EXISTENCE AND UNIQUENESS OF POSITIVE SOLUTIONS FOR FOURTH-ORDER m-point BOUNDARY VALUE PROBLEMS WITH TWO PARAMETERS

ROCKY MOUNTAIN JOURNAL OF MATHEMATICS Volume 43, Number 4, 213 EXISTENCE AND UNIQUENESS OF POSITIVE SOLUTIONS FOR FOURTH-ORDER m-point BOUNDARY VALUE PROBLEMS WITH TWO PARAMETERS XINAN HAO, NAIWEI XU AND LISHAN LIU ABSTRACT. This paper deals with the existence and uniqueness of positive solutions to fourth-order m-point boundary value problems with two parameters. The arguments are based upon a specially constructed cone and a fixed point theorem in a cone for a completely continuous operator, due to Krasnoselskii and Zabreiko. The results obtained herein generalize and complement the main results of [7, 1. 1. Introduction. In this paper, we will study the existence and uniqueness of positive solutions for the following fourth-order m-point boundary value problems with two parameters u (4) + αu βu = f(t, u), <t<1, (1.1) u() = a i u(ξ i ), u(1) = b i u(ξ i ), u () = a i u (ξ i ), u (1) = b i u (ξ i ), where α, β R, ξ i (, 1), a i,b i R + for i {1, 2,...,m 2} are given constants, R + =[, + ), and α, β, f satisfy 21 AMS Mathematics subject classification. Primary 34B1, 34B18. Keywords and phrases. Multipoint boundary value problem, positive solution, fixed point, cone, uniqueness. Research supported by the National Natural Science Foundation of China (Nos. 1171141, 11371221), the Natural Science Foundation of Shandong Province of China (No. ZR211AQ8), the Specialized Research Fund for the Doctoral Program of Higher Education of China (Nos. 212375111, 212375124), a Project of Shandong Province Higher Educational Science and Technology Program (No. J11LA6), Foundation of Qufu Normal University (No. BSQD2113) and the Program for Scientific Research Innovation Team in Colleges and Universities of Shandong Province. The first author is the corresponding author. Received by the editors on December 25, 28, and in revised form on January 5, 211. DOI:1.1216/RMJ-213-43-4-1161 Copyright c 213 Rocky Mountain Mathematics Consortium 1161

1162 X. HAO, N. XU AND L. LIU (H 1 ) f :[, 1 R + R + is continuous; (H 2 ) α, β R and α<2π 2, β α 2 /4, (α/π 2 )+(β/π 4 ) < 1. Multi-point boundary value problems (BVPs) for ordinary differential equations arise in a variety of areas of applied mathematics and physics. For instance, the vibrations of a guy wire of uniform cross-section and composed of N parts of different densities can be set up as a multi-point BVP in [12; also, many problems in the theory of elastic stability can be handled by multi-point problems in [13. Recently, the existence and multiplicity of positive solutions for second order or fourth order multipoint BVPs has been studied by many authors using the nonlinear alternative of Leray-Schauder, the method of upper and lower solutions, the Leggett-Williams fixed point theorem, the fixed point index theory and the fixed point theorem of expansion and compression type (in terms of norms) in a cone, see [2 5, 9, 11, 15 19 and references therein. In 23, Li [7 studied the existence of positive solutions for fourth order BVP with two parameters (1.2) { u (4) + αu βu = f(t, u), <t<1, u() = u(1) = u () = u (1) =. Under the assumptions of (H 1 ), (H 2 ) and some other conditions, the existence results of positive solutions are obtained. Very recently, in [1 the fourth order m-point BVP (1.1) is considered under some conditions concerning the first eigenvalue of the relevant linear operator. The existence of positive solutions is studied by means of fixed point index theory for the case: lim inf u + min t [,1 f(t, u) u >λ, lim sup u + max t [,1 f(t, u) u <λ, where λ is the first eigenvalue of the relevant linear operator. However, for another case: lim sup u + max t [,1 f(t, u) u <λ, lim inf min u + t [,1 f(t, u) u >λ, the author still does not know whether a similar result holds. Ma also established the uniqueness result for a kind of special case: f(t, u) = h(t)u r, r (, 1).

POSITIVE SOLUTIONS OF m-point BVPS 1163 Motivated by the above works and [1, 8, 14, we will consider the fourth-order m-point BVPs (1.1). It is clear that, when a i = b i =, BVP (1.1) reduces to the BVP (1.2). In this paper, we establish the existence and uniqueness of positive solutions for BVP (1.1). The main tool used in the proofs of existence results is a fixed point theorem in a cone due to Krasnoselskii and Zabreiko. We remark that our methods are entirely different from those used in [2 5, 7, 9 11, 15 19. Our results extend and complement corresponding ones in [7, 1. By a positive solution of (1.1), we mean a function u which is positive on (, 1) and satisfies the differential equation as well as the boundary conditions in (1.1). The following fixed point theorem in a cone, due to Krasnoselskii and Zabreiko [6, is of crucial importance in our proofs. Lemma 1.1. Let E be a real Banach space and W aconeofe. Suppose that A : (B R \B r ) W W is a completely continuous operator with <r<r,whereb ρ = {x E : x <ρ} for ρ>. If either: (1) Au u for each u B r W and Au u for each u B R W, or (2) Au u for each u B r W and Au u for each u B R W, then A has at least one fixed point on (B R \B r ) W. The remainder of this paper is organized as follows. In Section 2, we give some preliminary lemmas. Section 3 is devoted to the existence of positive solutions for (1.1). Section 4 is concerned with the uniqueness of positive solution for (1.1). 2. Preliminaries and lemmas. To state and prove the main result of this paper, we need the following lemmas. Lemma 2.1 [7. Let (H 2 ) hold. Then there exist unique ϕ 1,ϕ 2,ψ 1, and ψ 2 satisfying:

1164 X. HAO, N. XU AND L. LIU { ϕ 1 + λ 1 ϕ 1 =, ϕ 1 () =, ϕ 1 (1) = 1; { ψ 1 + λ 2 ψ 1 =, ψ 1 () =, ψ 1 (1) = 1; { ϕ 2 + λ 1 ϕ 2 =, ϕ 2 () = 1, ϕ 2 (1) = ; { ψ 2 + λ 2 ψ 2 =, ψ 2 () = 1, ψ 2 (1) =, respectively. And, on [, 1, ϕ 1,ϕ 2,ψ 1,ψ 2, ϕ 1 () >, ψ 1 () >, where λ 1,λ 2 are the roots for the polynomial equation λ 2 + αλ β =. That is, λ 1 = α + α 2 +4β 2, λ 2 = α α 2 +4β. 2 Notation. Set ρ 1 = ϕ 1 (), ρ 2 = ψ 1 (), Δ 1 = Σ a iϕ 1 (ξ i ) Σ a iϕ 2 (ξ i ) 1 Σ b iϕ 1 (ξ i ) 1 Σ b iϕ 2 (ξ i ), Δ 2 = Σ a iψ 1 (ξ i ) Σ a iψ 2 (ξ i ) 1 Σ b iψ 1 (ξ i ) 1 Σ b iψ 2 (ξ i ). (2.1) (2.2) G 1 (t, s) = 1 ρ 1 { ϕ1 (t)ϕ 2 (s) t s 1, ϕ 1 (s)ϕ 2 (t) s t 1, G 2 (t, s) = 1 ρ 2 { ψ1 (t)ψ 2 (s) t s 1, ψ 1 (s)ψ 2 (t) s t 1. Then G 1 (t, s) andg 2 (t, s) are the Green s function of the following linear BVP { u { (t)+λ 1 u(t) =, u (t)+λ 2 u(t) =, and u() = u(1) = ; u() = u(1) =, respectively.

POSITIVE SOLUTIONS OF m-point BVPS 1165 Lemma 2.2 [7. G i (t, s), ϕ i (t) and ψ i (t) (i =1, 2) have the following properties: (i) G i (t, s) >, t, s (, 1); (ii) G i (t, s) C i G i (s, s) for t, s [, 1; ϕ 1 (t) C 1, ϕ 2 (t) C 1, ψ 1 (t) C 2, ψ 2 (t) C 2 for t [, 1; (iii) G i (t, s) δ i G i (t, t)g i (s, s) for t, s [, 1; ϕ 1 (t) δ 1 G 1 (t, t), ϕ 2 (t) δ 1 G 1 (t, t), ψ 1 (t) δ 2 G 2 (t, t), ψ 2 (t) δ 2 G 2 (t, t) for t [, 1, where C i =1and δ i = w i /(sinh w i ) if λ i > ; C i =1and δ i =1if λ i =;C i =1/(sin w 1 ) and δ i = w i sin w i if π 2 <λ i < ; w i = λ i. In the rest of the paper, we make the following assumptions: (A 1 )Σ a iϕ 2 (ξ i ) < 1, Σ b iϕ 1 (ξ i ) < 1; (A 2 )Σ a iψ 2 (ξ i ) < 1, Σ b iψ 1 (ξ i ) < 1; (H 3 )Δ 1 < ; (H 4 )Δ 2 <. Lemma 2.3 [1. Let (A 1 ), (H 2 ) and (H 3 ) hold. Then, for any g C[, 1, BVP u + λ 1 u = g(t), <t<1, (2.3) u() = a i u(ξ i ), u(1) = b i u(ξ i ), has a unique solution where u(t) = G 1 (t, s)g(s) ds + A(g)ϕ 1 (t)+b(g)ϕ 2 (t), (2.4) A(g) = 1 Σ a i G 1(ξ i,s)g(s) ds Σ a iϕ 2 (ξ i ) 1 Δ 1 Σ b i G 1(ξ i,s)g(s) ds Σ b iϕ 2 (ξ i ), (2.5) B(g) = 1 Σ a iϕ 1 (ξ i ) Σ a i G 1(ξ i,s)g(s) ds Δ 1 Σ b iϕ 1 (ξ i ) 1 Σ b i G 1(ξ i,s)g(s) ds.

1166 X. HAO, N. XU AND L. LIU Furthermore, if g, the unique solution u of problem (2.3) satisfies u(t) for t [, 1. Lemma 2.4 [1. Let (A 2 ), (H 2 ) and (H 4 ) hold. Then, for any g C[, 1, BVP u + λ 2 u = g(t), <t<1, (2.6) u() = a i u(ξ i ), u(1) = b i u(ξ i ), has a unique solution where (2.7) (2.8) u(t) = C(g) = 1 Δ 2 Σ G 2 (t, s)g(s) ds + C(g)ψ 1 (t)+d(g)ψ 2 (t), a i G 2(ξ i,s)g(s) ds Σ b i G 2(ξ i,s)g(s) ds a iψ 2 (ξ i ) 1 Σ b iψ 2 (ξ i ), Σ D(g) = 1 Σ a iψ 1 (ξ i ) Σ a i G 2(ξ i,s)g(s) ds Δ 2 Σ b iψ 1 (ξ i ) 1 Σ b i G 2(ξ i,s)g(s) ds. Furthermore, if g, the unique solution u of problem (2.6) satisfies u(t) for t [, 1. Now notice that u (4) (t)+αu (t) βu(t) = so we can easily get: ( )( ) d2 dt 2 + λ 1 d2 dt 2 + λ 2 u(t), Lemma 2.5 [1. Let (H 2 ), (H 3 ) and (H 4 ) hold. Then, for any g C[, 1, BVP u (4) + αu βu = g(t), <t<1, (2.9) u() = a i u(ξ i ), u(1) = b i u(ξ i ), u () = a i u (ξ i ), u (1) = b i u (ξ i ),

POSITIVE SOLUTIONS OF m-point BVPS 1167 has a unique solution u(t) = = G 2 (t, s)h 2 (s) ds + C(h 2 )ψ 1 (t)+d(h 2 )ψ 2 (t) G 1 (t, s)h 1 (s) ds + A(h 1 )ϕ 1 (t)+b(h 1 )ϕ 2 (t), where G 1, G 2, A(g), B(g), C(g) and D(g) are defined as in (2.1), (2.2), (2.4), (2.5), (2.7) and (2.8), and h 2 (t) = h 1 (t) = G 1 (t, s)g(s) ds + A(g)ϕ 1 (t)+b(g)ϕ 2 (t), G 2 (t, s)g(s) ds + C(g)ψ 1 (t)+d(g)ψ 2 (t). In addition, if (A 1 ) and (A 2 ) hold and g, thenu(t) for t [, 1. Remark 2.1. For any g C[, 1 and g, A(g), B(g), C(g) and D(g) are all linear functions and nondecreasing in g. Let E = C[, 1, u = max u(t), t [,1 P = {u E : u(t), for all t [, 1}; then (E, ) is a Banach space and P is a cone in E. operator T : E E by Define an (2.1) (Tu)(t) = where (ef)(t) = G 2 (t, s)e(f)(s) ds+c(e(f))ψ 1 (t)+d(e(f))ψ 2 (t), G 1 (t, s)f(s, u(s)) ds + A(f)ϕ 1 (t)+b(f)ϕ 2 (t). From Lemma 2.4, we know that u is the nonzero fixed point of T in P equivalent to u is the positive solution of BVP (1.1). In addition, we

1168 X. HAO, N. XU AND L. LIU have from (A 1 ), (A 2 ), (H 1 ), (H 2 ), (H 3 )and(h 4 )thatt : P P is completely continuous. Define the linear operators L, L : E E by (2.11) (Lu)(t) = G 2 (t, s)(l 2 u)(s) ds + C(l 2 (u))ψ 1 (t)+d(l 2 (u))ψ 2 (t), (L u)(t) = where G 1 (t, s)(l 1 u)(s) ds + A(l 1 (u))ϕ 1 (t)+b(l 1 (u))ϕ 2 (t), (2.12) (l 2 u)(t) = (l 1 u)(t) = G 1 (t, s)u(s) ds + A(u)ϕ 1 (t)+b(u)ϕ 2 (t), G 2 (t, s)u(s) ds + C(u)ψ 1 (t)+d(u)ψ 2 (t). It follows from (A 1 ), (A 2 ), (H 2 ), (H 3 )and(h 4 )thatl, L : P P are completely continuous. Lemma 2.6 [1. Suppose that (A 1 ), (A 2 ), (H 2 ), (H 3 ) and (H 4 ) are satisfied. Then, for the operator L defined by (2.11), the spectral radius r(l) and L has a positive eigenfunction φ corresponding to its first eigenvalue λ =(r(l)) 1, that is, φ = λ Lφ. Remark 2.2. It follows from Lemmas 2.5 and 2.6 that L = L,the spectral radius r(l) =r(l )andl, L have the same positive eigenfunction φ corresponding to their first eigenvalue λ =(r(l)) 1 = (r(l )) 1,thatis,φ = λ Lφ, φ = λ L φ. Let C = M = κ = δ 1δ 2 C C 1 C 2 M G 2 (t, t)g 1 (t, t) dt + C(G 1 (s, s)) + D(G 1 (s, s)), G 2 (t, t) dt + C(1) + D(1), G 2 (t, t)φ (t) dt,

POSITIVE SOLUTIONS OF m-point BVPS 1169 and K = { u P : } φ (t)u(t) dt κ u. Clearly, M>, C >, κ>andk is a cone in E. Lemma 2.7. Suppose that (A 1 ), (A 2 ), (H 1 ), (H 2 ), (H 3 ) and (H 4 ) are satisfied. Then L(P ) K (in particular L(K) K). Proof. For u P, by (2.11), (2.12) and Lemmas 2.2 2.4, Lu(t), and (2.13) [ (Lu)(t) C 1 C 2 G 2 (s, s) ds + C(1) + D(1) [ G 1 (s, s)u(s) ds + A(u)+B(u) [ C 1 C 2 M G 1 (s, s)u(s) ds + A(u)+B(u), t [, 1. Then [ Lu C 1 C 2 M G 1 (s, s)u(s) ds + A(u)+B(u). On the other hand, Lemma 2.2 implies that (l 2 u)(t) = G 1 (t, s)u(s) ds + A(u)ϕ 1 (t)+b(u)ϕ 2 (t) [ δ 1 G 1 (t, t) G 1 (s, s)u(s)ds + A(u)+B(u),

117 X. HAO, N. XU AND L. LIU (2.14) [ (Lu)(t) δ 2 G 2 (t, t) [ δ 1 δ 2 G 2 (t, t) G 2 (s, s)(l 2 u)(s) ds + C(l 2 u)+d(l 2 u) G 1 (s, s)u(s) ds + A(u)+B(u) [ G 2 (s, s)g 1 (s, s) ds + C(G 1 (s, s)) + D(G 1 (s, s)) [ = δ 1 δ 2 C G 2 (t, t) G 1 (s, s)u(s) ds + A(u)+B(u), t [, 1. So (2.13) and (2.14) imply that and thus, Hence, L(P ) K. (Lu)(t) δ 1δ 2 C C 1 C 2 M G 2(t, t) Lu, t [, 1, φ (t)(lu)(t) dt κ Lu. Remark 2.3. It follows, by Lemma 2.7, that T (P ) K and, in particular, T (K) K. Remark 2.4. If u is the fixed point of T in P,then u(t) δ 1δ 2 C C 1 C 2 M G 2(t, t) u, t [, 1, so u(t) >, t (, 1). 3. Existence results for the BVP (1.1). Theorem 3.1. and (3.1) Assume that (A 1 ), (A 2 ), (H 1 ) (H 4 ) are satisfied, f = lim inf min x + t [,1 f = lim sup x + max t [,1 f(t, x) x f(t, x) x >λ, <λ,

POSITIVE SOLUTIONS OF m-point BVPS 1171 where λ is the first eigenvalue of L defined by (2.11). Then the BVP (1.1) has at least one positive solution. Proof. Byf >λ,thereexistsanε 1 > such that f(t, x) (λ + ε 1 )x, t 1, when x is sufficiently large. R 1 > such that By the continuity of f, thereexistsan f(t, x) (λ + ε 1 )x R 1, t [, 1, x [, + ). Then, for any u P and t [, 1, (ef)(t) = G 1 (t, s)f(s, u(s)) ds + A(f)ϕ 1 (t)+b(f)ϕ 2 (t) (λ + ε 1 )(l 2 u)(t) R 1 (l 2 1)(t), and (3.2) (Tu)(t) = G 2 (t, s)e(f)(s) ds + C(e(f))ψ 1 (t)+d(e(f))ψ 2 (t) (λ + ε 1 )(Lu)(t) R 1 (L1)(t), where 1 stands for the constant function 1(t) 1. Let Ω 1 = {u K : u Tu}. We shall show that Ω 1 is a bounded subset of K. Indeed, if u Ω 1, (3.2) implies that u(t) (λ + ε 1 )(Lu)(t) R 1 (L1)(t). Multiplying by φ (t) on both sides and integrating over [, 1, then using integration by parts and exchanging the integral sequence, we

1172 X. HAO, N. XU AND L. LIU deduce that u(t)φ (t) dt (λ + ε 1 ) (Lu)(t)φ (t) dt R 1 (L1)(t)φ (t) dt =(λ + ε 1 ) R 1 (Lφ )(t) dt =(λ + ε 1 ) u(t)(l φ )(t) dt r(l)u(t)φ (t) dt R 1 r(l)φ (t) dt, and so u(t)φ (t) dt R 1 φ (t) dt. ε 1 Hence, u R 1 φ (t) dt, for all u Ω 1. κε 1 This proves the boundedness of Ω 1. Taking r 1 > sup u Ω1 u, wesee that (3.3) Tu u, u B r1 K. By f <λ,thereexist<ε 2 <λ and <r 2 <r 1 such that f(t, x) (λ ε 2 )x, t [, 1, x [,r 2 ; then, for any u B r2 P and t [, 1, and (ef)(t) (λ ε 2 )(l 2 u)(t), (3.4) (Tu)(t) (λ ε 2 )(Lu)(t).

POSITIVE SOLUTIONS OF m-point BVPS 1173 Let Ω 2 = {u B r2 K : u Tu}. We claim that Ω 2 = {}. Indeed, if u Ω 2, by (3.4), u(t) (λ ε 2 )(Lu)(t), t [, 1. Multiplying by φ (t) on both sides and integrating over [, 1, then using integration by parts and exchanging the integral sequence, we get u(t)φ (t) dt (λ ε 2 ) (Lu)(t)φ (t) dt =(λ ε 2 ) =(λ ε 2 ) u(t)(l φ )(t) dt r(l)u(t)φ (t) dt, and so u(t)φ (t) dt =. u K implies u =. This proves that Ω 2 = {}. Therefore, (3.5) Tu u, u B r2 K. Now, (3.3) and (3.5) together with Lemma 1.1 imply that T has at least one fixed point u (B r1 \ B r2 ) K. From Remark 2.4, we know that u is a positive solution of BVP (1.1). Theorem 3.2. and (3.6) Assume that (A 1 ), (A 2 ), (H 1 ) (H 4 ) are satisfied, f = lim inf x + f = lim sup x + f(t, x) min >λ, x f(t, x) max <λ, t [,1 x t [,1 where λ is the first eigenvalue of L defined by (2.11). Then the BVP (1.1) has at least one positive solution. Proof. Sincef >λ, there exist ε 3 > andr 3 > such that f(t, x) (λ + ε 3 )x, t [, 1, x [,r 3.

1174 X. HAO, N. XU AND L. LIU Then, for any u B r3 P and t [, 1, we have and (ef)(t) (λ + ε 3 )(l 2 u)(t), (3.7) (Tu)(t) (λ + ε 3 )(Lu)(t). Let Ω 3 = {u B r3 K : u Tu}. We shall show that Ω 3 = {}. Indeed, if u Ω 3, it follows from (3.7) that u(t) (λ + ε 3 )(Lu)(t), t [, 1. Multiplying by φ (t) on both sides and integrating over [, 1, then using integration by parts on the right side and exchanging the integral sequence, we have u(t)φ (t) dt (λ + ε 3 ) =(λ + ε 3 ) =(λ + ε 3 ) (Lu)(t)φ (t) dt u(t)(l φ )(t) dt r(l)u(t)φ (t) dt, and so u(t)φ (t) dt =. u K implies that u =. This proves that Ω 3 = {}. Hence, (3.8) Tu u, u B r3 K. On the other hand, by f <λ,thereexistsa<ε 4 <λ such that f(t, x) (λ ε 4 )x, t 1, when x is sufficiently large. R 2 > such that By the continuity of f, thereexistsan f(t, x) (λ ε 4 )x + R 2, t [, 1, x [, + ).

POSITIVE SOLUTIONS OF m-point BVPS 1175 Then, for any u P and t [, 1, (ef)(t) (λ ε 4 )(l 2 u)(t)+r 2 (l 2 1)(t), and (3.9) Let (Tu)(t) (λ ε 4 )(Lu)(t)+R 2 (L1)(t). Ω 4 = {u K : u Tu}. We are going to prove that Ω 4 is a bounded subset of K. Indeed, if u Ω 4, (3.9) implies that u(t) (λ ε 4 )(Lu)(t)+R 2 (L1)(t). Multiplying by φ (t) on both sides and integrating over [, 1, then using integration by parts on the right side and exchanging the integral sequence, we get u(t)φ (t) dt (λ ε 4 ) (Lu)(t)φ (t) dt + R 2 (L1)(t)φ (t) dt =(λ ε 4 ) + R 2 (Lφ )(t) dt =(λ ε 4 ) u(t)(l φ )(t) dt r(l)u(t)φ (t) dt + R 2 r(l)φ (t) dt, and so u(t)φ (t) dt R 2 φ (t) dt ε 4.

1176 X. HAO, N. XU AND L. LIU u K implies that u R 2 φ (t) dt, for all u Ω 4. κε 4 This proves the boundedness of Ω 4. Taking r 4 > max{sup u Ω4 u, r 3 }, we then conclude that (3.1) Tu u, u B r4 K. Hence, (3.8), (3.1) and Lemma 1.1 imply that T has at least one fixed point on (B r4 \ B r3 ) K, which is a positive solution of BVP (1.1). Remark 3.1. In [1, the author only established the existence of positive solution for f >λ, f <λ, but for another case f <λ, f >λ, it is still not known whether a similar result holds. Hence, Theorem 3.1 is new for [1. Remark 3.2. λ = π 4 απ 2 β when a i = b i =(i =1, 2,...,). In such a case, Theorems 3.1 and 3.2 are reduced to Theorem 1.1 in [7, so our results extend the main results of [7. Moreover, our method is different from [7, 1. Remark 3.3. Since λ is the first eigenvalue of the linear problem corresponding to BVP (1.1), the positive result cannot be guaranteed when the strict inequality is weakened to nonstrict inequality. So our result is optimal. 4. Uniqueness of positive solutions for BVP (1.1). Theorem 4.1. Assume that (A 1 ), (A 2 ), (H 1 ) (H 4 ) are satisfied, and (F 1 ) for every t [, 1, f(t, ) is increasing in R +. (F 2 ) There is a function γ :(, 1) R + such that, for every t [, 1, x>, λ (, 1), one has γ(λ) >λand f(t, λx) γ(λ)f(t, x). Then the BVP (1.1) has at most one positive solution.

POSITIVE SOLUTIONS OF m-point BVPS 1177 Proof. Suppose u 1 and u 2 are two positive solutions of BVP (1.1). Then where Set u i (t) = (ef i )(t) = G 2 (t, s)e(f i )(s) ds + C(e(f i ))ψ 1 (t)+d(e(f i ))ψ 2 (t), G 1 (t, s)f i (s, u(s)) ds + A(f i )ϕ 1 (t)+b(f i )ϕ 2 (t), f i (t, u) =f(t, u i ), i =1, 2. μ =sup{μ >:u 1 (t) μu 2 (t), t [, 1}. Then μ > andu 1 (t) μ u 2 (t), t [, 1. We claim that μ 1. Suppose the contrary, μ < 1. Using (F 1 )and(f 2 ), for any t [, 1, we deduce that (ef 1 )(t) = G 1 (t, s)f(s, u 1 (s)) ds + A(f 1 )ϕ 1 (t)+b(f 1 )ϕ 2 (t) G 1 (t, s)f(s, μ u 2 (s)) ds + A(f )ϕ 1 (t)+b(f )ϕ 2 (t) [ γ(μ ) G 1 (t, s)f(s, u 2 (s)) ds + A(f 2 )ϕ 1 (t)+b(f 2 )ϕ 2 (t) = γ(μ )(ef 2 )(t), where Then f (t, u) =f(t, μ u 2 ). u 1 (t) = G 2 (t, s)e(f 1 )(s) ds + C(e(f 1 ))ψ 1 (t)+d(e(f 1 ))ψ 2 (t) [ γ(μ ) G 2 (t, s)e(f 2 )(s) ds + C(e(f 2 ))ψ 1 (t)+d(e(f 2 ))ψ 2 (t) = γ(μ )u 2 (t),

1178 X. HAO, N. XU AND L. LIU contradicting the definition of μ. So, μ 1andu 1 (t) μ u 2 (t) u 2 (t) fort [, 1. Similarly, u 2 (t) u 1 (t), t [, 1. Consequently, u 1 = u 2. Corollary 4.1. Assume that (A 1 ), (A 2 ), (H 1 ) (H 4 ), (F 1 ), (F 2 ) and (3.6) are satisfied. Then BVP (1.1) has exactly one positive solution. Remark 4.1. If we let f(t, u) =h(t)u r, r (, 1), it is easy to check that f(t, u) satisfies the conditions (F 1 )and(f 2 )forγ(λ) =λ r.then [1, Theorem 4.1 becomes a special case of Theorem 4.1 presented in this paper. Obviously, our result includes and improves the uniqueness result of [1. 5. Examples. We give three explicit examples to illustrate our main results. Example 5.1. We consider the BVP (1.1) with m =3,a 1 =, b 1 =1/2, ξ 1 =1/4, α = 1, β =andf(t, x) =μt ln(1 + x) +x 2, fixing μ>sufficiently small. BVP (1.1) becomes the fourth order three-point BVP { u (4) (t) u (t) =μt ln(1 + u(t)) + u 2 (t), t (, 1), (5.1) u() =, u(1) = 1 2 u( 1 4 ), u () =, u (1) = 1 2 u ( 1 4 ). For this case, ϕ 1 (t) =e/(e 2 1)e t e/(e 2 1)e t, ϕ 2 (t) = 1/(e 2 1)e t +e 2 /(e 2 1)e t, ψ 1 (t) =t, ψ 2 (t) =1 t, b 1 ϕ 1 (ξ 1 )=(1/2)e/(e 2 1)(e 1/4 e 1/4 ) < 1, b 1 ψ 1 (ξ 1 )=(1/2) (1/4) < 1, Δ 1 =(1/2)e/(e 2 1)(e 1/4 e 1/4 ) 1 <, Δ 2 = 7/8. On the other hand, by direct calculations, we can obtain that f =, f = μ. Therefore, the assumptions of Theorem 3.1 are satisfied. Thus, BVP (5.1) has at least one positive solution. Example 5.2. Consider the BVP (1.1) with m =3,a 1 =,b 1 =1/2, ξ 1 =1/4, α = 1, β =andf(t, x) =x 2 e x + μ sin x, fixingμ> sufficiently large. BVP (1.1) becomes the fourth order three-point BVP { u (4) (t) u (t) =u 2 (t)e u(t) + μ sin u(t), t (, 1), (5.2) u() =, u(1) = 1 2 u( 1 4 ), u () =, u (1) = 1 2 u ( 1 4 ).

POSITIVE SOLUTIONS OF m-point BVPS 1179 For this case, b 1 ϕ 1 (ξ 1 ) < 1, b 1 ψ 1 (ξ 1 ) < 1, Δ 1 =(1/2)e/(e 2 1)(e 1/4 e 1/4 ) 1 <, Δ 2 = 7/8. By direct calculations, we have f = μ, f =. Therefore, the assumptions of Theorem 3.2 are satisfied. Thus, BVP (5.2) has at least one positive solution. Example 5.3. Consider the BVP (1.1) with m =3,a 1 =,b 1 =1/2, ξ 1 =1/4, α = 1, β =andf(t, x) =(2+t)x 1/2. BVP (1.1) becomes the fourth order three-point BVP { u (4) (t) u (t) =(2+t)u 1/2 (t), t (, 1), (5.3) u() =, u(1) = 1 2 u( 1 4 ), u () =, u (1) = 1 2 u ( 1 4 ). For this case, b 1 ϕ 1 (ξ 1 ) < 1, b 1 ψ 1 (ξ 1 ) < 1, Δ 1 =(1/2)e/(e 2 1)(e 1/4 e 1/4 ) 1 <, Δ 2 = 7/8. By direct calculations, we have f =, f =. It is easy to check that f(t, x) satisfies the conditions (F 1 ) and (F 2 )forγ(λ) =λ 1/2, λ (, 1). Therefore, the assumptions of Corollary 4.1 are satisfied. Thus, BVP (5.3) has a unique positive solution. Acknowledgments. The authors are grateful to the referees for their valuable suggestions and comments. REFERENCES 1. A.R. Aftabizadeh, Existence and uniqueness theorems for fourth-order boundary value problems, J. Math. Anal. Appl. 116 (1986), 415 426. 2. C. Bai, D. Yang and H. Zhu, Existence of solutions for fourth order differential equation with four-point boundary conditions, Appl. Math. Lett. 2 (27), 1131 1136. 3. J.R. Graef, C. Qian and B. Yang, A three point boundary value problem for nonlinear fourth order differential equations, J. Math. Anal. Appl. 287 (23), 217 233. 4. C.P. Gupta, Solvability of a three-point nonlinear boundary value problem for a second order ordinary differential equation, J.Math.Anal.Appl.168 (1992), 54 551. 5. C.P. Gupta and S.I. Trofimchuk, Existence of a solution of a three-point boundary value problem and the spectral radius of a related linear operator, Nonlin. Anal. 34 (1998), 489 57. 6. M.A. Krasnoselskii and B.P. Zabreiko, Geometrical methods of nonlinear analysis, Springer-Verlag, Berlin, 1984. 7. Y. Li, Positive solutions of fourth-order boundary value problems with two parameters, J. Math. Anal. Appl. 281 (23), 477 484.

118 X. HAO, N. XU AND L. LIU 8. B. Liu, Positive solutions of fourth-order two-point boundary value problems, Appl. Math. Comp. 148 (24), 47 42. 9. X. Liu, J. Qiu and Y. Guo, Three positive solutions for second-order m-point boundary value problems, Appl. Math. Comp. 156 (24), 733 742. 1. H. Ma, Positive solution for m-point boundary value problems of fourth order, J. Math. Anal. Appl. 321 (26), 37 49. 11. R. Ma, Positive solutions for a nonlinear three-point boundary value problem, Electr. J. Diff. Equat. 34 (1999), 1 8. 12. M. Moshinsky, Sobre los problemas de condiciones a la frontiera en una dimension de caracteristicas discontinuas, Bol. Soc. Mat. Mex. 7 (195), 1 25. 13. S. Timoshenko, Theory of elastic stability, McGraw-Hill, New York, 1961. 14. R.A. Usmani, A uniqueness theorem for a boundary value problem, Proc. Amer. Math. Soc. 77 (1979), 329 335. 15. J.R.L. Webb, Positive solutions of some three-point boundary value problems via index theory, Nonlin. Anal. 47 (21), 4319 4332. 16. X. Xu, Existence and multiplicity of positive solutions for multi-parameter three-point differential equations system, J. Math. Anal. Appl. 324 (26), 472 49. 17. G. Zhang and J. Sun, Positive solutions of m-point boundary value problems, J. Math. Anal. Appl. 291 (24), 46 418. 18. X. Zhang and L. Liu, Positive solutions of four-order four-point boundary value problems with p-laplacian operator, J. Math. Anal. Appl. 336 (27), 1414 1423. 19., Positive solutions of four-order multi-point boundary value problems with bending term, Appl. Math. Comp. 194 (27), 321 332. School of Mathematical Sciences, Qufu Normal University, Qufu 273165, Shandong, P.R. China Email address: haoxinan24@163.com Shandong Water Conservation Professional Institute, Rizhao 276826, Shandong, P.R. China Email address: dahai9@126.com School of Mathematical Sciences, Qufu Normal University, Qufu 273165, Shandong, P.R. China and Department of Mathematics and Statistics, Curtin University of Technology, Perth, WA 6845, Australia Email address: lls@mail.qfnu.edu.cn