• No results found

Existence of solutions of elliptic boundary value problems with mixed type nonlinearities

N/A
N/A
Protected

Academic year: 2020

Share "Existence of solutions of elliptic boundary value problems with mixed type nonlinearities"

Copied!
11
0
0

Loading.... (view fulltext now)

Full text

(1)

R E S E A R C H

Open Access

Existence of solutions of elliptic boundary

value problems with mixed type

nonlinearities

Anmin Mao

*

, Yan Zhu and Shixia Luan

*Correspondence: [email protected] School of Mathematical Sciences, Qufu Normal University, Shandong, 273165, P.R. China

Abstract

We study the existence of a nontrivial solution of the following elliptic boundary value problem with mixed type nonlinearities:

u=f(x,u) in

,

u= 0 on

,

wheref(x,u) = –Ku+Wu. We consider the problem in a different case:

lim|u|→∞f(x,u)/u=∞, lim|u|→0f(x,u)/uis some constant. Assuming thatKsatisfies

the “pinching” condition, andWsatisfies a more general superquadratic growth condition than the well-known Ambrosetti-Rabinowitz condition usually used in literature, we obtain a nontrivial solution via the Mountain Pass Lemma.

MSC: 35J65; 35J20; 47J10

Keywords: pinching condition; Mountain Pass Lemma; Cerami condition

1 Introduction

In this paper, we shall be concerned with the elliptic boundary value problem in a different case

⎧ ⎨ ⎩

u=f(x,u) in,

u=  on∂, (P)

whereRN (N> ) is a bounded open domain with a smooth boundaryandfCR,R).

The existence of nontrivial weak solutions for (P) have been studied in many papers, see [–]. Su and Zhao in [] considered problem (P) for resonance case at infinity,

lim|u|→∞f(xu,u)=λk, whereλkis an eigenvalue of the linear boundary value problem

⎧ ⎨ ⎩

u=λu in,

u=  on∂, (P)

(2)

the existence of multiple nontrivial solutions for (P) are obtained by minimax methods and Morse theory. Ambrosetti and Rabinowitz in [] established the existence of a nontrivial solution for problem (P) by assuming the following conditions:

(f) f(x, ) = ,limu→f(xu,u)= , uniformly in a.e.x. (f) There exist two positive constantsaandbsuch that

f(x,u)≤a+b|u|p for some≤p<N+  N– ,∀uR

,x.

And the following well-known Ambrosetti-Rabinowitz condition ((AR) for short):

θ> ,R>  s.t.  <θF(x,u)≤uf(x,u), for all|u| ≥R,x,

whereF(x,u) =uf(x,s)ds.

Since then, the (AR) condition has been used extensively in many literature sources, see [–]. It is well known that the (AR) condition is quite natural and convenient not only to ensure that the Euler-Lagrange functional associated to problem (P) has a mountain pass geometry but also to guarantee that the Palais-Smale sequence of the Euler-Lagrange functional is bounded. LetEbe a Hilbert space andGC(E,R). Recall that the sequence

{un}nNEis said to be a Palais-Smale sequence ofGprovided that{G(un)}is bounded andG(un)→ asn→+∞, the functionGsatisfies the Palais-Smale condition ((PS) for short) if and only if any Palais-Smale sequence forGcontains a convergent subsequence. The functionGsatisfies the Cerami condition ((C) for short) if any sequence{un}nN in E satisfyingG(un) is bounded andG(un)( +un)→ asn→+∞has a convergent subsequence.

Without (AR), it becomes more complicated. Indeed, there are many functions which are superlinear, but it is not necessary to satisfy (AR) even if  <θ ≤. Willem and Zou stated the following examples:

f(x,u) =μ|u|μ–u+ (μ– )|u|μ–usinu+|u|μ–sinu, uR\ {},

whereμ> . Then it is easy to check that (AR) does not hold even for anyθ>μ–  > . On the other hand, in order to verify (AR), it usually is an annoying task to compute a primitive function off and sometimes it is almost impossible. For example,

f(x,u) =|u|u +e(+|sinu|)α+|cosu|α , uR,

whereα> .

Some authors have tried to drop or weaken the above superlinear condition (AR) in recent years, see [–, , ]. Miyagaki and Souto [] adapted some monotonicity argu-ments studying the existence of nontrivial weak solutions of (P).

The aim of the manuscript is to consider the problem in a different case:

lim|u|→∞f(x,u)/u=∞, lim|u|→f(x,u)/u is some constant. We study this problem

(3)

main difficulty when dealing with this problem is the lack of compactness of the Sobolev embedding theorem.

In this paper, here,F(t,u) :=uf(t,s)dsreplaced by –K+W, satisfy

(F) F(x,u) = –K(x,u) +W(x,u),K,W:×R→RareC-maps. (K) There are two positive constantsbandbsuch that

b|u|≤K(x,u)≤b|u|, for all(x,u)∈×R.

(K) There exists∈(, ]such that

K(x,u)≤Ku(x,u)uK(x,u), for all(x,u)∈×R.

(W) W(x,u)≥andWu(x,u) =o(|u|)as|u| →uniformly inx.

(W) W(x,u)/u→ ∞as|u| → ∞uniformly inx.

(W) SetW˜(x,u) := Wu(x,u)uW(x,u),W˜ (x,u) > ifu= , W˜(x,u)→ ∞as|u| → ∞ uniformly in x, and there exist r >  and σ > N/ such that |Wu(x,u)|σcW˜(x,u)|u|σ if|u| ≥r.

We will prove the following results.

Theorem . If assumptions (F), (K), (K) and (W)-(W) are satisfied, then problem (P)

has a nontrivial weak solution.

Remark 

(i) Our assumptions (W), (W) are weaker than (AR), and there is no monotone condition;

(ii) The condition (K) can be written in the form≤KKu((xx,,uu))u,∈(, ]which is weaker than the condition≤KKu((xx,,uu))u≤in [, ].

Example  Consider the functions

K(x,u) = +exp–|x| u, W(x,u) =

 –   +|x|

uln +u .

A straightforward computation shows thatKandW(x,u) satisfy the assumptions of The-orem ., but neitherF(x,u) norW(x,u) satisfy the (AR) condition.

Example  Consider the more general functions

Ku(x,u) =V(x)u, Wu(x,u) =g(x,u),

whereg(x,u) is of superlinear growth as|u| → ∞. A straightforward computation shows thatKandWsatisfy the assumptions of Theorem ..

(4)

2 Preliminary results

LetH:=H

() be the Sobolev space equipped with the inner product and the norm

(u,v) =

(∇u· ∇v+uv)dx, u= (u,u), u,vH.

And we denote the usualLp()-norm

up=

u(x)pdx

p

.

Our approach will be the variational techniques. Define the Euler-Lagrange functional associated to problem (P) given by

(u) =  

|∇u|dx

K(x,u) +W(x,u)dx, for alluH.

From the assumptions onf, it is standard to check that ∈Cwhose Gateaux derivative

is

(u)v=

u· ∇v dx

Ku(x,u)v+Wu(x,u)v

dx, for allu,vH.

Letη:H→[, +∞) be given by

η(u) :=

|∇u|+ K(x,u)dx

.

Hence

(u) =  η

(u) –

W(x,u)dx.

By (K),

(u)u

|∇u|dx+

K(x,u)dx

Wu(x,u)u dx.

By (K) and setb,:=min{, b},b,:=max{, b},

b,u≤η(u)≤b,u.

It is worth pointing out that if the functionK(x,u) is of the form V(x)uwithV(x)

C(,R) andinf

V(x)≥V>  thenηin a Hilbert spaceX={uH();

V(x)u<∞}

is equivalent to the norm · ; however, if the functionK(x,u) is not of the form V(x)u

,

ηis not a norm because of the lack of norm’s linear property.

Lemma . (see []) Let H be a real Banach space,C(H,R), satisfying () = .

Moreover,

(5)

(ii) there existseH\()such that (e)≤.

Then there exists a sequence{un} ∈H such that (un)(+un)→and (un)→cα as n→ ∞.

Lemma .(see []) Assume that||<∞,≤p, r<∞, fC(¯ ×R)and|f(x,u)| ≤ c( +|u|pr). Then, for every uLp(), f(x,u)Lr()and the operator A:Lp()Lr(),

uf(x,u)is continuous.

3 Proofs of theorems

First of all, we recall a property of the functionK(x,u), which is necessary to the proof of the geometric structure of theCfunctional .

Fact  Assume that (K) holds, then

K(x,u)≤K

x, u |u|

|u|, for all xand|u| ≥.

Proof DefineG:sK(x,s–u)s,s(, +∞)

G(s) = –Ku

x,s–u u ss

+

Kx,s–u s–

= –Ku

x,s–u s–us–+Kx,s–u s–

=s––Ku

x,s–u s–u+Kx,s–u .

By (K),G(s)≥, which impliesG(s) is non-decreasing. So, we have

K(x,u) =G()≤G(s) =K

x, u |u|

|u|, if|u|=s≥.

Next we discuss the geometric structure of theCfunctional onH.

Lemma . Under the assumptions of Theorem ., there are constantsρ,α> such that

|∂Bρ()≥α.

Proof From (W) and (W), as|u|>r, we have

Wu(x,u)

σ

c

Wu(x,u)uW(x,u)

|u|σ

cWu(x,u)|u|σ+,

wherecis a positive constant. Hence as|u|>r,

Wu(x,u)≤c|u|σ+/(σ–).

Fromσ>N/, we knowσ+ /(σ– ) < *– , so we can choose σ/(σ– )p< *. And

using (W) again, we observe that for any given>  there isc>  such that

(6)

and

W(x,u)≤|u|+c|u|p. (.)

It follows from (.) and the Sobolev embedding theorem that for alluH

W(x,u)dxu+cuppu+ccup, (.)

wherecis a positive constant. Then combining (K) and (.), we obtain

(u)≥ 

|∇u|dx+bu–

u+ccup

= 

|∇u|dx+ (b–)u–ccup≥min

, (b–)

u–ccup

set b=min{, (b–)}, it is clear thatb> . We chooseu=ρ= (  b

cc)

p– andα=

ρb

, then

(u)≥α.

Lemma . Under the assumptions of Theorem ., there exists eH\()such that

(e) < .

Proof Lete∈H\,M=maxx,|u|≤K(x,u) and A> (M+)e 

e

 . By (W), there exists

B>  such that

W(x,u)≥A|u|–B, for allx,uH. (.)

As< , by Fact , we have

η(ξe) =

|∇ξe|+ K(x,ξe)

dx

|∇e|ξdx+ 

{x;|ξe|≤}K(x,

ξe)dx+ 

{x;|ξe|≥}K(x,

ξe)dx

|∇e|ξdx+ M||+ M

{x;|ξe|≥}

|ξe|dx

|∇e|ξdx+ M||+ M

{x;|ξe|≥}|

ξe|dx

ξ( + M)e+ M||. (.)

Then, by inequalities (.) and (.), we get

(ξeo) =  η

e ) –

W(x,ξe)dx

 + Mξ

e

+M||–e–B||

=

 + Me

Ae 

(7)

By the choice ofA, we have (+Me–Ae) < , so there existsξ∈Rsuch that if

e=ξe, then

(e) < .

Suppose that the assumptions of Theorem . hold, we have Lemma . and Lemma .. Now it follows from Lemma . that there is a sequence{un} ⊂Hsuch that

(u

n) +un → and (un)→cα asn→ ∞. (.)

Lemma . Under the assumptions of Theorem ., the functional satisfies the (C)

con-dition.

Proof Let{un} ⊂Hbe such that

(un) is bounded and

 +un (un)→. (.)

By (K) we observe that for largen,

c≥ (un) –  

(u

n)un

=

K(x,un)dx

Ku(x,un)undx+

Wu(x,un)unW(x,un)dx

K(x,un) –

K(x,un)dx+

Wu(x,un)unW(x,un)dx

˜

W(x,un)dx. (.)

Arguing indirectly, assume as a contradiction thatun → ∞. Settingνn=un/un, then νn=  and since the embedding HLs fors∈[, *), we haveνnsγsνn=γs. Observe that, from (.), (K) and (K)

(u

n)un=

|∇un|dx+

Ku(x,un)undx

Wu(x,un)undx

|∇un|dx+

K(x,un)dx

Wu(x,un)undx

|∇un|dx+

b|un|

dx

Wu(x,un)undx

 |∇un|

dx+

b|un|

dx

Wu(x,un)undx

un

 b,–

Wu(x,unn un

dx

.

It follows that for any>  andnlarge enough,

Wu(x,unn un

dxb,

(8)

Set forr≥

h(r) :=infW˜ (x,u) :xanduRwith|u| ≥r.

By (W),h(r) >  for allr>  andh(r)→ ∞asr→ ∞. For ≤a<b, let

n(a,b) =

x:aun(x)<b

and

Cab=inf

˜ W(x,u)

u

xanduRwithaun(x)<b

.

SinceW˜(x,u) >  ifu= , one hasCb a>  and

˜

Wx,un(x) ≥Cabun(x)

for allxn(a,b).

It follows from (.) that

c≥

˜

W(x,un)dx

=

n(,a)

˜

W(x,un)dx+

n(a,b)

˜

W(x,un)dx+

n(b,∞)

˜

W(x,un)dx

n(,a)

˜

W(x,un)dx+Cab

n(a,b)

un(x)dx+h(b)n(b,∞). (.)

Setτ:= σ/(σ– ), sinceσ>N/, one seesτ∈(, *). Fix arbitrarilyτˆ∈(τ, *), using (.), n(b,∞)≤

c

h(b)→ asb→ ∞uniformly inn,

which implies by the Hölder inequality that

n(b,∞)

|νn|τdx

n(b,∞)

dx –ττ

n(b,∞)

|νn|τ τ τ dx

τ τ

γττn(b,∞)

–τ τ

→ (.)

asb→ ∞uniformly inn. Using (.) again, for any fixed  <a<b,

n(a,b)

|νn|dx=  un

n(a,b)

|un|dxc

Cb

aun → 

asn→ ∞. Let  <<b,

 , by (W), there existsa>  such that

Wu(x,u)≤ γ|

(9)

Consequently,

n(,a)

Wu(x,un) |un| |

νn|dx

n(,a) γ|

νn|dx γ

νn≤ (.)

for alln. By (W) and (.), we can take largebrso that

n(b,∞)

Wu(x,un) |un| |

νn|dx

n(b,∞)

|Wu(x,un)|σ |un|σ

dx

σ

n(b,∞) |νn|σ

dx

σ

cW˜(x,un)dx

σ

n(b,∞) |νn|τdx

τ

. (.)

Hence combining (.), (.) and (.), there isnsuch that

n(b,∞)

|Wu(x,un)||νn| |un|

dx≤(cc) 

σ

n(b,∞) |νn|τdx

τ

< (.)

fornn. Note that there isγ =γ() >  independent ofnsuch that

Wu(x,un)≤γ|un| forxn(a,b).

By (.)

n(a,b)

|Wu(x,un)||νn| |un|

dxγ

n(a,b)

|νn|dx< (.)

for allnn. Therefore, combining (.)-(.), we obtain fornn

|Wu(x,un)||νn| |un|

dx≤<b,  –

which contradicts (.). Hence{un}is bounded inH. Going if necessary to a subsequence, we assume that

unu inHfor someuH,

which impliesunua.e. in, because the imbeddingH()L() is compact. Hence

we haveunu→ and|( (un) – (u))(unu)| →. Using the Hölder inequality

fx,un(x) –f

x,u(x) un(x) –u(x) dx

fx,un(x) –f

x,u(x) q

q

un(x) –u(x)p

p

(p+q = ) forunuinLp(), and by (K), (K), (W), (W) we have

f(x,u)≤c +|u|p– =c +|u|

(10)

Then, by Lemma ., we havef(x,un(x))→f(x,u(x)) inLq(). Thus

fx,un(x) –f

x,u(x) un(x) –u(x) dx→

asn→+∞. Moreover, a straightforward computation shows that

(un) – (u) (unu) =∇(unu)

 –

fx,un(x) –f

x,u(x) un(x) –u(x) dx

it is clear that

(unu)

→. (.)

Finally,

unu→ inH.

This completes the proof.

Now, we are ready to prove Theorem ..

We will obtain a critical point of λby the use of a standard version of the Mountain

Pass Lemma (see []). It provides the minimax characterization for the critical value which is important for what follows. Therefore, we state this lemma precisely.

Lemma .(see []) Let H be a real Banach space and λ:HRbe a C-smooth

functional. If satisfies the following conditions:

(i) () = ,

(ii) every sequence{un}nN inHsuch that{ (un)}nNis bounded inRand (un)→

inHasn→+∞, contains a convergent subsequence ((PS) condition), (iii) there are constantsρ,α> such that |∂Bρ()≥α,

(iv) there is a constanteH\()such that (e)≤,

where Bρ()is an open ball in H of radiusρcentered at , then possesses a critical value

cαgiven by

c=inf

gsmax∈[,]

g(s) ,

where

=gC[, ],H :g() = ,g() =e}.

Now we are ready to give the proofs of Theorem ..

Proof of Theorem . Under conditions (F), (K), (K), (W)-(W), as shown in [], a

(11)

infgmaxs∈[,] (g(s)). Hence,uis a nontrivial solution of problem (P) satisfying (u) =c,

(u) = . The proof is done.

Competing interests

The authors declare that they have no competing interests.

Authors’ contributions

The paper is the result of joint work of all authors who contributed equally to the final version of the paper. All authors read and approved the final manuscript.

Acknowledgements

The authors are grateful to anonymous referees for detailed reading of the manuscript and valuable comments, which helped us improve this work. This work was supported by The National Natural Science Foundation of China (No. 11101237) and SNSFC ZR2012AM006.

Received: 14 May 2012 Accepted: 20 August 2012 Published: 31 August 2012 References

1. Willem, M: Minimax Theorems. Bikhäuser, Boston (1996)

2. Su, JB, Zhao, L: An elliptic resonance problem with multiple solutions. J. Math. Anal. Appl.319, 604-616 (2006) 3. Ambrosetti, A, Rabinowitz, PH: Dual variational methods in critical point theory and applications. J. Funct. Anal.14,

349-381 (1973)

4. Schechter, M, Zou, W: Superlinear problems. Pac. J. Math.214, 145-160 (2004)

5. Costa, DG, Magalhães, CA: Variational elliptic problems which are nonquadratic at infinity. Nonlinear Anal., Theory Methods Appl.23, 1401-1412 (1994)

6. Szulkin, A, Zou, W: Homoclinic orbits for asymptotically linear Hamiltonian system. J. Funct. Anal.187, 25-41 (2001) 7. Jeanjean, L: On the existence of bounded Palais-Smale sequences and application to a Landesman-Lazer type

problem set onRN. Proc. R. Soc. Edinb., Sect. A, Math.129, 787-809 (1999)

8. Miyagaki, OH, Souto, MAS: Superlinear problems without Ambrosetti and Rabinowitz growth condition. J. Differ. Equ.

12, 3628-3638 (2008)

9. Schechter, M: Superlinear elliptic boundary value problems. Manuscr. Math.86, 253-265 (1995)

10. Rabinowitz, PH, Su, JB, Wang, ZQ: Multiple solutions of superlinear elliptic equations. Rendiconti Lincei Matematicae Applicazioni18, 97-108 (2007)

11. Szulkin, A, Weth, T: Ground state solutions for some indefinite variational problems. J. Funct. Anal.257, 3802-3822 (2009)

12. Rabinowitz, PH: Minimax Methods in Critical Point Theory with Applications to Differential Equations, CBMS (1986) 13. Izydorek, M, Janczewska, J: Homoclinic solutions for a class of the second order Hamiltonian systems. J. Differ. Equ.

219, 375-389 (2005)

14. Coti-Zelati, V, Rabinowitz, P: Homoclinic type solutions for a semilinear elliptic PDE onRn. Commun. Pure Appl. Math.

46, 1217-1269 (1992)

15. Ding, YH, Luan, SX: Multiple solutions for a class of Schrödinger equations. J. Differ. Equ.207, 423-457 (2004) 16. Shi, SJ, Li, SJ: Existence of solutions for a class of semilinear elliptic equations with the Robin boundary value

condition. Nonlinear Anal., Theory Methods Appl.71, 3292-3298 (2009)

17. Bartsch, T, Ding, YH: On a nonlinear Schrödinger equation with periodic potential. Math. Ann.313, 15-37 (1999) 18. Li, SJ, Su, JB: Existence of multiple solutions of a two-point boundary value problem at resonance. Nonlinear Anal.,

Theory Methods Appl.10, 123-135 (1997)

19. Ding, YH, Lee, C: Multiple solutions of Schrödinger equations with indefinite linear part and super or asymptotically linear terms. J. Differ. Equ.222, 137-163 (2006)

20. Zhao, FK, Chen, J, Yang, MB: A periodic solution for a second-order asymptotically linear Hamiltonian system. Nonlinear Anal., Theory Methods Appl.70, 4021-4026 (2009)

doi:10.1186/1687-2770-2012-97

References

Related documents