Shah and Khan SpringerPlus (2016)5:1116 DOI 10.1186/s40064-016-2656-9
Open Access
RESEARCH
Multiple positive solutions to a coupled systems of nonlinear fractional differential equations Kamal Shah* and Rahmat Ali Khan *Correspondence:
[email protected] Department of Mathematics, University of Malakand, Dir(L), Chakdara, Khyber Pakhtunkhwa, Pakistan
Abstract In this article, we study existence, uniqueness and nonexistence of positive solution to a highly nonlinear coupled system of fractional order differential equations. Necessary and sufficient conditions for the existence and uniqueness of positive solution are developed by using Perov’s fixed point theorem for the considered problem. Further, we also established sufficient conditions for existence of multiplicity results for positive solutions. Also, we developed some conditions under which the considered coupled system of fractional order differential equations has no positive solution. Appropriate examples are also provided which demonstrate our results. Keywords: Fractional differential equations, Coupled system, Boundary condition, Generalized metric space, Contractions mapping Mathematics Subject Classification: 34A08, 34B37, 35R11
Background In recent years it is mainly proved that fractional differential equations are the best tools in the mathematical modeling of many phenomena in various field of physics, electrochemistry, viscoelasticity, control theory, image and signal processing etc, (see Hilfer 2000; Kilbas et al. 2006; Lakshmikantham et al. 2009; Podlubny 1999; Rossikhin and Shitikova 1997). Also, fractional differential equations are used in the modeling of various phenomena such as nonlinear oscillation of earthquake, Nutting’s law of charge transport in amorphous semi conductors, fluid dynamics, traffic model and non-Markorian diffusion process with memory. For the aforesaid application, see Mainardi (1995), Metzler and Klafer (2000), Scher and Montroll (1975), Wang et al. (2011). Moreover, considerable attention has been given to study mathematical epidemiology in term of fractional order models, this is due to the fact that fractional order models of epidemic disease are more realistic and provide great information as compared to classical order models. In last few years fractals and Chaos are also studied by using the tools of fractional differential equations. Besides from the above discussion, fractional order differential equations have also many applications in the fields of aerodynamics, fluid dynamics, physical chemistry, economics, polymerrheology, regular variation in thermodynamics, biophysics, blood flow phenomena, etc. Due to the important applications and uses of fractional © 2016 The Author(s). This article is distributed under the terms of the Creative Commons Attribution 4.0 International License (http://creativecommons.org/licenses/by/4.0/), which permits unrestricted use, distribution, and reproduction in any medium, provided you give appropriate credit to the original author(s) and the source, provide a link to the Creative Commons license, and indicate if changes were made.
Shah and Khan SpringerPlus (2016)5:1116
Page 2 of 20
differential equations, researchers of various field of mathematics, engineering, physics and computer science etc, gave much attention to study fractional differential equations. Another important application of fractional calculus has been found in condensed matter physics, where the fractional quantum Hall effect is one of the most attracting phenomena. For the present or planned technologies fractional order models are used by the implementation of an optical lattice setup. For detail, see Nielsen et al. (2013), Hagerstrom et al. (2012) and the reference there in. In networking systems it has been proved that several real networks in their degree of distribution obey a power-law. The presence of highly connected nodes in a scale-free network causes well known robustness against random failures. But on the other hand suffers from vulnerability to malicious attacks at their highly connected nodes. Fractional order models provide more realistic and accurate approach as compared to classical order models to study the afore said phenonmena, see Shang (2014). In last few years, the study of existence and uniqueness of solutions to boundary value problems for fractional order differential equations got much attention from many researchers and a number of research articles are available in the literature, we refer few of them in Deren (2015), Li et al. (2010), Khan and Shah (2015) and the reference therein. The iterative solutions to boundary and initial values problems of nonlinear fractional order differential equations were also studied by some authors (see Ahmad and Nieto 2008; Cui and Zou 2014; Shah et al. 2016 and the references therein). Moreover, existence and multiplicity of positive solutions to nonlinear boundary values problem of fractional order differential equations have been studied by many authors by using classical fixed point theorems, for example see Ahmad and Nieto (2009b), Bai (2008), Cui et al. (2012), Xu et al. (2009). Bai and Lü (2005), have studied the existence of multiple solutions for the following boundary value problems α D u(t) + f (t, u(t)) = 0, t ∈ (0, 1), 1 < α ≤ 2, u(0) = u(1) = 0. where Dα is the standard Riemann–Liouville fractional derivative of order α, and f : [0, 1] × [0, ∞) → [0, ∞) is continuous function. By means of classical fixed point theorems sufficient conditions were obtained for multiplicity of solutions. Recently, Goodrich (2010), considered the following class of nonlinear fractional differential equations with the given boundary conditions for multiplicity of positive solutions as
Dα u(t) + f (t, u(t)) = 0, (i)
u (0) = 0,
t ∈ (0, 1), n − 1 < α ≤ n,
0 ≤ i ≤ n − 2, and Dδ u(1) = 0,
2 ≤ δ ≤ n − 2.
where n > 3 and f : [0, 1] × [0, ∞) → [0, ∞) is continuous function. Many problems in applied sciences can be modeled as coupled system of differential equations with different type of boundary conditions. The coupled system of fractional order differential equations have many application in computer networking, see Li et al. (2015a, b), Suo et al. (2013). Boundary values problems for coupled systems with ordinary derivatives are well studied, however, coupled systems with fractional derivatives have attracted the attention quite recently. Most of the biological, physical, computer network model and chemical models etc, are in the form of coupled system (see Anastassiou et al. 2011; Chasnov 2009; Lia et al. 2015). Due to these important applications and uses of coupled
Shah and Khan SpringerPlus (2016)5:1116
Page 3 of 20
systems of fractional order differential equations, considerable attention was given to study coupled system for the existence, uniqueness and multiplicity of positive solutions, for detail we refer Miller and Ross (1993), Shah and Khan (2015), Su (2009), Shah et al. (2015) and the references therein. As Bai and Feng (2004), established sufficient conditions for existence of positive solution to a coupled system of fractional differential equations as given by the
Dα u(t) = f (t, v(t)); β
D v(t) = g(t, u(t));
0 < t < 1, 0 < t < 1,
where 1 < α, β < 1, f , g : [0, 1] × R → R are nonlinear continuous functions. Wang et al. (2010), developed sufficient conditions for existence and uniqueness of solution for the coupled system with three point boundary conditions of the form
Dα u(t) = f (t, v(t)), Dβ v(t) = g(t, u(t)); 0 < t < 1, u(0) = v(0) = 0; u(1) = au(ξ ), v(1) = bv(ξ ),
where 1 < α, β < 2, 0 ≤ a, b ≤ 1, 0 < ξ < 1, and f , g : [0, 1] × [0, ∞) → [0, ∞) are nonlinear continuous functions. Rehman and Khan (2010), established sufficient conditions for multiplicity results for positive solutions to the following coupled system of nonlinear boundary value problem of fractional differential equations as given by
α β D u(t) + a(t)f (v(t)) = 0; D v(t) + b(t)g(u(t)) = 0; 0 < t < 1, u(0) = u′ (0) = u′′ (0) = · · · = u(n−2) (0) = u(n−1) (0) = 0; u(1) = 0, v(0) = v′ (0) = v′′ (0) = · · · = v(n−2) (0) = v(n−1) (0) = 0; v(1) = 0,
where n − 1 < α, β ≤ n, > 0, f , g : [0, ∞) → [0, ∞) are continuous. Jalili and Samet (2014), studied existence and uniqueness as well as multiplicity of positive solutions to the following coupled system of boundary value problems of fractional differential equations α D u(t) + f (t, u(t), v(t)) = 0; t ∈ (0, 1); n − 1 < α ≤ n, Dβ v(t) + g(t, u(t), v(t)) = 0; t ∈ (0, 1); n − 1 < β ≤ n,
u(i) (0) = v(i) (0) = 0, i = 0, 1, 2 . . . n − 2, γ D u(1) = Dδ v(1) = 0,
where 2 < γ , δ ≤ n − 2, n > 3, f , g : I × [0, ∞) × [0, ∞) → [0, ∞) are continuous. The aim of this paper is to study the existence, uniqueness as well as non-existence conditions for positive solution to the following system of non-linear fractional order differential equations with four point boundary conditions α D u(t) + φ(t, u(t), v(t)) = 0; t ∈ (0, 1); n − 1 < α ≤ n, Dβ v(t) + ψ(t, u(t), v(t)) = 0; t ∈ (0, 1); n − 1 < β ≤ n, (1) u(0) = u′ (0) = u′′ (0) = · · · = u(n−2) (0) = 0; Dδ u(1) = Dδ u(η), v(0) = v′ (0) = v′′ (0) = · · · = v(n−2) (0) = 0; Dγ v(1) = µDγ v(ξ ).
Shah and Khan SpringerPlus (2016)5:1116
Page 4 of 20
where n > 3, α − δ ≥ 1, β − γ ≥ 1 and 0 < δ, γ ≤ n − 2, , µ ∈ (0, ∞), 0 < η, , ξ < 1, φ, ψ : [0, 1] × [0, ∞) × [0, ∞) → [0, ∞) are continuous functions and Dα , Dβ stand for Riemann–Liouville fractional derivative of order α, β respectively. Sufficient conditions are developed for uniqueness of solution of system (1), by using Perov’s fixed theorem. Moreover by means of some classical fixed point theorems of cone type, we develop necessary and sufficient conditions under which the considered system has at least one , two or more positive solutions. Also, we develop conditions for nonexistence of positive solution for system (1). We also provide some examples to illustrate our main results.
Preliminaries To proceed further, we recall some basic definitions and well known results of functional analysis, fixed point theory and fractional calculus (see i.e. Agarwal et al. 2004; Deimling 1985; Jalili and Samet 2014; Krasnoselskii 1964; Miller and Ross 1993; Podlubny 1993; Szilárd 2003; Zeidler 1986). Definition 1 The fractional integral of order α > 0 of a function y : (0, ∞) → R is defined by t 1 I α y(t) = (t − s)α−1 y(s)ds, Ŵ(α) 0 provided the integral is pointwise defined on (0, ∞). Definition 2 The Riemann–Liouville fractional derivative of order α > 0 of a function y ∈ C n [0, 1] is defined by n t 1 d α D0+ y(t) = (t − s)n−α−1 y(s)ds, Ŵ(n − α) dt 0 where n = ⌈α⌉ + 1, ⌈α⌉ represents integer part of α. The following results need in the sequel, Lemma 3 Ahmad and Nieto (2009a, b) Ci ∈ R, i = 0, 1, 2, . . . , n, n = [α] + 1, we have
Let
α>0
then
for
arbitrary
I α Dα y(t) = y(t) + C1 t α−1 + C2 t α−2 + . . . + Cn t α−n .
Lemma 4 Agarwal et al. (2004) Let X be a Banach space with P ⊆ X closed and convex. Let U be a relatively open subset of P with 0 ∈ U and A : U → U be a continuous and compact(completely continuous) mapping. Then either 1. The mapping A has a fixed point in U or 2. There exist u ∈ ∂U and ∈ (0, 1) with u = Au. Lemma 5 Podlubny (1993) Let P be a cone of real Banach space X and let 1 and 2 be two bounded open sets in X such that 0 ∈ 1 ⊂ 1 ⊂ 2. Let operator A : P ∩ (�2 \ �1 ) → P be completely continuous operator. If one of the following two conditions holds:
Shah and Khan SpringerPlus (2016)5:1116
(1) (2)
Page 5 of 20
�Au� ≤ �u� for all u ∈ P ∩ ∂�1 ; �Au� ≥ �u�, for all u ∈ P ∩ ∂�2 ; �Au� ≥ �u� for all u ∈ P ∩ ∂�1 ; �Au� ≤ �u�, for all u ∈ P ∩ ∂�2.
Then A has at least one fixed point in P ∩ (�2 \ �1 ). Definition 6 Jalili and Samet (2014) For a nonempty set X, the mapping d : X × X → Rn is called a generalized metric on X if for all x, y, and u, v ∈ X with v � = x, v �= u, v �= y satisfies
M1 . d (u, v) = 0Rn if and only if u = v, for all u, v ∈ X; d (u, v) = d(v, u), for all u, v ∈ X; (symmetric property) M2 . d (x, y) ≤ d(x, v) + d(v, u) + d(u, y), for all x, y, u, v, ∈ X (tetrahedral inequality). M3 . Moreover the pairs (X, d) is called generalized metric space. n×n }, the system of all n × n Definition 7 Jalili and Samet (2014) Let M = {Mn,n ∈ R+ matrices with positive element. For any matrix A the spectral radius is defined by ρ(A) = sup{|i |, i = 1, 2, . . . , n}, where i , (i = 1, 2, . . . , n) are the eigenvalues of the matrix A.
Lemma 8 Jalili and Samet (2014) Let (M, d) be a complete generalized metric space and let A : M → M be an operator such that there exists a matrix A ∈ M with
d(Au, Av) ≤ Ad(u, v), for all u, v ∈ M. if ρ(A) < 1, then A has a fixed point w ∈ M. Further, for any u0 the iterative sequence un+1 = Aun converges to w.
Main results This section is concerned to the existence, uniqueness and multiplicity results of positive solutions for boundary value problem (1). We begin with the following lemma. Lemma 9 Let y ∈ C[0, 1] then the boundary value problem
Dα u(t) + y(t) = 0; 0 < t < 1; n − 1 < α ≤ n,
(2)
u(0) = u′ (0) = · · · = u(n−2) (0) = 0; Dδ u(1) = Dδ u(η),
where 0 < η < 1, ∈ (0, ∞), 0 < δ ≤ n − 2, has a unique positive solution given by
u(t) =
1
Gα (t, s)y(s)ds,
0
where Gα (t, s) is a Green’s function given by Gα (t, s) α−1 (1 − s)α−δ−1 − (t − s)α−1 − d t α−1 (η − s)α−δ−1 ; 1 d1 t 1 d1 t α−1 (1 − s)α−δ−1 − (t − s)α−1 ; = d t α−1 (1 − s)α−δ−1 − d1 t α−1 (η − s)α−δ−1 ; Ŵ(α) 1 α−1 d1 t (1 − s)α−δ−1 ;
0 ≤ s ≤ t ≤ η ≤ 1, 0 ≤ η ≤ s ≤ t ≤ 1, 0 ≤ t ≤ η ≤ s ≤ 1, 0 ≤ η ≤ t ≤ s ≤ 1,
(3)
Shah and Khan SpringerPlus (2016)5:1116
where d1 =
Page 6 of 20
1 . 1−ηα−δ−1
Proof By applying I α and using Lemma 3, the general solution of linear boundary value problem (2) is given by
u(t) = −I α y(t) + c1 t α−1 + c2 t α−2 + · · · + cn t α−n .
(4)
With the help of boundary and initial conditions of Eq. (2) and Dδ [t α−1 ] = we get c2 = c3 = · · · = cn = 0 and using d1 = 1−η1α−δ−1 , we have
d1 c1 = Ŵ(α)
0
1
α−δ−1
(1 − s)
y(s)ds −
0
η
α−δ−1
(η − s)
y(s)ds
Ŵ(α) α−δ−1 , Ŵ(α−δ) t
and thus (4) becomes
η 1 d1 t α−1 α−δ−1 α−δ−1 u(t) = −I y(t) + (η − s) y(s)ds (1 − s) y(s)ds − Ŵ(α) 0 0 t 1 (t − s)α−1 y(s)ds =− Ŵ(α) 0 η 1 d1 t α−1 α−δ−1 α−δ−1 + (1 − s) y(s)ds − (η − s) y(s)ds Ŵ(α) 0 0 1 = Gα (t, s)y(s)ds, α
0
where Gα (t, s) is the Green’s function of linear boundary value problem (2). Similarly, we 1 can obtain v(t) = 0 Gβ (t, s)z(s)ds, where Gβ (t, s) is the Green’s function for the second equation of the system (1) and given by d t β−1 (1 − s)β−γ −1 − (t − s)β−1 − d2 µt β−1 (ξ − s)β−γ −1 ; 2 1 d2 t β−1 (1 − s)β−γ −1 − (t − s)β−1 ; Gβ (t, s) = d t β−1 (1 − s)β−γ −1 − d2 t β−1 µ(ξ − s)β−γ −1 ; Ŵ(β) 2 β−1 d2 t (1 − s)β−γ −1 ;
where d2 =
0 ≤ s ≤ t ≤ ξ ≤ 1, 0 ≤ ξ ≤ s ≤ t ≤ 1, 0 ≤ t ≤ ξ ≤ s ≤ 1, 0 ≤ ξ ≤ t ≤ s ≤ 1,
1 . 1−µξ β−γ −1
(5)
We recall the following lemma found in Jalili and Samet (2014) and Yang (2012). The proof is omitted because the proof is like the proof given in Goodrich (2010). Lemma 10 Let G(t, s) = (Gα (t, s), Gβ (t, s)) be the Green’s function of (1). This Green’s function G(t, s) has some properties given by
(P1 ) G(t, s) is continuous function on the unit square for all (t, s) ∈ [0, 1] × [0, 1]; (P2 ) G (t, s) ≥ 0 for all t, s ∈ [0, 1] and G(t, s) > 0 for all t, s ∈ (0, 1); (P3 ) max G(t, s) = G(1, s), ∀ s ∈ [0, 1]; 0≤t≤1
(P4 ) min G(t, s) ≥ γ (s)G(1, s) t∈[ω,1−ω]
for
γ = min{γα = ωα−1 , γβ = ωβ−1 }.
each
ω ∈ (0, 21 ), s ∈ (0, 1),
where
Shah and Khan SpringerPlus (2016)5:1116
Page 7 of 20
Further we define some fundamental results which will be used throughout in this paper. Let us define X = {u(t)|u ∈ C[0, 1]} endowed with the norm ||u|| = maxt∈[0,1] |u(t)|. Further the norm for the product space we define as ||(u, v)|| = ||u|| + ||v||. Obviously (X × X, || · ||) is a Banach space. Let J = [ω, 1 − ω], then, we define the cone P ⊂ X × X by
P = {(u, v) ∈ X × X : min[u(t) + v(t)] ≥ γ ||(u, v)||}. t∈J
Pr = {(u, v) ∈ P : �(u, v)� ≤ r}, ∂Pr = {(u, v) ∈ P : �(u, v)� = r}. Now inview of Lemma 9, we can write system (1) as an equivalent coupled system of integral equations given as
�1 u(t) = Gα (t, s)φ(s, u(s), v(s))ds, 0
(6)
�1 v(t) = Gβ (t, s)ψ(s, u(s), v(s))ds. 0
Let A : X × X → X × X be the operator defined as
A(u, v)(t) =
1
Gα (t, s)φ(s, u(s), v(s))ds, 0 = A1 (u, v)(t), A2 (u, v)(t) .
1
Gβ (t, s)ψ(s, u(s), v(s))ds
0
.
(7)
Then the fixed point of operator A coincides with the solution of coupled system (1). Theorem 11 Assume that f , g : [0, 1] × [0, ∞) × [0, ∞) → [0, ∞) are continuous. Then A(P) ⊂ P and A : P → P is completely continuous, where A is defined in (7). Proof To derive A(P) ⊂ P, let (u, v) ∈ P,then by Lemma 10, we have A(u, v) ∈ P. Further from property (P4 ) and for all t ∈ J , we get
A1 (u(t), v(t)) =
1
Gα (t, s)φ(s, u(s), v(s))ds ≥ γα
0
1
Gα (1, s)φ(s, u(s)), v(s)ds.
(8)
0
Also from (P3 ), we obtain
A1 (u(t), v(t)) =
1
Gα (t, s)φ(s, u(s), v(s))ds ≤
0
Thus from (8) and (9), we have
A1 (u(t), v(t)) ≥ γ �A1 (u, v)�, for all t ∈ J . Similarly, one can write that
1 0
Gα (1, s)φ(s, u(s)), v(s)ds.
(9)
Shah and Khan SpringerPlus (2016)5:1116
Page 8 of 20
A2 (u(t), v(t)) ≥ γ �A2 (u, v)�, for all t ∈ J . Thus A1 (u(t), v(t)) + A2 (u(t), v(t)) ≥ γ �A(u, v)�, for all t ∈ J , min[A1 (u(t), v(t)) + A2 (u(t), v(t))] ≥ γ �A(u, v)�. t∈J
Hence we have A(u, v) ∈ P ⇒ A(P) ⊂ P. Next by similar proof of Theorem 1 of Shah and Khan (2015) and using Arzela´ -Ascoli’s theorem, one can easily prove that A : P → P is completely continuous. Theorem 12 Assume that φ and ψ are continuous on [0, 1] × [0, ∞) × (0, ∞) → [0, ∞) , and there exist fi (t), gi (t), (i = 1, 2) : (0, 1) → [0, ∞) that satisfy
|φ(t, u, v) − φ(t, u, ¯ v¯ )| ≤ f1 (t)|u − u| ¯ + g1 (t)|v − v¯ |, for t ∈ (0, 1) and u, v, u, ¯ v¯ ≥ 0; (A1 )
|ψ(t, u, v) − ψ(t, u, ¯ v¯ )| ≤ f2 (t)|u − u| ¯ + g2 (t)|v − v¯ |, for t ∈ (0, 1) and u, u, u, ¯ v¯ ≥ 0; (A2 ) 2×2 (A3 ) } is a matrix given by ρ (A) < 1, where A ∈ {M2,2 ∈ R+
�
1
Gα (1, s)f1 (s)ds 0 � 1 Gβ (1, s)f2 (s)ds 0
1
�
Gα (1, s)g1 (s)ds . � 1 Gβ (1, s)g2 (s)ds 0
0
Then the system (1) has a unique positive solution (u, v) ∈ P. Proof Let us define a generalized metric d : X 2 × X 2 → R2 by ||u − u|| ¯ d((u, v), (u, ¯ v¯ )) = , for all (u, v), (u, ¯ v¯ ) ∈ X × X. ||v − v¯ || Obviously (X × X, d) is a generalized complete metric space. Then for any (u, v), (u, ¯ v¯ ) ∈ X × X and using property (P3 ) we get
|A1 (u, v)(t) − A1 (u, ¯ v¯ )(t)| ≤ max
t∈[0,1]
1
|Gα (t, s)|[|φ(s, u(s), v(s)) − φ(s, u(s), ¯ v¯ (s))|]ds
0
≤
1
≤
1
Gα (1, s)[f1 (s)�u − u� ¯ + g1 (s)�v − v¯ �]ds
0
f1 (s)Gα (1, s)ds�u − u� ¯ +
0
Similarly we can show that
|A2 (u, v) − A2 (u, ¯ v¯ )| ≤
0
+
1
f2 (s)Gβ (1, s)ds�u − u� ¯
0
1
g2 (s)Gβ (1, s)ds�v − v¯ �.
1 0
g1 (s)Gα (1, s)ds�v − v¯ �.
Shah and Khan SpringerPlus (2016)5:1116
Page 9 of 20
Thus we have
|A(u, v) − A(u, ¯ v¯ )| ≤ Ad((u, v), (u, ¯ v¯ )),
∀(u, v), (u, ¯ v¯ ) ∈ X × X,
where
�
1
Gα (1, s)f1 (s)ds A = �0 1 Gβ (1, s)f2 (s)ds 0
1
�
Gα (1, s)g1 (s)ds . � 1 Gβ (1, s)g2 (s)ds 0
0
As ρ(A) < 1, by means of Lemma 8, system (1) has a unique positive solution.
Theorem 13 Let φ and ψ are continuous on [0, 1] × [0, ∞) × (0, ∞) → [0, ∞) and there exist ci , di , ei (i = 1, 2) : (0, 1) → [0, ∞) satisfying:
φ (t, u(t), v(t)) ≤ c1 (t) + d1 (t)u(t) + e1 (t)v(t), t ∈ (0, 1), u, v ≥ 0; (A4 )
ψ (t, u(t), v(t)) ≤ c2 (t) + d2 (t)u(t) + e2 (t)v(t), t ∈ (0, 1), u, v ≥ 0; (A5 )
1 1 (A6 ) �1 = Gα (1, s)c1 (s)ds < ∞, �1 = Gα (1, s)[d1 (s) + e1 (s)]ds < 12 ; 0
0
1 1 (A7 ) �2 = Gβ (1, s)c2 (s)ds < ∞, �2 = Gβ (s, s)[d2 (s) + e2 (s)]ds < 12 . 0
0
Then the system (1) has at least one positive solution in
�1 (u, v) ∈ P : �(u, v)� ≤ r , where max 1 , − �1 2
�2 1 2
− �2
< r.
�1 , Proof Defined � = (u, v) ∈ P : �(u, v)� < r with max 1 2 − �1
�2 1 2
− �2
< r.
Also inview of Theorem 11, the operator A : → P is completely continuous. Let (u, v) ∈ �, such that (u, v) < r. Then, we have
1 �A1 (u, v)� = max Gα (t, s)φ(s, u(s), v(s))ds t∈[0,1] 0 1 1 ≤ Gα (1, s)d1 (s)|u(s)|ds + Gα (1, s)c1 (s)ds + 0
≤
0
1
Gα (1, s)c1 (s)ds + r
r ≤ �1 + r�1 < , 2
0
Gα (1, s)e1 (s)|v(s)|ds
0
0
1
1
Gα (1, s)[d1 (s) + e1 (s)]ds
similarly, A2 (u, v) < 2r , thus A(u, v) < r. Therefore, thank to Lemma 4, we have A(u, v) ∈ �. Therefore A : → . Let there exist ζ ∈ (0, 1) and (u, v) ∈ ∂� such that (u, v) = ζ A(u, v). Then inview of assumptions (A4 ), (A5 ) and using Property (P4 ) of Lemma 10, we obtain for all t ∈ [0, 1]
Shah and Khan SpringerPlus (2016)5:1116
|u(t)| ≤ ζ
Page 10 of 20
1
Gα (t, s)|φ(s, u(s), v(s))|ds
0
≤ζ
<ζ
r 2
1
Gα (1, s)c1 (s)ds + 0 ≤ ζ �1 + r�1
1
Gα (1, s)d1 (s)u(s)ds +
0
1
Gα (1, s)e1 (s)v(s)ds
0
r which implies that �u� < ζ , 2 similarly, one can show that v < ζ 2r . From which, we have �(u, v)� < ζ r, with ζ ∈ (0, 1) which is a contradiction that (u, v) ∈ ∂� as r = �(u, v)�. Thus by mean of Lemma 4, A has at least one fixed point (u, v) ∈ �. Next we use the following assumptions and notations:
φ , ψ : [0, 1] × [0, ∞) × [0, ∞) → [0, ∞) are continuous (C1 ) = ψ(t, 0, 0) = 0 uniformly with respect to t on [0, 1]. (C2 ) Gα (1, s), Gβ (1, s) defined in Lemma 10 satisfy 0<
1
Gα (1, s)ds < ∞, 0 <
0
1
and
φ(t, 0, 0)
Gβ (1, s)ds < ∞;
0
(C3 ) Let these limit hold φσ = φσ =
ψ(t, u, v) φ(t, u, v) , ψσ = , lim sup (u,v)→(σ ,σ ) t∈[0,1] u + v (u,v)→(σ ,σ ) t∈[0,1] u + v lim
sup
φ(t, u, v) ψ(t, u, v) , ψσ = lim inf , where σ ∈ {0, ∞}. (u,v)→(σ ,σ ) t∈[0,1] u + v (u,v)→(σ ,σ ) t∈[0,1] u + v lim
σα = max
t∈[0,1]
inf
1
Gα (t, s)ds, σβ = max
t∈[0,1]
0
1
Gβ (t, s)ds.
0
Theorem 14 If the assumptions (C1 ) − (C2 ) hold and one of the following conditions is also satisfied:
1−ω 1−ω (D1 ) φ0 γ 2 Gα (1, s)ds > 1, φ ∞ σα < 1 and ψ0 γ 2 Gβ (1, s)ds > 1, ψ ∞ σβ < 1. ω
ω
Moreover, φ0 = ψ0 = ∞ and φ ∞ = ψ ∞ = 0;
(D2 ) there exist two constants 1 , 2 with 0 < 1 ≤ 2 such that φ(t, ·, ·) and ψ(t, ·, ·) are nondecreasing on [0, 2 ] for all t ∈ [0, 1],
Shah and Khan SpringerPlus (2016)5:1116
Page 11 of 20
Then boundary value problem (1) has at least one positive solution.
−1 1−ω � 1 γα Gα (1, s)ds , φ(t, γα 1 , γβ 1 ) ≥ 2 ω
ψ(t, γα 1 , γβ 1 ) ≥
1 γβ 2
1−ω � ω
−1
Gβ (1, s)ds
2 2 and φ(t, 2 , 2 ) ≤ , ψ(t, 2 , 2 ) ≤ , forall t ∈ [0, 1]. 2σα 2σβ Proof As A defined in (7) is completely continuous.
Case I. Let the conditions of (D1 ) hold. Taking φ0 exists a constant ǫ1 > 0 such that
γα2
1−ω
Gα (1, s)ds
ω
> 1, then there
φ(t, u, v) ≥ (φ0 − r1 )(u(t) + v(t)), ψ(t, u, v) ≥ (ψ0 − r2 )(u(t) + v(t)), for all t ∈ [0, 1], u, v ∈ [0, ǫ1 ], where r1 > 0, and satisfies the conditions
γ2 (φ0 − r1 ) α 2
1−ω
Gα (1, s)ds ≥ 1, (ψ0 − r1 )
ω
γβ2 2
1−ω
Gβ (1, s)ds ≥ 1.
ω
So for t ∈ [0, 1], (u, v) ∈ ∂Pǫ1 , we have
A1 (u, v)(t) =
1
Gα (t, s)φ(s, u(s), v(s))ds ≥ γα
0
1
Gα (1, s)φ(s, u(s), v(s))ds
0
≥ (φ0 − r1 )
γα2 2
1
�(u, v)� . 2
Gα (1, s)ds�(u, v)� ≥
0
Analogously
A2 (u, v)(t) =
1
Gβ (t, s)ψ(s, u(s), v(s))ds ≥ γβ
0
≥ (ψ0 − r2 )
1
Gβ (1, s)φ(s, u(s), v(s))ds
0
γβ2 2
1
Gβ (1, s)ds�(u, v)� ≥
�(u, v)� . 2
0
Therefore, we have
�A(u, v)� ≥ �A1 (u, v)� + �A2 (u, v)� ≥ �(u, v)�.
(10)
Also for φ ∞ σα < 1 and ψ ∞ σβ < 1, there exists a constant say ǫ¯2 > 0 such that φ(t, u, v) ≤ (φ ∞ + r2 )(u + v), ψ(t, u, v) ≤ (ψ ∞ + r2 )(u + v), for t ∈ [0, 1], u, v ∈ (ǫ¯2 , ∞), where r2 > 0 satisfies the conditions σα (φ ∞ + r2 ) ≤ 1, σβ (ψ ∞ + r2 ) ≤ 1. Let K = maxt∈[0,1],u,v∈[0,ǫ¯2 ] φ(t, u, v), L = maxt∈[0,1],u,v∈[0,ǫ¯2 ] ψ(t, u, v), then φ(t, u, v) ≤ K +(φ ∞ + r2 )(u, v), ψ(t, u, v) ≤ L + (ψ ∞ + r2 )(u, v). Now setting
Shah and Khan SpringerPlus (2016)5:1116
Page 12 of 20
max{ǫ1 , ǫ¯2 , K σα (1−σα (φ ∞ +r2 ))−1 } ≤
ǫ2 ∞ −1 2 , max{ǫ1 , ǫ¯2 , Lσβ (1−σβ (ψ +r2 )) }
≤
ǫ2 2.
So for any t ∈ [0, 1], (u, v) ∈ ∂Pǫ2 , we obtain
A1 (u, v)(t) =
1
Gα (t, s)φ(s, u(s), v(s))ds ≤ γα
≤
1
Gα (1, s)(K + (φ ∞ + r2 )[u(s) + v(s)]ds
0
1
Gα (1, s)φ(s, u(s), v(s))ds
0
0
≤K
1
Gα (1, s)ds + (φ
0
∞
+ r2 )
1
Gα (1, s)ds�(u, v)�
0
ǫ2 ǫ2 ǫ2 − σα (φ ∞ + r2 ) + (φ ∞ + r2 )σα �(u, v)� < . < 2 2 2 Similarly A2 (u, v)(t) <
ǫ2 2,
as (u, v) ∈ ∂Pǫ2, thus, we have
(11)
A(u, v) < (u, v).
Case II. If assumptions in (D2 ) hold, then inview of definition of P for (u, v) ∈ ∂P1, we have �(u, v)� = 1 , for t ∈ J . Then from (D2 ), we have
A1 (u, v)(t) =
�1
1−ω �
Gα (t, s)φ(s, u(s), v(s))ds ≥ γα
Gα (1, s)φ(s, u(s), v(s))ds
ω
0
≥ γα
1−ω � ω
1 Gα (1, s)ds γα 2
Similarly one can also obtain A2 (u, v)(t) ≥
−1
1−ω � ω
Gα (1, s)ds
1 2 , for (u, v)
=
1 . 2
∈ ∂P1, we get
(12)
�A(u, v)� = �A1 (u, v)� + �A2 (u, v)� ≥ �(u, v)�.
Also for (u, v) ∈ ∂P2, we get that �(u, v)� = 2 for t ∈ [0, 1]. Then from (D2 ), one can get
A1 (u, v)(t) =
1
Gα (t, s)φ(s, u(s), v(s))ds ≤
0
≤
2 2σα
1
Gα (1, s)φ(s, u(s), v(s))ds
0
1
Gα (1, s)ds =
2 . 2
0
Similarly, one can also obtain that A2 (u, v)(t) ≤
2 2 , (u, v)
�A(u, v)� = �A1 (u, v)� + �A2 (u, v)� ≤ �(u, v)�.
∈ ∂P2. Hence, we have
(13)
Now inview of application of Lemma 5 to (10) and (11) or (12) and (13) implies that A has a fixed point (u1 , v1 ) ∈ P ǫ, or (u1 , v1 ) ∈ P¯ ǫi ,i (i = 1, 2) such that u1 (t) ≥ γα �u1 � > 0
Shah and Khan SpringerPlus (2016)5:1116
Page 13 of 20
and v1 (t) ≥ γβ �v1 � > 0, t ∈ [0, 1]. From which it follows that boundary value problem (1) has at least one positive solution. Theorem 15 Assume that (C1 ) − (C3 ) hold. Further the following conditions are also satisfied:
(D3 ) If φ 0 σα < 1; φ ∞ γα2
1−ω ω
Gα (1, s)ds > 1 and ψ 0 σβ < 1; ψ ∞ γβ2
1−ω
Gα (1, s)ds
ω
> 1,
Then the boundary value problem (1) has at least one positive solutions. Moreover, if φ 0 = ψ 0 = 0 and φ ∞ = ψ ∞ = ∞, then the the boundary value problem (1) has at least one positive solution. Proof Proof is similar as like the proof of Theorem 14. Theorem 16 Assume that (C1 ) − (C3 ) hold. Also the following conditions are also satisfied:
(D4 ) If φ0 γα2
1−ω
Gα (1, s)ds
ω
(1, s)ds) > 1, ψ∞
γβ2
is also hold;
1−ω
> 1, φ∞ γα2
Gα (1, s)ds
ω
(D5 ) there exists a > 0 such that maxt∈[0,1],(u,v)∈∂Pa φ(t, u, v) <
a 2σα
1−ω
Gα (1, s)ds
ω
> 1 and ψ0 γβ2
1−ω
Gβ
ω
> 1. Moreover, φ0 = ψ0 = φ∞ = ψ∞ = ∞
and maxt∈[0,1],(u,v)∈∂Pa ψ(t, u, v) <
a 2σβ .
Then
¯ v¯ ) the boundary value problem (1) has at least two positive solutions (u, v), (u, such that
(14)
0 < �(u, v)� < a < �(u, ¯ v¯ )�. Proof Let (D4 ) hold. Select ǫ, such that 0 < ǫ < ρ < . 1−ω 1−ω 2 2 φ0 γα Gα (1, s)ds > 1 and ψ0 γβ Gβ (1, s)ds > 1. ω
Now
if
ω
Then like the proof of Theorem 14, we have
(15)
�A(u, v)� ≥ |(u, v)�, for(u, v) ∈ ∂Pǫ . 1−ω 1−ω Gα (1, s)ds > 1 and ψ∞ γβ2 Gβ (1, s)ds > 1. Now, if φ∞ γα2 ω
ω
Then like the proof of Theorem 14, we have
�A(u, v)� ≥ �(u, v)�,
(16)
for (u, v) ∈ ∂P .
Also from (D5 ), (u, v) ∈ ∂Pρ , we get
A1 (u, v)(t) = ≤
1
Gα (t, s)φ(s, u(s), v(s))ds
0
0
1
Gα (1, s)φ(s, u(s), v(s))ds <
ρ 2σα
0
1
Gα (1, s)ds =
ρ . 2
Shah and Khan SpringerPlus (2016)5:1116
Page 14 of 20
Similarly, we have A1 (u, v)(t) <
�A(u, v)� < |(u, v)�,
ρ 2
as (u, v) ∈ ∂Pρ. Hence, we have
for (u, v) ∈ ∂Pρ .
(17)
Now by the use of Lemma 5 to (15) to (17) gives that A has a fixed point (u, v) ∈ ∂P ǫ,ρ ¯ v¯ ) ∈ ∂P ρ, . Hence it implies that the boundary value problem (1) and a fixed point in (u, ¯ v¯ ) such that �(u, v)� �= ρ and �(u, ¯ v¯ )� �= ρ. has at least two positive solutions (u, v), (u, Thus relation (14) holds. The proof is completed. Theorem 17 Assume that (C1 ) − (C3 ) hold and also the following conditions are satisfied:
σα φ0 < 1 and φ∞ σα < 1; σβ ψ0 < 1, and ψ∞ σβ < 1; (D6 ) (D7 ) there exist ρ > 0 such that −1 1−ω � ρ 2 γα Gα (1, s)ds , max φ(t, u, v) > t∈J ,(u,v)∈∂Pρ 2 ω
ρ max ψ(t, u, v) > γβ2 t∈J ,(u,v)∈∂Pρ 2
−1
1−ω � ω
Gβ (1, s)ds
such that Then the boundary value problem (1) has at least two positive solutions.
0 < �(u, v)� < ρ < �(u, ¯ v¯ )�. Proof Proof is like the proof of Theorem 16.
Similarly for multiplicity the following theorems can be easily deduced: Theorem 18 Let (C1 ) − (C3 ) hold. If there exist 2m positive numbers wk , w ˆ k , k = 1, 2 . . . m with w1 < γα w ˆ1 < w ˆ 1 < w2 < γα w ˆ2 < w ˆ 2 . . . wm < γα w ˆm < w ˆm ˆ1 < w ˆ 1 < w2 < γβ w ˆ2 < w ˆ 2 . . . wm < γβ w ˆm < w ˆ m such that and w1 < γβ w
1
(D8 ) φ (t, u(t), v(t)) ≥ wk γα Gα (1, s)ds
and
0
−1
, for (t, u, v) ∈ [0, 1] × [γα wk , wk ] × [γβ wk , wk ],
φ(t, u(t), v(t)) ≤ σα−1 w ˆ k , for (t, u, v) ∈ [0, 1] × [γα w ˆ k, w ˆ k ] × [γβ wk , wk ], k = 1, 2 . . . m,
ψ(t, u(t), v(t)) ≥ wk γβ (D9 )
�1 0
−1
Gβ (1, s)ds)
, for(t, u, v) ∈ [0, 1] × [γβ wk , wk ]
× [γα wk , wk ], andψ(t, u(t), v(t)) ≤ σβ−1 w ˆ k , for (t, u, v) ∈ [0, 1] × [γα wk , wk ] × [γβ w ˆ k, w ˆ k ], k = 1, 2 . . . m.
Shah and Khan SpringerPlus (2016)5:1116
Page 15 of 20
Then the boundary value problem (1) has at least m-positive solutions (uk , vk ), satisfying
wk ≤ �(uk , vk )� ≤ w ˆ k , k = 1, 2 . . . m. Theorem 19 Suppose that (C1 ) − (C3 ) holds. If there exist 2m positive numbers wk , w ˆ k , k = 1, 2 . . . m, with w1 < w ˆ 1 < w2 < w ˆ 2 . . . < wm < w ˆ m such that
(D10 ) φ and ψ are non-decreasing on [0, w ˆ m ] for all t ∈ [0, 1];
(D11 ) φ (t, u(t), v(t)) ≥ wk γα
1−ω
−1
Gα (1, s)ds)
ω
k = 1, 2 . . . m, ψ(t, u(t), v(t)) ≥ wk γβ ψ(t, u(t), v(t)) ≤
w ˆk , k = 1, 2 . . . m. σβ
, φ(t, u(t), v(t)) ≤
1−ω � ω
−1
Gβ (1, s)ds)
w ˆk , σα
,
Then the boundary value problem (1) has at least m-positive solutions (uk , vk ), satisfying wk ≤ �(uk , vk )� ≤ w ˆ k , v = 1, 2 . . . m.
Examples We conclude the paper with the following examples. Example 20 Consider the coupled system as follow
5 t + 1 �5� [Ŵ D 2 u(t) + |u(t)| + cos |v(t)|] = 0, t ∈ [0, 1], u, v ≥ 0, 4 2 2 D 52 v(t) + t + 1 [sin |u(t)| + v(t)] = 0, t ∈ [0, 1], u, v ≥ 0, 4 (i) (i) u (0) = v (0) = 0, i = 0, 1 � � � � 1 1 1 1 1 1 1 1 D 2 u(1) = D 2 u , D 2 v(1) = D 2 v . 2 2 2 2
Since φ(t, u(t), v(t)) =
(18)
t+1 5 t2 + 1 [Ŵ( )|u(t)| + cos |v(t)|], ψ(t, u(t), v(t)) = [sin |u(t)| + |v(t)|]. 4 2 4
Also as 0 < γ , δ < n − 2, n = [2.5] + 1 = 3 and γ = δ = 12 . Then
t+1 5 t+1 |u2 − u1 | + |v2 − v1 |, |φ(t, u2 , v2 ) − φ(t, u1 , v1 )| ≤ Ŵ( ) 2 4 4 t2 + 1 t2 + 1 |u2 − u1 | + |v2 − v1 |. |ψ(t, u2 , v2 ) − ψ(t, u1 , v1 )| ≤ 4 4
Shah and Khan SpringerPlus (2016)5:1116
where f1 (t) =
t+1 5 4 Ŵ( 2 ), g1 (t)
�1
Page 16 of 20
=
Gα (1, s)f1 (s)ds) 0 A= �1 Gβ (1, s)f2 (s)ds) 0
t+1 4 , f2 (t)
�1
= g2 (t) =
t2 + 1 , from which we have 4
Gα (1, s)g1 (s(ds) 5 36 0 = 22 �1 √ 135 π Gβ (1, s)g2 (s)ds) 0
5 √ 27 π . 22 √ 135 π
By standard calculation, we can obtain that ρ(A) = 0.2479 < 1, hence by the use of Theorem 13 boundary value problem (18) has a unique positive solution. Example 21 Consider the system of non-linear fractional differential equations.
� � 7 5 D 3 u(t) + a(t) u(t) + v(t) = 0, D 2 v(t) + b(t) 3 u(t) + v(t) = 0, u(0) = u′ (0) = 0, v(0) = v′ (0) = 0, � � � � 1 1 1 1 1 1 1 1 D 3 u(1) = D 3 u , D 2 v(1) = D 2 v , 3 2 2 2
t ∈ (0, 1),
(19)
√ √ where φ(t, u, v) = a(t) u(t) + v(t) and ψ(t, u, v) = b(t) 3 u(t) + v(t), δ = 13 , γ = 21 . and a(t), b(t) : [0, 1] → [0, ∞) are continuous.
φ(t, u, v) = ∞, similarly ψ 0 = ∞. Also by simple calculation we can u+v get that φ ∞ = 0 = ψ ∞ . Thus by Theorem 14, boundary value problem (19) has at least one positive solution. Now φ 0 = lim
(u,v)→0
Example 22 Consider the following boundary value problem
9 14 D 2 u(t) + (1 − t 2 )[u(t) + v(t)]2 = 0, D 3 v(t) + [u(t) + v(t)]3 = 0, (i) u (0) = v(i) = 0, i = 0, 1, 2, 3, � � � � 5 5 1 5 1 1 5 1 2 2 2 2 D u(1) = D u , D v(1) = D v . 4 4 3 3
t ∈ (0, 1),
(20)
from (20), we have γ = δ = 52 , as n = 5, 0 < γ , δ < 3. By simple calculation we obtain that φ 0 = ψ 0 = 0 and φ ∞ = ψ ∞ = ∞. Thus by Theorem 15, the boundary value problem (20) has a positive solution. Example 23 Consider the following boundary value problem 2 2 9 D 2 u(t) + (1 + t )[u (t) + v(t)] = 0, t ∈ (0, 1), 2 (4t + 4)σα 3 2 D 143 v(t) + (t + 1)[u(t) + v (t)] = 0, t ∈ (0, 1), 3 (4t + 4)σβ (i) (i) u (0) = v = 0, i = 0, 1, 2, 3, � � � � 5 5 1 5 1 1 5 1 D 2 u(1) = D 2 u , D 2 v(1) = D 2 v . 4 4 3 3
(21)
Shah and Khan SpringerPlus (2016)5:1116
Page 17 of 20
where γ = δ = 52 , as n = 5, 0 < γ , δ < 3. Also, by simple calculation we obtain that φ0 = ψ0 = ∞ and φ∞ = ψ∞ = ∞. Further for all (t, u, v) ∈ [0, 1] × [0, 1] × [0, 1], we have
σβ−1 (t 3 + 1)2 (t 2 + 1)2 σα−1 , ψ(t, u, v) ≤ φ(t, u, v) ≤ = = . 4(t 2 + 1)σα 2 4(t 3 + 1)σβ 2 Thus all the assumptions of Theorem 16 are satisfied and also a = 1. Hence by Theorem 16, the boundary value problem (21) has at least two positive solutions (u1 , v1 ) and (u2 , v2 ) which satisfy
0 < (u1 , v1 ) < 1 < (u2 , v2 ).
Non‑existence of positive solution In this section, we discuss the non-existence of positive solution to the coupled system (1) of fractional order differential equations. Theorem 24 Assume that (C1 ) − (C3 ) hold and φ(t, u, v) < (u,v) and 2σα (u,v) ψ(t, u, v) < 2σβ for all t ∈ [0, 1], u > 0, v > 0, then the boundary value problem (1) has no positive solution. Proof On contrary let (u, v) be the positive solution of boundary value problem (1) . Then (u, v) ∈ P for 0 < t < 1 and
�(u, v)� = �u� + �v� = max |u(t)| + max |v(t)| t∈[0,1]
≤ max
t∈[0,1]
1
t∈[0,1]
Gα (t, s)|φ(s, u(s), v(s))|ds + max
t∈[0,1]
0
<
1 0
1
Gβ (t, s)|ψ(s, u(s), v(s)|ds
0
�(u, v)� ds + Gα (1, s) 2σα
1
Gβ (1, s)
0
�(u, v)� ds 2σβ
⇒ �(u, v)� < �(u, v)�, which is contradiction. So boundary value problem (1) has no positive solution. Hence proof is completed. −1 1−ω γα2 Gα (1, s)ds , Theorem 25 Let(C1 ) − (C3 )holds and if φ(t, u(t), v(t)) > �(u,v)� 2 ω
ψ(t, u(t), v(t)) >
�(u,v)� 2
γβ2
1−ω ω
Gβ (1, s)ds
−1
for all t ∈ [0, 1], u > 0 and v > 0. Then
boundary value problem (1) has no positive solution. Proof Proof is just like the proof of Theorem 24, so we omit it.
Example 26 In this section we provide an example which illustrate the results of Theorems 24 and 25 respectively. Consider the system of non linear fractional differential equations:
Shah and Khan SpringerPlus (2016)5:1116
Page 18 of 20
5 (2u2 (t) + u(t))(20 + cos v(t)) = 0, t ∈ [0, 1], D 2 u(t) + u(t) + v(t) + 1 2 D 52 v(t) + (2v (t) + v(t))(20 + cos v(t)) = 0, t ∈ [0, 1], u(t) + v(t) + 1 (i) (i) u (0) = v (0) = 0, i = 0, 1, � � � � 1 1 1 1 1 1 1 1 D 2 u(1) = D 2 v(1) = . D2u D2v 10 2 10 2
(22)
Since (C1 ) − (C3 ) holds and also γ = δ = 12 , as n = 3, 0 < γ , δ < 1.
φ 0 = ψ 0 = 20, φ ∞ = ψ ∞ = 43, 20�(u, v)� < φ(t, u(t), v(t)) < 43�(u, v)� 20(u, v) < ψ(t, u(t), v(t)) < 43(u, v) and φ(t, u(t), v(t)) < 43(u, v) <
(u, v) , σα
where σα ≈ 0.876126 and σβ ≈ .475675 �(u, v)� ≈ 1.1413�(u, v)� implies that φ(t, u(t), v(t)) Case I: Now φ(t, u(t), v(t)) < σα < 43�(u, v)� ≈ 1.1413�(u, v)� and ψ(t, u(t), v(t)) < 43�(u, v)� ≈ 2.1022�(u, v)�. Thus by Theorem 24, boundary value problem (22) has no positive solution.
2 3
−1
Case II: Also φ(t, u(t), v(t)) > 20�(u, v)� > �(u, v)�γα2 � Gα (1, s)ds
ψ(t, u(t), v(t)) > 2�(u, v)� > �(u, v)�γβ2
2
�3 1 3
−1
Gβ (s, s)ds
1 3
≈ 15.8688�(u, v)� and
≈ 12.678�(u, v)�.
Then by Theorem 25, boundary value problem (22) has no solution.
Conclusion In this article, we have developed sufficient conditions for the multiplicity results of positive solutions to a highly nonlinear coupled system of fractional order differential equations. Our paper is the generalization of the work carried out in Goodrich (2010), Jalili and Samet (2014), Rehman and Khan (2010). In Jalili and Samet (2014), the authors studied the concerned coupled system with homogenous boundary conditions involving fractional order derivative, but we extended this work with taking non-homogenous boundary condition involved fractional order derivative of Riemann- Liouville type. By using classical fixed point theorems, we have successfully developed conditions under which the considered coupled system has multiple solutions. Moreover, uniqueness and non existence results have also been established. Numerous examples have been provided which justify the results developed by us. Authors’ contributions Both authors contribute the manuscript equally. Both authors read and approved the final manuscript. Acknowlegements We are really thankful to the anonymous referees for their valuable suggestions and comments which improved the quality of this Manuscript. We are also thankful to the editor for giving us some directions to produce motivation in this article for Refs. Nielsen et al. (2013), Shang (2014). The authors are also grateful to Springerplus for granting full fee waiver.
Shah and Khan SpringerPlus (2016)5:1116
Competing interests The authors declare that they have no competing interests. Received: 13 February 2016 Accepted: 18 June 2016
References Agarwal R, Meehan M, Regan DO (2004) Fixed points theory and applications. Cambridge University Press, Cambridge Ahmad B, Nieto J (2009) Existence of solutions for non local boundary value problems of higher order non linear fractional differential equations. Abstr Appl Anal 2009:494720-1–494720-9. doi:10.1155/2009/494720 Ahmad B, Nieto JJ (2008) Existence and approximation of solution for a class of non-linear impulsive fractional differential equation with anti-periodic boundary conditions. Nonlinear Anal 69:3291–3298 Ahmad B, Nieto JJ (2009) Existence result for a coupled system of non linear fractional differential equations with three point boundary conditions. Comput Math Appl 58:1838–1843 Anastassiou C, Perin R, Markram H, Koch C (2011) Ephaptic coupling of cortical neurons. Nat Neurosci 14:217–223 Bai C (2008) Triple positive solutions for a boundary value problem of nonlinear fractional differential equation. Electron J Qual Theory Differ Equ 24:1–10 Bai C, Feng J (2004) The existence of positive solution for singular coupled system of nonlinear fractional differential equation. Appl Math Comp 150:611–621 Bai Z, Lü H (2005) Positive solutions for boundary value problem of nonlinear fractional order differential equations. J Math Anal Appl 311:495–505 Chasnov JR (2009) Mathematical biology. Lecture notes. the Hong Kong University of Science and Technology Cui Z, Yu P, Mao Z (2012) Existence of solutions for nonlocal boundary value problems of non linear fractional differential equations. Adv Dyn Syst Appl 7(1):31–40 Cui Y, Zou Y (2014) Existence results and the monotone iterative technique for nonlinear fractional differential systems with coupled four-point Boundary Value Problems. Abstr Appl Anal 2014:1–6 Deimling K (1985) Nonlinear functional analysis. Springer, New York Deren FY (2015) Existence and nonexistence of positive solutions for nonlinear higher order boundary value problem with fractional boundary conditions in Banach spaces. Adv Differ Equ 2015:72 Goodrich C (2010) Existence of a positive solution to a class of fractional differential equations. Appl Math Lett 23:1050–1055 Hagerstrom A, Murphy T, Roy R, H övel P, Omelchenko I, Schöll E (2012) Experimental observation of chimeras in coupledmap lattices. Nat Phys 8:658–661 Hilfer R (2000) Applications of fractional calculus in physics. World Scientific, Singapore Jalili M, Samet B (2014) Existence of positive solutions to a coupled system of fractional differential equation. Math Methods Appl Sci. doi:10.1002/mma.3124 Khan RA, Shah K (2015) Existence and uniqueness of solutions to fractional order multi-point boundary value problems. Commun Appl Anal 19:515–526 Kilbas AA, Srivastava H, Trujillo J (2006) Theory and application of fractional differential equations, North Holland mathematics studies, vol 204. Elsevier, Amsterdam Krasnoselskii MA (1964) Positive solutions of operator equations. Noordhoff, Groningen Lakshmikantham V, Leela S, Vasundhara J (2009) Theory of fractional dynamic systems. Cambridge Academic Publishers, Cambridge Li CF, Luo XN, Zhou Y (2010) Existence of positive solutions of the boundary value problem for nonlinear fractional differential equations. Comput Math Appl 59:1363–1375 Li HL, Jiang YL, Wang ZL, Hu C (2015) Global stability problem for feedback control systems of impulsive fractional differential equations on networks. Neurocomputing 161:155–161 Li HL, Jiang YL, Wang ZL, Zhang L, Teng ZD (2015) Mittag–Leffler stability of coupled system of fractional-order differential equations on network. Appl Math Comput 270:269–277 Lia HL, Jiang YL, Wang Z, Zhang L, Teng Z (2015) Global Mittag–Leffler stability of coupled system of fractional-order differential equations on network. Appl Math Comp 270:269–277 Mainardi F (1995) Fractional diffusion waves in viscoelastic solids. In: Wegner JL, Norwood FR (eds) Nonlinear waves in solids. ASME/AMR, Fairfield, pp 93–97 Metzler R, Klafer J (2000) Boundary value problems for fractional diffusion equations. Phys A 278:107–125 Miller KS, Ross B (1993) An introduction to the fractional calculus and fractional differential equations. Wiley, New York Nielsen AEB, Sierra G, Cirac JI (2013) Fractional quantum Hall states in lattices: local models and physical implementation. Nat Commun 4:2864. doi:10.1038/ncomms3864 Podlubny I (1993) Fractional differential equations. Academic Press, New York Podlubny I (1999) Fractional differential equations. Academic Press, New York Rehman M, Khan RA (2010) Positive solutions to toupled system of fractional differential equations. Int J Nonlinear Sci 10(1):96–104 Rossikhin YA, Shitikova MV (1997) Applications of fractional calculus to dynamic problems of linear and nonlinear hereditary mechanics of solids. Appl Mech Rev 50:15–67 Scher H, Montroll E (1975) Anomalous tranit.time dispersion in amorphous solids. Phys Rev B 12:2455–2477 Shah K, khalil H, Khan RA (2015) Investigation of positive solution to a coupled system of impulsive boundary value problems for nonlinear fractional order differential equations. J Chaos Solitons Fractals 77:240–246 Shah K, Khalil H, Khan RA (2016) Upper and lower solutions to a coupled system of nonlinear fractional differential equationss. Prog Fract Differ Appl 2(1):1–10
Page 19 of 20
Shah and Khan SpringerPlus (2016)5:1116
Shah K, Khan RA (2015) Existence and uniqueness of positive solutions to a coupled system of nonlinear fractional order differential equations with anti periodic boundary conditions. Differ Equ Appl 7(2):245–262 Shang Y (2014) Vulnerability of networks: fractional percolation on random graphs. Phys Rev E 89:012813 Su X (2009) Boundary value problem for a coupled system of non linear fractional differential equations. Appl Math Lett 22:64–69 Suo JH, Sun JT, Zhang Y (2013) Stability analysis for impulsive coupled systems on networks. Neurocomputing 99:172–177 Szilárd A (2003) A note on Perov’s fixed point theorem. Fixed Point Theory 4(1):105–108 Wang J, Feckan M, Zhou Y (2011) On the new concept of solutions, existence results for impulsive fractional evolution equations. Dyn Partial Differ Equ 8:345–361 Wang H, Yu YG, Wen GG, Zhang S, Yu JZ (2015) Global stability analysis of fractional-order Hopfield neural networks. Neurocomputing 154:15–23 Wang J, Xiang M, Liu Z (2010) Positive solution to a non zero boundary value problem for a coupled system of non linear fractional differential equations. Int J Differ Equ 2010:186928-1–186928-12. doi:10.1155/2010/186928 Xu X, Jiang D, Yuan C (2009) Multiple positive solutions for the boundary value problem of a nonlinear fractional differential equation. Nonlinear Anal 71:4676–4688 Yang W (2012) Positive solution to non zero boundary values problem for a coupled system of nonlinear fractional differential equations. Comput Math Appl 63:288–297 Zeidler E (1986) Non linear functional analysis and its applications. Springer, New York
Page 20 of 20