Loading [MathJax]/jax/element/mml/optable/BasicLatin.js
Research article Special Issues

Positive solutions to a semipositone superlinear elastic beam equation

  • A semipositone fourth-order two-point boundary value problem is considered. In mechanics, the problem describes the deflection of an elastic beam rigidly fastened on the left and simply supported on the right. Under some conditions concerning the first eigenvalue corresponding to the relevant linear operator, the existence of nontrivial solutions and positive solutions to this boundary value problem is obtained. The main results are obtained by using the topological method and the fixed point theory of superlinear operators.

    Citation: Haixia Lu, Li Sun. Positive solutions to a semipositone superlinear elastic beam equation[J]. AIMS Mathematics, 2021, 6(5): 4227-4237. doi: 10.3934/math.2021250

    Related Papers:

    [1] Minggang Xia, Xingyong Zhang, Danyang Kang, Cuiling Liu . Existence and concentration of nontrivial solutions for an elastic beam equation with local nonlinearity. AIMS Mathematics, 2022, 7(1): 579-605. doi: 10.3934/math.2022037
    [2] Ebrahem A. Algehyne, Nifeen Hussain Altaweel, Mounirah Areshi, Faizan Ahmad Khan . Relation-theoretic almost ϕ-contractions with an application to elastic beam equations. AIMS Mathematics, 2023, 8(8): 18919-18929. doi: 10.3934/math.2023963
    [3] Xiao Wang, D. D. Hai . On a class of one-dimensional superlinear semipositone (p,q) -Laplacian problem. AIMS Mathematics, 2023, 8(11): 25740-25753. doi: 10.3934/math.20231313
    [4] Keyu Zhang, Yaohong Li, Jiafa Xu, Donal O'Regan . Nontrivial solutions for a fourth-order Riemann-Stieltjes integral boundary value problem. AIMS Mathematics, 2023, 8(4): 9146-9165. doi: 10.3934/math.2023458
    [5] Filippo Dell'Oro, Claudio Giorgi, Vittorino Pata . Steady states of elastically-coupled extensible double-beam systems. AIMS Mathematics, 2017, 2(1): 28-69. doi: 10.3934/Math.2017.1.28
    [6] Limin Guo, Jiafa Xu, Donal O'Regan . Positive radial solutions for a boundary value problem associated to a system of elliptic equations with semipositone nonlinearities. AIMS Mathematics, 2023, 8(1): 1072-1089. doi: 10.3934/math.2023053
    [7] Sung Woo Choi . Explicit characteristic equations for integral operators arising from well-posed boundary value problems of finite beam deflection on elastic foundation. AIMS Mathematics, 2021, 6(10): 10652-10678. doi: 10.3934/math.2021619
    [8] Sabbavarapu Nageswara Rao, Abdullah Ali H. Ahmadini . Multiple positive solutions for system of mixed Hadamard fractional boundary value problems with (p1,p2)-Laplacian operator. AIMS Mathematics, 2023, 8(6): 14767-14791. doi: 10.3934/math.2023755
    [9] Hasanen A. Hammad, Hassen Aydi, Mohra Zayed . Involvement of the topological degree theory for solving a tripled system of multi-point boundary value problems. AIMS Mathematics, 2023, 8(1): 2257-2271. doi: 10.3934/math.2023117
    [10] Shulin Zhang . Existence of nontrivial positive solutions for generalized quasilinear elliptic equations with critical exponent. AIMS Mathematics, 2022, 7(6): 9748-9766. doi: 10.3934/math.2022543
  • A semipositone fourth-order two-point boundary value problem is considered. In mechanics, the problem describes the deflection of an elastic beam rigidly fastened on the left and simply supported on the right. Under some conditions concerning the first eigenvalue corresponding to the relevant linear operator, the existence of nontrivial solutions and positive solutions to this boundary value problem is obtained. The main results are obtained by using the topological method and the fixed point theory of superlinear operators.



    The purpose of this paper is to investigate the existence of nontrivial solutions and positive solutions to the following nonlinear fourth-order two-point boundary value problem

    {u(4)(t)=λf(t,u(t)), 0t1,u(0)=u(0)=u(1)=u(1)=0,        (P)

    where λ is a positive parameter, f:[0,1]×RR is continuous. One function uC[0,1] is called a positive solution of problem (P) if u is a solution of (P) and u(t)>0,0<t<1.

    Fourth-order two-point boundary value problems appear in beam analysis (See [1,2,3,4,5,6,7,8,9,10,11,12,13,14,15]). The deflection of an elastic beam rigidly fastened on the left and simply supported on the right leads to problem (P). The existence and multiplicity of positive solutions for the elastic beam equations have been studied extensively when the nonlinear term satisfies

    f(t,u)0, u0, (1)

    see for example [1,2,3,6,7,8,9,11] and references therein. Agarwal and Chow [1] investigated the problem (P) by using contraction mapping and iterative methods. Bai [3] applied upper and lower solution method and Yao [11] used Guo-Krasnosel'skii fixed point theorem of cone expansion-compression type. However there are only a few papers concerned with the non-positone or semipositone elastic beam equations. Yao [12] considered the existence of positive solutions of semipositone elastic beam equations by using a special cone and the fixed point theorem of cone expansion-compression type. In this paper, we assume that there exists a constant b>0 such that f(t,u)b for all t[0,1] and u(t)R, which implies that problem (P) is semipositone. We obtain the existence of nontrivial solutions and positive solutions to the semipositone boundary value problem (P) under some conditions concerning the first eigenvalue corresponding to the relevant linear operator by the topological method and the fixed point theory of superlinear operators.

    In this section, we give some preliminaries required in our subsequent discussions.

    Let E be an ordered Banach space in which the partial ordering is induced by a cone PE, θ the zero element of E. For the concepts and properties about the cone we refer to [16,17,18].

    Lemma 1. (see [19]). Let ΩE be a bounded open set, θΩ, and A:¯ΩE a completely continuous operator. If

    Axμx, xΩ,μ1,

    then deg (IA,Ω,θ)=1.

    Lemma 2. (see [19]). Let ΩE be a bounded open set and A:¯ΩE a completely continuous operator. If there exists u0E{θ} such that

    uAuμu0, uΩ,μ0,

    then deg(IA,Ω,θ)=0.

    The following is the famous Leray-Schauder theorem.

    Lemma 3. (see [18]). Let ΩE be a bounded open set, θΩ and A:¯ΩE a completely continuous operator with Aθ=θ. Suppose that the Frechet derivative Aθ of A at θ exists and 1 is not an eigenvalue of Aθ. Then there exists r0>0 such that for any 0<r<r0,

    deg(IA,Tr,θ)=deg(IAθ,Tr,θ)=(1)κ,

    where κ is the sum of algebraic multiplicities for all eigenvalues of Aθ lying in the interval (0, 1) and Tr={xEx<r}.

    Let G(t,s) be the Green's function of homogeneous linear problem u(4)(t)=0,u(0)=u(0)=u(1)=u(1)=0, which can be explicitly given by

    G(t,s)={112(1t)s2[3(1s)(1t)2(3s)],  0st1,112(1s)t2[3(1t)(1s)2(3t)],  0ts1. (2)

    By (2) and Yao [13] we have

    (G1) G(t,s)=G(s,t), 0t,s1;

    (G2) 16s2(1s)t2(1t)G(t,s)14t2(1t), 0t,s1;

    By (G1)(G2),

    G(t,s)16s2(1s)t2(1t)14s2(1s)23t2(1t)14s2(1s)34t3(1t)234t3(1t)2G(s,τ)=34t3(1t)2G(τ,s),

    and we have

    (G3) G(t,s)34t3(1t)2G(τ,s), 0t,s,τ1.

    It is well known that the problem (P) is equivalent to the integral equation

    u(t)=λ10G(t,s)f(s,u(s))ds.

    Let

    (Au)(t)=10G(t,s)f(s,u(s))ds, (3)
    (Bu)(t)=10G(t,s)u(s)ds. (4)

    By the famous Krein-Rutmann theorem (See [16]) and similar to Lemma 3 in [20], we have

    Lemma 4. Suppose that the linear operator B is defined by (4). Then the spectral radius r(B)0 and B has a positive eigenfunction corresponding to its first eigenvalue λ1=r1(B).

    In the sequel we always take E=C[0,1] with the norm u=max and the cone P = \{u\in C[0, 1]\mid u(t)\geq 0, 0\leq t\leq 1\} .

    Now let us list the following conditions which will be used in this paper:

    (H _{1} ) There exists a constant \alpha > 0 satisfying

    \begin{align} \liminf\limits_{u\rightarrow +\infty}\frac{f(t, u)}{u}\geq\alpha, \ \forall t\in[0, 1]. \end{align} (5)

    (H _{2} ) There exists a constant \beta with 0 < \beta < \alpha satisfying

    \begin{align} \limsup\limits_{u\rightarrow 0}\left|\frac{f(t, u)}{u}\right|\leq\beta, \ \forall t\in[0, 1]. \end{align} (6)

    (H _{3} ) There exists a constant b > 0 such that

    \begin{align} f(t, u)\geq -b, \ \forall t\in[0, 1], u(t)\in \mathbb{R}. \end{align} (7)

    (H _{4} ) \lim\limits_{u\rightarrow +\infty} \frac{f(t, u)}{u} = +\infty .

    Theorem 5. Suppose that (H _{1} )-(H _{3} ) hold. Then for any \lambda\in(\frac{\lambda_{1}}{\alpha}, \frac{\lambda_{1}}{\beta}) , problem (P) has at least one nontrivial solution, where \lambda_{1} = r^{-1}(B) and B is given by (4).

    Proof. Let (Fu)(t) = f(t, u(t)) for all u\in E , then by (4), A = BF. Applying the Arzela-Ascoli theorem and a standard argument, we can prove that (\lambda A): E\rightarrow E is a completely continuous operator. It is known to all that the nonzero fixed points of the operator \lambda A are the nontrivial solutions of the boundary value problem (P).

    Now we show that there exists R_{0} > 0 , such that for any R > R_{0} ,

    \begin{align} u-\lambda Au\neq\mu u^{*}, \forall\mu\geq 0, u\in \partial T_{R}, \end{align} (8)

    where u^{*} is the positive eigenfunction of B corresponding to its first eigenvalue \lambda_{1} = r^{-1}(B) , T_{R} = \{u\in C[0, 1]\mid \|u\| < R\} is a bounded open subset of E .

    If (8) is not true, then there exist \mu _{0} > 0 (if \mu_{0} = 0 , then Theorem 5 holds) and u_{0}\in \partial T_{R} such that

    \begin{align} u_{0}-\lambda Au_{0} = \mu_{0} u^{*}. \end{align} (9)

    By (5) and (6), there exist l > 0, r_{0} > 0 and 0 < \varepsilon < {\rm min}\{\frac{\alpha\lambda-\lambda_{1}}{\lambda}, \frac{\lambda_{1}-\beta\lambda}{\lambda}\} such that

    \begin{align} f(t, u)\geq(\alpha-\varepsilon)u-l, \ \forall t\in [0, 1], u\geq 0, \end{align} (10)
    \begin{align} |f(t, u)|\leq(\beta+\varepsilon)|u|, \ \forall t\in [0, 1], |u|\leq r_{0}. \end{align} (11)

    It follows from (7) that

    \begin{align} f(t, u)\geq -b\geq(\alpha-\varepsilon)u-b, \forall t\in [0, 1], u\leq 0. \end{align} (12)

    Take \omega = {\rm max}\{b, l\} , then by (10) and (12) we have f(t, u)\geq(\alpha-\varepsilon)u-\omega, \forall t\in [0, 1], u\in R. Let \alpha_{1} = \alpha-\varepsilon , then

    \begin{align} \lambda Au = \lambda BFu\geq\lambda\alpha_{1}Bu-u_{1}, \ \forall u\in E, \end{align} (13)

    where u_{1} = \int_{0}^{1}G(t, s)\lambda\omega ds . Take

    \delta = \frac{2}{3}\lambda_{1}\int_{0}^{1}s^{2}(1-s)u^{*}(s)ds > 0.

    By (4) and (G _{2} ) we have

    \begin{array}{llllll} u^{*}(t)& = \lambda_{1}Bu^{*}(t) = \lambda_{1} \int_{0}^{1}G(t, s)u^{*}(s)ds\\ &\geq\lambda_{1} \int_{0}^{1}\frac{1}{6}s^{2}(1-s)t^{2}(1-t)u^{*}(s)ds\\ &\geq \frac{1}{6}\lambda_{1}t^{2}(1-t) \int_{0}^{1}s^{2}(1-s)u^{*}(s)ds\\ &\geq \frac{2}{3}\lambda_{1}G(t , s) \int_{0}^{1}s^{2}(1-s)u^{*}(s)ds\\ & = \delta G(t, s). \end{array}

    Let

    P_{1} = \left\{u\in P\mid\int_{0}^{1}u^{*}(t)u(t)dt\geq\lambda_{1}^{-1}\delta\|u\|\right\}.

    For any u\in P , we have

    \begin{array}{llll} \int_{0}^{1}u^{*}(t)(\lambda Bu)(t)dt& = \int_{0}^{1}u^{*}(t)[\int_{0}^{1}G(t, s)\lambda u(s)ds]dt\\ & = \int_{0}^{1}\lambda u(s)[\int_{0}^{1}G(s, t)u^{*}(t)dt]ds\\ & = \int_{0}^{1}\lambda u(s)Bu^{*}(s)ds\\ &\geq\lambda_{1}^{-1}\delta \int_{0}^{1}G(t, s)\lambda u(s)ds\\ & = \lambda_{1}^{-1}\delta(\lambda Bu)(t).\\ \end{array}

    And so \ \int_{0}^{1}u^{*}(t)(\lambda Bu)(t)dt\geq\lambda_{1}^{-1}\delta\|\lambda Bu\| , i.e., \ (\lambda B)(P)\subset P_{1}. Since Fu_{0}+\omega\in P , then (\lambda B)(Fu_{0}+\omega)\in P_{1} and \mu_{0}u^{*} = \mu_{0}\lambda_{1}Bu^{*}\in P_{1} . By (9) we have u_{0}+\lambda B\omega = \lambda Au_{0}+\mu_{0}u^{*}+\lambda B\omega = \lambda B(Fu_{0}+\omega)+\mu_{0}u^{*}\in P_{1}. Thus

    \begin{align} \int_{0}^{1}u^{*}(t)(u_{0}+\lambda B\omega)(t)dt \geq\lambda_{1}^{-1}\delta\|u_{0}+\lambda B\omega\|\geq\lambda_{1}^{-1}\delta\|u_{0}\|-\lambda_{1}^{-1}\delta\|\lambda B\omega\|. \end{align} (14)

    Take \varepsilon_{0} = \alpha_{1}\lambda r(B)-1 > 0 . By (13) we have

    \begin{array}{lll} \int_{0}^{1}u^{*}(t)(\lambda Au_{0})(t)dt&\geq& \int_{0}^{1}u^{*}(t)\alpha_{1}(\lambda Bu_{0})(t)dt- \int_{0}^{1}u^{*}(t)u_{1}(t)dt\\ & = &\alpha_{1}\lambda r(B) \int_{0}^{1}u^{*}(t)u_{0}(t)dt- \int_{0}^{1}u^{*}(t)u_{1}(t)dt\\ & = & \int_{0}^{1}u^{*}(t)u_{0}(t)dt+\varepsilon_{0} \int_{0}^{1}u^{*}(t)(u_{0}+\lambda B\omega)(t)dt\\ && -\varepsilon_{0} \int_{0}^{1}u^{*}(t)(\lambda B\omega)(t)dt - \int_{0}^{1}u^{*}(t)u_{1}(t)dt.\\ \end{array}

    Take R_{0} = \frac{\lambda_{1}}{\varepsilon_{0}\delta}[\varepsilon_{0}\delta\lambda_{1}^{-1}\|\lambda B\omega\| +\varepsilon_{0}\int_{0}^{1}u^{*}(t)(\lambda B\omega)(t)dt+\int_{0}^{1}u^{*}(t)u_{1}(t)dt]. For any \|u_{0}\| = R > R_{0} , by (14) we have

    \begin{array}{lll} \int_{0}^{1}u^{*}(t)(\lambda Au_{0}-u_{0})(t)dt&\geq&\varepsilon_{0}[\lambda_{1}^{-1}\delta\|u_{0}\|-\lambda_{1}^{-1}\delta\|\lambda B\omega\|]\\ &&-\varepsilon_{0} \int_{0}^{1}u^{*}(t)(\lambda B\omega)(t)dt- \int_{0}^{1}u^{*}(t)u_{1}(t)dt\\ & > &\varepsilon_{0}\delta\lambda_{1}^{-1}R_{0}-\varepsilon_{0}\delta\lambda_{1}^{-1}\|\lambda B\omega\|\\ &&-\varepsilon_{0} \int_{0}^{1}u^{*}(t)(\lambda B\omega)(t)dt- \int_{0}^{1}u^{*}(t)u_{1}(t)dt\\ & = &0.\\ \end{array}

    But we see from (9) that

    \int_{0}^{1}u^{*}(t)(u_{0}-\lambda Au_{0})(t)dt = \int_{0}^{1}u^{*}(t)\mu_{0}u^{*}(t)dt\geq 0,

    which is a contradiction. So (8) is true. By Lemma 2 we have

    \begin{align} {\rm deg}(I-\lambda A, T_{R}, \theta) = 0. \end{align} (15)

    Next we show that

    \begin{align} (\lambda A)u\neq \mu u, \ \forall u\in\partial T_{r}, \mu\geq 1, \end{align} (16)

    where 0 < r < {\rm min}\{r_{0}, R_{0}\} . Assume on the contrary that there exist u_{0}\in\partial T_{r} and \mu_{0}\geq 1 such that (\lambda A)u_{0} = \mu_{0}u_{0} . Since \lambda A has no fixed point on \partial T_{r} , we have \mu_{0} > 1 . Let B_{1} = \lambda(\beta+\varepsilon)B , then r(B_{1}) < 1 . By (11) we have |\lambda Au_{0}|\leq(\beta+\varepsilon)\lambda B|u_{0}| = B_{1}|u_{0}| , then \mu_{0}|u_{0}|\leq B_{1}|u_{0}|, and therefore

    \begin{align} \mu_{0}^{n}|u_{0}|\leq B_{1}^{n}|u_{0}|. \end{align} (17)

    Let D = \{v\mid v\geq|u_{0}|\} . It follows from (17) that \{\mu_{0}^{-n}B_{1}^{n}|u_{0}|\mid n = 1, 2, \cdots\}\subset D. And u_{0}\in\partial T_{r} and \theta\in T_{r} imply that d = d(\theta, D) > 0 . Then one can have that

    \|B_{1}^{n}\|\geq\frac{1}{\|u_{0}\|}\|B_{1}^{n}u_{0}\|\geq\frac{d}{\|u_{0}\|}\mu_{0}^{n}, \qquad n = 1, 2, \cdots,

    which shows

    r(B_{1}) = \lim\limits_{n\rightarrow\infty}(\|B_{1}^{n}\|)^{1/n} \geq\lim\limits_{n\rightarrow\infty}\left(\frac{d}{\|u_{0}\|}\mu_{0}^{n}\right)^{1/n} = \mu_{0} > 1.

    This contradicts r(B_{1}) < 1. So (16) holds. By Lemma 1, we have

    \begin{align} {\rm deg}(I-\lambda A, T_{r}, \theta) = 1. \end{align} (18)

    By (15) and (18), we have

    {\rm deg}(I-\lambda A, T_{R}\backslash\overline{T_{r}}, \theta) = {\rm deg}(I-\lambda A, T_{R}, \theta)-{\rm deg}(I-\lambda A, T_{r}, \theta) = -1.

    Then \lambda A has at least one fixed point in T_{R}\backslash\overline{T_{r}} . This means that problem (P) has at least one nontrivial solution.

    Remark 6. For any \alpha > 0 and \beta = 0, the proof of Theorem 5 is valid. Then for any \lambda > 0 , problem (P) has at least one nontrivial solution.

    Corollary 7. Suppose that (H _{1} ) and (H _{2} ) hold. Assume there exists a constant b^{*} > 0 such that

    f(t, u)\geq-\frac{b^{*}}{K}, \ \forall t\in[0, 1], u\geq-\frac{\lambda_{1}b^{*}}{\beta},

    where K = \max_{t\in[0, 1]}\int_{0}^{1}G(t, s)ds . Then for any \lambda\in(\frac{\lambda_{1}}{\alpha}, \frac{\lambda_{1}}{\beta}) problem (P) has at least one nontrivial solution.

    Proof. Let

    f_{1}(t, u) = \left\{ \begin{array}{lll} f(t, u), & \forall u\geq-\frac{\lambda_{1}b^{*}}{\beta}, & t\in [0, 1], \ \ \ \ \\ f(t, -\frac{\lambda_{1}b^{*}}{\beta}), & \forall u < -\frac{\lambda_{1}b^{*}}{\beta}, & t\in [0, 1], \end{array} \right.
    (A_{1}u)(t) = \int_{0}^{1}G(t, s)f_{1}(s, u(s))ds.

    Then all conditions of Theorem 5 hold for f_{1} . By Theorem 5, \lambda A_{1} has at least one nonzero fixed point v^{*}(t) , and

    v^{*}(t) = \lambda\int_{0}^{1}G(t, s)f_{1}(s, v^{*}(s))ds\geq-\lambda\frac{b^{*}}{K}\int_{0}^{1}G(t, s)ds\geq-\frac{\lambda_{1}b^{*}}{\beta}.

    Thus

    v^{*}(t) = \lambda\int_{0}^{1}G(t, s)f_{1}(s, v^{*}(s))ds = \lambda\int_{0}^{1}G(t, s)f(s, v^{*}(s))ds = \lambda Av^{*}(t).

    This indicates v^{*}(t) is a nontrivial solution of problem (P).

    Theorem 8. Suppose that (H _{1} ) and (H _{3} ) hold. Let f(t, 0)\equiv 0, \forall t\in[0, 1] and

    \begin{align} \lim\limits_{u\rightarrow 0}\frac{f(t, u)}{u} = \rho. \end{align} (19)

    Then for any \lambda\in(\frac{\lambda_{1}}{\alpha}, +\infty) and \lambda\neq\frac{\lambda_{1}}{\rho} , problem (P) has at least one nontrivial solution.

    Proof. From the proof of Theorem 5, if (H _{1} ) and (H _{3} ) hold, then there exists R_{0} > 0 such that for any R > R_{0} and \lambda > \frac{\lambda_{1}}{\alpha} (15) holds.

    Since f(t, 0)\equiv 0, \forall t\in[0, 1] , then A\theta = \theta . By (19) we have that the Frechet derivative A_{\theta}^{'} of A at \theta exists and (A_{\theta}^{'}u)(t) = \int_{0}^{1}G(t, s)\rho u(s)ds. Notice that \lambda\neq\frac{\lambda_{1}}{\rho} , then 1 is not an eigenvalue of \lambda A_{\theta}^{'} . By Lemma 3 there exists r_{0} > 0 , for any 0 < r < {\rm min}\{r_{0}, R_{0}\} ,

    \begin{align} {\rm deg}(I-\lambda A, T_{r}, \theta) = {\rm deg}(I-\lambda A_{\theta}^{'}, T_{r}, \theta) = (-1)^{\kappa}\neq 0, \end{align} (20)

    where \kappa is the sum of algebraic multiplicities for all eigenvalues of \lambda A_{\theta}^{'} lying in the interval (0, 1).

    By (15) and (20) \lambda A has at least one nonzero fixed point. Thus problem (P) has at least one nontrivial solution.

    In many realistic problems, the positive solution is more significant. In this section we will study this question.

    Theorem 9. Suppose that (H _{4} ) holds. Then there exists \lambda^{*} > 0 such that for any 0 < \lambda < \lambda^{*} problem (P) has at least one positive solution.

    Proof. Let D = [0, 1], D_{0} = [t_{1}, t_{2}]\subset(0, 1)\subset D, \eta = \min_{t_{1}\leq t\leq t_{2}}\frac{3}{4}t^{3}(1-t)^{2} > 0 . By (H _{4} ) there exist b_{1} > 0 and R_{1} > 0 such that

    f(t, u)\geq -b_{1}, \forall t\in[0, 1], u\geq 0,
    \begin{align} f(t, u)\geq \eta^{-1}Nb_{1}, \forall t\in[t_{1}, t_{2}], u\geq R_{1}, \end{align} (21)

    where N > \frac{1-(t_{2}-t_{1})}{t_{2}-t_{1}} is a natural number. Let

    f_{2}(t, u) = \left\{ \begin{array}{ll} f(t, u), & u\geq 0, \\ f(t, -u), & u < 0. \end{array} \right.

    Then

    \begin{align} f_{2}(t, u)\geq -b_{1}, \forall t\in[0, 1], u\in \mathbb{R}. \end{align} (22)

    Let

    (A_{2}u)(t) = \int_{0}^{1}G(t, s)f_{2}(s, u(s))ds.

    Obviously, A_{2}: E\rightarrow E is a completely continuous operator.

    From Remark 6 and the proof of Theorem 5 there exists R_{0} > 0 , for any R > R_{0} ,

    \begin{align} {\rm deg}(I-\lambda A_{2}, T_{R}, \theta) = 0, \forall\lambda > 0. \end{align} (23)

    Take 0 < r < R_{0} . Let m = \max_{0\leq t\leq 1, |u|\leq r}|f_{2}(t, u)| , M = \max_{0\leq s, t\leq 1}G(t, s) , \overline{\lambda} = r(mM)^{-1} . For any 0 < \lambda < \overline{\lambda}, u\in \partial T_{r} , we have

    \|\lambda A_{2}u\| = \max\limits_{0\leq t\leq 1}|\int_{0}^{1}\lambda G(t, s)f_{2}(s, u(s))ds| < \overline{\lambda}Mm = r = \|u\|.

    Thus

    \begin{align} {\rm deg}(I-\lambda A_{2}, T_{r}, \theta) = 1, \ \forall \ 0 < \lambda < \overline{\lambda}. \end{align} (24)

    From (23) and (24) we have that for any 0 < \lambda < \overline{\lambda} , there exists u_{\lambda}\in C[0, 1] with \|u_{\lambda}\| > r such that u_{\lambda} = \lambda A_{2}u_{\lambda}. Now we show

    \begin{align} \lim\limits_{\lambda\rightarrow 0^{+}, \ u_{\lambda} = \lambda A_{2}u_{\lambda}, \ \|u_{\lambda}\| > r}\|u_{\lambda}\| = +\infty. \end{align} (25)

    In fact, if (25) doesn't hold, then there exist \lambda_{n} > 0, u_{\lambda_{n}}\in C[0, 1] such that \lambda_{n}\rightarrow 0, r < \|u_{\lambda_{n}}\| < c (c > 0 is a constant), and

    \begin{align} u_{\lambda_{n}} = \lambda_{n} A_{2}u_{\lambda_{n}}. \end{align} (26)

    Since A_{2} is completely continuous, then \{u_{\lambda_{n}}\} has a subsequence (assume without loss of generality that it is \{u_{\lambda_{n}}\} ) converging to u_{*}\in C[0, 1]. Let n\rightarrow+\infty in (26), we have u_{*} = \theta , which is a contradiction of \|u_{\lambda_{n}}\| > r > 0 . Then (25) holds.

    Next we show that there exists R = R(\overline{\lambda}) > 0 such that if 0 < \lambda_{0}\leq\overline{\lambda}, \|u_{0}\|\geq R and u_{0} = \lambda_{0}Au_{0} , then u_{0}(t)\geq 0. Take

    \begin{align} R = R(\overline{\lambda}) = {\rm max}\{2\eta^{-1}R_{1}, \ 2\eta^{-1}\overline{\lambda}b_{1}M, \ 2\overline{\lambda}b_{1}M\}. \end{align} (27)

    Assume that 0 < \lambda_{0}\leq\overline{\lambda}, \|u_{0}\|\geq R and u_{0} = \lambda_{0}Au_{0}. Take any \overline{t}\in[t_{1}, t_{2}] , by (G _{3} ) we have that for any \tau\in[0, 1] ,

    \begin{align} \begin{array}{lll} u_{0}(\overline{t})& = \lambda_{0} \int_{0}^{1}G(\overline{t}, s)[f_{2}(s, u_{0}(s))+b_{1}]ds- \lambda_{0} \int_{0}^{1}b_{1}G(\overline{t}, s)ds\\ &\geq\lambda_{0}\eta \int_{0}^{1}G(\tau, s)f_{2}(s, u_{0}(s))ds-\overline{\lambda}b_{1}M\\ & = \eta u_{0}(\tau)-\overline{\lambda}b_{1}M.\\ \end{array} \end{align} (28)

    On account of the continuity of u_{0} , there exists t^{*}\in[0, 1] such that u_{0}(t^{*}) = \|u_{0}\|. Take \tau = t^{*} in (28), by (27) we have

    \begin{array}{lll} u_{0}(\overline{t})&\geq\eta\|u_{0}\|-\overline{\lambda}b_{1}M\geq\eta R-\overline{\lambda}b_{1}M\\ & = \frac{\eta}{2} R+\frac{\eta}{2} R-\overline{\lambda}b_{1}M\geq\frac{\eta}{2} R\geq R_{1}.\\ \end{array}

    Thus u_{0}(\overline{t})\geq R_{1}, for any \overline{t}\in[t_{1}, t_{2}] . By (21) we have

    \begin{align} f_{2}(s, u_{0}(s))\geq \eta^{-1}Nb_{1}, \forall s\in D_{0} = [t_{1}, t_{2}]. \end{align} (29)

    It follows from (G _{1} ) and (G _{3} ) that for any s\in[t_{1}, t_{2}] and t, \tau\in[0, 1]

    \begin{align} G(t, s) = G(s, t)\geq\eta G(\tau, t) = \eta G(t, \tau). \end{align} (30)

    Take D_{i}\subset D \ (i = 1, 2, \cdots, N) such that {\rm mes}D_{i} = {\rm mes}D_{0} , \bigcup\limits_{i = 1}^{N}D_{i}\supset D\backslash D_{0} . By (29) and (30) we have that for any t\in D, s\in D_{0}, \tau\in D_{i} \ (i = 1, 2, \cdots, N),

    \begin{align} \frac{1}{N}G(t, s)f_{2}(s, u_{0}(s))\geq b_{1}G(t, \tau). \end{align} (31)

    Notice that {\rm mes}D_{i} = {\rm mes}D_{0}(i = 1, 2, \cdots, N), then

    \frac{1}{N}\int_{D_{0}}G(t, s)f_{2}(s, u_{0}(s))ds\geq\int_{D_{i}}b_{1}G(t, \tau)d\tau, \ i = 1, 2, \cdots, N.

    Thus

    \begin{align} \begin{array}{lll} \int_{D_{0}}G(t, s)f_{2}(s, u_{0}(s))ds&\geq\sum\limits_{i = 1}^{N} \int_{D_{i}}b_{1}G(t, \tau)d\tau\\ &\geq \int_{D\backslash D_{0}}b_{1}G(t, \tau)d\tau\\ & = \int_{D\backslash D_{0}}b_{1}G(t, s)ds.\\ \end{array} \end{align} (32)

    By (32) and (21) we have that for any t\in[0, 1] ,

    u_{0}(t) = \lambda_{0}\int_{D_{0}}G(t, s)f_{2}(s, u_{0}(s))ds+\lambda_{0}\int_{D\backslash D_{0}}G(t, s)f_{2}(s, u_{0}(s))ds\geq 0.

    For R in (27), by (25), there exists \lambda^{*} < \overline{\lambda} such that if 0 < \lambda\leq\lambda^{*}, \|u_{\lambda}\|\geq r and u_{\lambda} = \lambda A_{2}u_{\lambda}, then \|u_{\lambda}\|\geq R , thus u_{\lambda}(t)\geq 0. By the definition of A_{2} and f_{2} we have

    u_{\lambda}(t) = \lambda\int_{0}^{1}G(t, s)f_{2}(s, u_{\lambda}(s))ds = \lambda\int_{0}^{1}G(t, s)f(s, u_{\lambda}(s))ds = \lambda Au_{\lambda}(t).

    So u_{\lambda}(t) is a positive solution of problem (P).

    Remark 10. In Theorem 9 we obtain the existence of positive solutions for the semipositone boundary value problem (P) without that assuming (1) holds.

    Remark 11. Since we only study the existence of positive solutions for the boundary value problem (P), which is irrelevant to the value of f(t, u) when u\leq 0 , we only suppose that f(t, u) is bounded below when u\geq 0 . The nonlinear term f(t, u) may be unbounded from below when u\leq 0 .

    Example 12. Consider the fourth-order boundary value problem

    \left\{ \begin{array}{cc} u^{(4)}(t) = \lambda \left[(\sqrt{t}+1)u^{3}-\sqrt[3]{u}\right], \ 0\leq t\leq 1, \\ u(0) = u^{'}(0) = u(1) = u^{''}(1) = 0.\ \ \ \ \ \ \ \ \ \ \ \end{array} \right.\ \ \ \ \ \ \ \ ({\rm P_{1}})

    In this example, f(t, u) = (\sqrt{t}+1)u^{3}-\sqrt[3]{u} , then

    \lim\limits_{u\rightarrow +\infty}\frac{f(t, u)}{u} = \lim\limits_{u\rightarrow +\infty}\left((\sqrt{t}+1)u^{2}-\frac{1}{\sqrt[3]{u^{2}}}\right) = +\infty,

    which means that (H _{4}) holds. By Theorem 9 there exists \lambda^{*} > 0 such that for any 0 < \lambda < \lambda^{*} BVP ({\rm P_{1}}) has at least one positive solution.

    Remark 13. In Example 12, the nonlinear term f doesn't satisfy (1) and (H _{3}) , but the existence of positive solutions of BVP ({\rm P_{1}}) is obtained by using our result.

    This work is supported by the Foundation items: National Natural Science Foundation of China (11501260), Natural Science Foundation of Jiangsu Higher Education Institutions(18KJB180027), and Postgraduate Research Innovation Program of Jiangsu Province(KYCX20_82).

    The authors declare that there is no conflict of interests regarding the publication of this article.



    [1] R. P. Agarwal, Y. M. Chow, Iterative method for fourth order boundary value problem, J. Comput. Appl. Math., 10 (1984), 203–217. doi: 10.1016/0377-0427(84)90058-X
    [2] Z. B. Bai, H. Y. Wang, On positive solutions of some nonlinear four-order beam equations, J. Math. Anal. Appl., 270 (2002), 357–368. doi: 10.1016/S0022-247X(02)00071-9
    [3] Z. B. Bai, The upper and lower solution method for some fourth-order boundary value problems, Nonlinear Anal.-Theor.,, 67 (2007), 1704–1709.
    [4] G. Bonanno, B. D. Bella, A boundary value problem for fourth-order elastic beam equations, J. Math. Anal. Appl., 343 (2008), 1166–1176. doi: 10.1016/j.jmaa.2008.01.049
    [5] G. Bonanno, B. D. Bella, D. O'Regan, Non-trivial solutions for nonlinear fourth-order elastic beam equations, Comput. Math. Appl., 62 (2011), 1862–1869. doi: 10.1016/j.camwa.2011.06.029
    [6] R. Graef, B. Yang, Positive solutions of a nonlinear fourth order boundary value problem, Communications on Applied Nonlinear Analysis, 14 (2007), 61–73.
    [7] C. P. Gupta, Existence and uniqueness results for the bending of an elastic beam equation, Appl. Anal., 26 (1988), 289–304. doi: 10.1080/00036818808839715
    [8] P. Korman, Uniqueness and exact multiplicity of solutions for a class of fourth-order semilinear problems, P. Roy. Soc. Edinb. A, 134 (2004), 179–190. doi: 10.1017/S0308210500003140
    [9] B. D. Lou, Positive solutions for nonlinear elastic beam models, International Journal of Mathematics and Mathematical Sciences, 27 (2001), 365–375. doi: 10.1155/S0161171201004203
    [10] R. Y. Ma, L. Xu, Existence of positive solutions of a nonlinear fourth-order boundary value problem, Appl. Math. Lett., 23 (2010), 537–543. doi: 10.1016/j.aml.2010.01.007
    [11] Q. L. Yao, Positive solutions for eigenvalue problems of four-order elastic beam equations, Appl. Math. Lett., 17 (2004), 237–243. doi: 10.1016/S0893-9659(04)90037-7
    [12] Q. L. Yao, Existence of n solutions and/or positive solutions to a semipositone elastic beam equation, Nonlinear Anal.-Theor., 66 (2007), 138–150. doi: 10.1016/j.na.2005.11.016
    [13] Q. L. Yao, positive solutions of nonlinear elastic beam equation rigidly fastened on the left and simply supported on the right, Nonlinear Anal.-Theor., 69 (2008), 1570–1580. doi: 10.1016/j.na.2007.07.002
    [14] C. B. Zhai, R. P. Song, Q. Q. Han, The existence and the uniqueness of symmetric positive solutions for a fourth-order boundary value problem, Comput. Math. Appl., 62 (2011), 2639–2647. doi: 10.1016/j.camwa.2011.08.003
    [15] X. P. Zhang, Existence and iteration of monotone positive solutions for an elastic beam equation with a corner, Nonlinear Anal.-Real, 10 (2009), 2097–2103. doi: 10.1016/j.nonrwa.2008.03.017
    [16] K. Deimling, Nonlinear Functional Analysis, Springer-Verlag, Berlin-Heidelberg-Newyork, 1985.
    [17] D. J. Guo, V. Lakshmikanthan, Nonlinear Problems in Abstract Cones, Academic press, San Diego, 1988.
    [18] D. J. Guo, Nonlinear Functional Analysis, second edn., Shandong Science and Technology Press, Jinan, 2001 (in Chinese).
    [19] H. Amann, Fixed point equations and nonlinear eigenvalue problems in ordered Banach spaces, SIAM Rev., 18 (1976), 620–709. doi: 10.1137/1018114
    [20] G. W. Zhang, J. X. Sun, Positive solutions of m-point boundary value problems, J. Math. Anal. Appl., 291 (2004), 406–418. doi: 10.1016/j.jmaa.2003.11.034
  • This article has been cited by:

    1. Limin Guo, Jiafa Xu, Donal O'Regan, Positive radial solutions for a boundary value problem associated to a system of elliptic equations with semipositone nonlinearities, 2023, 8, 2473-6988, 1072, 10.3934/math.2023053
  • Reader Comments
  • © 2021 the Author(s), licensee AIMS Press. This is an open access article distributed under the terms of the Creative Commons Attribution License (http://creativecommons.org/licenses/by/4.0)
通讯作者: 陈斌, bchen63@163.com
  • 1. 

    沈阳化工大学材料科学与工程学院 沈阳 110142

  1. 本站搜索
  2. 百度学术搜索
  3. 万方数据库搜索
  4. CNKI搜索

Metrics

Article views(2992) PDF downloads(258) Cited by(1)

Other Articles By Authors

/

DownLoad:  Full-Size Img  PowerPoint
Return
Return

Catalog