APP下载

Multiple solutions for a class of fourth-order two-point boundary value problems

2017-09-11XILijingHUANGYisheng

关键词:四阶临界点边值问题

XI Lijing,HUANG Yisheng

(1.School of Mathematics and Physics,SUST,Suzhou 215009,China;2.School of Mathematical Sciences,Soochow University,Suzhou 215006,China)

Multiple solutions for a class of fourth-order two-point boundary value problems

XI Lijing1,2,HUANG Yisheng2

(1.School of Mathematics and Physics,SUST,Suzhou 215009,China;2.School of Mathematical Sciences,Soochow University,Suzhou 215006,China)

We studied the existence of multiple solutions for the following fourth-order two-point boundary value problem u(4)(x)=λf(u(x)),x∈(0,1);u(0)=u(1)=u″(0)=u″(1)=0,where λ>0 is a parameter and f:R→R is a continuous function.By means of variational methods and the theory of critical points,we obtained the results on the existence of two nonzero solutions and infinitely multiple solutions

fourth-order boundary value problem;(PS)condition;critical point

1 Introduction and main results

Owing to the importance of higher order differential equations in Physics,existence and multiplicity of solutions for problems like the following fourth-order boundary value problem

have been extensively studied,see for examples,[1-6]and the references therein.In particular,in[1],by applying the Morse theory,the authors obtained two nonzero solutions in C4[0,1]for Problem (1) if f∈C1satisfying f(0)=0,there exists a natural number m≥1 such that m4π4<f′(0)<(m+1)4π4,and there exist α and β∈R1with α<π4/2 such thatIn[2],the authors got a multiplicity result of solutions for Problem(1),where to guarantee the existence of three solutions by using the Amann three-solution theorem the authors assumed that f is increasing and imposed some conditions to ensure the existence of sub-solutions and super-solutions.In addition,in[1]and[3],the authors obtained the existence of infinitely many solutions for Problem (1)if f is an odd function.And in[4],by employing the cone expansion or compression fixed point theorem,authors also got two results on existence of infinitely many solutions for a problem like (1) with a parameter if f is monotone increasing.

In the present paper,motivated by[3,7-8],we consider existence of multiple solutions for the boundary value problemwhere λ>0 is a parameter and f:R→R is a continuous function.Firstly,by using the Mountain Pass Theorem (cf.[9]),we establish result on the existence of two nonzero solutions of Problem (2) when λ=1.Then,based on a general variational principle of Ricceri,we prove that Problem (2) has infinitely many solution if the nonlinear term f has an suitable oscillation condition.Here we introduce more general conditions on f to guarantee two nonzero the existence of solutions which are different from that on f in[1-2] (see Theorems 1 for more details).Moreover,unlike in[1],[3]and[4],to obtain the existence of infinitely many solutions for Problem (2),we assume that f is neither odd nor monotone (see Theorems 2 and 3 for more details).

Theorem 1 Assume that the following conditions hold:

(H1)there exist constants θ∈(0,1) and α>0 such that|f(s)|≤α(1+|s|θ) for all s∈R;

(H2)there exist three positive constants a,b,l withsuch that F(s)≥0 for all s∈(0,b),and F(s)≥ls2for all s∈(a,b);

Then,Problem (1) has at least two nontrivial solutions in C4[0,1].

Note that(H3) implies that f(0)=0,then u(x)=0 is a solution of Problem (1),which is called the trivial solution.

Theorem 2 Assume that the following conditions hold:

(H4)F(s)≥0 for all s>0,and

(H5)there exists a sequence{bn}in (0,∞) withand bn→∞ such that f(s)≤0 for all s∈,bn],and

Then

(i)if B<+∞,for every λ>π4/(2B),Problem (2) has infinitely many solutions in C4[0,1];

(ii)if B=+∞,for every λ>0,Problem (2) has infinitely many solutions in C4[0,1].

Theorem 3 Assume that the following conditions hold:

(H6)F(s)≥0 for all s>0,and

(H7)there exists a sequence{bn}in (0,∞) withbn+1<bnand bn→0 such that f(s)≤0 for all s∈,bn],and

Then

(i)if B<+∞,for every λ>π4/B,Problem (2) has infinitely many solutions in C4[0,1]whose norms converge to 0;

(ii)if B=+∞,for every λ>0,Problem (2) has infinitely many solutions in C4[0,1]whose norms converge to 0.

2 Variational setting

In this section,we will see that Problem (2)has a variational structure.Let C[0,1]denote the usual real Banach space with the norm||u||C=maxx∈[0,1]|u(x)|.By L2[0,1]we denote the real Hilbert space with the inner productand the induced norm

It is well known that to find solutions of Problem (2) in C4[0,1]is equivalent to find solutions of the following integral equation in C[0,1]

where G:[0,1]×[0,1]→[0,1]is the Green function for-u″(x)=0 (x∈[0,1]) subject to u(0)=u(1)=0,i.e.

It is easy to see that G is nonnegative,continuous and max(x,s)∈[0,1]×[0,1]G(x,s)=1/4.

We respectively define operators K:L2[0,1]→C[0,1]and T:C[0,1]→C[0,1]as

and

Then to find solutions of integral equation (3) in C[0,1]is equivalent to find solutions of operator equation

in C[0,1].For the linear operator K,we have the following properties.

Lemma 1 (cf.[3])

(i)K:L2[0,1]→C[0,1]is a linear completely continuous operator,and K:L2[0,1]→L2[0,1]is also a linear completely continuous operator;

(ii)(Ku,v)=(u,Kv) for all u and v in L2[0,1];

(iii)the operator equation u=λK2Tu has a solution in C[0,1]if and only if the operator equation v=λKTKv has a solution in L2[0,1].

Remark 1

(a)By the definition of K,we can obtain that Ku≠0 for all u∈L2[0,1]with u≠0.Therefore,Ku1≠Ku2for all u1,u2∈L2[0,1]if u1≠u2.

(b)It is easy to see that the eigenvalues of K2are 1/(n4π4) and the corresponding eigenvectors are sinnπx,n=1,2,…

(c)K2is also linear compact,symmetric and the norm||K2||=1/π4.Moreover,we have||K||=1/π2and

Remark 2 For any u∈L2[0,1],it follows from the definition K that

Therefore,

(i)I is Fréchet differentiable on C[0,1]and I′(u)w=(Tu,w)for all u,w∈C[0,1];

(ii)I◦K is Fréchet differentiable on L2[0,1]and (I◦K)′(v)=KTKv for all v∈L2[0,1].

Let us introduce a functional J:L2[0,1]→R as

Then,according to Lemma 2,J∈C1(L2[0,1],R) and

It follows from Lemma 1 that Problem (2) has a nontrivial solution in C4[0,1]if and only if the functional J has a nontrivial critical point in L2[0,1].More precisely,if v∈L2[0,1]is a critical point of J,then u=Kv is a solution of Problem (2) in C4[0,1].Thus Problem (2) has a variational structure.

In the end of this section,let us recall the Palais-Smale condition for a functional defined on a Banach space.

Definition 1 Let X be a real Banach space.A functonal J∈C1(X,R) is said to satisfy the Palais-Smale condition (the (PS) condition for short) if any sequence{um}⊂X for which J(um) is bounded and J′(um)→0 as m→∞ possesses a convergent subsequence.

3 Proof of main results

Now we are going to prove our main results.

Proof of Theorem 1 By the preceding section,it is only necessary to prove the functional

has at least two nonzero critical points in L2[0,1].

First of all,we prove that there exists v1∈L2[0,1]such that

By (H1),we obtain

where α′>0 is a constant.Therefore,by using inequalities (7) and (5),for all v∈L2[0,1]we have

In addition,it follows from Lemma 1 that KTK:L2[0,1]→L2[0,1]is completely continuous,therefore J is a weakly semicontinuous functional on L2[0,1].Then we can find an element v1∈L2[0,1]such that

Secondly,we will prove that J(v1)=infv∈L2[0,1]J(v)<0.Indeed,chooseand let v0(x)=t0sinπx for all x∈[0,1],then v0(x)∈(aπ2,bπ2) when x∈(1/4,3/4) and |v0|2=t02/2.Moreover,it follows from Remark 1 that

Then,by (H2),we have

From the choice of l,we know J(v1)=infv∈L2[0,1]J(v)≤J(v0)<0.

Finally,we will use the Mountain Pass Theorem to find the second nonzero critical point of J.For the sake of clarity,we divide the following proof into two steps.

Step 1 We prove that J satisfies the (PS) condition.Let{vn}be a sequence in L2[0,1]such that

We claim that{vn}is bound sequence.In fact,suppose the contrary,then there wound be a subsequence{vni}of{vn}such that||vni||→∞ as i→∞.Since J(v) is coercive,J(vni)→∞ as||vni||→∞ which clearly contradicts the fact that|J(vn)|≤β for all n∈N.Then the properties of K and T guarantee that{vn}has a convergent subsequence.This shows that J satisfies the (PS) condition.

Step 2 We check that the functional J has the Mountain Pass geometry.By (H3),there exists ε∈(0,1)and δ>0 such that

So

Now we choose ρ>0 such that ρ<min{4δ,||v1||}andThen,it follows from (6),(8) and (4)that for all v∈∂Bρ={v∈L2[0,1]:||v||=ρ},

On the other hand,it is obvious from the choice of ρ that||v1||>ρ and J(v1)<0=J(0).Thus,according to the Mountain Pass Theorem (cf.[9,Theorem 2.2]),there exists v2∈L2[0,1]such that J′(v2)=0 and J(v2)≥η>0.Since J(v2)>0>J(v1),it is clear that v2≠v1.The proof is completed.

In the following,we will prove the existence of infinitely many solutions for Problem (2).Our main tool is a general variational principle of Ricceri (cf.[10,Theorem 2.5]),which we recall here,stated in the most suitable form for our purposes.

Lemma 3 Let X be a reflexive real Banach space,and let Φ,Ψ:X→R be two sequentially weakly lower semicontinuous and Gâteaux differentiable functionals.Assume also that Φ is strongly continuous and coercive.For every r>infXΦ,set

Then,the following conclusions holds:

(A)If γ<+∞,then for every λ∈(0,1/γ),either

(A1)Φ-λΨ possesses a global minimum;or

(A2)there exists a sequence{un}of critical points of the functional Φ-λΨ such that

(B)If δ<+∞,then for every λ∈(0,1/δ),either

(B1)Φ-λΨ possesses a local minimum which is also a global minimum of Φ;or

(B2)there exists a sequence{un}of pairwise distinct critical points of theweakly converging to a global minimum of Φ.

Proof of Theorem 2 For every v∈L2[0,1],Φ and Ψ are defined by

thus the energy functional becomes J(v)=Φ(v)-λΨ(v).Then we only need to prove that Φ(v)-λΨ(v) has infinitely many critical points in L2[0,1].

It is obvious that Φ and Ψ are Gâteaux differentiable and sequentially weakly lower semicontinuous.Obviously,Φ is strong continuous and coercive.

In our case,we can define the function φ of Lemma 3 by putting for every r>0

We claim that

Clearly, γ≥0.On the other hand,by (H6),we can deduce that there is a sequence{an}in (0,+∞) with an→+∞such that

Let rn=8an2so that rn→+∞.Then for every u∈L2[0,1]with Φ(u)≤rn,we have

Take v0(x)=0 for all x∈[0,1],then||v0||=0<2rn,it follows from (10) that

This,together with (9),implies

Then γ=0.

In the sequel,we will show that

Indeed,Let wn(x)=π2bnsinπx for all x∈[0,1].Then

Moreover,

Thanks to (H7),F(s) is nonincreasing for,bn].Thus,according to (H6) and (11),we have

Therefore,for n∈N,

If B<+∞,let λ>π4/2B and є∈(π4/2λB,1),then by (H7) there exists Nє∈N such that

for all n>Nє.Note that the choose of є,it follows from (12) that for all n>Nє,

Therefore,the (A1)in Lemma 3 can be excluded.Thus,(A2) holds,i.e.there exists a sequence{vn}of critical points of J=Φ-λΨ such that

If B=+∞,let M>π4/2λ,then by (H7) there exists NM∈N such that

for all n>NM.This yields that for all n>NM,

so that

Then,(A2) of Lemma 3 holds for any λ>0,that is,for every λ>0 there exists a sequence{vn}of critical points of J=Φ-λΨ such that.This concludes the proof.

Proof of Theorem 3 We take Φ,Ψ and wnas in the proof of Theorem 2.In a similar way we can prove

and for n large enough,

Since||wn||→0 as n→∞,0 is not a local minimum of Φ-λΨ.Therefore,the (B2) of Lemma 3 ensures that there exists a sequence{vn}of critical points of J=Φ-λΨ such thatSo Problem (2) has infinitely many solutions in C4[0,1]whose norms converge to 0.

[1]HAN G,XU Z.Multiple solutions of some nonlinear fourth-order beam equations[J].Nonlinear Anal,2008,68:3646-3656.

[2]HAN G,LI F.Multiple solutions of some fourth-order boundary value problems[J].Nonlinear Anal,2007,66:2591-2603.

[3]LI F,ZHANG Q,LIANG Z.Existence and multiplicity of solutions of a kind of fourth-order boundary value problem[J].Nonlinear Anal,2005,62:803-816.

[4]BAI Z,WANG H.On positive solutions of some nonlinear four-order beam equations[J].J Math Anal Appl,2002,270:357-368.

[5]GRAEF J R,QIAN C,YANG B.Multiple symmetric positive solution of a class of boundary value problem for higher order ordinary differential equations[J].Proc Amer Math Soc,2003,131:577-585.

[6]LIU B.Positive solutions of fourth-order two point boundary value problems[J].Appl Math Comput,2004,148:407-420.

[7]KRISTÁLY A,VARGA C.On a class of quasilinear eigenvalue problem in RN[J].Math Nachr,2005,278(15):1756-1765.

[8]KRISTÁLY A,MOROSANU G,TERSIAN S.Quasilinear elliptic problem in RNinvolving oscillatory nonlinearities[J].J Differential Equations,2007,235:366-375.

[9]RABINOWITZ P H.Minimax Methods in Critical Point Theory with Applications to Differential Equations[M].Providence:CMBS Regional Conference Series in Mathematics,Amer Math Soc,1986.

[10]RICCERI B.A general variational principal and some of its applications[J].J Comput Appl Math,2000,113:401-410.

一类四阶两点边值问题的多解性

席莉静1,2, 黄毅生2
(1.苏州科技大学 数理学院,江苏 苏州 215009;2.苏州大学 数学科学学院,江苏 苏州 215006)

研究了四阶两点边值问题 u(4)(x)=λf(u(x)),x∈(0,1);u(0)=u(1)=u″(0)=u″(1)=0 解的存在性,其中λ>0是参数,f:R→R是连续函数。利用变法的方法和临界点理论,获得了两个非平凡解及无穷多个解存在性的结果。

四阶边值问题;(PS)条件;临界点

2016-10-13

国家自然科学基金资助项目(11471235);江苏省自然科学基金资助项目(BK20150281);苏州科技大学自然科学基金资助项目(XKZ201614);苏州科技大学博士科研启动基金资助项目

席莉静(1971-),女,山西大同人,副教授,博士,研究方向:非线性分析。

O175.8MR(2010) Subject Classification:35J35;35J40;58E05

A

2096-3289(2017)03-0008-07

责任编辑:谢金春

猜你喜欢

四阶临界点边值问题
四阶p-广义Benney-Luke方程的初值问题
基于临界点的杭州湾水体富营养化多年变化研究
带有积分边界条件的奇异摄动边值问题的渐近解
具衰退记忆的四阶拟抛物方程的长时间行为
一类具有p-Laplacian算子的分数阶微分方程反周期边值问题解的存在唯一性
超越生命的临界点
非线性m点边值问题的多重正解
四阶累积量谱线增强方法的改进仿真研究
基于四阶累积量和简化粒子群的盲分离算法
Banach空间分数阶微分方程边值问题解的存在性