This Provisional PDF corresponds to the article as it appeared upon acceptance. Fully formatted PDF and full text (HTML) versions will be made available soon. The Sturm-Liouville BVP in Banach space Advances in Difference Equations 2011, 2011:65 doi:10.1186/1687-1847-2011-65 Su Hua H. Su (jnsuhua@163.com) Lishan Liu L. Liu (lls@mail.qfnu.edu.cn) Xinjun Wang X. Wang (wangxj566@sina.com) ISSN 1687-1847 Article type Research Submission date 12 October 2011 Acceptance date 21 December 2011 Publication date 21 December 2011 Article URL http://www.advancesindifferenceequations.com/content/2011/1/65 This peer-reviewed article was published immediately upon acceptance. It can be downloaded, printed and distributed freely for any purposes (see copyright notice below). For information about publishing your research in Advances in Difference Equations go to http://www.advancesindifferenceequations.com/authors/instructions/ For information about other SpringerOpen publications go to http://www.springeropen.com Advances in Difference Equations © 2011 Hua et al. ; licensee Springer. 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. The Sturm–Liouville BVP in Banach space Hua Su ∗ 1 , Lishan Liu 2 and Xinjun Wang 3 1 School of Mathematic and Quantitative Economics, Shandong University of Finance and Economics, Jinan Shandong 250014, China 2 School of Mathematical Sciences, Qufu Normal University, Qufu Shandong 273165, China 3 School of Economics, Shandong University, Jinan Shandong 250014, China ∗ Corresponding author: jnsuhua@163.com Email addresses: LL: lls@mail.qfnu.edu.cn XW: wangxj566@sina.com Abstract We consider the existence of sin gle and mul t iple positive solutions for fourth-order Sturm–Liouville b oundary value problem in Banach space. The suffici ent condit ion for the existence of single and multiple positive solutions is obtained by fixed t heorem of strict set contraction operator in the frame of the ODE technique. Our results s ignificantly extend and improve many known results including sing ular and nonsingular cases. 1 Introduction The boundary value problems (BVPs) for ordinary differential equations play a very important role in both theory and application. They are used to describe a large number of physical, biological, and chemical phenomena. In this article, we will study the existence of positive solutions for the following fourth-order 1 nonlinear Sturm–Liouville BVP in a rea l Banach space E 1 p(t) (p(t)u ′′′ (t)) ′ = f(u(t)), 0 < t < 1, α 1 u(0) −β 1 u ′ (0) = 0, γ 1 u(1) + δ 1 u ′ (1) = 0, α 2 u ′′ (0) −β 2 lim t→0+ p(t)u ′′′ (t) = 0, γ 2 u ′′ (1) + δ 2 lim t→1− p(t)u ′′′ (t) = 0, (1.1) where α i , β i , δ i , γ i ≥ 0 (i = 1, 2) are constants such that ρ 1 = β 1 γ 1 +α 1 γ 1 +α 1 δ 1 > 0, B(t, s) = s t dτ p(τ) , ρ 2 = β 2 γ 2 + α 2 γ 2 B(0, 1) + α 2 δ 2 > 0, and p ∈ C 1 ((0, 1), (0, +∞)). Moreover p may be singular at t = 0 and/or 1. BVP (1.1) is often referred to as the deformation of an elastic beam under a variety of boundary conditions, for detail, see [1–17]. For example, BVP (1.1) subject to Lidstone boundar y value conditions u(0) = u(1) = u ′′ (0) = u ′′ (1) = 0 are used to model such phenomena as the deflection of elastic beam simply supported at the endpoints, see [1, 3, 5, 7–11, 13–14]. We notice that the above articles use the co mpletely continuous operator and require f satisfies some growth condition or assumptions of monotonicity which are essential for the technique used. The aim of this article is to consider the existence of positive solutions for the more general Sturm– Liouville BVP by using the properties of strict set contraction operator. Here, we allow p have singularity at t = 0, 1, a s far as we know, there were fewer works to be done. This article attempts to fill part of this gap in the literature. This article is organized as follows. In Section 2, we first present some properties of the Green functions that are used to define a positive operator. Next we approximate the singular fourth-order BVP to singular second-order BVP by constructing an integral operator. In Section 3, the sufficient condition for the existence of single and multiple positive so lutions for B VP (1.1) will be established. In Sectio n 4, we give one example as the application. 2 Preliminaries and lemmas In this article, we all suppose that (E, · 1 ) is a real Banach spac e . A nonempty closed convex subset P in E is said to be a cone if λP ∈ P for λ ≥ 0 and P {−P } = {θ}, where θ denotes the zero element of E. The cone P defines a partial ordering in E by x ≤ y iff y − x ∈ P . Recall the cone P is said to be normal 2 if there exists a positive cons tant N such that 0 ≤ x ≤ y implies x 1 ≤ N y 1 . The cone P is normal if every order interval [x, y] = {z ∈ E|x ≤ z ≤ y} is bounded in norm. In this article, we assume that P ⊆ E is normal, and witho ut loss of generality, we may assume that the normality of P is 1. Let J = [0, 1], and C(J, E) = {u : J → E | u(t) continuous}, C i (J, E) = {u : J → E | u(t) is i-o rder continuously differentiable}, i = 1, 2, . . . . Fo r u = u(t) ∈ C(J, E), let u = max t∈J u(t) 1 , then C(J, E) is a Banach space with the norm · . Definition 2.1 A function u(t) is said to be a positive solution of the BVP (1.1), if u ∈ C 2 ([0, 1], E) C 3 ((0, 1), E) satisfies u(t) ≥ 0, t ∈ (0, 1], pu ′′′ ∈ C 1 ((0, 1), E) and the BVP (1.1), i.e., u ∈ C 2 ([0, 1], P) C 3 ((0, 1), P) and u(t) ≡ θ, t ∈ J. We notice that if u(t) is a positive solution of the BVP (1.1) and p ∈ C 1 (0, 1), then u ∈ C 4 (0, 1). Now we denote that H(t, s) and G(t, s) are the Green functions for the following boundary value problem −u ′′ = 0, 0 < t < 1, α 1 u(0) −β 1 u ′ (0) = 0, γ 1 u(1) + δ 1 u ′ (1) = 0 and 1 p(t) (p(t)v ′ (t)) ′ = 0, 0 < t < 1, α 2 v(0) − lim t→0+ β 2 p(t)v ′ (t) = 0, γ 2 v(1) + lim t→1− δ 2 p(t)v ′ (t) = 0, respectively. It is well known that H(t, s) and G(t, s) can be written by H(t, s) = 1 ρ 1 (β 1 + α 1 s)(δ 1 + γ 1 (1 −t)), 0 ≤ s ≤ t ≤ 1, (β 1 + α 1 t)(δ 1 + γ 1 (1 −s)), 0 ≤ t ≤ s ≤ 1 (2.1) and G(t, s) = 1 ρ 2 (β 2 + α 2 B(0, s)) (δ 2 + γ 2 B(t, 1)) , 0 ≤ s ≤ t ≤ 1, (β 2 + α 2 B(0, t)) (δ 2 + γ 2 B(s, 1)) , 0 ≤ t ≤ s ≤ 1, (2.2) where ρ 1 = γ 1 β 1 + α 1 γ 1 + α 1 δ 1 > 0, B(t, s) = s t dτ p(τ) , ρ 2 = α 2 δ 2 + α 2 γ 2 B(0, 1) + β 2 γ 2 > 0. 3 It is easy to verify the following properties of H(t, s) and G(t, s) (I) G(t, s) ≤ G(s, s) < +∞, H(t, s) ≤ H(s, s) < +∞; (II) G(t, s) ≥ ρG(s, s), H(t, s) ≥ ξH(s, s), for any t ∈ [a, b] ⊂ (0, 1), s ∈ [0, 1], where ρ = min δ 2 + γ 2 B(b, 1) δ 2 + γ 2 B(0, 1) , β 2 + α 2 B(0, a) β 2 + α 2 B(0, 1) , ξ = min δ 1 + γ 1 (1 −b) δ 1 + γ 1 , β 1 + α 1 a β 1 + α 1 . Throughout this article, we adopt the following assumptions (H 1 ) p ∈ C 1 ((0, 1), (0, +∞)) and satisfies 0 < 1 0 ds p(s) < +∞, 0 < λ = 1 0 G(s, s)p(s)ds < +∞. (H 2 ) f (u) ∈ C(P \ {θ}, P ) and there exists M > 0 such that for any bounded set B ⊂ C(J, E), we have α(f(B(t))) ≤ M α(B(t)), 2Mλ < 1. (2.3) where α(·) denote the Kuratowski measure of noncompactness in C(J, E). The following lemma s play an important role in this article. Lemma 2.1 [17]. Let B ⊂ C[J, E] be bounded and equicontinuous on J, then α(B) = sup t∈J α(B(t)). Lemma 2.2 [16]. Let B ⊂ C(J, E) be bounded and equicontinuous on J , let α(B) is continuous on J and α J u(t)dt : u ∈ B ≤ J α(B(t))dt. Lemma 2.3 [16]. Let B ⊂ C(J, E) be a bounded set on J. Then α(B(t)) ≤ 2α(B). Now we define a n integral operator S : C(J, E) → C(J, E ) by Sv(t) = 1 0 H(t, τ)v(τ)dτ. (2.4) Then, S is linear continuous operator and by the expresse d of H(t, s), we have (Sv) ′′ (t) = −v(t), 0 < t < 1, α 1 (Sv)(0) −β 1 (Sv) ′ (0) = 0, γ 1 (Sv)(1) + δ 1 (Sv) ′ (1) = 0. (2.5) 4 Lemma 2.4. The Sturm–Liouville BVP (1.1) has a positive solution if and only if the following integral- differential boundary value problem has a positive solution of 1 p(t) (p(t)v ′ (t)) ′ + f(Sv(t)) = 0, 0 < t < 1, α 2 v(0) − lim t→0+ β 2 p(t)v ′ (t) = 0, γ 2 v(1) + lim t→1− δ 2 p(t)v ′ (t) = 0, (2.6) where S is given in (2.4). Proof In fact, if u is a positive solution of (1.1), let u = Sv, then v = −u ′′ . This implies u ′′ = −v is a solution of (2.6). Conversely, if v is a positive solution of (2.6). Let u = Sv, by (2 .5), u ′′ = (Sv) ′′ = −v. Thus, u = Sv is a positive solution of (1.1). This completes the proof of Lemma 2.1. So, we only ne ed to concentrate our study on (2.6). Now, for the given [a, b] ⊂ (0, 1), ρ as above in (II), we introduce K = {u ∈ C(J, P ) : u(t) ≥ ρu(s), t ∈ [a, b], s ∈ [0, 1]}. It is easy to check that K is a cone in C[0 , 1]. Further, for u(t) ∈ K, t ∈ [a, b], we have by normality of cone P with normal constant 1 that u(t) 1 ≥ ρu. Next, we define an operator T given by T v(t) = 1 0 G(t, s)p(s)f(Sv(s))ds, t ∈ [0, 1], (2.7) Clearly, v is a solution of the BVP (2.6) if and only if v is a fixed point of the operator T . Through direct calculation, by (II) and for v ∈ K, t ∈ [a, b], s ∈ J, we have T v(t) = 1 0 G(t, s)p(s)f(Sv(s))ds ≥ ρ 1 0 G(s, s)p(s)f(Sv(s))ds = ρT v(s). So, this implies that T K ⊂ K. Lemma 2.5. Assume that (H 1 ), (H 2 ) hold. Then T : K → K is s trict set contraction. Proof Firstly, The continuity of T is easily obtained. In fact, if v n , v ∈ K and v n → v in the sup norm, then for any t ∈ J, we get T v n (t) −T v(t) 1 ≤ f(Sv n (t)) −f(Sv(t)) 1 1 0 G(s, s)p(s)ds, 5 so, by the co ntinuity of f, S, we have T v n − T v = sup t∈J T v n (t) −T v(t) 1 → 0. This implies tha t T v n → T v in the sup norm, i.e., T is continuous. Now, let B ⊂ K is a bounded set. It follows from the the continuity of S and (H 2 ) that there exists a positive number L such that f (Sv(t)) 1 ≤ L for any v ∈ B. Then, we can get T v(t) 1 ≤ Lλ < ı, ∀ t ∈ J, v ∈ B. So, T (B) ⊂ K is a bounded set in K. Fo r any ε > 0, by (H 1 ), there exists a δ ′ > 0 such that δ ′ 0 G(s, s)p(s) ≤ ε 6L , 1 1−δ ′ G(s, s)p(s) ≤ ε 6L . Let P = max t∈[δ ′ ,1−δ ′ ] p(t). It follows from the continuity of G(t, s) on [0, 1] × [0, 1] that there exists δ > 0 such that |G(t, s) −G(t ′ , s)| ≤ ε 3P L , |t −t ′ | < δ, t, t ′ ∈ [0, 1]. Consequently, when |t −t ′ | < δ, t, t ′ ∈ [0, 1], v ∈ B, we have T v(t) − T v(t ′ ) 1 = 1 0 (G(t, s) −G(t ′ , s))p(s)f(Sv(s))ds 1 ≤ δ ′ 0 |G(t, s) −G(t ′ , s)|p(s)f(Sv(s)) 1 ds + 1−δ ′ δ ′ |G(t, s) −G(t ′ , s)|p(s)f(Sv(s)) 1 ds + 1 1−δ ′ |G(t, s) −G(t ′ , s)|p(s)f(Sv(s)) 1 ds ≤ 2L δ ′ 0 G(s, s)p(s)ds + 2 L 1 1−δ ′ G(s, s)p(s)ds +P L 1 0 |G(t, s) −G(t ′ , s)|ds ≤ ε. 6 This implies tha t T (B) is equicontinuous set on J . Therefore, by Lemma 2.1, we have α(T (B)) = sup t∈J α(T (B)(t)). Without loss of generality, by condition (H 1 ), we may assume that p(t) is singular at t = 0, 1. So, There exists {a n i }, { b n i } ⊂ (0, 1), {n i } ⊂ N with {n i } is a s trict increasing sequence and lim i→+ı n i = +ı such that 0 < ··· < a n i < ··· < a n 1 < b n 1 < ··· < b n i < ··· < 1 ; p(t) ≥ n i , t ∈ (0, a n i ] ∪[b n i , 1), p(a n i ) = p(b n i ) = n i ; lim i→+ı a n i = 0, lim i→+ı b n i = 1. (2.9) Next, we let p n i (t) = n i , t ∈ (0, a n i ] ∪[b n i , 1); p(t), t ∈ [a n i , b n i ]. Then, from the above discussion we know that (p) n i is continuous on J for every i ∈ N and p n i (t) ≤ p(t); p n i (t) → p(t), ∀ t ∈ (0, 1), as i → +ı. Fo r any ε > 0, by (2.9) and (H 1 ), there exists a i 0 such that for a ny i > i 0 , we have that 2L a n i 0 G(s, s)p(s)ds < ε 2 , 2L 1 b n i G(s, s)p(s)ds < ε 2 . (2.10) Therefore, for any bounded set B ⊂ C[J, E], by (2.4), we have S(B) ⊂ B. In fact, if v ∈ B, there exists D > 0 such that v ≤ D, t ∈ J. Then by the properties of H(t, s), we can have Sv(t) 1 ≤ 1 0 H(t, τ)v(τ) 1 dτ ≤ D 1 0 H(t, τ)dτ ≤ D, i.e., S(B) ⊂ B. Then, by Lemmas 2.2 and 2.3, (H 2 ), the above discussion and no te that p n i (t) ≤ p(t), t ∈ (0, 1), as 7 t ∈ J, i > i 0 , we know that α(T (B)(t)) = α 1 0 G(t, s)p(s)f(Sv(s))ds ∈ B ≤ α 1 0 G(t, s)[p(s) −p n i (s)]f(Sv(s))ds ∈ B +α 1 0 G(t, s)p n i (s)f(Sv(s))ds ∈ B ≤ 2L a n i 0 G(s, s)p(s)ds + 2L 1 b n i G(s, s)p(s)ds + 1 0 α (G(t, s)p n i (s)f(Sv(s)) ∈ B) ds ≤ ε + 1 0 G(s, s)p(s)α (f(Sv(s)) ∈ B) ds ≤ ε + 2Mλα(B). Since the randomness of ε, we get α(T (B)(t)) ≤ 2M λα(B), t ∈ J. (2.11) So, it follows fr om (2.8) (2.11) that for any bounded set B ⊂ C[J, E], we have α(T (B)) ≤ 2Mλα(B). And note that 2M λ < 1, we have T : K → K is a strict set contraction. The proof is completed. Remark 1. When E = R, (2.3) naturally ho ld. In this c ase, we may take M as 0, consequently, T : K → K is a completely continuous operator. So, our condition (H 1 ) is weaker than those of the above mention articles. Our main tool of this article is the following fixed point theorem of cone. Theorem 2.1 [16]. Supp ose that E is a Banach space, K ⊂ E is a cone, let Ω 1 , Ω 2 be two bounded open sets of E such that θ ∈ Ω 1 , Ω 1 ⊂ Ω 2 . Let operator T : K ∩ (Ω 2 \ Ω 1 ) −→ K be strict set contraction. Suppose that one of the following two conditions hold, (i) T x ≤ x, ∀ x ∈ K ∩∂Ω 1 , T x ≥ x, ∀ x ∈ K ∩ ∂ Ω 2 ; 8 (ii) T x ≥ x, ∀ x ∈ K ∩∂Ω 1 , T x ≤ x, ∀ x ∈ K ∩ ∂ Ω 2 . Then, T has at least one fixed point in K ∩(Ω 2 \ Ω 1 ). Theorem 2.2 [16]. Suppose E is a real Ba nach space, K ⊂ E is a cone, let Ω r = {u ∈ K : u ≤ r}. Let operator T : Ω r −→ K be completely continuous and satisfy T x = x, ∀ x ∈ ∂Ω r . Then (i) If T x ≤ x, ∀ x ∈ ∂Ω r , then i(T, Ω r , K) = 1; (ii) If T x ≥ x, ∀ x ∈ ∂Ω r , then i(T, Ω r , K) = 0. 3 The main results Denote f 0 = lim x 1 →0 + f (x) 1 x 1 , f ∞ = lim x 1 →∞ f (x) 1 x 1 . In this section, we will give our main results. Theorem 3.1. Suppose that conditions (H 1 ), (H 2 ) hold. Assume that f also satisfy (A 1 ): f(x) ≥ ru ∗ , ξ 1 0 H(τ, τ)x(τ)dτ 1 ≤ x 1 ≤ r; (A 2 ): f(x) ≤ Ru ∗ , 0 ≤ x 1 ≤ R, where u ∗ and u ∗ satisfy ρ b a G(s, s)p(s)u ∗ (s)ds 1 ≥ 1, u ∗ (s) 1 1 0 G(s, s)p(s)ds ≤ 1. Then, the b oundary value problem (1.1) has a positive solution. Proof of Theorem 3.1. Without loss of generality, we suppose that r < R. For any u ∈ K, we have u(t) 1 ≥ ρu, t ∈ [a, b]. (3.1) we define two open s ubs ets Ω 1 and Ω 2 of E Ω 1 = {u ∈ K : u < r}, Ω 2 = {u ∈ K : u < R} Fo r u ∈ ∂Ω 1 , by (3.1), we have r = u ≥ u(s) 1 ≥ ρu = ρr, s ∈ [a, b]. (3.2) 9 [...]... Now, by Lemma 2.4 we see that u1 = Sv1 , u2 = Sv2 are two position solutions of BVP (1.1) The proof of Theorem 3.2 is complete 4 Application In this section, in order to illustrate our results, we consider some examples Now, we consider the following concrete second-order singular BVP (SBVP) Example 4.1 Consider the following SBVP ′ 1 √ ′′′ 1 1 3 3 √ tu (t) + 160 u 2 + u 3 = θ, 0 < t < 1, 3 ... fourth-order two point boundary value problem Appl Math Comput 148, 407–420 (2004) 14 Wei, Z: Positive solutions of singular dirichlet boundary value problems Chin Ann Math 20(A), 543–552 (2007) (in Chinese) 15 Wong, PJY, Agarwal, RP: Eigenvalue of lidstone boundary value problems Appl Math Comput 104, 15–31 (1999) 16 Guo, DJ, Lakshmikantham, V: Nonlinear Problems in Abstract Cone Academic Press, Inc., New... Fourth-order two-point boundary value problems: estimates by two side bounds Nonlinear Anal 8, 107–144 (1984) 3 Agarwal, RP, Chow, MY: Iterative methods for a fourth order boundary value problem J Comput Appl Math 10, 203–217 (1984) 4 Agarwal, RP: On the fourth-order boundary value problems arising in beam analysis Diff Integral Equ 2, 91–110 (1989) 5 Ma, RY: Positive solutions of fourth-order two point boundary... solutions of sublinear Sturm–Liouville problems with changing sign nonlinearity Comput Math Appl 58, 1808–1815 (2009) 11 Yang, J, Wei, Z, Liu, K: Existence of symmetric positive solutions for a class of Sturm–Liouville-like boundary value problems Appl Math Comput 214, 424–432 (2009) 12 Xu, F, Su, H, Positive solutions of fourth-order nonlinear singular boundary value problems Nonlinear Anal 68(5),... monotone increasing a of f (u) on [0, ∞), then f (u) ≥ f (1.05) = 326.4, 1.05 ≤ u ≤ 3 Therefore, as 1 1 ρξr = ρξ 0 H(τ, τ ) u dτ ≤ ξ so we have H(τ, τ )u(τ )dτ , 0 1 ∗ f (u) ≥ ru , ξ 0 H(τ, τ )u(τ )dτ ≤ u ≤ r, then conditions (A1 ) holds Then by Theorem 3.2, SBVP (4.1) has at least two positive solutions u1 , u2 and 0 < u1 < 3 < u2 Competing interests We declare that we have no significant competing financial,... were supported financially by the Shandong Province Natural Science Foundation (ZR2009AQ004), NSFC (11026108, 11071141) and XW was supported by the Shandong Province planning Foundation of Social Science (09BJGJ14), Shandong Province Natural Science Foundation (Z2007A04) References 1 Ma, RY, Wang, HY: On the existence of positive solutions of fourth order ordinary differential equation Appl Anal 59, 225–231... fourth (and higher) order singular boundary value problems J Math Anal Appl 161, 78–116 (1991) 7 Gupta, CP: Existence and uniqueness theorems for a bending of an elastic beam equation at resonance J Math Anal Appl 135, 208–225 (1988) 8 Gupta, CP: Existence and uniqueness theorems for a bending of an elastic beam equation Appl Anal 26, 289–304 (1988) 9 Yang, YS: Fourth order two-point boundary value problem... two positive solutions u1 , u2 and 0 < u1 < 3 < u2 Competing interests We declare that we have no significant competing financial, professional, or personal interests that might have in uenced the performance or presentation of the work described in this manuscript 13 Authors’ contributions All authors contributed equally to the manuscript and read and approved the final manuscript Acknowledgments The... (3.4) Therefore, by (3.2), (3.3), Lemma 2.5 and r < R, we have that T has a fixed point v ∈ (Ω2 \ Ω1 ) Obviously, v is positive solution of problem (2.6) Now, by Lemma 2.4 we see that u = Sv is a position solution of BVP (1.1) The proof of Theorem 3.1 is complete Theorem 3.2 Suppose that conditions (H1 ), (H2 ) and (A1 ) in Theorem 3.1 hold Assume that f also satisfy (A3 ): f0 = 0; (A4 ): f∞ = 0 Then,... K) = 1 (3.8) Finally, set Ωr = {u ∈ K : u < r}, For any u ∈ ∂Ωr , by (A2 ), Lemma 2.2 and also similar to the latter proof of Theorem 3.1, we can also have Tu ≥ u , ∀ u ∈ ∂Ωr Then by Theorem 2.2, we have i(T, Ωr , K) = 0 (3.9) Therefore, by (3.5), (3.7), (3.8), and ρ∗ < r < ρ∗ , we have i(T, Ωr \ Ωρ∗ , k) = −1, i(T, Ωρ∗ \ Ωr , k) = 1 Then T have fixed point v1 ∈ Ωr \ Ωρ∗ , and fixed point v2 ∈ Ωρ∗ \ . below). For information about publishing your research in Advances in Difference Equations go to http://www.advancesindifferenceequations.com/authors/instructions/ For information about other SpringerOpen. many known results including sing ular and nonsingular cases. 1 Introduction The boundary value problems (BVPs) for ordinary differential equations play a very important role in both theory and. we approximate the singular fourth-order BVP to singular second-order BVP by constructing an integral operator. In Section 3, the sufficient condition for the existence of single and multiple positive