0,y0=y′0=y′1=0,where (14)Fit,x1,x2,x3,x4≔f¯it,x1,x2,x3,x4,t∈0,1,x1,x2,x3,x4≥0,f¯it,0,0,0,0,t∈0,1,x1,x2,x3,x4<0,and f¯i(t,x1,x2,x3,x4)=fi(t,x1,x2,x3,x4)+M,f¯i:[0,1]×R+4→R+ are continuous, and(15)wt≔M∫01G1t,sds=MΓαtα-1α-1-tαα∀t∈0,1.
(ii) If xt,yt is a solution of (13) and x(t)≥w(t), y(t)≥w(t),t∈[0,1], then (x∗(t),y∗(t))=(x(t)-w(t),y(t)-w(t)) is a positive solution of (7).
Proof.
If (x∗(t),y∗(t)) is a positive solution of (7) then (note w(t)=M∫01G1t,sds) we obtain x∗(0)+w(0)=x∗′(0)+w′0=x∗′1+w′1=0 and (16)D0+αx∗t+wt+F1t,∫01G1t,sx∗sds,∫01G2t,sx∗sds,∫01G1t,sy∗sds,∫01G2t,sy∗sds=D0+αx∗t+D0+αwt+f1t,∫01G1t,sx∗sds,∫01G2t,sx∗sds,∫01G1t,sy∗sds,∫01G2t,s×y∗sds+M=D0+αwt+M=D0+αM∫01G1t,sds+M=-M+M=0.Similarly, we have (17)D0+αy∗t+wt+F2t,∫01G1t,sx∗sds,∫01G2t,sx∗sds,∫01G1t,sy∗sds,∫01G2t,sy∗sds=-M+M=0;that is, x∗t+wt,y∗t+wt satisfies (13). Therefore, (i) holds. Similarly, it is easy to prove (ii). This completes the proof.
From Lemma 5, to obtain a positive solution of (7), we only need to find solutions xt, yt of (13) satisfying (t)≥w(t), y(t)≥w(t), t∈[0,1]. If x(t), y(t) are solutions of (13), then x(t), y(t)satisfy(18)xt=∫01G1t,sF1s,∫01G1s,τxτ-wτdτ,∫01G2s,τxτ-wτdτ,∫01G1s,τyτ-wτdτ,∫01G2s,τyτ-wτdτds,yt=∫01G1t,sF2s,∫01G1s,τxτ-wτdτ,∫01G2s,τxτ-wτdτ,∫01G1s,τyτ-wτdτ,∫01G2s,τyτ-wτdτds.
Let E≔C[0,1], x≔maxt∈0,1xt, P≔{x∈E:xt≥0,t∈[0,1]}. Then (E,·) is a real Banach space, and P is a cone on E. We denote Bρ≔x∈E:x<ρ for ρ>0. Now, note that u, v solve (1) if and only if x≔-D0+αu, y≔-D0+αv are fixed points of operator (19)Aix,yt≔∫01G1t,sFis,∫01G1s,τxτ-wτdτ,∫01G2s,τxτ-wτdτ,∫01G1s,τ×yτ-wτdτ,∫01G2s,τyτ-wτdτds,Ax,yt=A1,A2x,ytfor x,y∈E.Therefore, if (x,y) is a positive fixed for A with x(t)≥w(t), y(t)≥w(t) for t∈[0,1], then (x∗,y∗)=(x-w,y-w) is a positive solution for (1). Moreover, from the continuity of Gi and Fi(i=1,2), we know that Ai:P×P→P, A:P×P→P×P are continuous and completely continuous operators.
Lemma 6.
Let P0≔{x∈P:x(t)≥tα-1x, t∈[0,1]}. Then P0 is a cone in E and A(P×P)⊂P02.
Proof.
From (9) for t∈[0,1] we have (20)A1x,yt=∫01G1t,sF1s,∫01G1s,τxτ-wτdτ,∫01G2s,τxτ-wτdτ,∫01G1s,τ×yτ-wτdτ,∫01G2s,τyτ-wτdτds≤1Γα∫01s1-sα-2F1s,∫01G1s,τxτ-wτdτ,∫01G2s,τxτ-wτdτ,∫01G1s,τ×yτ-wτdτ,∫01G2s,τyτ-wτdτds.Also from (9) and the above inequality, for every (x,y)∈P×P, we obtain (21)A1x,yt=∫01G1t,sF1s,∫01G1s,τxτ-wτdτ,∫01G2s,τxτ-wτdτ,∫01G1s,τ×yτ-wτdτ,∫01G2s,τyτ-wτdτds≥tα-1Γα∫01s1-sα-2F1s,∫01G1s,τxτ-wτdτ,∫01G2s,τxτ-wτdτ,∫01G1s,τ×yτ-wτdτ,∫01G2s,τyτ-wτdτds≥tα-1A1x,yfor all t∈[0,1]. Similarly, A2(x,y)(t)≥tα-1A2(x,y). Therefore A(P×P)⊂P02. This completes the proof.
To obtain a positive solution of (1), we seek a positive fixed point (x∗,y∗) of A with x∗≥w, y∗≥w (note mean that x∗t=A1x∗,y∗t, y∗(t)=A2(x∗,y∗)(t) for t∈[0,1]). From Lemma 6, we have x∗,y∗∈P0. For x∗∈P0 we have (22)x∗t-wt=x∗t-M∫01G1t,sds=x∗t-MΓαtα-1α-1-tαα=x∗t-Mtα-1Γα1α-1-tα≥x∗t-x∗tx∗Mα-1Γα.As a result, x∗(t)≥w(t) for t∈[0,1] if x∗≥M/α-1Γα≔k5. Similarly, if y∗≥k5, we have y∗(t)≥w(t), for t∈[0,1].
Lemma 7 (see [20]).
Let Ω⊂E be a bounded open set and A:Ω¯∩P→P a continuous or completely continuous operator. If there exists u0∈P∖{0} such that u-Au≠μu0 for all μ≥0 and u∈∂Ω∩P, then i(A,Ω∩P,P)=0, where i denotes the fixed point index on P.
Lemma 8 (see [20]).
Let Ω⊂E be a bounded open set with 0∈Ω. Suppose A:Ω¯∩P→P is a continuous or completely continuous operator. If u≠μAu for all u∈∂Ω∩P and 0≤μ≤1, then i(A,Ω∩P,P)=1.
3. Main Results
Let K≔α/Γα≥maxt,s∈[0,1]G1t,s+G2t,s, In the sequel, we use c1,c2,… and d1,d2,… to stand for different positive constants. Now, we list our assumptions on Fi(i=1,2).
(H2) There exist h,g∈C(R+,R+) such that
h is concave and strictly increasing on R+ (and limx→+∞h(x)=+∞);
there exist c1>0, d1>1/k12k1+k32, for all t,x1,x2,y1,y2∈[0,1]×R+4 such that (23)F1t,x1,x2,y1,y2≥d1hy1+y2-c1,F2t,x1,x2,y1,y2≥gx1+x2-c1;
h(K2g(x))≥K2x-c1 for x∈R+.
(H3) For all (t,x1,x2,y1,y2)∈[0,1]×[0,k5]4, there is a constant M1∈(0,k5k2-1) such that (24)Fit,x1,x2,y1,y2≤M1,i=1,2.
(H4) There exist β, γ∈C(R+,R+) such that
β is convex and strictly increasing on R+ (and limx→+∞β(x)=+∞);
for all (t,x1,x2,y1,y2)∈[0,1]×R+4, (25)F1t,x1,x2,y1,y2≤βy1+y2,F2t,x1,x2,y1,y2≤γx1+x2;
there exist d2>0 such that β(K2γ(x))≤K2x+d2, for x∈R+.
(H5) There exist Q:[0,1]→R, θ∈(0,1], t0∈[θ,1], for all (t,x1,x2,y1,y2)∈[θ,1]×0,k54, such that (26)fit,x1,x2,y1,y2+M≥Qt,i=1,2,where (27)∫θ1G1t0,sQsds>Mα-1Γα.
Theorem 9.
Suppose that (H1)–(H3) hold. Then (1) has at least one positive solution.
Proof.
We first prove that there exists R>k5 such that(28)x,y≠Ax,y+λϕ,ϕ,∀x,y∈∂BR∩P×P,λ≥0,where ϕ∈P0 is a given function. Suppose there exist (x,y)∈∂BR∩(P×P), λ≥0 with (x,y)=A(x,y)+λ(ϕ,ϕ), then x(t)≥A1(x,y)(t), y(t)≥A2(x,y)(t) for t∈[0,1]. From (i), (ii) of (H2) we have(29)xt≥∫01G1t,sF1s,∫01G1s,τxτ-wτdτ,∫01G2s,τxτ-wτdτ,∫01G1s,τyτ-wτdτ,∫01G2s,τyτ-wτdτds≥∫01G1t,sd1h∫01G1s,τ+G2s,τyτ-wτdτ-c1ds≥d1∫01G1t,sh∫01G1s,τ+G2s,τyτdτds-d1∫01G1t,sh∫01G1s,τ+G2s,τwτdτds-c1∫01G1t,sds≥d1∫01G1t,sh∫01G1s,τ+G2s,τyτdτds-c2=d1∫01G1t,sh∫01G1s,τ+G2s,τKKyτdτds-c2≥d1∫01G1t,s∫01G1s,τ+G2s,τKhKyτdτds-c2.From (ii) of (H2) we have(30)Kyt≥K∫01G1t,sF2s,∫01G1s,τxτ-wτdτ,∫01G2s,τxτ-wτdτ,∫01G1s,τ×yτ-wτdτ,∫01G2s,τyτ-wτdτds≥K∫01G1t,sg∫01G1s,τ+G2s,τxτ-wτdτds-Kc1∫01G1t,sds=K∫01G1t,sg∫01G1s,τ+G2s,τxτ-wτdτds-c3.From (30) and (i) of (H2) we obtain(31)hKyt+c3≥hK∫01G1t,sg∫01G1s,τ+G2s,τxτ-wτdτds.This together with (iii) of (H2) yields(32)hKyt≥hKyt+c3-hc3≥hK∫01G1t,sg∫01G1s,τ+G2s,τxτ-wτdτds-hc3≥∫01hG1t,sKK2g∫01G1s,τ+G2s,τxτ-wτdτds-hc3≥∫01G1t,sKhK2g∫01G1s,τ+G2s,τxτ-wτdτds-hc3≥∫01KG1t,s∫01G1s,τ+G2s,τxτ-wτdτds-c4.Then (32) is substituted into (29) and we obtain(33)xt≥d1∫01G1t,s∫01G1s,τ+G2s,τK∫01KG1τ,r∫01G1r,l+G2r,l×xl-wldldr-c4dτds-c2≥d1∫01G1t,s∫01G1s,τ+G2s,τ∫01G1τ,r∫01G1r,l+G2r,l×xl-wldldrdτds-c5≥d1∫01G1t,s∫01G1s,τ+G2s,τ∫01G1τ,r∫01G1r,l+G2r,l×xldldrdτds-c6.Multiplying by φ(t) for (33) and integrating over [0,1], we have(34)∫01φtxtdt≥d1k12k1+k32∫01φtxtdt-c6k2Γα,using the fact that(35)k1φs≤∫01G1t,sφtdt≤k2φs,k1+k3φs≤∫01G1t,s+G2t,sφtdt≤k2+k4φs,which can be derived from (12) in Lemma 4. From (34) we obtain(36)∫01φtxtdt≤c6k2Γαd1k12k1+k32-1.Note that x∈P0 (note that x=A1(x,y)+λϕ and A1(x,y)∈P0 from Lemma 6 and ϕ∈P0) and we have(37)∫01φttα-1xdt≤∫01φtxtdt≤c6k2Γαd1k12k1+k32-1,x≤c6k2d1k13k1+k32-k1.From (29) we have(38)d1∫01G1t,s∫01G1s,τ+G2s,τKhKyτdτds≤xt+c2≤xt+c2≤c6k2d1k13k1+k32-k1+c2.Multiplying by φ(t) and integrating over [0,1] we obtain (39)∫01xt+c2φtdt≥∫01φtd1∫01G1t,s∫G1s,τ+G2s,τKhKyτdτdsdt≥d1∫01k1φs∫01G1s,τ+G2s,τKhKyτdsdt≥d1k1k1+k3K∫01φthKytdt.Consequently, we have (40)∫01φthKytdt≤Kd1k1k1+k3∫01c6k2d1k13k1+k32-k1+c2φtdt=KΓαk2d1k1k1+k3c6k2d1k13k1+k32-k1+c2≔N1.Note that we may assume y(t)≢0 for t∈[0,1]. Then y>0 and h(Ky)>0. For y∈P0, we have (41)k1ΓαKy≤∫01Kφtytdt=KyhKy∫01φtKytKyhKydt≤KyhKy∫01φthKytdt≤KyhKyN1.Hence, h(Ky)≤N1/k1Γα. Note limz→+∞h(z)=+∞, and thus there exists N2≥0 such that Ky≤N2. Therefore if (x,y)∈∂BR∩(P×P), λ≥0 with (x,y)=A(x,y)+λ(ϕ,ϕ) then x≤c6k2/d1k13k1+k32-k1 and y≤N2/K. Thus if we take R>maxk5,c6k2/d1k13k1+k32-k1,N2/K then (28) is true. Lemma 7 implies(42)iA,BR∩P×P,P×P=0.Let x,y∈∂Bk5∩P. From (H3) we have (43)A1x,yt=∫01G1t,sF1s,∫01G1s,τxτ-wτdτ,∫01G2s,τxτ-wτdτ,∫01G1s,τ×yτ-wτdτ,∫01G2s,τyτ-wτdτds≤∫01G1t,sM1ds≤∫01s1-sα-2ΓαM1ds<Mα-1Γα=k5=x,so A1(x,y)<x. Similarly A2(x,y)<y. Hence A(x,y)<(x,y) for x,y∈∂Bk5∩P. Thus (44)x,y≠λAx,y,∀x,y∈∂Bk5∩P×P,λ∈0,1.It follows from Lemma 8 that(45)iA,Bk5∩P×P,P×P=1.From (42) and (45) we have (46)iA,BR∖B¯k5∩P×P,P×P=0-1=-1.Therefore the operator A has at least one fixed point in (BR∖B¯k5)∩(P×P) and so (1) has at least a positive solution. This completes the proof.
Theorem 10.
Suppose that (H1), (H4), and (H5) hold. Then (1) has at least one positive solution.
Proof.
We first show that there exists R>k5 such that(47)x,y≠λAx,y,∀x,y∈∂BR∩P×P,λ∈0,1.Suppose there exist (x,y)∈∂BR∩(P×P), λ∈[0,1] with (x,y)=λA(x,y), then x(t)≤A1(x,y)(t), y(t)≤A2(x,y)(t) for t∈[0,1]. From (i), (ii) of (H4), we have(48)xt≤∫01G1t,sF1s,∫01G1s,τxτ-wτdτ,∫01G2s,τxτ-wτdτ,∫01G1s,τ×yτ-wτdτ,∫01G2s,τyτ-wτdτds≤∫01G1t,sβ∫01G1s,τ+G2s,τyτ-wτdτds=∫01G1t,sβ∫01G1s,τ+G2s,τKKyτ-wτdτds≤∫01G1t,s∫01G1s,τ+G2s,τKβKyτdτds.From (ii) of (H4), we get(49)Kyt≤K∫01G1t,sF2s,∫01G1s,τxτ-wτdτ,∫01G2s,τxτ-wτdτ,∫01G1s,τ×yτ-wτdτ,∫01G2s,τyτ-wτdτds≤K∫01G1t,sγ∫01G1s,τ+G2s,τxτ-wτdτds.From (49) and (i), (iii) of (H4) we have(50)βKyt≤βK∫01G1t,sγ∫01G1s,τ+G2s,τxτ-wτdτds≤∫01βG1t,sKK2γ∫01G1s,τ+G2s,τxτ-wτdτds≤∫01G1t,sKβK2γ∫01G1s,τ+G2s,τxτ-wτdτds≤K∫01G1t,s∫01G1s,τ+G2s,τxτ-wτdτ+d2ds≤K∫01G1t,s∫01G1s,τ+G2s,τxτ-wτdτds+d3.Then substitute (50) into (48) and we obtain(51)xt≤∫01G1t,s∫01G1s,τ+G2s,τKβKyτdτds≤∫01G1t,s∫01G1s,τ+G2s,τKK∫01G1τ,r∫01G1r,l+G2r,l×xl-wldldr+d3dτds≤∫01G1t,s∫01G1s,τ+G2s,τ∫01G1τ,r∫01G1r,l+G2r,l×xl-wldldrdτds+d4.Multiplying by φ(t) for (51) and integrating over [0,1], from (35), we obtain(52)∫01φtxtdt≤k22k2+k42∫01φtxtdt+Γαk2d4.Consequently, we have (53)∫01φtxtdt≤Γαk2d41-k22k2+k42.Note that x∈P0 (note x=λA1(x,y) and A1(x,y)∈P0) and we have (54)∫01φttα-1xtdt≤∫01φtxtdt≤Γαk2d41-k22k2+k42,x≤k2d4k1-k1k22k2+k42.From (50) and Lemma 3 we have (55)βKyt≤K∫01G1t,s∫01G1s,τ+G2s,τxτ-wτdτds+d3≤Kx∫01G1t,s∫01G1s,τ+G2s,τdτds+d3≤Kx∫01s1-sα-2∫011Γατ1-τα-2+α-1sα-3τ1-τα-2dτds+d3≤Kxα2α-12Γα2+π23-2αΓα-1Kxα-1Γα-1/2α2Γα2+d3,so β(Ky(t)) is bounded. Note limz→+∞β(z)=+∞, and thus there exists N3≥0 such that Ky≤N3. Therefore if (x,y)∈∂BR∩(P×P),λ∈[0,1] with (x,y)=λA(x,y) then x≤k2d4/k1-k1k22k2+k42 and y≤N3/K. Thus if we take R>maxk5,k2d4/k1-k1k22k2+k42,N3/K then (47) is true. Lemma 8 implies(56)iA,BR∩P×P,P×P=1.Let Bk5≔{x∈P:x<k5} and consider x∈∂Bk5∩P. It follows from (H5) that (57)A1x,yt0=∫01G1t0,sF1s,∫01G1s,τxτ-wτdτ,∫01G2s,τxτ-wτdτ,∫01G1s,τyτ-wτdτ,∫01G2s,τyτ-wτdτds≥∫θ1G1t0,sF1s,∫01G1s,τxτ-wτdτ,∫01G2s,τxτ-wτdτ,∫01G1s,τyτ-wτdτ,∫01G2s,τyτ-wτdτds=∫θ1G1t0,sf1s,∫01G1s,τxτ-wτdτ,∫01G2s,τxτ-wτdτ,∫01G1s,τyτ-wτdτ,∫01G2s,τyτ-wτdτ+Mds≥∫θ1G1t0,sQsds>Mα-1Γα,and, hence, A1(x,y)≥A1x,yt0>x. Similarly A2(x,y)≥A2(x,y)(t0)>y for y∈∂Bk5∩P. Therefore A(x,y)>(x,y) for all x,y∈∂Bk5∩(P×P). Thus (58)x,y≠Ax,y+λϕ,ϕ,∀x,y∈∂Bk5∩P×P,ϕ∈P,λ∈0,1.It follows from Lemma 7 that(59)iA,Bk5∩P×P,P×P=0.From (56) and (59), we have (60)iA,BR∖B¯k5∩P×P,P×P=1-0=1.
Therefore the operator A has at least one fixed point on (BR∖B¯k5)∩(P×P) and so (1) has at least one positive solution, which completes the proof.
Conflicts of Interest
The authors declare that they have no conflicts of interest.
Acknowledgments
The research is supported by the National Natural Science Foundation of China (Grant no. 11601048), Natural Science Foundation of Chongqing (Grant no. cstc2016jcyjA0181), the Science and Technology Research Program of Chongqing Municipal Education Commission (Grant no. KJ1703050), and Natural Science Foundation of Chongqing Normal University (Grant nos. 16XYY24, 15XLB011).
JiangJ.LiuL.WuY.Positive solutions to singular fractional differential system with coupled boundary conditions201318113061307410.1016/j.cnsns.2013.04.009MR3072528Zbl1329.340102-s2.0-84878954981DahalR.DuncanD.GoodrichC.Systems of semipositone discrete fractional boundary value problems201420347349110.1080/10236198.2013.856073MR3173559HendersonJ.LucaR.Existence of positive solutions for a system of semipositone fractional boundary value problems20162212810.14232/ejqtde.2016.1.22HaoX.WangH.LiuL.CuiY.Positive solutions for a system of nonlinear fractional nonlocal boundary value problems with parameters and p-Laplacian operator20172017118210.1186/s13661-017-0915-52-s2.0-85038125725LucaR.TudoracheA.Positive solutions to a system of semipositone fractional boundary value problems20142014117910.1186/1687-1847-2014-1792-s2.0-84924551185QiT.LiuY.CuiY.Existence of solutions for a class of coupled fractional differential systems with nonlocal boundary conditions20172017670386010.1155/2017/67038602-s2.0-85025105335QiT.LiuY.ZouY.Existence result for a class of coupled fractional differential systems with integral boundary value conditions201710074034404510.22436/jnsa.010.07.52WangY.LiuL.ZhangX.WuY.Positive solutions of an abstract fractional semipositone differential system model for bioprocesses of HIV infection201525831232410.1016/j.amc.2015.01.080MR3323070WangY.LiuL.WuY.Positive solutions for a class of higher-order singular semipositone fractional differential systems with coupled integral boundary conditions and parameters2014201426810.1186/1687-1847-2014-268MR3359149YangW.Positive solutions for nonlinear semipositone fractional q-difference system with coupled integral boundary conditions201424470272510.1016/j.amc.2014.07.039MR3250614YuanC.Two positive solutions for (n-1, 1)-type semipositone integral boundary value problems for coupled systems of nonlinear fractional differential equations201217293094210.1016/j.cnsns.2011.06.008Zbl1248.35225ZhangY.BaiZ.FengT.Existence results for a coupled system of nonlinear fractional three-point boundary value problems at resonance20116141032104710.1016/j.camwa.2010.12.053MR2770508Zbl1217.340312-s2.0-79651475091ChengW.XuJ.CuiY.Positive solutions for a system of nonlinear semipositone fractional q-difference equations with q-integral boundary conditions201710084430444010.22436/jnsa.010.08.35ZhangK.XuJ.O'ReganD.Positive solutions for a coupled system of nonlinear fractional differential equations2015388166216722-s2.0-8492810862910.1002/mma.3178Zbl1322.34015DingY.XuJ.WeiZ.Positive solutions for a system of discrete boundary value problem20153831207122110.1007/s40840-014-0063-32-s2.0-84930618184KilbasA. A.SrivastavaH. M.TrujilloJ. J.Preface200620471010.1016/S0304-0208(06)80001-02-s2.0-77956684069PodlubnyI.Fractional differential equations1999El-ShahedM.Positive solutions for boundary value problem of nonlinear fractional differential equation20072007810.1155/2007/1036810368MR2353784Zbl1149.260122-s2.0-36248933572XuJ.WeiZ.DingY.Positive solutions for a boundary-value problem with Riemann-Liouville fractional derivative*20125244624762-s2.0-8487094936510.1007/s10986-012-9187-zZbl1266.34017GuoD.LakshmikanthamV.1988Boston, MA, USAAcademic Press