Home Mathematics On sufficient “local” conditions for existence results to generalized p(.)-Laplace equations involving critical growth
Article Open Access

On sufficient “local” conditions for existence results to generalized p(.)-Laplace equations involving critical growth

  • Ky Ho and Inbo Sim EMAIL logo
Published/Copyright: September 3, 2022

Abstract

In this article, we study the existence of multiple solutions to a generalized p ( ) -Laplace equation with two parameters involving critical growth. More precisely, we give sufficient “local” conditions, which mean that growths between the main operator and nonlinear term are locally assumed for p ( ) -sublinear, p ( ) -superlinear, and sandwich-type cases. Compared to constant exponent problems (e.g., p -Laplacian and ( p , q ) -Laplacian), this characterizes the study of variable exponent problems. We show this by applying variants of the mountain pass theorem for p ( ) -sublinear and p ( ) -superlinear cases and constructing critical values defined by a minimax argument in the genus theory for sandwich-type case. Moreover, we also obtain a nontrivial nonnegative solution for sandwich-type case changing the role of parameters. Our work is a generalization of several existing works in the literature.

MSC 2010: 35B33; 35J20; 35J25; 35J62; 46E35

1 Introduction and main results

In this article, we are concerned with the existence and multiplicity of solutions to the following problem:

(1.1) { div a ( x , u ) = λ f ( x , u ) + θ b ( x ) u t ( x ) 2 u in Ω , u = 0 on Ω ,

where Ω is a bounded domain in R N with a Lipschitz boundary Ω ; a : Ω × R N R N is a Carathéodory function which generalizes ξ p ( x ) 2 ξ , where p C + ( Ω ¯ ) with

C + ( Ω ¯ ) { r C ( Ω ¯ ) : 1 < r min x Ω ¯ r ( x ) r + max x Ω ¯ r ( x ) < } ;

f : Ω × R R is a Carathéodory function having subcritical growth; t C + ( Ω ¯ ) with p + < t t ( x ) p ( x ) for all x Ω ¯ such that

C { x Ω ¯ : t ( x ) = p ( x ) } ,

where p ( x ) N p ( x ) N p ( x ) if p + < N and p ( x ) p 0 ( x ) for any fixed p 0 C + ( Ω ¯ ) with p ( x ) < p 0 ( x ) for all x Ω ¯ if p + N ; b L ( Ω ) and b ( x ) > 0 for a.e. x Ω ; and λ , θ are parameters.

Throughout this article, we always assume the following conditions on a :

  1. a ( x , ξ ) = a ( x , ξ ) for a.e. x Ω and all ξ R N .

  2. There exists a Carathéodory function A : Ω × R N R , continuously differentiable with respect to its second variable, such that A ( x , 0 ) = 0 for a.e. x Ω and a ( x , ξ ) = ξ A ( x , ξ ) for a.e. x Ω and all ξ R N .

  3. a ( x , ξ ) C [ 1 + ξ p ( x ) 1 ] for a.e. x Ω and all ξ R N , where C is a positive constant.

  4. ξ p ( x ) a ( x , ξ ) ξ p ( x ) A ( x , ξ ) for a.e. x Ω and all ξ R N .

  5. 0 < [ a ( x , ξ ) a ( x , η ) ] ( ξ η ) for a.e. x Ω and all ξ , η R N , ξ η .

The operator satisfying (A0)–(A4) is of Leray-Lions-type and typical examples are

div ( u p ( x ) 2 u ) and div [ ( 1 + u 2 ) ( p ( x ) 2 ) / 2 u ] ,

which we call the p ( ) -Laplacian and the generalized mean curvature operator, respectively. Another remarkable example is

div i = 1 n u p i ( x ) 2 u ,

which satisfies (A0)–(A4) with p ( x ) max 1 i n p i ( x ) , and we call this operator the multiple p ( ) -Laplacian. A special case of the multiple p ( ) -Laplacian is the operator div ( u p 2 u + u q 2 u ) with 1 < q p < , which is called the ( p , q ) -Laplacian and has received a great attention from mathematicians in the last decade.

The study of p ( ) -Laplacian was initiated to describe some materials (e.g., electrorheological fluids) that are inadequate to deal with in a constant p ( ) . Indeed, Rǎdulescu [27] gave a couple of examples that required the setting of p ( ) -Laplacian problems that portray phenomenon strongly depending on the position. Thus, it is natural to ask for sufficient effects of the ration of p ( ) and growth of a nonlinear term in a local sense. It is worth mentioning that the set C is the other condition of locality. Meanwhile, a local effect was also assumed in this study when p ( ) is even constant in the sequel papers [4,5,6]. More precisely, the authors were concerned with the “local” growth of nonlinearities, investigating the existence, nonexistence, and multiplicity of solutions for the family of problems

div ( u p 2 u ) = f λ ( x , u ) in Ω , u > 0 in Ω , u = 0 on Ω ,

where f is locally “ p -sublinear” at 0 and is locally “ p -superlinear” at , namely (roughly speaking):

lim s 0 + f λ ( x , s ) s p 1 =

for x in a subdomain Ω 1 of Ω and

lim s f λ ( x , s ) s p 1 =

for x in a subdomain Ω 2 of Ω . We emphasize that such local assumptions are effective for a nonautonomous nonlinear term and does not allow higher or lower growth than p , respectively. Recently, Komiya and Kajikiya [20] studied the existence of infinitely many solutions to problem (1.1) when a ( x , ξ ) = ξ p 2 ξ + ξ q 2 ξ with 1 < q p < and θ = 0 . Indeed, they obtained a sequence of solutions with C 1 -norms converging to 0 (resp. ) when f satisfies a locally q -sublinear condition, i.e., inf x B ε ( x 0 ) F ( x , τ ) τ q as τ 0 (resp. a locally p -superlinear condition, i.e., inf x B ε ( x 0 ) F ( x , τ ) τ p as τ ) for some ball B ε ( x 0 ) Ω . Here and in what follows, F ( x , τ ) 0 τ f ( x , s ) d s and B ε ( x 0 ) denotes a ball in R N centered at x 0 with radius ε . It is worth mentioning that they did not treat the case of growth between p and q , which we call sandwich-type growth for the ( p , q ) -problem. These results have recently been obtained for problem (1.1) under (A0)–(A4) in [12]. Note that both [20] and [12] only considered subcritical problems. The existence of infinitely many solutions to critical problems involving the p -Laplacian for this kind of locally p -superlinear condition was investigated in [28].

Regarding critical and sandwich-type growth, in [15] we considered problem (1.1) of the form:

(1.2) div ( u p 2 u + u q 2 u ) = λ m ( x ) u s 2 u + θ b ( x ) u p 2 u in Ω , u = 0 on Ω ,

where 1 < q < s < p and the weight m is possibly sign-changing. We obtained the existence of nontrivial nonnegative solutions to [1] for a fixed λ and some range of parameter θ (depending on λ ). In [1], Baldelli et al. obtained the multiplicity of solutions for problem (1.2) when Ω = R N in a reverse way, that is, for θ fixed and small, they find some range of λ . The subcritical problem of sandwich-type growth involving variable exponent was studied by Mihailescu-Rǎdulescu in [25]. In that article, the authors studied (1.1) of the form:

(1.3) div ( u p ( x ) 2 u + u q ( x ) 2 u ) = λ u s ( x ) 2 u in Ω , u = 0 on Ω ,

with p , q , s C + ( Ω ¯ ) satisfying q + < s s + < p and s + < q ( x ) for all x Ω ¯ and showed that there exists λ > 0 such that problem (1.3) admits a nontrivial solution for any λ [ λ , ) .

Motivated by [1,4,5, 6,11,15,20,25,28], our aim is to show the existence of multiple solutions to problem (1.1), which may include a critical term and generalize previous results, in particular [15,25,28] when f satisfies one of three types of growth conditions: locally p ( ) -sublinear, locally p ( ) -superlinear, and sandwich-type. Moreover, we also obtain a nontrivial nonnegative solution for the sandwich-type case by changing the role of the parameters. We look for solutions to problem (1.1) in W 0 1 , p ( ) ( Ω ) , that is the completion of C c ( Ω ) with respect to the norm

u u L p ( ) ( Ω ) .

Note that if p + < N and p is logarithmic Hölder continuous (notation: p C 0 , 1 log t ( Ω ¯ ) ), namely,

p ( x ) p ( y ) C log x y for all x , y Ω ¯ with 0 < x y < 1 2 ,

then the following critical imbedding holds

(1.4) W 0 1 , p ( ) ( Ω ) L p ( ) ( Ω )

(see Section 2 for the definitions and properties of the variable exponent Lebesgue-Sobolev spaces). Since we mainly focus on critical problems, throughout this article we always assume p C 0 , 1 log t ( Ω ¯ ) when p + < N in our main results.

Before giving a statement of the main results, we note that under ( A1 ) , we have that for a.e. x Ω and for all ξ R N ,

(1.5) A ( x , ξ ) = 0 1 a ( x , t ξ ) ξ d t .

Thus, ( A2 ) yields

(1.6) A ( x , ξ ) C ˜ [ 1 + ξ p ( x ) ] for a.e. x Ω and all ξ R N ,

where C ˜ is a positive constant. Also, from ( A2 ) we obtain

(1.7) a ( x , ξ ) p ( x ) p ( x ) 1 C ¯ [ 1 + ξ p ( x ) ]

for a.e. x Ω and all ξ R N . In what follows, for r C ( Ω ¯ ) with r ( x ) 1 for all x Ω and an open set D Ω , denote

r D inf x D r ( x ) , r D + sup x D r ( x ) ,

and

L + r ( ) ( D ) { u L r ( ) ( D ) : u > 0 a.e. in D } .

First, let us start with the locally p ( ) -sublinear case. In this case, we additionally assume for the operator a that:

  1. For each ball B Ω , there exist constants C B , δ B > 0 , and q B [ 1 , ) such that

    a ( x , ξ ) C B ξ q B 1 for a.e. x B and all ξ < δ B .

It is clear that the multiple p ( ) -Laplacian and the generalized mean curvature operator above also satisfy ( A5 ). Note that ( A3 ) and ( A5 ) yield q B p B and moreover, from ( A5 ) and (1.5) we obtain

(1.8) A ( x , ξ ) C B q B ξ q B for a.e. x B and all ξ < δ B .

Assume that f satisfies the following conditions.

  1. f : Ω × [ ε 0 , ε 0 ] R is a Carathéodory function with an ε 0 > 0 such that f is odd with respect to the second variable and d ( ) sup τ ε 0 f ( , τ ) L + σ ( ) ( Ω ) with σ C + ( Ω ¯ ) satisfying σ ( x ) > max 1 , N p ( x ) for all x Ω ¯ .

  2. There exist a ball B Ω and m L + 1 ( B ) such that

    lim τ 0 F ( x , τ ) m ( x ) τ q B = uniformly for a.e. x B ,

    where q B is determined as in ( A5 ) .

The existence result for the locally p ( ) -sublinear case is stated as follows.

Theorem 1.1

Let (A0)–(A5), (F1), and (F2) hold. Then, for any λ > 0 and θ 0 , problem (1.1) admits a sequence of nontrivial solutions { u n } n = 1 converging to zero in W 0 1 , p ( ) ( Ω ) L ( Ω ) .

Next, we study problem (1.1) with a nonlinearity f of locally p ( ) -superlinear-type growth. In this case, we need the following additional assumption on the exponent p :

  1. There exists w L + p ( ) p ( ) p ( ) ( Ω ) such that

    (1.9) μ 1 inf φ C c ( Ω ) Ω φ p ( x ) d x Ω w ( x ) φ p ( x ) d x ( 0 , ) .

Obviously, condition (P) automatically holds true when p ( ) is constant due to the Hölder and Sobolev inequalities. For variable exponent case, (P) is fulfilled if there is a vector x 0 R N { 0 } such that for any x Ω , h ( τ ) p ( x + τ x 0 ) is monotone for τ I x = { τ : x + τ x 0 Ω } (by taking w L + ( Ω ) ), see [10, Theorems 3.2–3.3]. It is worth pointing out that μ 1 given by (1.9) is zero if there is an open ball B ε 0 ( x 0 ) Ω such that p ( x 0 ) < (or > ) p ( x ) for all x B ε 0 ( x 0 ) , see [10, Proof of Theorem 3.1].

Now, we state conditions on f for the locally p ( ) -superlinear case.

  1. f : Ω × R R is a Carathéodory function such that f is odd with respect to the second variable.

  2. There exist functions r j , a j with r j C + ( Ω ¯ ) , inf x Ω ¯ [ t ( x ) r j ( x ) ] > 0 , a j L + t ( ) t ( ) r j ( ) b r j t r j , Ω ( j = 1 , , m 0 ), and max 1 j m 0 r j + > p such that

    f ( x , τ ) j = 1 m 0 a j ( x ) τ r j ( x ) 1 for a.e. x Ω and all τ R .

  3. There exist an open ball B Ω and a function m L + 1 ( B ) such that

    lim τ F ( x , τ ) m ( x ) τ p B + = uniformly for a.e. x B .

  4. For w given by (P) , there exist α [ p + , t ) and e L + 1 ( Ω ) such that

    α F ( x , τ ) τ f ( x , τ ) α p + p + w ( x ) τ p ( x ) + e ( x ) for a.e. x Ω and all τ R .

The existence result for the locally p ( ) -superlinear case is stated as follows.

Theorem 1.2

Let (P), (A0)–(A4), and (F3)–(F6) hold. For a given λ ( 0 , μ 1 ] , there exists a sequence { θ n } n = 1 with 0 < θ n + 1 < θ n for all n N such that for any θ ( θ n + 1 , θ n ) , problem (1.1) admits at least n distinct pairs of nontrivial solutions.

In the last part, we study problem (1.1) with the nonlinearity f of locally sandwich-type growth. Precisely, we consider problem (1.1) of the form:

(1.10) div a ( x , u ) = λ m ( x ) u s ( x ) 2 u + θ b ( x ) u t ( x ) 2 u in Ω , u = 0 on Ω ,

where a satisfies (A0)–(A5) and the following assumptions hold:

  1. s C + ( Ω ¯ ) with s + < p and t p + < t s .

  2. m L t ( ) t ( ) s ( ) b s t s , Ω .

Furthermore, the subcritical nonlinearity satisfies the following local condition:
  1. There exists a ball B Ω such that q B < s B and meas { x B : m ( x ) > 0 } > 0 , where q B is defined by B as in (A5).

Note that under (L1), from (1.6) and (1.8) we find positive constants C B and C p such that

(1.11) A ( x , ξ ) C B ξ q B + C p ξ p B + for a.e. x B and for all ξ R N .

Let us start to study the sandwich case by investigating the multiplicity of solutions to problem (1.10). The following multiple existence result is a counterpart of [1, Theorem 1] for the bounded domain case involving Leray-Lions-type operators.

Theorem 1.3

Let (A0)–(A5), (S), (W), and (L1) hold. Assume furthermore that m ( x ) > 0 for a.e. x B . Then there exists { θ k } k = 1 with 0 < θ k < θ k + 1 for all k N such that for each k N , for θ ( 0 , θ k ) given, there exist λ , λ > 0 with λ < λ such that for any λ ( λ , λ ) , problem (1.10) admits at least k pairs of distinct solutions with negative energy.

It is worth pointing out that [1, Theorem 1] actually asserted the existence of a sequence of solutions provided the L norm of K ( x ) θ b ( x ) is sufficiently small. However, we found that their argument can only produce a finite number of solutions in the same way as the statement of Theorem 1.3 and discussed it with the authors of [1] about this issue; please see Remark 4.4 for more details.

Finally, we investigate a nontrivial nonnegative solution to problem (1.10) changing the role of parameters. For this purpose, in addition to (L1) we assume that

  1. There exists ϕ C c ( B ) such that B m ( x ) s ( x ) ϕ + s ( x ) d x > 0 with ϕ + max { ϕ , 0 } and for each τ > 0 , there exists s B ( τ ) [ s B , s B + ] such that

    B m ( x ) s ( x ) ϕ + s ( x ) τ s ( x ) d x = τ s B ( τ ) B m ( x ) s ( x ) ϕ + s ( x ) d x .

Clearly, (L2) automatically holds true if either m ( x ) > 0 a.e. in B or s ( ) is constant in B and meas { x B : m ( x ) > 0 } > 0 (see, e.g., [18, Proposition 4.2]). Denote by A the set of all ϕ C c ( B ) satisfying (L2) and define

(1.12) λ inf ϕ A max η [ s B , s B + ] C ( η ) B m ( x ) s ( x ) ϕ + s ( x ) d x B ϕ p B + d x η q B p B + q B B ϕ q B d x p B + η p B + q B ,

where

C ( η ) p B + q B p B + η p B + η p B + q B p B + q B η q B η q B p B + q B C p η q B p B + q B C B p B + η p B + q B for η [ s B , s B + ] ,

where C p and C B are constants taken from (1.11). Clearly, λ is well defined and λ [ 0 , ) . The existence of a nontrivial nonnegative solution for the sandwich case is shown in the next theorem.

Theorem 1.4

Let (A0)–(A5), (S), (W), (L1), and (L2) hold. Then for each given λ > λ , there exists θ > 0 such that for any θ [ 0 , θ ) , problem (1.10) has a nontrivial nonnegative solution.

Remark 1.5

It is worth pointing out that if b ( x ) u t ( x ) 2 u is replaced by a subcritical term ω ( x ) u r ( x ) 2 u , where s + < p p + < r and r ( x ) < p ( x ) for all x Ω ¯ and ω L + p ( ) p ( ) r ( ) ( Ω ) , then we can obtain one more nontrivial nonnegative solution to problem (1.10) without assuming any combined condition on m and ω by applying the mountain pass theorem. If r ( ) is constant, it suffices to assume ω L p ( ) p ( ) r ( ) ( Ω ) and meas { x Ω : ω ( x ) > 0 } > 0 instead of ω L + p ( ) p ( ) r ( ) ( Ω ) (see Theorem 4.6).

Outline of the article. We organize the article as follows: In Section 2, we review the definition and properties of Lebesgue-Sobolev spaces with variable exponents and introduce notations that are used throughout this article. In Section 3, we investigate the existence of infinitely many solutions for the locally p ( ) -sublinear case and the multiplicity of solutions for the locally p ( ) -superlinear case by giving proofs of Theorems 1.1 and 1.2. Section 4 is devoted to the study of the existence of k pairs of solutions with an arbitrarily given number k and a nontrivial nonnegative solution to problem (1.1) when the nonlinear f is of a (locally) sandwich type by giving a proof of Theorems 1.3 and 1.4. In this section, we also discuss problem (1.10) when the growth is subcritical. Finally, in Section 5, we comment on the regularity of solutions to problem (1.1) obtained in the previous sections.

2 Notions and preliminary results

Let Ω be a bounded Lipschitz domain in R N . Denote

C + ( Ω ¯ ) { h C ( Ω ¯ ) : 1 < inf x Ω ¯ h ( x ) sup x Ω ¯ h ( x ) < } ,

and for h C + ( Ω ¯ ) , denote

h + sup x Ω ¯ h ( x ) and h inf x Ω ¯ h ( x ) .

For p C + ( Ω ¯ ) and for a Lebesgue measurable and positive a.e. function w : Ω R , define the weighted variable exponent Lebesgue space L p ( ) ( w , Ω ) as

L p ( ) ( w , Ω ) u : Ω R is measurable , Ω w ( x ) u ( x ) p ( x ) d x <

endowed with the Luxemburg norm

u L p ( ) ( w , Ω ) inf λ > 0 : Ω w ( x ) u ( x ) λ p ( x ) d x 1 .

Set L + p ( ) ( w , Ω ) { u L p ( ) ( w , Ω ) : u > 0 a.e. in Ω } . When w 1 , we write L p ( ) ( Ω ) , L + p ( ) ( Ω ) , and u L p ( ) ( Ω ) in place of L p ( ) ( w , Ω ) , L + p ( ) ( w , Ω ) , and u L p ( ) ( w , Ω ) , respectively. Some basic properties of L p ( ) ( w , Ω ) are listed in the next three propositions.

Proposition 2.1

[9,21] The space L p ( ) ( Ω ) is a separable, uniformly convex Banach space, and its dual space is L p ( ) ( Ω ) , where 1 / p ( x ) + 1 / p ( x ) = 1 . For any u L p ( ) ( Ω ) and v L p ( ) ( Ω ) , we have

Ω u v d x 2 u L p ( ) ( Ω ) v L p ( ) ( Ω ) .

Proposition 2.2

[7] Define ρ : L p ( ) ( w , Ω ) R as

ρ ( u ) Ω w ( x ) u ( x ) p ( x ) d x , u L p ( ) ( w , Ω ) .

Then, for all u L p ( ) ( w , Ω ) we have

  1. u L p ( ) ( w , Ω ) < 1 ( = 1 , > 1 ) if and only if ρ ( u ) < 1 ( = 1 , > 1 ) , respectively;

  2. if u L p ( ) ( w , Ω ) > 1 , then u L p ( ) ( w , Ω ) p ρ ( u ) u L p ( ) ( w , Ω ) p + ;

  3. if u L p ( ) ( w , Ω ) < 1 , then u L p ( ) ( w , Ω ) p + ρ ( u ) u L p ( ) ( w , Ω ) p .

Consequently, we have

u L p ( ) ( w , Ω ) p 1 ρ ( u ) u L p ( ) ( w , Ω ) p + + 1 , u L p ( ) ( w , Ω ) .

Proposition 2.3

[19] Let p C + ( Ω ¯ ) and q C ( Ω ¯ ) such that p q C + ( Ω ¯ ) . For all u L p ( ) q ( ) ( w , Ω ) , it holds that

  1. if u L p ( ) q ( ) ( w , Ω ) 1 , then u L p ( ) q ( ) ( w , Ω ) q u q L p ( ) ( w , Ω ) u L p ( ) q ( ) ( w , Ω ) q + ;

  2. if u L p ( ) q ( ) ( w , Ω ) < 1 , then u L p ( ) q ( ) ( w , Ω ) q + u q L p ( ) ( w , Ω ) u L p ( ) q ( ) ( w , Ω ) q .

Consequently, we have

u L p ( ) q ( ) ( w , Ω ) q 1 u q L p ( ) ( w , Ω ) u L p ( ) q ( ) ( w , Ω ) q + + 1 , u L p ( ) q ( ) ( w , Ω ) .

Define

W 1 , p ( ) ( Ω ) { u L p ( ) ( Ω ) : u L p ( ) ( Ω ) }

endowed with the norm

u W 1 , p ( ) ( Ω ) u L p ( ) ( Ω ) + u L p ( ) ( Ω ) .

We recall the following crucial imbeddings on W 1 , p ( ) ( Ω ) .

Proposition 2.4

[7,9,21] The space W 1 , p ( ) ( Ω ) is a reflexive separable Banach space. Moreover, the following assertions hold:

  1. if q C + ( Ω ¯ ) with q ( x ) < p ( x ) for all x Ω ¯ , then

    W 1 , p ( ) ( Ω ) L q ( ) ( Ω ) ;

  2. if p C 0 , 1 log t ( Ω ¯ ) and p + < N , then

    W 1 , p ( ) ( Ω ) L p ( ) ( Ω ) .

The space W 0 1 , p ( ) ( Ω ) defined in Section 1 is in fact the closure of C c ( Ω ) in W 1 , p ( ) ( Ω ) due to the following Poincaré-type inequality.

Proposition 2.5

[9] There exists a positive constant C such that

u L p ( ) ( Ω ) C u L p ( ) ( Ω ) , u W 0 1 , p ( ) ( Ω ) .

Next, we present a compact result that will be frequently utilized in the next sections. Let t C + ( Ω ¯ ) and b : Ω R be measurable and positive a.e. in Ω such that W 0 1 , p ( ) ( Ω ) L t ( ) ( b , Ω ) . Let m L t ( ) t ( ) σ ( ) b σ t σ , Ω for some σ C ( Ω ¯ ) satisfying 1 σ ( x ) < t ( x ) for all x Ω ¯ . Then, m u σ L 1 ( Ω ) for all u W 0 1 , p ( ) ( Ω ) in view of Proposition 2.1. Furthermore, we have the following.

Proposition 2.6

Let t , b , and σ be as above. Then, for u n u in W 0 1 , p ( ) ( Ω ) as n , we have

(2.1) Ω m ( x ) u n σ ( x ) d x Ω m ( x ) u σ ( x ) d x as n ,

(2.2) Ω m ( x ) u n u σ ( x ) d x 0 as n ,

and

(2.3) Ω m ( x ) u n σ ( x ) 1 u n u d x 0 as n .

Consequently, if m L + t ( ) t ( ) σ ( ) b σ t σ , Ω , then

W 0 1 , p ( ) ( Ω ) L σ ( ) ( m , Ω ) .

Proof

Clearly, (2.1) follows if we can prove that

(2.4) Ω v n d x 0 as n ,

where v n m ( u n σ ( x ) u σ ( x ) ) . To this end, we first note that up to a subsequence, v n 0 a.e. in Ω as n . Then, by the Vitali convergence theorem it suffices to show that for a given and arbitrary ε > 0 , there exists δ = δ ( ε ) > 0 such that for any measurable subset Q Ω with meas ( Q ) < δ ,

(2.5) Q v n d x < ε .

Indeed, let ε > 0 be given and arbitrary. Since { u n } n = 1 is bounded in W 0 1 , p ( ) ( Ω ) we have that { u n } n = 1 is bounded in L t ( ) ( b , Ω ) and hence

(2.6) M ε + 2 [ 2 + u L t ( ) ( b , Ω ) σ + + sup n N u n L t ( ) ( b , Ω ) σ + ] ( 0 , ) .

On the other hand, the assumption m L t ( ) t ( ) σ ( ) b σ t σ , Ω implies that there exist δ = δ ( ε ) > 0 such that for any measurable subset Q Ω with meas ( Q ) < δ ,

(2.7) Q m ( x ) b ( x ) σ ( x ) t ( x ) t ( x ) t ( x ) σ ( x ) d x < ε M t t σ + ( < 1 ) .

Invoking Proposition 2.2 and using (2.6) and (2.7) we have

Q v n d x = Q m ( x ) u n σ ( x ) u σ ( x ) d x 2 m b σ t L t ( ) t ( ) σ ( ) ( Q ) b u n σ u σ L t ( ) σ ( ) ( Q ) 2 m b σ t L t ( ) t ( ) σ ( ) ( Q ) [ 2 + u n L t ( ) ( b , Q ) σ + + u L t ( ) ( b , Q ) σ + ] M Q m ( x ) b ( x ) σ ( x ) t ( x ) t ( x ) t ( x ) σ ( x ) d x 1 t t σ + < ε , n N .

That is, we have proved (2.5), and hence, in turn we obtain (2.4) and (2.1). The proof for (2.2) is similar and we omit it. To see (2.3), by Propositions 2.1 and 2.2 we have

Ω m ( x ) u n σ ( x ) 1 u n u d x 2 m σ 1 σ u n σ 1 L σ ( ) σ ( ) 1 ( Ω ) m 1 σ u n u L σ ( ) ( Ω ) 2 1 + Ω m ( x ) u n σ ( x ) d x σ 1 σ + max Ω m ( x ) u n u σ ( x ) d x 1 σ + , Ω m ( x ) u n u σ ( x ) d x 1 σ .

Combining this with (2.1) and (2.2) we obtain (2.3).□

We conclude this section with a concentration-compactness principle for variable exponent spaces. Let p C 0 , 1 log t ( Ω ¯ ) with p + < N and let t C + ( Ω ¯ ) with p ( x ) < t ( x ) p ( x ) for all x Ω ¯ such that

C { x Ω ¯ : t ( x ) = p ( x ) } .

Let b L + ( Ω ) . Then, it follows from (1.4) that

(2.8) S b inf u W 0 1 , p ( ) ( Ω ) { 0 } u u L t ( ) ( b , Ω ) ( 0 , ) .

Let ( Ω ¯ ) denote the space of Radon measures on Ω ¯ ; that is, the dual of C ( Ω ¯ ) . By the Riesz representation theorem, for each μ ( Ω ¯ ) , there is a unique signed Borel measure on Ω ¯ , still denoted by μ , such that

μ , f = Ω ¯ f d μ , f C ( Ω ¯ ) .

We identify L 1 ( Ω ) with a subspace of ( Ω ¯ ) through the imbedding T : L 1 ( Ω ) ( Ω ¯ ) defined by

T u , f = Ω u f d x , u L 1 ( Ω ) , f C ( Ω ¯ )

(see e.g., [3, p. 116]). The next theorem is the concentration-compactness principle for variable exponent spaces, which plays a key role in our arguments in the next sections.

Theorem 2.7

(cf. [2,14]) Let { u n } n = 1 be a bounded sequence in W 0 1 , p ( ) ( Ω ) such that

u n u in W 0 1 , p ( ) ( Ω ) , u n p ( x ) μ in ( Ω ¯ ) , b u n t ( x ) ν in ( Ω ¯ ) .

Then, there exist { x i } i C of distinct points and { ν i } i , { μ i } i ( 0 , ) , where is at most countable, such that

ν = b u t ( x ) + i ν i δ x i , μ u p ( x ) + i μ i δ x i , S b ν i 1 / t ( x i ) μ i 1 / p ( x i ) , i .

Other notations:

  • E : the space W 0 1 , p ( ) ( Ω ) endowed with the norm u u L p ( ) ( Ω )

  • E : the dual of E

  • , : the duality pairing between a Banach space (from the context) and its dual

  • B ρ : the open ball in E centered at 0 with radius ρ

  • B ρ ( x 0 ) : the open ball in R N centered at x 0 with radius ρ

  • Q : the Lebesgue measure of Q R N .

3 The locally p ( ) -sublinear and locally p ( ) -superlinear cases

In this section, we will prove Theorems 1.1 and 1.2 by employing variants of the mountain pass theorem.

3.1 The locally p ( ) -sublinear case

The existence of a sequence of small solutions to problem (1.1) for the locally p ( ) -sublinear case, Theorem 1.1, can be obtained easily by adapting [12, Proof of Theorem 1.1], which employed a critical points theorem obtained in [17].

Proof of Theorem 1.1

Let λ > 0 and θ 0 . Consider the problem

div a ( x , u ) = f ˜ ( x , u ) in Ω , u = 0 on Ω ,

where f ˜ ( x , u ) λ f ( x , u ) + θ b ( x ) u t ( x ) 2 u . Note that f ˜ also satisfies ( F 1 ) and ( F 2 ) in place of f and hence, by repeating the argument used in [12, Proof of Theorem 1.1] in which we apply [29, Theorem 1.1] instead of [12, Proposition A.1], the desired conclusion follows.□

Remark 3.1

It is clear that Theorem 1.1 remains valid for any t C + ( Ω ¯ ) , and therefore, we do not need to assume p C 0 , 1 log t ( Ω ¯ ) when p + < N in this theorem.

3.2 The locally p ( ) -superlinear case

In this subsection, we will prove Theorem 1.2 by using the idea of [28]. We will restrict our proof to the case p + < N since the proof for the case p + N is similar and simpler. We will make use of the following abstract result for symmetric C 1 functionals.

Proposition 3.2

[28] Let X = V W , where X is a real Banach space and V is finite dimensional. Suppose that I C 1 ( X , R ) is an even functional satisfying I ( 0 ) = 0 and

  1. there exist constants ρ , β > 0 such that I ( u ) β for all u B ρ W ;

  2. there exist a subspace X ˜ of X and L > 0 such that dim V < dim X ˜ < and sup u X ˜ I ( u ) < L ;

  3. considering L given by ( I 2 ) , I satisfies the (PS) c condition for any c [ 0 , L ] .

Then I possesses at least dim X ˜ dim V pairs of nontrivial critical points.

To determine solutions to problem (1.1), we will apply Proposition 3.2 for X = E and I : E R defined as

(3.1) I ( u ) Ω A ( x , u ) d x λ Ω F ( x , u ) d x θ Ω b ( x ) t ( x ) u t ( x ) d x , u E .

It is clear that under the assumption of Theorem 1.2, I is of class C 1 and a critical point of I is a weak solution to problem (1.1). Moreover, I is even on E and I ( 0 ) = 0 . In order to verify condition ( I 3 ) of Proposition 3.2, we need the following lemma.

Lemma 3.3

Let (P), (F4), and (F6) hold. Then, for λ ( 0 , μ 1 ] and θ > 0 , I satisfies the ( P S ) c condition for all c R satisfying

(3.2) c < e L 1 ( Ω ) λ α + 1 α 1 t S b N min θ 1 h + 1 , θ 1 h 1

with e , α given by (F6), S b given by (2.8), and h ( x ) t ( x ) p ( x ) for x Ω ¯ .

Proof

Let λ ( 0 , μ 1 ] and θ > 0 and let { u n } n = 1 be a (PS) c -sequence for I with c satisfying (3.2). We first claim that { u n } n = 1 is bounded in E . Indeed, by ( A3 ) we have

I ( u n ) 1 α I ( u n ) , u n 1 p + 1 α Ω a ( x , u n ) u n d x λ α Ω [ α F ( x , u n ) f ( x , u n ) u n ] d x + θ Ω 1 α 1 t ( x ) b ( x ) u n t ( x ) d x .

From this, ( A3 ) , and (F6) we obtain

(3.3) I ( u n ) 1 α I ( u n ) , u n 1 p + 1 α Ω u n p ( x ) d x λ ( α p + ) α p + Ω w ( x ) u n p ( x ) d x λ e L 1 ( Ω ) α + θ 1 α 1 t Ω b ( x ) u n t ( x ) d x .

Note that by (P) it holds that

(3.4) ( α p + ) μ 1 p + Ω w ( x ) u n p ( x ) d x α p + p + Ω u n p ( x ) d x , n N .

Combining (3.3) with (3.4) and recalling I ( u n ) c and I ( u n ) 0 we deduce that for n large,

c + 1 + u n ( μ 1 λ ) ( α p + ) μ 1 α p + Ω u n p ( x ) d x λ e L 1 ( Ω ) α + θ 1 α 1 t Ω b ( x ) u n t ( x ) d x .

Thus,

(3.5) Ω b ( x ) u n t ( x ) d x C 1 ( 1 + u n ) , n N .

Here and in the remaining proof, C i ( i N ) denotes a positive constant independent of n . On the other hand, we deduce from ( A3 ) , (F4) , and Proposition 2.2 that for n large,

(3.6) 1 p + ( u n p 1 ) I ( u n ) + λ Ω F ( x , u n ) d x + θ Ω b ( x ) t ( x ) u n t ( x ) d x c + 1 + λ j = 1 m 0 Ω a j ( x ) r j ( x ) u n r j ( x ) d x + θ t Ω b ( x ) u n t ( x ) d x .

By the Young’s inequality, we have

a j u n r j ( x ) t ( x ) r j ( x ) t ( x ) a j b r j ( x ) t ( x ) t ( x ) t ( x ) r j ( x ) + r j ( x ) t ( x ) b u n t ( x ) .

Plugging this into (3.6) and recalling the assumption of a j , we arrive at

(3.7) u n p C 2 1 + Ω b ( x ) u n t ( x ) d x .

From (3.5) and (3.7) we obtain the boundedness of { u n } n = 1 in E since p > 1 . Then by Theorem 2.7, there exist μ , ν ( Ω ¯ ) , { x i } i C , and { μ i } i , { ν i } i ( 0 , ) , where is at most countable, such that up to a subsequence we have

(3.8) u n ( x ) u ( x ) a.e. x Ω ,

(3.9) u n u in E ,

(3.10) u n p ( x ) μ u p ( x ) + i μ i δ x i in ( Ω ¯ ) ,

(3.11) b u n t ( x ) ν = b u t ( x ) + i ν i δ x i in ( Ω ¯ ) ,

(3.12) S b ν i 1 / t ( x i ) μ i 1 / p ( x i ) , i ,

where δ x i is the Dirac mass at x i . We claim that = . Suppose, on the contrary, that there exists i . Arguing similarly to that obtained (3.5), we have

c + o n ( 1 ) = I ( u n ) 1 α I ( u n ) , u n e L 1 ( Ω ) λ α + 1 α 1 t θ Ω b ( x ) u n t ( x ) d x .

Passing to the limit as n in the last inequality and using (3.11), we obtain

(3.13) c e L 1 ( Ω ) λ α + 1 α 1 t θ ν i .

Next, we will prove that

(3.14) μ i θ ν i .

To this end, let ϕ C c ( R N ) be such that 0 ϕ 1 , ϕ 1 on B 1 2 ( 0 ) and supp ( ϕ ) B 1 ( 0 ) . For ε > 0 , define ϕ i , ε ( x ) ϕ x x i ε for x R N . Fixing such an ε , we obtain that

Ω a ( x , u n ) ( ϕ i , ε u n ) d x = I ( u n ) , ϕ i , ε u n + λ Ω f ( x , u n ) ϕ i , ε u n d x + θ Ω b ( x ) u n t ( x ) ϕ i , ε d x ,

i.e.,

(3.15) Ω ϕ i , ε a ( x , u n ) u n d x = I ( u n ) , ϕ i , ε u n Ω u n a ( x , u n ) ϕ i , ε d x + λ Ω ϕ i , ε f ( x , u n ) u n d x + θ Ω ϕ i , ε b ( x ) u n t ( x ) d x .

Let δ > 0 be arbitrary and fixed. By the Young’s inequality, we have

u n a ( x , u n ) ϕ i , ε δ a ( x , u n ) p ( x ) p ( x ) 1 + C ( p , δ ) ϕ i , ε p ( x ) u n p ( x ) ,

where C ( p , δ ) is a positive constant independent of n and ε . Using this estimation and ( A3 ) , we deduce from (3.15) that

Ω ϕ i , ε u n p ( x ) d x I ( u n ) , ϕ i , ε u n + C δ + C ( p , δ ) Ω ϕ i , ε p ( x ) u n p ( x ) d x + λ Ω ϕ i , ε f ( x , u n ) u n d x + θ Ω ϕ i , ε b ( x ) u n t ( x ) d x ,

where C sup n N Ω a ( x , u n ) p ( x ) p ( x ) 1 d x [ 0 , ) due to (1.7) and the boundedness of { u n } n = 1 in E . Then, by passing to the limit as n in the last estimation and invoking Propositions 2.4 and 2.6 to taking into account (3.9), the fact that I ( u n ) 0 and the boundedness of { ϕ i , ε u n } n = 1 in E , we arrive at

(3.16) Ω ¯ ϕ i , ε d μ C δ + C ( p , δ ) Ω ϕ i , ε p ( x ) u p ( x ) d x + λ Ω ϕ i , ε j = 1 m 0 a j ( x ) u r j ( x ) d x + θ Ω ¯ ϕ i , ε d ν .

Note that u L p ( ) ( Ω ) in view of Proposition 2.4; hence, by Proposition 2.1,

(3.17) Ω ϕ i , ε p ( x ) u p ( x ) d x 2 u p L p ( ) p ( ) ( B ε ( x i ) Ω ) ϕ i , ε p L N p ( ) ( B ε ( x i ) Ω ) .

On the other hand, invoking Proposition 2.2 we have

ϕ i , ε p L N p ( ) ( B ε ( x i ) Ω ) p 1 + B ε ( x i ) Ω ϕ i , ε N d x 1 + B 1 ( 0 ) ϕ ( y ) N d y .

From this and the fact that u L p ( ) ( Ω ) it follows from (3.17) that

lim ε 0 + Ω ϕ i , ε p ( x ) u p ( x ) d x = 0 .

Clearly, lim ε 0 + Ω ϕ i , ε j = 1 m 0 a j ( x ) u r j ( x ) d x = 0 . By letting ε 0 + in (3.16) and taking into account the last two limits we arrive at

μ i C δ + θ ν i .

Thus, we obtain (3.14) since δ > 0 was chosen arbitrarily. Combining (3.14) and (3.12) and noting t ( x i ) = N p ( x i ) N p ( x i ) , we have

(3.18) θ ν i μ i S b N θ p ( x i ) t ( x i ) p ( x i ) .

Combining (3.13) and (3.18) gives

c e L 1 ( Ω ) λ α + 1 α 1 t S b N min θ 1 h + 1 , θ 1 h 1 ,

which contradicts to (3.2). That is, = and hence,

Ω b ( x ) u n t ( x ) d x Ω b ( x ) u t ( x ) d x .

From this and (3.8) we deduce that u n u in L t ( ) ( b , Ω ) in view of the Brézis-Lieb lemma (see e.g., [14, Lemma 3.6]). Thus, by Propositions 2.1 and 2.2 we easily obtain

(3.19) Ω b ( x ) u n t ( x ) 2 u n ( u n u ) d x 0 .

On the other hand, in view of (F4) and Proposition 2.6 it follows from (3.9) that

(3.20) Ω f ( x , u n ) ( u n u ) d x 0 .

Using (3.19), (3.20), the boundedness of { u n } n = 1 in E , and the fact that I ( u n ) 0 in E , we obtain

Ω a ( x , u n ) ( u n u ) d x = I ( u n ) , u n u + λ Ω f ( x , u n ) ( u n u ) d x + θ Ω b ( x ) u n t ( x ) 2 u n ( u n u ) d x 0 ;

hence, u n u in E due to the ( S + ) property of Leray-Lions operators with variable growths (see [23, Theorem 4.1]). In other words, I satisfies the (PS) c condition and the proof is complete.□

In order to verify condition ( I 2 ) of Proposition 3.2, we will take X ˜ from a sequence { E k } k = 1 of linear subspaces of E , which is constructed as follows. For each k N , define

(3.21) E k span { φ 1 , φ 2 , , φ k } ,

where φ n is an eigenfunction corresponding to the nth eigenvalue of the following eigenvalue problem with B in (F5) :

Δ u = μ u in B , u = 0 on B ,

which is extended to Ω by putting φ n ( x ) = 0 for x Ω B .

Lemma 3.4

Let (F4) and (F5) hold. Let λ > 0 and let θ > 0 . Then, for each k N , there exists R k > 1 independent of θ such that

(3.22) I ( u ) < 0 , u E k B R k .

Proof

Let k N . Since all norms on E k are equivalent, we can find δ k > 0 such that

(3.23) δ k u u L p B + ( m , B ) , u E k .

Let γ k > 0 be positive constant such that

(3.24) λ γ k δ k p B + C ˜ > 0

with C ˜ in (1.6). Then, by (F5) we find T k > 0 such that

F ( x , τ ) γ k m ( x ) τ p B + , for a.e. x B and all τ T k .

Hence,

(3.25) F ( x , τ ) γ k m ( x ) τ p B + sup τ T k F ( x , τ ) , for a.e. x B and all τ R .

From (1.6) and (3.23)–(3.25), for u E k B R k with R k > 1 large enough we have

I ( u ) C ˜ B ( u p ( x ) + 1 ) d x λ B F ( x , u ) d x C ˜ ( u p B + + B ) λ γ k u L p B + ( m , B ) p B + + λ B sup τ T k F ( x , τ ) d x ( λ γ k δ k p B + C ˜ ) u p B + + λ B sup τ T k F ( x , τ ) d x + C ˜ B ( λ γ k δ k p B + C ˜ ) R k p B + + λ B sup τ T k F ( x , τ ) d x + C ˜ B < 0 ,

proving (3.22). Here we have used the fact that sup τ T k F ( , τ ) L 1 ( B ) due to (F4) . Obviously, R k above can be chosen independently of θ .□

Finally, to verify the remaining conditions of Proposition 3.2, we introduce a sequence { V k } k = 1 of finite dimensional linear subspaces of E as follows. Let { e n } n = 1 E be a Schauder basis of E and let { f n } n = 1 E be such that for each n N ,

f n , u = α n for u = j = 1 α j e j E .

For each k N , define

V k { u E : f n , u = 0 , n > k }

and

W k { u E : f n , u = 0 , n k } .

Then, E = V k W k and dim V k = k . Define

(3.26) δ k sup v W k v 1 max 1 j m 0 v L r j ( ) ( a j , Ω ) .

The next lemma was proved in [12].

Lemma 3.5

The sequence { δ k } k = 1 above satisfies 0 < δ k + 1 δ k for all k N and

lim k δ k = 0 .

We are now in a position to complete the proof of Theorem 1.2.

Proof of Theorem 1.2

Let λ ( 0 , μ 1 ] be given and let θ > 0 . We will show that conditions (I1)–(I3) of Proposition 3.2 are fulfilled with X = E and I given by (3.1). First, we will show that I satisfies ( I 1 ) . By Lemma 3.5, we find k 0 N such that δ k 0 ( 0 , 1 ) . By using ( A3 ) and (F4) we have that for any u W k 0 with u > 1 ,

(3.27) I ( u ) 1 p + ( u p 1 ) λ j = 1 m 0 1 r j ( u L r j ( ) ( a j , Ω ) r j + + 1 ) θ t Ω b u t ( x ) d x .

Note that for u E with u > 1 , it follows from Proposition 2.2 and (2.8) that

(3.28) Ω b u t ( x ) d x ( 1 + S b t + ) u t + .

Moreover, by considering v = z / z for z W k 0 { 0 } we deduce from the definition (3.26) of δ k that

(3.29) max 1 j m 0 z L r j ( ) ( a j , Ω ) δ k 0 z , z W k 0 .

Invoking (3.28) and (3.29), we obtain from (3.27) that for all u W k 0 with u > 1 ,

(3.30) I ( u ) 1 p + u p m 0 λ δ k 0 r r u r m 0 λ r + 1 p + ( 1 + S b t + ) θ t u t + ,

where r min 1 j m 0 r j and r max 1 j m 0 r j + . Let ρ k 0 > 0 be such that

m 0 λ δ k 0 r r ρ k 0 r = 1 2 p + ρ k 0 p i.e., ρ k 0 = r 2 m 0 p + λ δ k 0 r 1 r p .

By Lemma 3.5, we can fix k 0 N from the beginning such that

0 < δ k 0 < 1 < ρ k 0 and 1 2 p + ρ k 0 p m 0 λ r + 1 p + > 1 4 p + ρ k 0 p .

Thus, (3.30) yields

I ( u ) 1 4 p + ρ k 0 p ( 1 + S b t + ) θ t ρ k 0 t + , u B ρ k 0 W k 0 .

Therefore, by choosing V V k 0 , W W k 0 , and θ 0 t ρ k 0 p t + 4 p + ( 1 + S b t + ) > 0 , we have that for any θ ( 0 , θ 0 ) ,

I ( u ) β , u B ρ W

with ρ = ρ k 0 and β = ( 1 + S b t + ) ρ k 0 t + t ( θ 0 θ ) > 0 . That is, I satisfies ( I 1 ) in Proposition 3.2.

Next, we will show that I satisfies ( I 2 ) in Proposition 3.2. To this end, let { E k } k = 1 be defined by (3.21). For each k N we have

(3.31) sup u E k I ( u ) = max u E k u R k I ( u )

due to the fact that I ( 0 ) = 0 and Lemma 3.4. Note that by using (1.6) and (F4) , we have that for any u E with u R k ,

I ( u ) < C ˜ Ω ( u p ( x ) + 1 ) d x λ Ω F ( x , u ) d x C ˜ u p + + C ˜ ( Ω + 1 ) + λ j = 1 m 0 Ω a j ( x ) r j ( x ) u r j ( x ) d x C ˜ u p + + C ˜ ( Ω + 1 ) + j = 1 m 0 λ r j ( u L r j ( ) ( a , Ω ) r j + + 1 ) C ˜ u p + + λ r j = 1 m 0 C j r j + u r j + + m 0 λ r + C ˜ ( Ω + 1 ) C ˜ R k p + + λ r j = 1 m 0 C j r j + R k r j + + m 0 λ + C ˜ ( Ω + 1 ) L k ,

where C j ( j = 1 , , m 0 ) denotes the imbedding constant for E L r j ( ) ( a j , Ω ) . From this and (3.31) we obtain

(3.32) sup u E k I ( u ) < L k , k N .

Finally, let { θ k } k = 1 ( 0 , θ 0 ) be such that for any k N ,

(3.33) L k 0 + k < e L 1 ( Ω ) λ α + 1 α 1 t S b N min { θ k 1 h + 1 , θ k 1 h 1 } , θ k + 1 < θ k .

Then, for any θ ( θ k + 1 , θ k ) we have

L k 0 + k < e L 1 ( Ω ) λ α + 1 α 1 t S b N min θ 1 h + 1 , θ 1 h 1

and hence, I satisfies the (PS) c for any c [ 0 , L k 0 + k ] in view of Lemma 3.3. Thus, I satisfies ( I 2 ) and ( I 3 ) with X ˜ = E k 0 + k and L = L k 0 + k . So, I admits at least dim X ˜ dim V = k pairs of nontrivial critical points in view of Proposition 3.2; hence, problem (1.1) has at least k pairs of nontrivial solutions. The proof is complete.□

Remark 3.6

We would like to discuss problem (1.1) without a parameter in the subcritical term as follows:

(3.34) div a ( x , u ) = f ( x , u ) + θ b ( x ) u t ( x ) 2 u in Ω , u = 0 on Ω .

The existence of a sequence of nontrivial solutions to problem (3.34) for the p ( ) -sublinear case was obtained in Theorem 1.1 (for all θ 0 ). For the p ( ) -superlinear case, by repeating the proof of Theorem 1.2, we find a sequence { θ n } n = 1 with 0 < θ n + 1 < θ n for all n N such that for any θ ( θ n + 1 , θ n ) , problem (3.34) admits at least n distinct pairs of nontrivial solutions under (P), (A0)–(A4), (F3)–(F5), and (F6) , replaced by

  1. For w given by (P) , there exist α [ p + , t ) and e L + 1 ( Ω ) such that

    α F ( x , τ ) τ f ( x , τ ) ( α p + ) μ 1 p + w ( x ) τ p ( x ) + e ( x ) for a.e. x Ω and all τ R .

The result remains valid under (A0)–(A4), (F3)–(F5), and (F6) replaced by
  1. There exist α [ p + , t ) and functions σ C + ( Ω ¯ ) with σ + < p , w L + t ( ) t ( ) σ ( ) b σ t σ , Ω , and e L + 1 ( Ω ) such that

    α F ( x , τ ) τ f ( x , τ ) w ( x ) τ σ ( x ) + e ( x ) for a.e. x Ω and all τ R .

In this case, Theorem 1.2 generalizes of [28, Theorem A] that treated for the p -Laplacian.

4 The sandwich case

Throughout this section we always assume that (A0)–(A5), (S) , (W) , and (L1) hold. As in Section 3.2, we will restrict our proof to the case p + < N . To determine (nonnegative) solutions of problem (1.10), we introduce the functionals J , Ψ : E R defined as

(4.1) J ( u ) = Ω A ( x , u ) d x λ Ω m ( x ) s ( x ) u + s ( x ) d x θ Ω b ( x ) t ( x ) u + t ( x ) d x

and

(4.2) Ψ ( u ) = Ω A ( x , u ) d x λ Ω m ( x ) s ( x ) u s ( x ) d x θ Ω b ( x ) t ( x ) u t ( x ) d x

for u E . Here and in the rest of the article, both ϕ + and ϕ + stand for max { ϕ , 0 } . Under the given assumptions, J , Ψ are of class C 1 ( E , R ) and a critical point of J (resp. Ψ ) is a nonnegative solution (resp. a solution) to problem (1.10). We will make use of the critical point theorems and genus theory to determine critical points of J and Ψ . In order to verify the (PS) condition for J and Ψ , we need the next lemma.

Lemma 4.1

The following assertions hold.

  1. For θ = 0 and λ R , J , and Ψ satisfy the (PS) c condition for all c R .

  2. For θ > 0 and λ R , J , and Ψ satisfy the (PS) c condition for all c R satisfying

    (4.3) c < 1 p + 1 t S b N min θ 1 h + 1 , θ 1 h 1 K max θ 1 + 1 , θ 1 1 max λ + + 1 , λ 1 ,

    where K = K ( p , s , t , m , b , Ω ) > 0 , h ( x ) t ( x ) p ( x ) , and ( x ) t ( x ) s ( x ) for x Ω ¯ .

Proof

We only prove the conclusions for J since those of Ψ are proved similarly. We first show that for λ R , θ 0 , and c R , any (PS) c -sequence of J is bounded in E . Indeed, let { u n } n = 1 be a (PS) c -sequence of J . Then, for n large we have

c + 1 + u n J ( u n ) 1 t J ( u n ) , u n 1 p + 1 t Ω u n p ( x ) d x λ 1 s 1 t Ω m ( x ) u n s ( x ) d x .

On the other hand, by invoking Propositions 2.12.3 and utilizing (2.8) we have

Ω u n p ( x ) d x u n p 1

and

Ω m ( x ) u n s ( x ) d x 2 m b 1 L t ( ) t ( ) s ( ) ( b , Ω ) u n s L t ( ) s ( ) ( b , Ω ) 2 m b 1 L t ( ) t ( ) s ( ) ( b , Ω ) ( u n L t ( ) ( b , Ω ) s + + 1 ) 2 m b 1 L t ( ) t ( ) s ( ) ( b , Ω ) ( S b s + u n s + + 1 ) .

From the last three estimations we deduce the boundedness of { u n } n = 1 in E since 1 < s + < p . Hence, up to a subsequence we have

(4.4) u n u in E as n .

(i) Case θ = 0 , λ R . Since

Ω a ( x , u n ) ( u n u ) d x = J ( u n ) , u n u + λ Ω m ( x ) ( u n + ) s ( x ) 1 ( u n u ) d x ,

we easily obtain from (4.4) and Proposition 2.6 that

Ω a ( x , u n ) ( u n u ) d x 0 as n .

Then, as in the proof of Lemma 3.3 we obtain u n u in E as n .

(ii) Case θ > 0 , λ R . By Theorem 2.7 again, we infer from the boundedness of { u n } n = 1 that up to a subsequence,

(4.5) u n ( x ) u ( x ) a.e. x Ω ,

(4.6) u n p ( x ) μ u p ( x ) + i μ i δ x i in ( Ω ¯ ) ,

(4.7) b u n t ( x ) ν = b u t ( x ) + i ν i δ x i in ( Ω ¯ ) ,

(4.8) S b ν i 1 / t ( x i ) μ i 1 / p ( x i ) , i .

We claim that = . Suppose by contradiction that there exists i . Arguing as that obtained (3.18), we obtain

(4.9) θ ν i S b N min θ 1 h + 1 , θ 1 h 1 .

On the other hand, we have

c = J ( u n ) 1 p + J ( u n ) , u n + o n ( 1 ) λ 1 s 1 p + Ω m ( x ) u n s ( x ) d x + θ 1 p + 1 t Ω b ( x ) u n t ( x ) d x + o n ( 1 ) .

Passing to the limit as n in the last estimation and invoking Proposition 2.6 again and using (4.7), we obtain

c λ 1 s 1 p + Ω m ( x ) u s ( x ) d x + θ 1 p + 1 t Ω b ( x ) u t ( x ) d x + ν i .

Invoking (4.9) the preceding inequality gives

c 1 p + 1 t θ Ω b ( x ) u t ( x ) d x λ 1 s 1 p + Ω m ( x ) u s ( x ) d x + k ( θ ) ,

where k ( θ ) 1 p + 1 t S b N min θ 1 h + 1 , θ 1 h 1 . From Proposition 2.1, we have

(4.10) Ω m u s ( x ) d x = Ω b m b 1 u s ( x ) d x 2 m b 1 L ( ) ( ) 1 ( b , Ω ) u s L ( ) ( b , Ω ) ,

where ( x ) t ( x ) s ( x ) . From the last two inequalities, we obtain

(4.11) c a 1 θ Ω b ( u s ( x ) ) ( x ) d x b 1 λ u s L ( ) ( b , Ω ) + k ( θ ) ,

where a 1 1 p + 1 t > 0 and b 1 2 1 s 1 p + m b 1 L ( ) ( ) 1 ( b , Ω ) > 0 .

  • If u s L ( ) ( b , Ω ) 1 , then (4.11) and Proposition 2.2 yield

c a 1 θ ξ b 1 λ ξ + k ( θ ) g 1 ( ξ ) with ξ = u s L ( ) ( b , Ω ) 1 .

Thus,

c min ξ 0 g 1 ( ξ ) = g 1 b 1 λ a 1 θ 1 1 ,

i.e.,

c k ( θ ) ( 1 ) ( ) 1 a 1 1 1 b 1 1 θ 1 1 λ 1 .

c a 1 θ ξ + b 1 λ ξ + k ( θ ) g 2 ( ξ ) with ξ = u s L ( ) ( b , Ω ) [ 0 , 1 ) .

Thus,

c min ξ 0 g 2 ( ξ ) = g 2 b 1 λ + a 1 θ 1 + 1 ,

i.e.,

c k ( θ ) ( + 1 ) ( + ) + + 1 a 1 1 + 1 b 1 + + 1 θ 1 + 1 λ + + 1 .

Therefore, in any case, we obtain

c k ( θ ) K max θ 1 + 1 , θ 1 1 max λ + + 1 , λ 1 ,

where K max * { + , } ( 1 ) ( ) 1 a 1 1 1 b 1 1 > 0 . This contradicts to (4.3) and hence, = . It follows that Ω b ( x ) u n t ( x ) d x Ω b ( x ) u t ( x ) d x as n . We then argue as in the proof of Lemma 3.3 to conclude that u n u in E as n . The proof is complete.□

4.1 Multiplicity of solutions

In this subsection, we borrow the idea from [1] to prove Theorem 1.3, which asserts the existence of k pairs of solutions to problem (1.10) for an arbitrarily given number k . To seek such a k pairs of solutions, we determine the critical points of the functional Ψ defined in (4.2). Throughout this subsection, in addition to (S) , (W) , and (L1) we always assume that m ( x ) > 0 a.e. in B . We will frequently make use of the following estimates that can be easily obtained by using (3.28), (4.10), and Proposition 2.3: there exists a constant C 1 > 1 such that for all u E with u > 1 ,

(4.12) Ω m ( x ) u s ( x ) d x C 1 s + u s + and Ω b ( x ) u t ( x ) d x C 1 t + u t + .

Let λ , θ > 0 to be specified later. By ( A3 ) , we have

Ψ ( u ) Ω 1 p ( x ) u p ( x ) d x λ Ω m ( x ) s ( x ) u s ( x ) d x θ Ω b ( x ) t ( x ) u t ( x ) d x 1 p + Ω u p ( x ) d x λ s Ω m ( x ) u s ( x ) d x θ t Ω b ( x ) u t ( x ) d x , u E .

By invoking Proposition 2.2 then using (4.12) we deduce from the last inequality that

Ψ ( u ) 1 p + u p C 1 s + λ s u s + C 1 t + θ t u t + for all u E with u 1 .

Thus,

(4.13) Ψ ( u ) g λ ( u ) for all u E with u 1 ,

where g λ C [ 0 , ) is given by

g λ ( τ ) 1 p + τ p C 1 s + λ s τ s + C 1 t + θ t τ t + .

Rewrite g λ ( τ ) = C 1 s + τ s + s ( h ( τ ) λ ) with

h ( τ ) a 0 τ p s + b 0 θ τ t + s + ,

where

(4.14) a 0 ( p + C 1 s + ) 1 s > 0 , b 0 ( t ) 1 s C 1 t + s + > 0 .

It is clear that

max τ 0 h ( τ ) = h ( p s + ) a 0 ( t + s + ) b 0 θ 1 t + p = c 0 θ s + p t + p > 0 ,

where c 0 a 0 t + s + t + p b 0 s + p t + p p s + t + s + p s + t + p t + p t + s + > 0 . Thus, for λ ( 0 , λ 0 ( θ ) ) with

(4.15) λ 0 ( θ ) c 0 θ s + p t + p ,

g λ ( τ ) has only positive roots T 0 and T 1 with

(4.16) 0 < T 0 < T ( p s + ) a 0 ( t + s + ) b 0 θ 1 t + p = d 0 θ 1 t + p < T 1 ,

where d 0 ( p s + ) a 0 ( t + s + ) b 0 1 t + p > 0 . Obviously, on [ 0 , ) the function g λ ( τ ) is negative only on ( 0 , T 0 ) ( T 1 , ) . It is clear that

(4.17) T 0 > λ a 0 1 p s +

and for θ ( 0 , θ ) with

(4.18) θ c 0 t + p p s + a 0 t + p s + p ,

we have that λ 0 ( θ ) > a 0 . From the above analysis, we infer that for θ ( 0 , θ ) and λ ( a 0 , λ 0 ( θ ) ) , g λ ( τ ) has only positive roots T 0 and T 1 with 1 < T 0 < T < T 1 .

Let θ ( 0 , θ ) . For each λ ( a 0 , λ 0 ( θ ) ) , we define the truncated functional T λ : E R as

T λ ( u ) Ω A ( x , u ) d x λ Ω m ( x ) s ( x ) u s ( x ) d x ϕ ( u ) θ Ω b ( x ) t ( x ) u t ( x ) d x , u E ,

where ϕ C c ( R ) , 0 ϕ ( τ ) 1 for all τ R , ϕ ( τ ) = 1 for τ T 0 , and ϕ ( τ ) = 0 for τ T 1 . Note that u ϕ ( u ) is of class C 1 ( E , R ) due to [22, Corollary 3.3] and therefore, T λ C 1 ( E , R ) . Moreover, it holds that

(4.19) T λ ( u ) Ψ ( u ) for all u E ,

(4.20) T λ ( u ) = Ψ ( u ) for all u E with u < T 0 ,

and

(4.21) T λ ( u ) = Ω A ( x , u ) d x λ Ω m ( x ) s ( x ) u s ( x ) d x for all u E with u > T 1 .

We have the following.

Lemma 4.2

Let θ ( 0 , θ ) and let λ ( a 0 , λ 0 ( θ ) ) . Then, T λ ( u ) < 0 implies that u < T 0 and hence, T λ ( u ) = Ψ ( u ) .

Proof

Suppose that T λ ( u ) < 0 . Then, we have Ψ ( u ) < 0 due to (4.19). Suppose on the contrary that u T 0 . Note that T 0 > 1 , and hence, (4.13) gives g λ ( u ) < 0 . Thus, u > T 1 and hence; T λ ( u ) = Ω A ( x , u ) d x λ Ω m ( x ) s ( x ) u s ( x ) d x < 0 due to (4.21). On the other hand, the last inequality combining with ( A3 ) and (4.12) gives

1 p + u p 1 s C 1 s + λ u s + .

This yields

T 0 u p + C 1 s + λ s 1 p s + ,

which contradicts to (4.17). Thus, u < T 0 and hence, T λ ( u ) = Ψ ( u ) due to (4.20). The proof is complete.□

Next, by employing genus theory we will show that for any given number k N , T λ admits at least k pairs of critical points { ± u n } n = 1 k with T λ ( ± u n ) < 0 provided θ is sufficiently small. Let Σ denote the family of sets A E { 0 } such that A is closed in E and symmetric with respect to the origin. For A Σ , define the genus of A to be n (denoted by γ ( A ) = n ) if there is a map φ C ( A , R n { 0 } ) and n is the smallest integer with this property. When there does not exist a finite such n , set γ ( A ) = . Finally, set γ ( ) = 0 . For each k N , define

Σ k { A Σ : γ ( A ) k } .

Clearly, ( H B ρ ) Σ k for any k -dimensional subspace H of E and for any ρ > 0 (see [26, Proposition 7.7]). Thus, we can define

(4.22) c k ( λ , θ ) inf A Σ k sup u A T λ ( u ) .

Clearly, c k ( λ , θ ) c k + 1 ( λ , θ ) for all k N . We will show that with λ , θ specified later, c k ( λ , θ ) are negative critical values of T λ and hence, of Ψ . To this end, let E k be a k -dimensional subspace of E given by (3.21) with B in ( L 1 ) . Since all norms on E k are mutually equivalent, we find δ k > 1 such that

(4.23) δ k 1 max { u L p B + ( B ) , u L q B ( B ) } u δ k u L s ( ) ( m , B ) , u E k .

It is clear that we can choose { δ k } k = 1 such that δ k < δ k + 1 for all k N . Recalling a 0 given by (4.14), we have the following.

Lemma 4.3

There exists { λ k } k = 1 with max { 1 , a 0 } < λ k < λ k + 1 for all k N such that for each k N , it holds that

< c k ( λ , θ ) < 0

for any λ > λ k and θ > 0 .

Proof

Let k N be arbitrary and fixed and let λ , θ be positive real numbers. By the definition of T λ and the fact that s + < p , we easily see that T λ is bounded from below and hence,

c k ( λ , θ ) > .

Let u E k with u = τ > δ k (hence, u L s ( ) ( m , B ) > 1 due to (4.23)). We have

T λ ( u ) C p u L p B + ( B ) p B + + C B u L q B ( B ) q B λ s + u L s ( ) ( m , B ) s B

in view of (1.11) and Proposition 2.2. From this and (4.23) we obtain

(4.24) T λ ( u ) C p ( δ k τ ) p B + + C B ( δ k τ ) q B λ s + ( δ k 1 τ ) s B = τ q B h ¯ λ ( τ ) ,

where

h ¯ λ ( τ ) α k τ p B + q B + β k γ k λ τ s B q B

with α k C p δ k p B + , β k C B δ k q B and γ k ( s + ) 1 δ k s B . Denote

(4.25) T ( s B q B ) γ k λ ( p B + q B ) α k 1 p B + s B .

Then, for

λ > C 1 ( B ) δ k 2 s B

with C 1 ( B ) C B ( p B + q B ) p B + s B p B + s B p B + q B C p ( p B + q B ) s B q B s B q B p B + q B s + , it holds that

(4.26) h ¯ λ ( T ) = β k p B + s B p B + q B s B q B p B + q B s B q B p B + s B α k s B q B p B + s B γ k p B + q B p B + s B λ p B + q B p B + s B < 0 .

From (4.25) we have

(4.27) T > δ k λ > C 2 ( B ) δ k 2 p B +

with C 2 ( B ) ( p B + q B ) C p s + s B q B . Set

(4.28) λ k C 3 ( B ) δ k 2 p B + ,

where C 3 ( B ) max { 1 , a 0 , C 1 ( B ) , C 2 ( B ) } with a 0 given by (4.18). Then, we have that λ k > max { 1 , a 0 } since δ k > 1 . Moreover, for λ > λ k it holds that T > δ k due to (4.27) and therefore,

T λ ( u ) ( T ) q B h ¯ λ ( T ) ε k < 0 , u B T E k

in view of (4.24) and (4.26). Thus, B T E k { u E : T λ ( u ) ε k } T λ ε k . Note that T λ ε k Σ due to ( A0 ) and the definition of T λ . Hence, we arrive at

γ ( T λ ε k ) γ ( B T E k ) = k

(see [26, Proposition 7.7]). In other words, T λ ε k Σ k and thus,

c k ( λ , θ ) = inf A Σ k sup u A T λ ( u ) sup u T λ ε k T λ ( u ) ε k < 0 .

Finally, it is obvious that this sequence { λ k } k = 1 satisfies λ k < λ k + 1 for all k N . The proof is complete.□

Let { λ k } k = 1 be defined as in (4.28). Note that with λ 0 ( θ ) defined by (4.15) we have

λ k = λ 0 ( θ ) = c 0 θ s + p t + p θ = [ c 0 1 C 3 ( B ) ] t + p s + p δ k 2 p B + ( t + p ) p s + .

Moreover, for θ ( 0 , 1 ) and λ > 1 , the inequality

(4.29) 0 < 1 p + 1 t S b N min θ 1 h + 1 , θ 1 h 1 K max θ 1 + 1 , θ 1 1 max λ + + 1 , λ 1

is equivalent to

(4.30) λ < K 1 1 p + 1 t S b N 1 θ h + ( h + 1 ) = e 0 θ h + ( h + 1 ) λ ¯ 0 ( θ )

with e 0 K 1 1 p + 1 t S b N 1 . We also note that

λ k = λ ¯ 0 ( θ ) θ = [ e 0 1 C 3 ( B ) ] ( h + 1 ) h + δ k 2 p B + ( h + 1 ) h + .

Thus, by taking

(4.31) θ k min 1 , θ , [ c 0 1 C 3 ( B ) ] t + p s + p , [ e 0 1 C 3 ( B ) ] ( h + 1 ) h + δ k κ

with κ max 2 p B + ( t + p ) p s + , 2 p B + ( h + 1 ) h + and setting

(4.32) λ ( θ ) min { λ 0 ( θ ) , λ ¯ 0 ( θ ) } ,

we deduce from the definitions of λ 0 ( θ ) and λ ¯ 0 ( θ ) in (4.15) and (4.30) that θ k > 0 is independent of λ and for any θ ( 0 , θ k ) ,

θ < min { 1 , θ } and λ k < λ ( θ ) .

We are now ready to prove Theorem 1.3.

Proof of Theorem 1.3

Let k N be given and define λ k and θ k as in (4.28) and (4.31), respectively. Let θ ( 0 , θ k ) be given and let λ ( λ k , λ ( θ ) ) with λ ( θ ) given by (4.32). Let { c n ( λ , θ ) } n = 1 be defined as in (4.22); for simplicity of notation, we write c n in place of c n ( λ , θ ) . By the definitions of { θ n } n = 1 and { λ n } n = 1 , for any n { 1 , , k } we have that 0 < θ < θ n < 1 and λ n < λ < λ ( θ ) as well. By Lemma 4.3, it follows that < c n < 0 . From this and (4.29)–(4.30) we infer that T λ satisfies the (PS) c n condition in view of Lemmas 4.1 and 4.2. Suppose that c n = = c k = c for some n { 1 , , k } . Then, it follows from the fact that T λ satisfies the (PS) c n condition that K c { u E \ { 0 } : T λ ( u ) = 0 and T λ ( u ) = c } is a compact set. By a standard argument using the deformation lemma and properties of genus (see, e.g., [1, Proof of Theorem 1]) we obtain

γ ( K c ) k n + 1 .

This implies that either { c n } n = 1 k are k distinct critical values of T λ or K c k admits infinitely many points. From this and the fact that c 1 c 2 c k < 0 , Ψ admits at least k pairs of critical points { ± u n } n = 1 k with Ψ ( ± u n ) < 0 for n = 1 , , k in view of Lemma 4.2. Consequently, problem (1.10) admits at least k pair of solutions { ± u n } n = 1 k with negative energy. The proof is complete.□

Remark 4.4

As we saw in the proof of Theorem 1.3, the range of θ to obtain at least k pair of solutions is ( 0 , θ k ) . Since θ k 0 as k , we cannot conclude that problem (1.10) admits a sequence of solution with a certain range of θ . This similarly happens with the proof of [1, Theorem 1]. Indeed, [1, equation (70)] gives the range of K of the form K < C d j α for some positive constants C , α independent of d j , where d j is given by [1, equation (67)] and possibly satisfies that d j 0 as j .

4.2 Nontrivial nonnegative solutions

In this subsection, we will prove Theorem 1.4. We will apply the Ekeland variational principle to determine critical points of the functional J given by (4.1), which are in turn nonnegative solutions to problem (1.10). Throughout this subsection, in addition to (S) , (W) , and (L1) , we always assume that (L2) holds.

We start with the following geometry of J .

Lemma 4.5

For each given λ > λ , there exists θ ˜ > 0 such that for any θ ( θ ˜ , θ ˜ ) , there exist r , ρ > 0 such that

inf u B r J ( u ) ρ > 0 > inf u B r J ( u ) .

Proof

Fix λ > λ with λ given by (1.12) and let θ R . Since λ > λ we can find ϕ A such that

(4.33) max η [ s B , s B + ] C ( η ) B m ( x ) s ( x ) ϕ + s ( x ) d x B ϕ p B + d x η q B p B + q B B ϕ q B d x p B + η p B + q B < λ .

Fix such ϕ . By (1.11) and (L2) , we have

J ( τ ϕ ) C p B ϕ p B + d x τ p B + + C B B ϕ q B d x τ q B λ B m ( x ) s ( x ) ϕ + s ( x ) d x τ s B ( τ ) θ B b ( x ) t ( x ) ( τ ϕ + ) t ( x ) d x , τ > 0 ,

where s B ( τ ) is determined by ϕ and τ as in (L2) . Thus,

(4.34) J ( τ ϕ ) y λ ( τ ) B b ( x ) t ( x ) ( τ ϕ + ) t ( x ) d x θ , τ 0 ,

where

y λ ( τ ) α 1 τ p B + β 1 λ τ s B ( τ ) + γ 1 τ q B

with α 1 C p B ϕ p B + d x > 0 , β 1 B m ( x ) s ( x ) ϕ + s ( x ) d x > 0 , and γ 1 C B B ϕ q B d x > 0 . We claim that there exists τ 0 = τ 0 ( λ ) ( 0 , ) such that

(4.35) y λ ( τ 0 ) < 0 .

Indeed, we write

(4.36) y λ ( τ ) = τ q B y ¯ λ ( τ ) for τ > 0 ,

where y ¯ λ ( τ ) α 1 τ p B + q B β 1 λ τ s B ( τ ) q B + γ 1 for τ > 0 . It is clear that s B ( ) : ( 0 , ) [ s B , s B + ] is a continuous function, and hence, there exist τ 0 > 0 such that

(4.37) τ 0 = s B ( τ 0 ) q B p B + q B α 1 1 β 1 λ 1 p B + s B ( τ 0 ) .

We have

y ¯ λ ( τ 0 ) = γ 1 p B + s B ( τ 0 ) p B + q B s B ( τ 0 ) q B p B + q B s B ( τ 0 ) q B p B + s B ( τ 0 ) α 1 q B s B ( τ 0 ) p B + s B ( τ 0 ) β 1 p B + q B p B + s B ( τ 0 ) λ p B + q B p B + s B ( τ 0 ) .

Thus, y ¯ λ ( τ 0 ) < 0 due to (4.33); hence,

(4.38) y λ ( τ 0 ) < 0 ,

due to (4.36). Then, from (4.38) and (4.34) we have

(4.39) J ( τ 0 ϕ ) y λ ( τ 0 ) B b ( x ) t ( x ) ( τ 0 ϕ + ) t ( x ) d x θ < 0 ,

provided

(4.40) θ > θ ¯ ,

where θ ¯ y λ ( τ 0 ) B b ( x ) t ( x ) ( τ 0 ϕ + ) t ( x ) d x > 0 . Set

(4.41) r λ max 1 + τ 0 ϕ , ( 2 p + C 1 s + ( s ) 1 λ ) 1 p s +

and

(4.42) θ ˜ min θ ¯ , t r λ p t + 2 p + C 1 t + ,

where C 1 is given in (4.12). Then, for any θ ( θ ˜ , θ ˜ ) , by choosing r = r λ and ρ = C 1 t + r λ t + t ( θ ˜ θ ) and utilizing (4.12) and (4.41)–(4.42) we have

J ( u ) 1 p + u p C 1 t + λ s u s + C 1 t + θ t u t + 1 2 p + u p C 1 t + θ t u t + = C 1 t + r λ t + t ( θ ˜ θ ) , u B r ,

i.e.,

(4.43) J ( u ) ρ , u B r .

Finally, note that τ 0 ϕ B r by the choice of r and (4.41); hence, (4.39) yields

inf u B r J ( u ) J ( τ 0 ϕ ) < 0 .

This and (4.43) complete the proof.□

We are now in a position to give a proof of Theorem 1.4.

Proof of Theorem 1.4

Let λ > λ be arbitrary and fixed. Note that for θ ( 0 , 1 ) , the right-hand side of (4.3) can be rewritten as

C 2 θ 1 h + 1 C 3 ( λ ) θ 1 1 = θ 1 h + 1 C 2 C 3 ( λ ) θ 1 h + 1 1 1 ,

where C 2 , C 3 ( λ ) > 0 are positive constants independent of θ . Note that 1 h + 1 1 1 > 0 due to the assumption t p + < t s . Hence, we can choose θ ( 0 , min { 1 , θ ˜ } ) , where θ ˜ is as in Lemma 4.5, such that

(4.44) 0 < 1 p + 1 t S b N min θ 1 h + 1 , θ 1 h 1 K max { θ 1 + 1 , θ 1 1 } max λ + + 1 , λ 1 , θ ( 0 , θ ) .

Let θ [ 0 , θ ) . By Lemma 4.5, there exist r , ρ > 0 such that

(4.45) inf u B r J ( u ) ρ > 0 > inf u B r J ( u ) c .

Then, arguing as in [13, Proof of Theorem 3.1] in which the Ekeland variational principle was applied, we find a (PS) c -sequence { u n } n = 1 . On the other hand, from (4.44) and (4.45), J satisfies the (PS) c condition in view of Lemma 4.1, and hence u n u in E as n . Thus,

J ( u ) = 0 and J ( u ) = c < 0 .

That is, u is a nontrivial nonnegative solution to problem (1.10).□

Let us conclude this section by pointing out that Theorem 1.4 remains valid for subcritical problem. More precisely, consider the following problem:

(4.46) div a ( x , u ) = λ m ( x ) u s ( x ) 2 u + μ ω ( x ) u r ( x ) 2 u in Ω , u = 0 on Ω ,

where (A0)–(A5) hold; p , q , s , r C + ( Ω ¯ ) such that s + < p p + < r and r ( x ) < p ( x ) for all x Ω ¯ ; and the weights m and ω satisfy the following assumption.

  1. m L p ( ) p ( ) s ( ) ( Ω ) , ω L + p ( ) p ( ) r ( ) ( Ω ) and there exists a ball B Ω satisfying

    1. q B < s B ;

    2. { x B : m ( x ) > 0 } > 0 ;

    3. there exists ϕ C c ( B ) such that B m ( x ) s ( x ) ϕ + s ( x ) d x > 0 and for each τ > 0 , there exists s B ( τ ) [ s B , s B + ] such that

      B m ( x ) s ( x ) ϕ + s ( x ) τ s ( x ) d x = τ s B ( τ ) B m ( x ) s ( x ) ϕ + s ( x ) d x .

Theorem 4.6

Let (A0)–(A5) and (A) hold with s + < p p + < r and r ( x ) < p ( x ) for all x Ω ¯ . Then, for each given λ > λ with λ given by (1.12), there exists μ > 0 such that for any μ ( μ , μ ) , problem (4.46) has a nontrivial nonnegative solution and has two nontrivial nonnegative solutions for any μ ( 0 , μ ) . Moreover, if r ( ) is constant in Ω , it suffices to assume ω L p ( ) p ( ) r ( ) ( Ω ) and { x Ω : ω ( x ) > 0 } > 0 instead of ω L + p ( ) p ( ) r ( ) ( Ω ) .

This result can be proved using the same method as in [15, Theorem 1.2] so we omit it.

5 Comment on the regularity of solutions

Let functions d in (F1), a j in (F4) or m in ( W ) or m , ω in ( A ) be in the class L ( Ω ) . Then by [16], all aforementioned solutions are of class L ( Ω ) . Moreover, we have that:

  1. when p C 1 ( Ω ¯ ) and A ( x , u , u ) a ( x , u ) additionally satisfies assumption ( A k ) in [8], for instance, the p ( ) -Laplacian or the generalized mean curvature operator, any aforementioned solution is also of class C 1 , β ( Ω ¯ ) and the convergence in Theorem 1.1 is in the C 1 ( Ω ¯ ) -topology (see [8, Theorem 1.2]);

  2. when p ( ) is constant and A ( x , u , u ) a ( x , u ) additionally satisfies conditions (1.10a)–(1.10c) in [24], for instance, the case of the multiple p -Laplacian, any aforementioned solution is also of class C 1 , β ( Ω ¯ ) and the convergence in Theorem 1.1 is in the C 1 ( Ω ¯ ) -topology (see [24, Theorem 1.7 and the comments below Theorem 1.7]).

  1. Funding information: Ky Ho was supported by University of Economics Ho Chi Minh City, Vietnam. Inbo Sim was supported by the National Research Foundation of Korea Grant funded by the Korea Government (MEST) (NRF-2021R1I1A3A0403627011).

  2. Conflict of interest: The authors state no conflict of interest.

References

[1] L. Baldelli, Y. Brizi, and R. Filippucci, Multiplicity results for (p,q)-Laplacian equations with critical exponent in RN and negative energy, Calc. Var. Partial Differential Equations 60 (2021), no. 8, 30. 10.1007/s00526-020-01867-6Search in Google Scholar

[2] J. F. Bonder and A. Silva, Concentration-compactness principle for variable exponent spaces and applications, Electron. J. Differ. Equ. 141 (2010), 1–18. Search in Google Scholar

[3] H. Brezis, Functional Analysis, Sobolev Spaces and Partial Differential Equations, Universitext, Springer, New York, 2011. 10.1007/978-0-387-70914-7Search in Google Scholar

[4] D. G. de Figueiredo, J.-P. Gossez, and P. Ubilla, Local superlinearity and sublinearity for indefinite semilinear elliptic problems, J. Funct. Anal. 199 (2003), no. 2, 452–467. 10.1007/978-3-319-02856-9_36Search in Google Scholar

[5] D. G. de Figueiredo, J.-P. Gossez, and P. Ubilla, Multiplicity results for a family of semilinear elliptic problems under local superlinearity and sublinearity, J. Eur. Math. Soc. 8 (2006), no. 2, 269–286. 10.4171/JEMS/52Search in Google Scholar

[6] D. G. de Figueiredo, J.-P. Gossez, and P. Ubilla, Local “superlinearity” and “sublinearity” for the p-Laplacian, J. Funct. Anal. 257 (2009), 721–752. 10.1007/978-3-319-02856-9_43Search in Google Scholar

[7] L. Diening, P. Harjulehto, P. Hästö, and M. Růžička, Lebesgue and Sobolev spaces with variable exponents, Lecture Notes in Mathematics. 2017, Springer-Verlag, Heidelberg, 2011. 10.1007/978-3-642-18363-8Search in Google Scholar

[8] X. Fan, Global C1,α regularity for variable exponent elliptic equations in divergence form, J. Differ. Equ. 235 (2007), 397–417. 10.1016/j.jde.2007.01.008Search in Google Scholar

[9] X. Fan and D. Zhao, On the spaces Lp(x)(Ω) and Wm,p(x)(Ω), J. Math. Anal. Appl. 263 (2001), no. 2, 424–446. 10.1006/jmaa.2000.7617Search in Google Scholar

[10] X. Fan, Q. Zhang, and D. Zhao, Eigenvalues of p(x)-Laplacian Dirichlet problem, J. Math. Anal. Appl. 302 (2005), 306–317. 10.1016/j.jmaa.2003.11.020Search in Google Scholar

[11] G. M. Figueiredo, Existence and multiplicity of solutions for a class of p&q elliptic problems with critical exponent, Math. Nachr. 286 (2013), no. 11–12, 1129–1141. 10.1002/mana.201100237Search in Google Scholar

[12] H. H. Ha, K. Ho, and I. Sim, Infinitely many solutions for a generalized p(⋅) Laplace equation involving Leray-Lions-type operators, Math. Methods Appl. Sci. 45 (2022), no. 14, 8450–8473.10.1002/mma.7246Search in Google Scholar

[13] K. Ho and I. Sim, Existence and multiplicity of solutions for degenerate p(x)-Laplace equations involving concave-convex type nonlinearities with two parameters, Taiwanese J. Math. 19 (2015), no. 5, 1469–1493. 10.11650/tjm.19.2015.5187Search in Google Scholar

[14] K. Ho and I. Sim, On degenerate p(x)-Laplace equations involving critical growth with two parameters, Nonlinear Anal. 132 (2016), 95–114. 10.1016/j.na.2015.11.003Search in Google Scholar

[15] K. Ho and I. Sim, An existence result for (p,q)-Laplace equations involving sandwich-type and critical growth, Appl. Math. Lett. 111 (2021), 106646. 10.1016/j.aml.2020.106646Search in Google Scholar

[16] K. Ho, Y.-H. Kim, P. Winkert, and C. Zhang, The boundedness and Hölder continuity of weak solutions to elliptic equations involving variable exponents and critical growth, J. Differ. Equ. 313 (2022), 503–532. 10.1016/j.jde.2022.01.004Search in Google Scholar

[17] R. Kajikiya, A critical point theorem related to the symmetric mountain pass lemma and its applications to elliptic equations, J. Funct. Anal. 225 (2005), 352–370. 10.1016/j.jfa.2005.04.005Search in Google Scholar

[18] B. Kawohl, M. Lucia, and S. Prashanth, Simplicity of the principal eigenvalue for indefinite quasilinear problems, Adv. Differ. Equ. 12 (2007), no. 4, 407–434. 10.57262/ade/1355867457Search in Google Scholar

[19] Y.-H. Kim and I. Sim, Existence of solutions and positivity of the infimum eigenvalue for degenerate elliptic equations with variable exponents, Discrete Contin. Dyn. Syst. Supplement (2013), 695–707. Search in Google Scholar

[20] Y. Komiya and R. Kajikiya, Existence of infinitely many solutions for the (p,q)-Laplace equation, NoDEA Nonlinear Differential Equations Appl. 23 (2016), no. 49, 1–23. 10.1007/s00030-016-0402-1Search in Google Scholar

[21] O. Kováčik and J. Rákosník, On spaces Lp(x) and Wk,p(x), Czechoslovak Math. J. 41 (1991), 592–618. 10.21136/CMJ.1991.102493Search in Google Scholar

[22] J. Lang, O. Méndez, Stability of the Norm-Eigenfunctions of the p(⋅)-Laplacian, Integr. Equ. Oper. Theory 85 (2016), 245–257. 10.1007/s00020-015-2275-9Search in Google Scholar

[23] V. K. Le, On a sub-supersolution method for variational inequalities with Leray-Lions operators in variable exponent spaces, Nonlinear Anal. 71 (2009), 3305–3321. 10.1016/j.na.2009.01.211Search in Google Scholar

[24] G. M. Lieberman, The natural generalization of the natural conditions of Ladyzhenskaya and Uralatseva for elliptic equations, Commun. Partial Differ. Equ. 16 (1991), 311–361. 10.1080/03605309108820761Search in Google Scholar

[25] M. Mihailescu and V. Rǎdulescu, Continuous spectrum for a class of nonhomogeneous differential operators, Manuscripta Math. 125 (2008), 157–167. 10.1007/s00229-007-0137-8Search in Google Scholar

[26] P. H. Rabinowitz, Minimax methods in critical point theory with applications to differential equations, CBMS Regional Conference Series in Mathematics, vol. 65, American Mathematical Society, Providence, RI, 1986. 10.1090/cbms/065Search in Google Scholar

[27] V. Rǎdulescu, Nonlinear elliptic equations with variable exponent: old and new, Nonlinear Anal. 121 (2015), 336–369. 10.1016/j.na.2014.11.007Search in Google Scholar

[28] E. A. B. Silva and M. S. Xavier, Multiplicity of solutions for quasilinear elliptic problems involving critical Sobolev exponents, Ann. Inst. H. Poincaré Anal. Non Linéaire 20 (2003), 341–358. 10.1016/s0294-1449(02)00013-6Search in Google Scholar

[29] C. Yu and D. Ri, Global L∞-estimates and Hölder continuity of weak solutions to elliptic equations with the general nonstandard growth conditions, Nonlinear Anal. 156 (2017), 144–166. 10.1016/j.na.2017.02.019Search in Google Scholar

Received: 2022-04-11
Revised: 2022-07-15
Accepted: 2022-07-17
Published Online: 2022-09-03

© 2023 Ky Ho and Inbo Sim, published by De Gruyter

This work is licensed under the Creative Commons Attribution 4.0 International License.

Articles in the same Issue

  1. Regular Articles
  2. On sufficient “local” conditions for existence results to generalized p(.)-Laplace equations involving critical growth
  3. On the critical Choquard-Kirchhoff problem on the Heisenberg group
  4. On the local behavior of local weak solutions to some singular anisotropic elliptic equations
  5. Viscosity method for random homogenization of fully nonlinear elliptic equations with highly oscillating obstacles
  6. Double-phase parabolic equations with variable growth and nonlinear sources
  7. Logistic damping effect in chemotaxis models with density-suppressed motility
  8. Bifurcation diagrams of one-dimensional Kirchhoff-type equations
  9. Standing wave solution for the generalized Jackiw-Pi model
  10. Weak-strong uniqueness and energy-variational solutions for a class of viscoelastoplastic fluid models
  11. Eigenvalue inequalities for the buckling problem of the drifting Laplacian of arbitrary order
  12. Homoclinic solutions for a differential inclusion system involving the p(t)-Laplacian
  13. Equivalence between a time-fractional and an integer-order gradient flow: The memory effect reflected in the energy
  14. Bautin bifurcation with additive noise
  15. Small solitons and multisolitons in the generalized Davey-Stewartson system
  16. Nonstationary Poiseuille flow of a non-Newtonian fluid with the shear rate-dependent viscosity
  17. A global compactness result with applications to a Hardy-Sobolev critical elliptic system involving coupled perturbation terms
  18. On a strongly damped semilinear wave equation with time-varying source and singular dissipation
  19. Singularity for a nonlinear degenerate hyperbolic-parabolic coupled system arising from nematic liquid crystals
  20. Stability of stationary solutions to the three-dimensional Navier-Stokes equations with surface tension
  21. Uniform decay estimates for the semi-linear wave equation with locally distributed mixed-type damping via arbitrary local viscoelastic versus frictional dissipative effects
  22. Symmetry and nonsymmetry of minimal action sign-changing solutions for the Choquard system
  23. Periodic and quasi-periodic solutions of a four-dimensional singular differential system describing the motion of vortices
  24. Multiple nontrivial solutions of superlinear fractional Laplace equations without (AR) condition
  25. Existence and blow-up of solutions in Hénon-type heat equation with exponential nonlinearity
  26. Existence and concentration behavior of positive solutions to Schrödinger-Poisson-Slater equations
  27. On the fractional Korn inequality in bounded domains: Counterexamples to the case ps < 1
  28. Gradient estimates for nonlinear elliptic equations involving the Witten Laplacian on smooth metric measure spaces and implications
  29. Dynamical analysis of a reaction–diffusion mosquito-borne model in a spatially heterogeneous environment
  30. Existence and multiplicity of solutions for a quasilinear system with locally superlinear condition
  31. Nontrivial solution for Klein-Gordon equation coupled with Born-Infeld theory with critical growth
  32. Modeling Wolbachia infection frequency in mosquito populations via a continuous periodic switching model
  33. Infinitely many localized semiclassical states for nonlinear Kirchhoff-type equation
  34. Fujita-type theorems for a quasilinear parabolic differential inequality with weighted nonlocal source term
  35. Approximations of center manifolds for delay stochastic differential equations with additive noise
  36. Periodic solutions to a class of distributed delay differential equations via variational methods
  37. Groundstate for the Schrödinger-Poisson-Slater equation involving the Coulomb-Sobolev critical exponent
  38. Existence of nontrivial solutions for the Klein-Gordon-Maxwell system with Berestycki-Lions conditions
  39. Global Sobolev regular solution for Boussinesq system
  40. Normalized solutions for the p-Laplacian equation with a trapping potential
  41. Nonlinear elliptic–parabolic problem involving p-Dirichlet-to-Neumann operator with critical exponent
  42. Blow-up for compressible Euler system with space-dependent damping in 1-D
  43. High energy solutions of general Kirchhoff type equations without the Ambrosetti-Rabinowitz type condition
  44. On the dynamics of grounded shallow ice sheets: Modeling and analysis
  45. A survey on some vanishing viscosity limit results
  46. Blow-up for logarithmic viscoelastic equations with delay and acoustic boundary conditions
  47. Generalized Liouville theorem for viscosity solutions to a singular Monge-Ampère equation
  48. Front propagation in a double degenerate equation with delay
  49. Positive solutions for a class of singular (pq)-equations
  50. Higher integrability for anisotropic parabolic systems of p-Laplace type
  51. The Poincaré map of degenerate monodromic singularities with Puiseux inverse integrating factor
  52. On a system of multi-component Ginzburg-Landau vortices
  53. Existence and concentration of solutions to Kirchhoff-type equations in ℝ2 with steep potential well vanishing at infinity and exponential critical nonlinearities
  54. Multiplicity results for fractional Schrödinger-Kirchhoff systems involving critical nonlinearities
  55. On double phase Kirchhoff problems with singular nonlinearity
  56. Estimates for eigenvalues of the Neumann and Steklov problems
  57. Nodal solutions with a prescribed number of nodes for the Kirchhoff-type problem with an asymptotically cubic term
  58. Dirichlet problems involving the Hardy-Leray operators with multiple polars
  59. Incompressible limit for compressible viscoelastic flows with large velocity
  60. Characterizations for the genuine Calderón-Zygmund operators and commutators on generalized Orlicz-Morrey spaces
  61. Non-local gradients in bounded domains motivated by continuum mechanics: Fundamental theorem of calculus and embeddings
  62. Noncoercive parabolic obstacle problems
  63. Touchdown solutions in general MEMS models
  64. Existence and nonexistence of nontrivial solutions for the Schrödinger-Poisson system with zero mass potential
  65. Short-time existence of a quasi-stationary fluid–structure interaction problem for plaque growth
  66. Fine bounds for best constants of fractional subcritical Sobolev embeddings and applications to nonlocal PDEs
  67. Symmetries of Ricci flows
  68. Asymptotic behavior of solutions of a second-order nonlinear discrete equation of Emden-Fowler type
  69. On the topological gradient method for an inverse problem resolution
  70. Supersolutions to nonautonomous Choquard equations in general domains
  71. Uniform complex time heat Kernel estimates without Gaussian bounds
  72. Global existence for time-dependent damped wave equations with nonlinear memory
  73. Normalized solutions for a critical fractional Choquard equation with a nonlocal perturbation
  74. Reversed Hardy-Littlewood-Sobolev inequalities with weights on the Heisenberg group
  75. Lamé system with weak damping and nonlinear time-varying delay
  76. Global well posedness for the semilinear edge-degenerate parabolic equations on singular manifolds
  77. Blowup in L1(Ω)-norm and global existence for time-fractional diffusion equations with polynomial semilinear terms
  78. Boundary regularity results for minimisers of convex functionals with (p, q)-growth
  79. Parametric singular double phase Dirichlet problems
  80. Special Issue on Nonlinear analysis: Perspectives and synergies
  81. Editorial to Special issue “Nonlinear analysis: Perspectives and synergies”
  82. Identification of discontinuous parameters in double phase obstacle problems
  83. Global boundedness to a 3D chemotaxis-Stokes system with porous medium cell diffusion and general sensitivity
  84. On regular solutions to compressible radiation hydrodynamic equations with far field vacuum
  85. On Cauchy problem for fractional parabolic-elliptic Keller-Segel model
  86. The evolution of immersed locally convex plane curves driven by anisotropic curvature flow
  87. Stability of combination of rarefaction waves with viscous contact wave for compressible Navier-Stokes equations with temperature-dependent transport coefficients and large data
  88. On concave perturbations of a periodic elliptic problem in R2 involving critical exponential growth
  89. Existence of solutions for a double-phase variable exponent equation without the Ambrosetti-Rabinowitz condition
Downloaded on 7.12.2025 from https://www.degruyterbrill.com/document/doi/10.1515/anona-2022-0269/html
Scroll to top button