Skip to content

Advertisement

Open Access

Nontrivial solution for asymmetric ( p , 2 ) -Laplacian Dirichlet problem

Boundary Value Problems20142014:241

https://doi.org/10.1186/s13661-014-0241-0

Received: 29 July 2014

Accepted: 3 November 2014

Published: 13 November 2014

Abstract

We consider a class of particular ( p , 2 ) -Laplacian Dirichlet problems with a right-hand side nonlinearity which exhibits an asymmetric growth at +∞ and −∞. Namely, it is linear at −∞ and superlinear at +∞. However, it need not satisfy the Ambrosetti-Rabinowitz condition on the positive semi-axis. Some existence results for a nontrivial solution are established by the mountain pass theorem and a variant version of the mountain pass theorem in the general case 2 < p < N . Similar results are also established by combining the mountain pass theorem and a variant version of the mountain pass theorem with the Moser-Trudinger inequality in the case of p = N .

Keywords

asymmetric Dirichlet problemsubcritical exponential growthone side resonance

1 Introduction

Let Ω be a bounded domain in R N ( N > 2 ) with smooth boundary Ω. We consider the following quasilinear elliptic boundary problem:
{ p u ( x ) μ u = a ( x ) | u | s 2 u + f ( x , u ) in  Ω , u = 0 on  Ω ,
(1.1)

where 2 < p < and 1 < s < p , p denotes the p-Laplacian operator defined by p u = div ( | u | p 2 u ) , μ > 0 is a real parameter, a ( x ) L ( Ω ) and f ( x , t ) C ( Ω ¯ × R ) .

It is known that the nontrivial solutions of problem (1.1) are equivalent to the corresponding nonzero critical points of the C 1 -energy functional
I ( u ) = 1 p Ω | u | p d x + μ 2 Ω | u | 2 d x 1 s Ω a ( x ) | u | s d x Ω F ( x , u ) d x
(1.2)

for all u W 0 1 , p ( Ω ) , where F ( x , t ) = 0 t f ( x , s ) d s .

For the case of p > 2 , a ( x ) 0 and μ > 0 , there has been an increasing interest in looking for the existence of solutions of (1.1). Using the following conditions,
μ m < f ( x , 0 ) < μ m + 1 , F ( x , t ) < λ 1 p | t | p + C , x Ω ,

where m 1 and C is a constant, the authors in [1], [2] prove that (1.1) has at least two nontrivial solutions by the three critical point theorems. Here and in the sequel, 0 < μ 1 < μ 2 < denote the eigenvalues of − in H 0 1 ( Ω ) , and λ 1 is the first eigenvalue of p in W 0 1 , p ( Ω ) (see [3]). For Eq. (1.1) with the right-hand side having p-linear growth at infinity, i.e., lim | t | f ( x , t ) | t | p 2 t = λ σ ( p ) , the spectrum of p in W 0 1 , p ( Ω ) , the papers [4], [5] get the existence of a nontrivial solution. In [6], the author extends the results in [1], [2] under the general asymptotically linear condition.

The main purpose of this paper is to establish existence results of a nontrivial solution for problem (1.1) with 2 < p N when the nonlinearity term f ( x , ) exhibits an asymmetric behavior as t R approaches +∞ and −∞. More precisely, we assume that for a.e. x Ω , f ( x , ) grows superlinear at +∞, while at −∞ it has a linear growth. In case of 1 < p < N , μ = 0 and a ( x ) 0 , equations with nonlinearities which are superlinear in one direction and linear in the other were investigated by Arcoya and Villegas [7], de Figueiredo and Ruf [8], Perera [9]. All three works express the superlinear growth at +∞ using the Ambrosetti-Rabinowitz condition ((AR)-condition, for short). Recall that a function f : Ω × R R is said to satisfy the (AR)-condition in the positive direction if there exist μ > p and M > 0 such that
0 < μ F ( x , t ) t f ( x , t ) for all  t M  and a.e.  x Ω ,

where F ( x , t ) = 0 t f ( x , s ) d s . Since Ambrosetti and Rabinowitz proposed the mountain pass theorem in their celebrated paper [10], critical point theory has become one of the main tools for finding solutions to elliptic equations of variational type. When we apply the mountain pass theorem, the (AR)-condition usually plays an important role in verifying that the functional I has a ‘mountain pass’ geometry and showing that a related ( PS ) c sequence is bounded.

By simple calculation, it is easy to see that the previous one side (AR)-condition implies that lim t + F ( x , t ) t p = + . That is, f ( x , t ) must be superlinear with respect to | t | p 2 t at positive infinity. Recently, Motreanu et al.[11], Papageorgiou and Papageorgiou [12] and Papageorgiou and Smyrlis [13] studied asymmetric problem (1.1) with nonlinearity f not satisfying the (AR)-condition on the positive semi-axis when μ = 0 and a ( x ) 0 . Nevertheless, all of the above-mentioned works involve the nonlinear term f ( x , u ) of a subcritical (polynomial) growth, say,

(SCP): there exist positive constants c 1 and c 2 and q 0 ( p 1 , p 1 ) such that
| f ( x , t ) | c 1 + c 2 | t | q 0 for all  t R  and  x Ω ,

where p = N p / ( N p ) denotes the critical Sobolev exponent. One of the main reasons to assume this condition (SCP) is that they can use the Sobolev compact embedding W 0 1 , p L q ( Ω ) , 1 q < p .

In this paper, we always assume that μ = 1 in (1.1). Under the motivation of Lam and Lu [14], our first main results will be to study problem (1.1) in the improved subcritical polynomial growth
( SCPI ) : lim t f ( x , t ) t p 1 = 0 uniformly on  x Ω ,

which is much weaker than (SCP). Note that in this case, we do not have the Sobolev compact embedding anymore. Our work again is to study asymmetric problem (1.1) without the (AR)-condition in the positive semi-axis. In fact, this condition was studied by Liu and Wang in [15] in the case of Laplacian (i.e., p = 2 ) by the Nehari manifold approach. However, we will use the mountain pass theorem and a suitable version of the mountain pass theorem to get the nontrivial solution to problem (1.1) in the general case 2 < p < N . Our results are different from those in [11]–[13] and our proof of the compactness condition is skillful.

Let us now state our results: Suppose that f ( x , t ) C ( Ω ¯ × R ) and satisfies:

(H1): lim t 0 f ( x , t ) | t | p 2 t = f 0 uniformly for a.e. x Ω , where f 0 [ 0 , + ) ;

(H2): lim t f ( x , t ) | t | p 2 t = l uniformly for a.e. x Ω , where l [ 0 , + ] ;

(H3): lim t + f ( x , t ) t p 1 = + uniformly for a.e. x Ω ;

(H4): f ( x , t ) | t | p 2 t is nonincreasing with respect to t 0 for a.e. x Ω .

Let λ 1 be the first eigenvalue of ( p , W 0 1 , p ( Ω ) ) and ϕ 1 ( x ) > 0 for every x Ω be the λ 1 eigenfunction. Throughout this paper, we denote by | | p the L p ( Ω ) norm and the norm of u in W 0 1 , p ( Ω ) will be defined by
u = ( Ω | u | p d x ) 1 p .

Theorem 1.1

Let 2 < p < N and assume that f has the improved subcritical polynomial growth on Ω (condition (SCPI)) and satisfies (H1)-(H3). If f 0 < λ 1 < l < , then there exists m = m ( f 0 , s , f , N , Ω ) such that for all a ( x ) L ( Ω ) with | a | < m , problem (1.1) has at least a nontrivial solution if l is not any of the eigenvalues of p on W 0 1 , p ( Ω ) .

Remark 1.1

In view of conditions (H2) and (H3), problem (1.1) is called asymmetric. Hence, Theorem 1.1 is completely different from the results contained in [4]–[6].

Theorem 1.2

Let 2 < p < N and assume that f has the improved subcritical polynomial growth on Ω (condition (SCPI)) and satisfies (H1)-(H3). If f 0 < λ 1 = l and lim t [ f ( x , t ) t p F ( x , t ) ] = + uniformly for a.e. x Ω , then there exists m = m ( f 0 , s , f , N , Ω ) such that for all a ( x ) L ( Ω ) and a ( x ) < 0 with | a | < m , problem (1.1) has at least a nontrivial solution.

Remark 1.2

When l = λ 1 , problem (1.1) is called resonant at negative infinity. This case is completely new. Here, we also give an example for f ( x , t ) . It satisfies our conditions (H1)-(H3) and (SCPI).

Example A

Define
f ( x , t ) = { g ( t ) | t | p 2 t , t 0 , g ( t ) | t | p 2 t + h ( t ) , t > 0 ,

where g ( t ) C ( R ) , g ( 0 ) = 0 ; g ( t ) 0 , t R ; h ( t ) C [ 0 , + ) ; lim t + 0 h ( t ) t p 1 = 0 ; lim t + h ( t ) t p 1 = 0 ; lim t + h ( t ) t p 1 = + . Moreover, there exists t 0 > 0 such that g ( t ) λ 1 for all | t | t 0 .

Theorem 1.3

Let 2 < p < N and assume that f has the improved subcritical polynomial growth on Ω (condition (SCPI)) and satisfies (H1)-(H4). If f 0 < λ 1 and l = + , then there exists m = m ( f 0 , s , f , N , Ω ) such that for all a ( x ) L ( Ω ) with | a | < m , problem (1.1) has at least a nontrivial solution.

In case of p = N , we have p = + . In this case, every polynomial growth is admitted, but one knows easy examples that W 0 1 , n ( Ω ) L ( Ω ) . Hence, one is led to look for a function g ( s ) : R R + with maximal growth such that
sup u W 0 1 , N , u 1 Ω g ( u ) d x < .

It was shown by Trudinger [16] and Moser [17] that the maximal growth is of exponential type. So, we must redefine the subcritical (exponential) growth in this case as follows.

(SCE):f has subcritical (exponential) growth on Ω, i.e., lim t | f ( x , t ) | exp ( α | t | N N 1 ) = 0 uniformly on x Ω for all α > 0 .

When p = N and f has the subcritical (exponential) growth ( S C E ) , our work is still to study asymmetric problem (1.1) without the (AR)-condition in the positive semi-axis. To our knowledge, this case is completely new. Our results are as follows.

Theorem 1.4

Let p = N and assume that f has the subcritical exponential growth on Ω (condition (SCE)) and satisfies (H1)-(H3). If f 0 < λ 1 < l < , then there exists m = m ( f 0 , s , f , N , Ω ) such that for all a ( x ) L ( Ω ) with | a | < m , problem (1.1) has at least a nontrivial solution if l is not any of the eigenvalues of N on W 0 1 , N ( Ω ) .

Remark 1.3

In view of conditions (H2), (H3) and (SCE), problem (1.1) is called asymmetric subcritical exponential problem. Hence, Theorem 1.4 is completely new.

Theorem 1.5

Let p = N and assume that f has the subcritical exponential growth on Ω (condition (SCE)) and satisfies (H1)-(H3). If f 0 < λ 1 = l and lim t [ f ( x , t ) t N F ( x , t ) ] = + uniformly for a.e. x Ω , then there exists m = m ( f 0 , s , f , N , Ω ) such that for all a ( x ) L ( Ω ) and a ( x ) < 0 with | a | < m , problem (1.1) has at least one nontrivial solution.

Remark 1.4

When l = λ 1 , problem (1.1) is called resonant at negative infinity. This case is new and completely different from the results contained in [14].

Theorem 1.6

Let p = N and assume that f has the subcritical exponential growth on Ω (condition (SCE)) and satisfies (H1)-(H4). If f 0 < λ 1 and l = + , then there exists m = m ( f 0 , s , f , N , Ω ) such that for all a ( x ) L ( Ω ) with | a | < m , problem (1.1) has at least one nontrivial solution.

2 Preliminaries and some lemmas

Definition 2.1

Let ( E , E ) be a real Banach space with its dual space ( E , E ) and I C 1 ( E , R ) . For c R , we say that I satisfies the ( PS ) c condition if for any sequence { x n } E with
I ( x n ) c , D I ( x n ) 0 in  E ,
there is a subsequence { x n k } such that { x n k } converges strongly in E. Also, we say that I satisfies the ( C ) c condition if for any sequence { x n } E with
I ( x n ) c , D I ( x n ) E ( 1 + x n E ) 0 ,

there is a subsequence { x n k } such that { x n k } converges strongly in E.

We have the following version of the mountain pass theorem (see [18]).

Proposition 2.1

Let E be a real Banach space and suppose that I C 1 ( E , R ) satisfies the condition
max { I ( 0 ) , I ( u 1 ) } α < β inf u = ρ I ( u )
for some α < β , ρ > 0 and u 1 E with u 1 > ρ . Let c β be characterized by
c = inf γ Γ max 0 t 1 I ( γ ( t ) ) ,
where Γ = { γ C ( [ 0 , 1 ] , E ) , γ ( 0 ) = 0 , γ ( 1 ) = u 1 } is the set of continuous paths joining 0 and  u 1 . Then there exists a sequence { u n } E such that
I ( u n ) c β and ( 1 + u n ) I ( u n ) E 0 as  n .

Lemma 2.1

Let 2 < p < N and ϕ 1 > 0 be a λ 1 -eigenfunction with ϕ 1 = 1 and assume that (H1)-(H3) and (SCPI) hold. If f 0 < λ 1 < l + , then there exists m = m ( f 0 , s , f , N , Ω ) such that for all a ( x ) L ( Ω ) with | a | < m , we have:
  1. (i)

    There exist ρ , α > 0 such that I ( u ) α for all u W 0 1 , p ( Ω ) with u = ρ .

     
  2. (ii)

    I ( t ϕ 1 ) as t + .

     

Proof

By (SCPI) and (H1)-(H3), if l ( λ 1 , + ) , for any ε > 0 , there exist A 1 = A 1 ( ε ) , B 1 = B 1 ( ε ) such that for all ( x , s ) Ω × R ,
F ( x , s ) 1 p ( f 0 + ε ) | s | p + A 1 | s | p ,
(2.1)
F ( x , s ) 1 p ( l ε ) | s | p B 1 if  l ( λ 1 , + ) .
(2.2)
Choose ε > 0 such that ( f 0 + ε ) < λ 1 . By (2.1), the Poincaré inequality and the Sobolev inequality, | u | p p K u p ,
I ( u ) 1 p Ω | u | p d x a s Ω | u | s d x Ω F ( x , u ) d x 1 p Ω | u | p d x a s Ω | u | s d x 1 p Ω [ ( f 0 + ε ) | u | p + A 1 | u | p ] d x 1 p ( 1 f 0 + ε λ 1 ) u p a K s s u s A 1 K p u p .
Set
ρ = ( a ( p s ) K s s A 1 K p ) 1 p s , m = ( λ 1 f 0 ε p λ 1 ) p s p p s K s ( A 1 K p ) s p p p [ ( p s p p ) s p p s + ( p s p p ) p p p s ] s p p p .

So, part (i) holds if we choose u = ρ > 0 and a < m .

On the other hand, if l ( λ 1 , + ) , take ε > 0 such that l ε > λ 1 . By (2.2), we have
I ( t ϕ 1 ) t p p ϕ 1 p + t 2 Ω | ϕ 1 | 2 d x t s s Ω a ( x ) ( ϕ 1 ) s d x l ε p | ϕ 1 | p p + B | Ω | .
Since l ε > λ 1 and ϕ 1 = 1 , it is easy to see that
I ( t ϕ 1 ) 1 p ( 1 l ε λ 1 ) t p + t 2 Ω | ϕ 1 | 2 d x t s s Ω a ( x ) ( ϕ 1 ) s d x + B 1 | Ω | as  t + ,

and part (ii) is proved. □

Lemma 2.2

(see [16], [17])

Let u W 0 1 , N ( Ω ) , then exp ( | u | N N 1 ) L q ( Ω ) for all 1 q < . Moreover,
sup u W 0 1 , N ( Ω ) , u 1 Ω exp ( α | u | N N 1 ) d x C ( Ω ) for  α α N .

The inequality is optimal: for any growth exp ( α | u | N N 1 ) with α > α N , the corresponding supremum is +∞.

Lemma 2.3

Let p = N and ϕ 1 > 0 be a λ 1 -eigenfunction with ϕ 1 = 1 and assume that (H1)-(H3) and (SCE) hold. If f 0 < λ 1 < l + , then there exists m = m ( f 0 , s , f , N , Ω ) such that for all a ( x ) L ( Ω ) with | a | < m , we have:
  1. (i)

    There exist ρ , α > 0 such that I ( u ) α for all u W 0 1 , N ( Ω ) with u = ρ .

     
  2. (ii)

    I ( t ϕ 1 ) as t + .

     

Proof

By (SCE) and (H1)-(H3), if l ( λ 1 , + ) , for any ε > 0 , there exist A 1 = A 1 ( ε ) , B 1 = B 1 ( ε ) , κ > 0 and q > N such that for all ( x , s ) Ω × R ,
F ( x , s ) 1 N ( f 0 + ε ) | s | N + A 1 exp ( κ | s | N N 1 ) | s | q ,
(2.3)
F ( x , s ) 1 N ( l ε ) | s | N B 1 if  l ( λ 1 , + ) .
(2.4)
Choose ε > 0 such that ( f 0 + ε ) < λ 1 . By (2.3), the Holder inequality and the Moser-Trudinger embedding inequality, we get
I ( u ) 1 N u N f 0 + ε N | u | N N a s Ω | u | s d x A 1 Ω exp ( κ | u | N N 1 ) | u | q d x 1 N ( 1 f 0 + ε λ 1 ) u N a s Ω | u | s d x A 1 ( Ω exp ( κ r u N N 1 ( | u | u ) N N 1 ) d x ) 1 r ( Ω | u | r q d x ) 1 r 1 N ( 1 f 0 + ε λ 1 ) u N a K s s u s C u q ,
where r > 1 sufficiently close to 1, u σ and κ r σ N N 1 < α N . Set
ρ = ( a ( N s ) K s s C ) 1 q s , m = ( λ 1 f 0 ε N λ 1 ) q s q N s K s ( C ) s N q N [ ( N s q N ) s N q s + ( N s q N ) q N q s ] s q q N .

So, part (i) holds if we choose u = ρ > 0 and a < m .

On the other hand, if l ( λ 1 , + ) , taking ε > 0 such that l ε > λ 1 and using (2.4), we have
I ( t ϕ 1 ) 1 N ( 1 l ε λ 1 ) | t | N + t 2 Ω | ϕ 1 | 2 d x t s s Ω a ( x ) ( ϕ 1 ) s d x + B 1 | Ω | as  t + .

Thus part (ii) is proved. By exactly slight modification to the proof above, we can prove (ii) if l = + . □

Lemma 2.4

For the functional I defined by (1.2), if u n ( x ) 0 a.e. x Ω , n N and
I ( u n ) , u n 0 as  n ,
then there exists a subsequence, still denoted by { u n } , such that
I ( t u n ) 1 + t p n p + ( t p p t s s ) Ω a ( x ) | u n | s d x + I ( u n ) for all  t 0  and  n N .

Proof

By I ( u n ) , u n 0 as n , for a suitable subsequence, we may assume that
1 n < I ( u n ) , u n = u n p + u n 2 Ω a ( x ) | u n | s d x Ω f ( x , u n ( x ) ) u n d x < 1 n
(2.5)

for all n, where denotes the norm of H 0 1 ( Ω ) .

We claim that for any 0 t and n N ,
I ( t u n ) ( 1 2 u n 2 1 p u n 2 ) + ( t p p t s s ) Ω a ( x ) | u n | s d x + t p p n + Ω { 1 p f ( x , u n ( x ) ) u n F ( x , u n ( x ) ) } d x .
(2.6)
Indeed, for any t 0 , at fixed x Ω and n N , if we set
h ( t ) = 1 p t p f ( x , u n ) u n ( x ) F ( x , t u n ( x ) ) ,
then
h ( t ) = t p 1 f ( x , u n ) u n ( x ) f ( x , t u n ) u n ( x ) = t p 1 u n ( x ) { f ( x , u n ) f ( x , t u n ( x ) ) / t p 1 } = { 0 for  0 < t 1 , 0 for  t 1 by  ( H 4 ) ,
hence
h ( t ) h ( 1 ) for all  t 0 .
Therefore,
I ( t u n ) = 1 p t p u n p + 1 2 t 2 u n 2 t s s Ω a ( x ) | u n | s d x Ω F ( x , t u n ( x ) ) d x < 1 2 t 2 u n 2 + 1 p t p { 1 n u n 2 + Ω f ( x , u n ( x ) ) u n ( x ) d x } t s s Ω a ( x ) | u n | s d x Ω F ( x , t u n ( x ) ) d x ( 1 2 t 2 u n 2 t p p u n 2 ) + t p p n + ( t p p t s s ) Ω a ( x ) | u n | s d x + Ω { 1 p t p f ( x , u n ( x ) ) u n ( x ) F ( x , t u n ( x ) ) } d x ( 1 2 u n 2 1 p u n 2 ) + t p p n + ( t p p t s s ) Ω a ( x ) | u n | s d x + Ω { 1 p f ( x , u n ( x ) ) u n ( x ) F ( x , u n ( x ) ) } d x ,

and our claim (2.6) is proved.

On the other hand,
I ( u n ) = 1 p u n p + 1 2 u n 2 1 s Ω a ( x ) | u n | s d x Ω F ( x , u n ( x ) ) d x 1 2 u n 2 + 1 p { 1 n u n 2 + Ω f ( x , u n ( x ) ) u n ( x ) d x } Ω F ( x , u n ( x ) ) d x ,
that is,
Ω { 1 p f ( x , u n ( x ) ) u n ( x ) F ( x , u n ( x ) ) } d x 1 p n + 1 p u n 2 1 2 u n 2 + I ( u n ) .
(2.7)
Combining (2.6) and (2.7), we find that
I ( t u n ) 1 + t p n p + ( t p p t s s ) Ω a ( x ) | u n | s d x + I ( u n ) for all  t 0  and  n N .
(2.8)

 □

3 Proofs of the main results

Here, we only prove Theorems 1.1-1.4. Others follow these results.

Proof of Theorem 1.1

By Lemma 2.1, the geometry conditions of mountain pass theorem hold. So, we only need to verify condition ( PS ) . Let { u n } W 0 1 , p ( Ω ) be a ( PS ) sequence such that for every n N ,
| 1 p Ω | u n | p d x + 1 2 Ω | u n | 2 d x t s s Ω a ( x ) | u n | s d x Ω F ( x , u n ) d x | c
(3.1)
and
| Ω | u n | p 2 u n v d x + Ω u n v d x Ω a ( x ) | u n | s 2 u n v Ω f ( x , u n ) v d x | ε n v
(3.2)

for all v W 0 1 , p ( Ω ) , where c > 0 is a positive constant and { ε n } R + is a sequence which converges to zero.

Step 1. In order to prove that { u n } has a convergent subsequence, we first show that it is a bounded sequence. To do this, we argue by contradiction assuming that for a subsequence, which we denote by { u n } , we have
u n + as  n .
Without loss of generality, we can assume u n > 1 for all n N and define z n = u n u n . Obviously, z n = 1 , n N , and then it is possible to extract a subsequence (denoted also by { z n } ) such that
z n z 0 in  W 0 1 , p ( Ω ) ,
(3.3)
z n z 0 in  L p ( Ω ) ,
(3.4)
z n ( x ) z 0 ( x ) a.e.  x Ω ,
(3.5)
| z n ( x ) | q ( x ) a.e.  x Ω ,
(3.6)
where z 0 W 0 1 , p ( Ω ) and q L p ( Ω ) . Dividing both sides of (3.2) by u n p 1 , we obtain
| Ω | z n | p 2 z n v d x + u n 2 p Ω z n v d x Ω a ( x ) | u n | s 2 u n v u n p 1 Ω f ( x , u n ) u n p 1 v d x | ε n u n p 1 v
for all v W 0 1 , p ( Ω ) . Passing to the limit we deduce from (3.3) that
lim n Ω f ( x , u n ) u n p 1 v d x = Ω | z 0 | p 2 z 0 v d x
(3.7)

for all v W 0 1 , p ( Ω ) .

Now we claim that z 0 ( x ) 0 for a.e. x Ω . To verify this, let us observe that by choosing v = z 0 + = max { z 0 , 0 } in (3.7), we have
lim n Ω + f ( x , u n ) u n p 1 z 0 d x = Ω + | z 0 | p d x < + ,
(3.8)
where Ω + = { x Ω | z 0 ( x ) > 0 } . On the other hand, from (H2) and (H3), it implies
f ( x , u n ( x ) ) u n p 1 z 0 ( x ) ( l | q ( x ) | p 2 q ( x ) K 1 ) z 0 ( x ) a.e.  x Ω
for some positive constant K 1 > 0 . Moreover, using lim n u n ( x ) = + for a.e. x Ω + , (3.5) and the superlinearity of f (see (H3)), we also deduce
lim n f ( x , u n ( x ) ) u n p 1 z 0 ( x ) = lim n f ( x , u n ( x ) ) u n p 1 z n ( x ) p 1 z 0 ( x ) = + a.e.  x Ω + .
Therefore, if | Ω + | > 0 , by Fatou’s lemma, we will obtain that
lim n Ω + f ( x , u n ( x ) ) u n p 1 z 0 ( x ) d x = + ,

which contradicts (3.8). Thus | Ω + | = 0 and the claim is proved.

Clearly, z 0 ( x ) 0 . By (H2), there exists c > 0 such that | f ( x , u n ) | | u n | p 1 c for a.e. x Ω . By using the Lebesgue dominated convergence theorem in (3.7), we have
Ω | z 0 | p 2 z 0 v d x Ω l | z 0 | p 2 z 0 v d x = 0
(3.9)

for all v W 0 1 , p ( Ω ) . This contradicts our assumption, i.e., l is not any of the eigenvalues of p on W 0 1 , p ( Ω ) .

Step 2. Now, we prove that { u n } has a convergent subsequence. In fact, we can suppose that
u n u in  W 0 1 , p ( Ω ) , u n u in  L q ( Ω ) , 1 q < p , u n ( x ) u ( x ) a.e.  x Ω .
Now, since f has the subcritical growth on Ω, for every ε > 0 , we can find a constant C ( ε ) > 0 such that
f ( x , s ) C ( ε ) + ε | s | p 1 , ( x , s ) Ω × R ,
then
| Ω f ( x , u n ) ( u n u ) d x | C ( ε ) Ω | u n u | d x + ε Ω | u n u | | u n | p 1 d x C ( ε ) Ω | u n u | d x + ε ( Ω ( | u n | p 1 ) p p 1 d x ) p 1 p ( Ω | u n u | p ) 1 p C ( ε ) Ω | u n u | d x + ε C ( Ω ) .
Similarly, since u n u in W 0 1 , p ( Ω ) , Ω | u n u | d x 0 . Since ε > 0 is arbitrary, we can conclude that
Ω ( f ( x , u n ) f ( x , u ) ) ( u n u ) d x 0 as  n .
(3.10)
By (3.2), we have
I ( u n ) I ( u ) , ( u n u ) 0 as  n .
(3.11)
From (3.10) and (3.11), we obtain
Ω ( | u n | p 2 u n | u | p 2 u ) ( u n u ) 0 as  n .
Using an elementary inequality
2 2 p | b a | p | b | p 2 b | a | p 2 a , b a , a , b R N ,
we can imply that
u n u in  L p ( Ω ) .

So we have u n u in W 0 1 , p ( Ω ) , which means that I satisfies ( PS ) . □

Proof of Theorem 1.2

Since l = λ 1 , obviously, Lemma 2.1(i) holds. We only need to show that Lemma 2.1(ii) holds. Let u = t ϕ 1 . Using condition (H3), we have
I ( t ϕ 1 ) = 1 p t p Ω | ϕ 1 | p d x + t 2 2 Ω | ϕ 1 | 2 d x t s s Ω a ( x ) | ϕ 1 | s d x Ω F ( x , t ϕ 1 ) d x = 1 p t p Ω | ϕ 1 | p d x + t 2 2 Ω | ϕ 1 | 2 d x t s s Ω a ( x ) | ϕ 1 | s d x t p M Ω | ϕ 1 | p d x + C
as t + , where M is a positive constant large enough. By Proposition 2.1, there exists a sequence { u n } W 0 1 , p ( Ω ) such that
I ( u n ) = 1 p u n p + 1 2 u n 2 1 s Ω a ( x ) | u n | s d x Ω F ( x , u n ) d x = c + ( 1 ) ,
(3.12)
( 1 + u n ) I ( u n ) W 0 1 , p 0 as  n .
(3.13)
Clearly, (3.13) implies that
I ( u n ) , u n = u n p + u n 2 Ω a ( x ) | u n | s d x Ω f ( x , u n ( x ) ) u n d x = ( 1 ) .
(3.14)
To complete our proof, we first need to verify that { u n } is bounded in W 0 1 , p ( Ω ) . Similar to the proof of Theorem 1.1, we have z 0 ( x ) 0 , x Ω , z 0 ( x ) 0 and
Ω | z 0 | p 2 z 0 v d x Ω l | z 0 | p 2 z 0 v d x = 0
for all v W 0 1 , p ( Ω ) . By the maximum principle (see [19]), z 0 < 0 is an eigenfunction of λ 1 , then | u n ( x ) | for a.e. x Ω . By our assumptions, we have
lim n ( f ( x , u n ( x ) ) u n ( x ) p F ( x , u n ( x ) ) ) = +
uniformly in x Ω , which implies that
Ω ( f ( x , u n ( x ) ) u n ( x ) p F ( x , u n ( x ) ) ) d x + as  n .
(3.15)
On the other hand, (3.14) implies that
p I ( u n ) I ( u n ) , u n p c as  n .
Thus
Ω ( f ( x , u n ) u n p F ( x , u n ) ) d x as  n ,

which contradicts (3.15). Hence { u n } is bounded. According to Step 2 of the proof of Theorem 1.1, we have u n u in W 0 1 , p ( Ω ) , which means that I satisfies ( C ) c . □

Proof of Theorem 1.3

By Lemma 2.1 and Proposition 2.1, (3.12)-(3.14) hold. We still can prove that { u n } is bounded in W 0 1 , p ( Ω ) . Assume u n + as n . Similar to the proof of Theorem 1.1, we have z 0 ( x ) 0 and when z 0 ( x ) < 0 , u n = z n u n as n . Let
s n = 2 p c p u n , w n = s n u n = 2 p c p u n u n .
(3.16)
Since { w n } is bounded in W 0 1 , p ( Ω ) , it is possible to extract a subsequence (denoted also by { w n } ) such that
w n w 0 in  W 0 1 , p ( Ω ) , w n w 0 in  L p ( Ω ) , w n ( x ) w 0 ( x ) a.e.  x Ω , | w n ( x ) | h ( x ) a.e.  x Ω ,

where w 0 W 0 1 , p ( Ω ) and h L p ( Ω ) .

If u n + as n , then w 0 ( x ) 0 . In fact, letting Ω = { x Ω : w 0 ( x ) < 0 } and noticing l = + , from (H3) we have that
f ( x , u n ) | u n | p 2 u n M uniformly for all  x Ω ,
where M is a large enough constant. Therefore, by (3.14) and (3.16), we have
2 p c = lim n w n p = lim n Ω f ( x , u n ) | u n | p 2 u n | w n | p d x lim n Ω f ( x , u n ) | u n | p 2 u n | w n | p d x M lim n Ω | w 0 | p d x .
So w 0 0 for a.e. x Ω . But, if w 0 0 , then Ω F ( x , w n ) d x 0 . Hence
I ( w n ) = 1 p w n p + 1 2 w n 2 + ( 1 ) 2 c + ( 1 ) .
(3.17)
On the other hand, by u n as n , we have s n 0 as n . From Lemma 2.4 and (3.12), we get
I ( w n ) = I ( s n u n ) 1 + ( s n ) p n p + I ( u n ) c as  n .

Obviously, it contradicts (3.17). So { u n } is bounded in W 0 1 , p ( Ω ) . According to Step 2 of the proof of Theorem 1.1, we have u n u in W 0 1 , p ( Ω ) , which means that I satisfies ( C c ) . □

Proof of Theorem 1.4

By Lemma 2.3, the geometry conditions of mountain pass theorem hold. So, we only need to verify condition ( PS ) . Similar to Step 1 of the proof of Theorem 1.1, we easily know that the ( PS ) sequence { u n } is bounded in W 0 1 , N ( Ω ) . Next, we prove that { u n } has a convergent subsequence. Without loss of generality, suppose that
u n β , u n u in  W 0 1 , N ( Ω ) , u n u in  L q ( Ω ) , q 1 , u n ( x ) u ( x ) a.e.  x Ω .
Now, since f has the subcritical exponential growth (SCE) on Ω, we can find a constant C β > 0 such that
| f ( x , t ) | C β exp ( α N 2 β N N 1 | t | N N 1 ) , ( x , t ) Ω × R .
Thus, by the Moser-Trudinger inequality (see Lemma 2.2),
| Ω f ( x , u n ) ( u n u ) d x | C ( Ω exp ( α N β N N 1 | u n | N N 1 ) d x ) 1 2 | u n u | 2 C ( Ω exp ( α N β N N 1 u n N N 1 | u n u n | N N 1 ) d x ) 1 2 | u n u | 2 C | u n u | 2 0 .

Similar to the last proof of Theorem 1.1, we have u n u in W 0 1 , N ( Ω ) , which means that I satisfies ( PS ) . □

Declarations

Acknowledgements

This work was supported by the National NSF (Grant No. 11101319) of China and Planned Projects for Postdoctoral Research Funds of Jiangsu Province (Grant No. 1301038C).

Authors’ Affiliations

(1)
School of Mathematics and Statistics, Tianshui Normal University, Tianshui, P.R. China
(2)
School of Mathematical Sciences, Institute of Mathematics, Nanjing Normal University, Nanjing, P.R. China

References

  1. Chang K-C: Morse theory on Banach space and its applications to partial differential equations. Chin. Ann. Math., Ser. B 1983, 4: 381-399.Google Scholar
  2. do Ó JM: Existence of solutions for quasilinear elliptic equations. J. Math. Anal. Appl. 1997, 207: 104-126. 10.1006/jmaa.1997.5270MathSciNetView ArticleGoogle Scholar
  3. Lindqvist P:On the equation div ( | u | p 2 u ) + λ | u | p 2 u = 0 . Proc. Am. Math. Soc. 1990, 109: 157-164.MathSciNetGoogle Scholar
  4. Cingolani S, Degiovanni M: Nontrivial solutions for p -Laplace equations with right-hand side having p -linear growth at infinity. Commun. Partial Differ. Equ. 2005, 30: 1191-1203. 10.1080/03605300500257594MathSciNetView ArticleGoogle Scholar
  5. Cingolani S, Vannella G: Marino-Prodi perturbation type results and Morse indices of minimax critical points for a class of functionals in Banach space. Ann. Mat. Pura Appl. 2007, 186: 155-183. 10.1007/s10231-005-0176-2MathSciNetView ArticleGoogle Scholar
  6. Sun MZ: Multiplicity solutions for a class of the quasilinear elliptic equations at resonance. J. Math. Anal. Appl. 2012, 386: 661-668. 10.1016/j.jmaa.2011.08.030MathSciNetView ArticleGoogle Scholar
  7. Arcoya D, Villegas S: Nontrivial solutions for a Neumann problem with a nonlinear term asymptotically linear at −∞ and superlinear at +∞. Math. Z. 1995, 219: 499-513. 10.1007/BF02572378MathSciNetView ArticleGoogle Scholar
  8. de Figueiredo DG, Ruf B: On a superlinear Sturm-Liouville equation and a related bouncing problem. J. Reine Angew. Math. 1991, 421: 1-22.MathSciNetGoogle Scholar
  9. Perera K: Existence and multiplicity results for a Sturm-Liouville equation asymptotically linear at −∞ and superlinear at +∞. Nonlinear Anal. 2000, 39: 669-684. 10.1016/S0362-546X(98)00228-4MathSciNetView ArticleGoogle Scholar
  10. Ambrosetti A, Rabinowitz PH: Dual variational methods in critical point theory and applications. J. Funct. Anal. 1973, 14: 349-381. 10.1016/0022-1236(73)90051-7MathSciNetView ArticleGoogle Scholar
  11. Motreanu D, Motreanu VV, Papageorgiou NS: Multiple solutions for Dirichlet problems which are superlinear at +∞ and (sub-)linear at −∞. Commun. Pure Appl. Anal. 2009, 13: 341-358.MathSciNetGoogle Scholar
  12. Papageorgiou EH, Papageorgiou NS: Multiplicity of solutions for a class of resonant p -Laplacian Dirichlet problems. Pac. J. Math. 2009, 241: 309-328. 10.2140/pjm.2009.241.309View ArticleGoogle Scholar
  13. Papageorgiou NS, Smyrlis G: A multiplicity theorem for Neumann problems with asymmetric nonlinearity. Ann. Mat. Pura Appl. 2010, 189: 253-272. 10.1007/s10231-009-0108-7MathSciNetView ArticleGoogle Scholar
  14. Lam N, Lu G: N -Laplacian equations in R N with subcritical and critical growth without the Ambrosetti-Rabinowitz condition. Adv. Nonlinear Stud. 2013, 13: 289-308.MathSciNetGoogle Scholar
  15. Liu ZL, Wang ZQ: On the Ambrosetti-Rabinowitz superlinear condition. Adv. Nonlinear Stud. 2004, 4: 563-574.MathSciNetGoogle Scholar
  16. Trudinger NS: On imbeddings in to Orlicz spaces and some applications. J. Math. Mech. 1967, 17: 473-483.MathSciNetGoogle Scholar
  17. Moser J: A sharp form of an inequality by N. Trudinger. Indiana Univ. Math. J. 1971, 20: 1077-1092. 10.1512/iumj.1971.20.20101View ArticleGoogle Scholar
  18. Costa DG, Miyagaki OH: Nontrivial solutions for perturbations of the p -Laplacian on unbounded domains. J. Math. Anal. Appl. 1995, 193: 737-755. 10.1006/jmaa.1995.1264MathSciNetView ArticleGoogle Scholar
  19. Vázquez JL: A strong maximum principle for some quasilinear elliptic equations. Appl. Math. Optim. 1984, 12: 191-202. 10.1007/BF01449041MathSciNetView ArticleGoogle Scholar

Copyright

© Pei and Zhang; licensee Springer. 2014

This article is published under license to BioMed Central Ltd.Open Access This is an Open Access article distributed under the terms of the Creative Commons Attribution License (http://creativecommons.org/licenses/by/4.0), which permits unrestricted use, distribution, and reproduction in any medium, provided the original work is properly credited.

Advertisement