Home Modified quasilinear equations with strongly singular and critical exponential nonlinearity
Article Open Access

Modified quasilinear equations with strongly singular and critical exponential nonlinearity

  • Reshmi Biswas , Sarika Goyal and Konijeti Sreenadh EMAIL logo
Published/Copyright: June 8, 2024

Abstract

In this article, we study global multiplicity result for a class of modified quasilinear singular equations involving the critical exponential growth:

Δ u Δ ( u 2 ) u = λ ( α ( x ) u q + f ( x , u ) ) in Ω , u > 0 in Ω , u = 0 on Ω ,

where Ω is a smooth bounded domain in R 2 , 0 < q < 3 , and α : Ω ( 0 , + ) such that α L ( Ω ) . The function f : Ω × R R is continuous and enjoys critical exponential growth of the Trudinger-Moser type. Using a sub-super solution method, we show that there exists some Λ * > 0 such that for all λ ( 0 , Λ * ) , the problem has at least two positive solutions, for λ = Λ * , the problem achieves at least one positive solution for λ > Λ * , and the problem has no solutions.

MSC 2010: 35J20; 35J60

1 Introduction and statement of main results

In this article, we study the existence, nonexistence, and multiplicity of the positive solutions for the following modified quasilinear equation:

(P⁎) Δ u Δ ( u 2 ) u = λ ( α ( x ) u q + f ( x , u ) ) in Ω , u > 0 in Ω , u = 0 on Ω ,

where Ω is a smooth bounded domain in R 2 , 0 < q < 3 and the function f : Ω × R R is defined as f ( x , s ) = g ( x , s ) exp ( s 4 ) , where g C ( Ω ¯ × R ) satisfies some appropriate assumptions described later. We also have the following assumption on the function α : Ω R :

  1. α L ( Ω ) and α 0 inf x Ω α ( x ) > 0 .

The study on the equations driven by the modified quasilinear operator Δ u Δ ( u 2 ) u is quite popular for long because of their wide range of applications in the modelling of the physical phenomenon such as in plasma physics and fluid mechanics [6], in dissipative quantum mechanics [18], etc. Solutions of such equations (called soliton solutions) are related to the existence of standing wave solutions for quasilinear Schrödinger equations of the form

(1.1) i u t = Δ u + V ( x ) u h 1 ( u 2 ) u C Δ h 2 ( u 2 ) h ( u 2 ) u , x R N ,

where V is a potential function, C is a real constant, and h 1 and h 2 are real valued functions. Equations of the form (1.1) appear in the study of mathematical physics. Each different type of the function h 2 represents different physical phenomenon. For example, if h 2 ( s ) = s , then (1.1) is used in the modelling of the super fluid film equation in plasma physics [23]. When h 2 ( s ) = 1 + s 2 , (1.1) attributes to the study of self-channelling of a high-power ultra short laser in matter [32]. Because of the quasilinear term Δ ( u 2 ) u , present in the problems of type ( P * ), the natural energy functional associated to such problems is not well defined. Hence, we have a restriction in applying the variational method directly for studying such problems. To overcome this shortcoming, researchers developed several methods and arguments, such as the perturbation method [25,28], a constrained minimization technique [26,27,31,33], and a change of variables [8,1113,21,22].

Without the quasilinear term Δ ( u 2 ) u , the problem ( P * ) goes back to the original semilinear equation:

(1.2) Δ u = λ α ( x ) u q + f ( x , u ) in Ω , u > 0 in Ω , u = 0 on Ω .

Such kind of equations have significant applications in the physical modelling related to the boundary layer phenomenon for viscous fluids, non-Newtonian fluids, etc. If f 0 , (1.2) becomes purely singular problem, for which existence, uniqueness, non-existence and regularity results are extensively studied in [9,20] with suitable α ( x ) and for different ranges of q .

One of the main features of problem ( P * ) is that the nonlinear term f ( x , s ) enjoys the critical exponential growth with respect to the following Trudinger-Moser inequality (see [29]):

Theorem 1

(Trudinger-Moser inequality) Let Ω be a smooth bounded domain in R 2 . Then for u H 0 1 ( Ω ) and p > 0 , we have

exp ( p u 2 ) L 1 ( Ω ) .

Moreover,

sup u 1 Ω exp ( p u 2 ) d x < +

if and only if p 4 π .

Here, the Sobolev space H 0 1 ( Ω ) and the corresponding norm are defined in Section 2. The study on the critical growth exponential problems associated to this inequality were initiated with the work of Adimurthi [1] and de Figueiredo et al. [10]. Furthermore, the problem of type ( P * ) without the singular term, i.e., the equation

(1.3) Δ u Δ ( u 2 ) u + V ( x ) u = f ( x , u ) in R 2 ,

where V : R N R is a continuous potential and f : R 2 × R R is a continuous function with some suitable assumption and is having critical exponential growth ( exp ( p s 4 ) ), was studied by do Ó et al. [11] for the first time. Note that, unlike as in the case of semilinear critical exponential problem involving Laplacian, where the critical exponential growth is given as exp ( s 2 ) , in problem ( P * ) and in (1.3), the form of critical exponential nonlinearity f is considered as exp ( s 4 ) because of the quasilinear term Δ ( u 2 ) u .

Motivated by the seminal work of Ambrosetti et al. [3], many researchers studied the global multiplicity results for singular-convex problems. In [19], the author proved the global multiplicity result for (1.2), considering critical polynomial growth as f ( x , s ) = s N + 2 N 2 , and 0 < q < 1 in a general smooth bounded domain Ω R N , N > 2 . Then, in the critical dimension N = 2 , Adimurthi and Giacomoni [2] discussed the global multiplicity of H 0 1 solutions for (1.2) by taking the optimal range of q as 1 < q < 3 and considered f to be having critical exponential growth as f ( x , s ) = g ( x , s ) exp ( 4 π s 2 ) , where g is some appropriate continuous function. Later, in [17], Giacomoni and Saoudi showed the similar results as in [19] for (1.2) while considering f ( x , s ) = s N + 2 N 2 , 0 < q < 3 and adding a smooth perturbation with sub-critical asymptotic behaviour at + . In [14,15], Dhanya et al. studied the global multiplicity result for the problem of type (1.2) with critical growth nonlinearities combined with a discontinuous function multiplied with the singular term u q , 0 < q < 3 .

On the other hand, for modified quasilinear Schrödinger equations involving singular nonlinearity, there are a few results in the existing literature. Authors in [5,30,34] discussed existence of multiple solutions of such equations with singular nonlinearity u q , 0 < q < 1 in combination with some polynomial type perturbation. Very recently, in [35], the authors studied the global multiplicity results for the equations of type ( P * ) in a smooth bounded domain Ω R N , N > 2 , for the first time, by assuming critical polynomial growth f ( x , s ) = b ( x ) s 3 N + 2 N 2 , where b is a sign changing continuous function.

Inspired by all these aforementioned works, in this article, we investigate the existence, multiplicity, and non-existence results (i.e., the global multiplicity result) for problem ( P * ) driven by modified quasilinear operator involving singular and critical exponential nonlinearity, which was an open question. The main mathematical difficulties we face in studying problem ( P * ) occur in three-folds as following:

  • The modified quasilinear term Δ ( u 2 ) u prohibits the natural energy functional corresponding to problem ( P * ) to be well defined for all u H 0 1 ( Ω ) (defined in Section 2).

  • The critical exponential nature of f induces non-compactness of the Palais-Smale sequences.

  • The growth of the singular nonlinearity falls into the range 0 < q < 3 , which again prevents the associated energy functional to problem ( P * ) to be well defined for all u H 0 1 ( Ω ) .

First, we transform the problem ( P * ) by using a change of variable as in [8,35]. Then, by applying variational and sub-super solution methods on the transformed problem, we show that there exists some Λ * > 0 such that for the range of the parameter 0 < λ < Λ * , the transformed problem has at least two positive solutions, for λ = Λ * , it achieves at least one positive solution, and for λ > Λ * , it has no solutions. In Section 3, we prove the existence of a non-trivial weak solution for the range 0 < λ Λ * and non-existence of solutions for λ > Λ * , by constructing a suitable sub-super solution argument. We also prove the asymptotic boundary behaviour of such solutions. Then in Section 4, we investigate the existence of a second solution in a cone around the first solution for 0 < λ < Λ * by using the Ekeland variational principle and a mountain pass argument with a min-max level. There we find the first critical level, below which we prove the compactness condition of the Palais-Smale sequences. This gives rise to the existence of a second positive solution. In this whole process, we prove many technically involved delicate estimates due to the several complexities present in the problem ( P * ). We would like to mention that to the best of our knowledge, there is no work in the literature addressing the question of global multiplicity of positive solutions involving modified quasilinear operator and singular and exponential nonlinearity. In this article, we study the global multiplicity results for the problem ( P * ) up to the optimal range for the singularity ( 0 < q < 3 ).

We now state all the hypotheses imposed on the continuous function f : Ω × R R , given by f ( x , s ) = g ( x , s ) exp ( s 4 ) :

  1. g C 1 ( Ω ¯ × R ) such that for each x Ω ¯ , g ( x , s ) = 0 for all s 0 ; g ( x , s ) > 0 for all s > 0 and f ( x , s ) s 3 is non-decreasing in s > 0 , for all x Ω ¯ .

  2. Critical growth assumption: For any ε > 0 ,

    lim s + sup x Ω ¯ g ( x , s ) exp ( ε s 4 ) = 0 and lim s + inf x Ω ¯ g ( x , s ) exp ( ε s 4 ) = + .

  3. There exists a constant τ > 4 such that 0 < τ F ( x , s ) f ( x , s ) s , for all ( x , s ) Ω × ( 0 , + ) .

  4. There exists a constant M 1 > 0 such that F ( x , s ) M 1 ( 1 + f ( x , s ) ) for all s > 0 .

Example 2

Consider f ( x , s ) = g ( x , s ) e s 4 , where g ( x , s ) = t a 0 + 1 exp ( d 0 s r ) , if s > 0 0 , if s 0 for some a 0 > 0 , 0 < d 0 4 π and 1 r < 4 . Then f satisfies all the conditions from ( f 1 ) ( f 4 ) .

Remark 3

From the condition ( f 1 ) , we deduce

0 lim s 0 + f ( x , s ) s lim s 0 + f ( x , s ) s 3 s 2 lim s 0 + f ( x , 1 ) s 2 = 0 uniformly in  x Ω ,

since g is continuous, which implies that

(1.4) lim s 0 f ( x , s ) s = 0 uniformly in  x Ω .

Moreover, the condition ( f 1 ) yields that g ( x , s ) is non-decreasing in s , and hence, g ( x , s ) g s ( x , s ) 0 so that

f ( x , s ) f s ( x , s ) = ( g ( x , s ) + 4 s 3 g ( x , s ) ) exp ( s 4 ) 4 s 3 g ( x , s ) exp ( s 4 ) = 4 s 3 f ( x , s ) .

Thus, for any M 0 > 2 , there exists L > 0 such that

(1.5) f ( x , s ) M 0 f ( x , s ) L for all s > 0 .

Now for any ϕ C ( Ω ¯ ) with ϕ > 0 in Ω , we define the set

C ϕ { u C 0 ( Ω ¯ ) there exists c 0 such that u ( x ) c ϕ for all x Ω }

equipped with the norm u C ϕ ( Ω ¯ ) u ϕ . Next, we define the following open convex set of C ϕ ( Ω ) as follows:

C ϕ + ( Ω ) = u C ϕ ( Ω ) inf x Ω u ( x ) ϕ ( x ) > 0 .

That is, the set C ϕ + ( Ω ) is consisting of all those functions u C ( Ω ) such that C 1 ϕ u C 2 ϕ in Ω for some C 1 , C 2 > 0 . Let us also define the distance function

δ ( x ) dist ( x , Ω ) = inf y Ω x y , for any  x Ω ¯ .

We consider the following eigenvalue problem :

(1.6) Δ u = λ u in Ω , u > 0 in Ω , u = 0 on Ω .

Let φ 1 , Ω H 0 1 ( Ω ) be a positive solution (first eigenfunction) of the aforementioned equation corresponding to the first eigenvalue λ ˜ 1 , Ω with φ 1 , Ω < 1 . We recall that φ 1 , Ω C 1 , θ ( Ω ¯ ) for some θ ( 0 , 1 ) and φ 1 , Ω C δ + ( Ω ) . For more properties of φ 1 , Ω , one can refer to [16]. Then, we define the function φ q as follows:

φ q = φ 1 , Ω if 0 < q < 1 , φ 1 , Ω log 2 φ 1 , Ω 1 q + 1 if q = 1 , φ 1 , Ω 2 q + 1 if q > 1 .

Now we state the main results of this article:

Theorem 4

Let Ω R 2 be a smooth bounded domain. Suppose that the hypotheses ( f 1 ) ( f 4 ) and ( α 1 ) hold. Then there exists Λ * > 0 such that for every λ ( 0 , Λ * ] , problem ( P * ) has at least one solution in H 0 1 ( Ω ) C φ q + ( Ω ) , and for λ > Λ * , problem ( P * ) has no solutions.

Theorem 5

Let Ω R 2 be a smooth bounded domain. Assume that the hypotheses ( f 1 ) ( f 4 ) and ( α 1 ) hold. Then for every λ ( 0 , Λ * ) , problem ( P * ) has at least two solutions in H 0 1 ( Ω ) C φ q + ( Ω ) .

2 Variational framework

For u : Ω R , measurable function, and for 1 p + , we define the Lebesgue space L p ( Ω ) as follows:

L p ( Ω ) = u : Ω R measurable Ω u p d x < +

equipped with the usual norm denoted by u p . Now the Sobolev space H 0 1 ( Ω ) is defined as

H 0 1 ( Ω ) = u L 2 ( Ω ) Ω u 2 d x < +

endowed with the norm

u Ω u 2 d x 1 2 .

Since Ω R 2 is a smooth bounded domain, we have the continuous embedding H 0 1 ( Ω ) L p ( Ω ) for p [ 1 , + ) . Moreover, the embedding H 0 1 ( Ω ) u exp ( u β ) L 1 ( Ω ) is compact for all β [ 1 , 2 ) and is continuous for β = 2 . Consequently, the map : H 0 1 ( Ω ) L p ( Ω ) for p [ 1 , + ) , defined by ( u ) exp ( u 2 ) is continuous with respect to the H 0 1 -norm topology. The natural energy function associated to problem ( P * ) is given as follows:

(2.1) I λ ( u ) = 1 2 Ω ( 1 + 2 u 2 ) u 2 d x λ 1 1 q Ω α ( x ) u 1 q d x λ Ω F ( x , u ( x ) ) d x if q 1 ; 1 2 Ω ( 1 + 2 u 2 ) u 2 d x λ Ω α ( x ) log u d x λ Ω F ( x , u ( x ) ) d x if q = 1 .

Observe that, the functional I λ is not well defined in H 0 1 ( Ω ) because of the nature of the singularity ( 0 < q < 3 ) as well as, due to the fact that Ω u 2 u 2 d x is not finite for all u H 0 1 ( Ω ) . So, it is difficult to apply variational methods directly in our problem ( P * ). To get rid of this inconvenience, first, we apply the following change of variables introduced in [8], namely, w h 1 ( u ) , where h is defined by

(2.2) h ( s ) = 1 ( 1 + 2 h ( s ) 2 ) 1 2 in [ 0 , + ) , h ( s ) = h ( s ) in ( , 0 ] .

Now we gather some properties of h , which we follow throughout in this article. For the detailed proofs of such results, one can see [8,11].

Lemma 6

The function h satisfies the following properties:

  1. h is uniquely defined, C and invertible;

  2. h ( 0 ) = 0 ;

  3. 0 < h ( s ) 1 for all s R ;

  4. 1 2 h ( s ) s h ( s ) h ( s ) for all s > 0 ;

  5. h ( s ) s for all s R ;

  6. h ( s ) 2 1 4 s 1 2 for all s R ;

  7. lim s + h ( s ) s 1 2 = 2 1 4 ;

  8. h ( s ) h ( 1 ) s for s 1 and h ( s ) h ( 1 ) s 1 2 for s 1 ;

  9. h ( s ) = 2 h ( s ) ( h ( s ) ) 4 , for all s 0 and h ( s ) < 0 when s > 0 , h ( s ) > 0 when s < 0 .

  10. the function ( h ( s ) ) γ h ( s ) s is strictly increasing for γ 3 ;

  11. lim s 0 h ( s ) s = 1 ;

  12. h ( s ) h ( s ) < 1 2 for all s R ;

  13. the function h ( s ) γ h ( s ) is decreasing for all s > 0 , where γ > 0 .

After employing the change of variable w = h 1 ( u ) in (2.1), we define the new functional J λ : H 0 1 ( Ω ) R as

(2.3) J λ ( w ) = 1 2 Ω w 2 d x λ 1 1 q Ω α ( x ) h ( w ) 1 q d x λ Ω F ( x , h ( w ) ) d x if q 1 ; 1 2 Ω w 2 d x λ Ω α ( x ) log h ( w ) d x λ Ω F ( x , h ( w ) ) d x if q = 1 .

Remark 7

The functional J λ ( w ) is well defined for w H 0 1 ( Ω ) C φ q + ( Ω ) . Indeed, let us define the set D ( J λ ) { w H 0 1 ( Ω ) : J λ ( w ) < + } . Now we intend to show that this set is non-empty. Let 1 < q < 3 . Then using the fact that h ( s ) is increasing in s > 0 , φ 1 , Ω C δ + ( Ω ) and Lemma 6- ( h 8 ) , we have

h ( w ) > h C 1 φ 1 , Ω 2 q + 1 > h C 2 δ 2 q + 1 > C 2 h ( 1 ) δ 2 q + 1 if C 2 δ 2 q + 1 < 1 , h ( 1 ) if C 2 δ 2 q + 1 1 ,

where C 2 is a positive constant. Thus, for any ψ H 0 1 ( Ω ) and w C φ q + ( Ω ) , we deduce the following by applying Lemma 6- ( h 3 ) , Hölder’s inequality, Hardy’s inequality, and the Sobolev embedding:

(2.4) Ω h ( w ) q h ( w ) ψ d x = Ω { x : C 2 δ ( x ) 2 q + 1 < 1 } h ( w ) q h ( w ) ψ d x + Ω { x : C 2 δ ( x ) 2 q + 1 1 } h ( w ) q h ( w ) ψ d x C 3 Ω d x δ 2 ( q 1 ) ( q + 1 ) 1 2 Ω ψ 2 δ 2 d x 1 2 + C 4 h ( 1 ) q Ω ψ d x < C 5 ψ < + ,

since 2 ( q 1 ) q + 1 < 1 for 1 < q < 3 , where C 3 , C 4 , C 5 are positive constants. For the case 0 < q 1 , arguing is a similar manner as mentioned earlier, and following [19,35], we obtain (2.4). In view of this, we obtain that the set D ( J λ ) for 0 < q < 3 .

Next, we check for the Gateaux differentiability of the functional J λ . For 0 < q < 1 and w H 0 1 ( Ω ) with w c δ , using the idea of [19] combining with the properties of h described in Lemma 6, we can show that the functional J λ is Gateaux differentiable at w . For the range 1 q < 3 , we have the following lemma regarding the similar property of J λ .

Lemma 8

Let S { w H 0 1 ( Ω ) : w 1 w w 2 } , where w 1 C φ q + ( Ω ) and w 2 H 0 1 ( Ω ) . Then J λ is Gateaux differentiable at w in the direction v w for v , w S .

Proof

We have to prove that

lim t 0 J λ ( w + t ( v w ) ) J λ ( w ) t = Ω w ( v w ) d x λ Ω α ( x ) h ( w ) q h ( w ) ( v w ) d x λ Ω f ( x , h ( w ) ) h ( w ) ( v w ) d x .

For the first term and third term in the right-hand side of the aforementioned expression, the proof follows in a standard way. So, we are left to show for the second term, i.e., for the singular term. Since S is a convex set, for any t ( 0 , 1 ) , w + t ( v w ) S . Let us define the function H : H 0 1 ( Ω ) R as follows:

H ( w ) = 1 1 q Ω h ( w ) 1 q d x if q 1 Ω log ( h ( w ) ) d x if q = 1 .

Now using mean value theorem, for any 0 < q < 3 , it follows that

(2.5) H ( w + t ( v w ) ) H ( w ) t = Ω h ( w + t θ ( v w ) ) q h ( w + t θ ( v w ) ) ( v w ) d x

for some θ ( 0 , 1 ) . Since w + t θ ( v w ) S , using Lemma 6- ( h 13 ) , it follows that

h ( w + t θ ( v w ) ) q h ( w + t θ ( v w ) ) ( v w ) d x h ( w 1 ) q h ( w 1 ) ( v w ) d x .

Now recalling (2.4), we obtain

Ω h ( w 1 ) q h ( w 1 ) ( v w ) d x < + .

Thus, by applying the Lebesgue dominated convergence theorem and passing to the limit t 0 + in (2.5), we have

(2.6) lim t 0 H ( w + t ( v w ) ) H ( w ) t = Ω h ( w ) q h ( w ) ( v w ) d x .

This completes the proof.□

Thus, using the properties of the functions f , h and the aforementioned results, it can be derived that (2.3) is the associated energy functional to the following problem:

(2.7) Δ w = λ ( α ( x ) h ( w ) q h ( w ) + f ( x , h ( w ) ) h ( w ) ) in Ω , w > 0 in Ω , w = 0 on Ω .

Moreover, by applying Lemma 6 and following the idea as in [36, Proposition 2.2], one can show that if w is a solution to (2.7), then u = h ( w ) is a solution to problem ( P * ). Thus, our main objective is now reduced to proving the existence of solutions to the new transformed equation (2.7).

Definition 9

A function w H 0 1 ( Ω ) is said to be a weak solution to (2.7) if for every compact set K Ω , there exists a constant m K > 0 such that w > m K holds in K , and for every ϕ H 0 1 ( Ω ) , we have

(2.8) Ω w ϕ d x λ Ω α ( x ) h ( w ) q h ( w ) ϕ d x λ Ω f ( x , h ( w ) ) h ( w ) ϕ d x = 0 .

In the next lemma, we discuss some comparison type result related to our problem (2.7).

Lemma 10

Let w 1 , w 2 H 0 1 ( Ω ) C φ q + ( Ω ) satisfy

Δ w 1 γ ( x ) h ( w 1 ) q h ( w 1 ) , x Ω ; Δ w 2 γ ( x ) h ( w 2 ) q h ( w 2 ) , x Ω ,

where γ L ( Ω ) with γ > 0 . Then w 1 w 2 a.e. in Ω .

Proof

The proof of this lemma follows in a similar fashion as in [35, Lemma 2.2].□

Notations. In the next subsequent sections, we make use of the following notations:

  • If u is a measurable function, we denote the positive and negative parts by u + = max { u , 0 } and u = max { u , 0 } , respectively.

  • For any function f , supp f = { x : f ( x ) 0 } .

  • If A is a measurable set in R 2 , we denote the Lebesgue measure of A by A .

  • The arrows , denote weak convergence, strong convergence, respectively.

  • The arrow denotes continuous embedding.

  • B r denotes the ball of radius r > 0 centred at 0 H 0 1 ( Ω ) .

  • B r ¯ denotes the closure of the ball B r with respect to H 0 1 ( Ω ) -norm topology.

  • B r denotes the boundary of the ball B r .

  • B r ( x ) denotes the ball of radius r > 0 centred at x H 0 1 ( Ω ) .

  • For any p > 1 , p p p 1 denotes the conjugate of p .

  • c , C 0 , C 1 , C 2 , , C ˜ 1 , C ˜ 2 , , C and C ˜ denote positive constants which may vary from line to line.

3 Proof of Theorem 4: Existence and non-existence results

In this section, to prove the existence and non-existence results for the problem (2.7), we first need to study the existence and regularity result for the following purely singular problem:

(3.1) Δ w = λ α ( x ) h ( w ) q h ( w ) in Ω , w > 0 in Ω , w = 0 on Ω .

Now, we have the following result for problem (3.1).

Theorem 11

Let Ω R 2 be a smooth bounded domain. Assume that λ > 0 , q > 0 , α satisfies ( α 1 ) and the function h is defined in (2.2). Then,

  1. the problem (3.1) has a unique solution for each λ > 0 , say w λ ̲ , in H 0 1 ( Ω ) C φ q + ( Ω ) , for q < 3 ;

  2. the solution w λ ̲ C 1 ( Ω ¯ ) if q < 1 , w λ ̲ C 1 ε ( Ω ¯ ) for any small ε > 0 if q = 1 and w λ ̲ C ( Ω ¯ ) if q < 3 ;

  3. the map λ w λ ̲ is non-decreasing and continuous from R + to C ( Ω ¯ ) .

Proof

Let us set ρ ( s ) λ α ( x ) h ( s ) q h ( s ) . Then ρ ( w ) verifies the hypotheses of Proposition 4.1 in [2]. Hence, using [2, Proposition 4.1], we can infer that there exists a unique solution to (3.1), say w λ ̲ , such that w λ ̲ H 0 1 ( Ω ) for 1 < q < 3 . Next, following the proof of [9, Theorem 2.2], there exist two positive constants c 1 c 1 ( λ , q ) 1 , and c 2 c 2 ( λ , q ) such that

(3.2) c 1 ( q , λ ) δ ( x ) w λ ̲ c 2 ( q , λ ) δ ( x )     if   0 < q < 1 ,

(3.3) c 1 ( q , λ ) δ ( x ) 2 q + 1 w λ ̲ c 2 ( q , λ ) δ ( x ) 2 q + 1 if 1 < q < 3 .

Furthermore, for the case q = 1 , again recalling [9, Theorem 2.2], we can find that there exists a constant c ( λ ) > 0 , and for any ε > 0 small enough, there exists a constant c ε ( λ ) > 0 such that

(3.4) c ( λ ) δ ( x ) w λ ̲ c ε ( λ ) δ ( x ) 1 ε .

This, in combination with standard elliptic regularity theory, implies that w λ ̲ C φ q + ( Ω ) . Thus, ( i ) follows.

Again, using [2, Proposition 4.1] (also see [9, Theorem 2.2]), we obtain ( i i ) .

Finally, by using ( i ) ( i i ) and the maximum principle, we obtain ( i i i ) . This completes the proof.□

Remark 12

One can check that (3.1) is the transformed form (with the transformation u = h ( w ) ) of the following purely singular problem with the modified quasilinear operator corresponding to the problem ( P * ) :

(3.5) Δ u Δ ( u ) 2 u = λ α ( x ) u q in Ω , u > 0 in Ω , u = 0 on Ω ,

where λ , α , Ω are as in Theorem 11. So, by the properties of h and following the idea of the proof of the Proposition 2.2 in [36], we can deduce that (3.5) has a solution h ( w λ ̲ ) for every λ > 0 , which satisfies all the properties in Theorem 11.

From the assumption ( f 2 ) and (1.4), we obtain that for any ε > 0 , r 1 , there exist C ˜ ( ε ) and C ( ε ) > 0 such that

(3.6) f ( x , s ) ε s + C ˜ ( ε ) s r 1 exp ( ( 1 + ε ) s 4 ) for all ( x , s ) Ω × R ,

(3.7) F ( x , s ) ε s 2 + C ( ε ) s r exp ( ( 1 + ε ) s 4 ) for all ( x , s ) Ω × R .

For any w H 0 1 ( Ω ) , in light of the Sobolev embedding, we have w L q ( Ω ) for all q [ 1 , + ) .

Let us define the set Q as follows:

Q = { λ > 0 : the problem (2.7) has a weak solution in H 0 1 ( Ω ) } ,

and let Λ * sup Q . Then we have the following result:

Lemma 13

Assume that the conditions in Theorem 4hold and let h be defined as in (2.2). Then the set Q is non-empty.

Proof

First, we consider the case 0 < q < 1 .

Using (3.7), Lemma 6- ( h 5 ) , ( h 6 ) with the Sobolev embedding and Hölder’s inequality, from (2.3), we obtain

(3.8) J λ ( w ) 1 2 w 2 λ q 1 Ω α ( x ) h ( w ) 1 q d x λ ε Ω h ( w ) 2 d x λ C ( ε ) Ω w r exp ( ( 1 + ε ) h ( w ) 4 ) d x 1 2 w 2 λ α q 1 Ω w 1 q d x λ ε Ω w 2 d x λ C ( ε ) Ω w r exp ( 2 ( 1 + ε ) w 2 ) d x 1 2 w 2 λ α q 1 C 1 w 1 q λ ε C 2 w 2 λ C 3 ( ε ) w r p r Ω exp ( 2 p ( 1 + ε ) w 2 ) d x 1 p

(3.9) 1 2 λ ε C 2 w 2 λ α q 1 C 1 w 1 q λ C 4 ( ε ) w r Ω exp 2 p ( 1 + ε ) w 2 . w 2 w 2 d x 1 p ,

for any ε > 0 , p > 1 and r > 2 . Choose w = r 0 with 0 < r 0 < 1 sufficiently small, 0 < ε < 1 2 C 2 sufficiently small and p > 1 very near to 1 such that 2 ( 1 + ε ) p r 0 < 4 π . Then by using Theorem 1, from (3.7) (3.9), we obtain

1 2 w 2 λ Ω F ( x , h ( w ) ) d x 2 δ 0 for all w B r 0 ; 1 2 w 2 λ Ω F ( x , h ( w ) ) d x 0 for all w B r 0 .

Now we can choose λ = λ 0 > 0 sufficiently small so that the last two relations yield that

J λ 0 B r 0 δ 0 > 0 .

Set m 0 inf w B r 0 J λ 0 ( w ) . Since for t > 0 very small and w 0 , from (2.3), we have

J λ 0 ( t w ) 1 2 t w 2 C λ 0 t w 1 q ,

which implies that m 0 < 0 . Let { w k } B r 0 be a minimizing sequence such that as k + ,

J λ 0 ( w k ) m 0 ; w k w 0 weakly in H 0 1 ( Ω ) ; w k w 0 strongly in L p ( Ω ) , p 1 and w k ( x ) w 0 ( x ) pointwise a.e. in Ω .

Now, without loss of generality, let us assume that w k 0 due to the fact that J λ ( w ) = J λ ( w ) . Then, using the Sobolev and the Hölder’s inequality, one can easily deduce that for 0 < q < 1 ,

(3.10) Ω w k 1 q d x Ω w 0 1 q d x and Ω w k w 0 1 q d x 0 as  k + .

Next, by the mean value theorem, there exists w ˜ k in between w 0 and w k such that using h ( w ˜ k ) 1 and (3.10), we deduce

Ω α ( x ) h ( w k ) 1 q d x Ω α ( x ) h ( w 0 ) 1 q d x + Ω α ( x ) h ( w k ) h ( w 0 ) 1 q d x Ω α ( x ) h ( w 0 ) 1 q d x + Ω α ( x ) w k w 0 1 q h ( w ˜ k ) 1 q d x Ω α ( x ) h ( w 0 ) 1 q d x + α Ω w k w 0 1 q d x Ω α ( x ) h ( w 0 ) 1 q d x + o ( 1 ) ,

as k + . Similarly, we obtain Ω α ( x ) h ( w 0 ) 1 q d x Ω α ( x ) h ( w k ) 1 q d x + o ( 1 ) as k + . Therefore, from the last two relations, we obtain

(3.11) Ω α ( x ) h ( w k ) 1 q d x Ω α ( x ) h ( w 0 ) 1 q d x as k + .

Next, using Lemma 6- ( h 4 ) , (3.6) and then borrowing the similar argument as in the third and fourth terms in (3.8), for sufficiently small ε > 0 and for any p > 1 , we deduce

Ω f ( x , h ( w k ) ) h ( w k ) w k d x Ω f ( x , h ( w k ) ) h ( w k ) d x Ω ( ε h ( w k ) 2 + C ˜ ( ε ) h ( w k ) r exp ( ( 1 + ε ) h ( w k ) 4 ) ) d x ε C ˜ 2 w k 2 + C ˜ 4 ( ε ) w k r Ω exp 2 p ( 1 + ε ) w k 2 . w k 2 w k 2 d x 1 p .

In the last relation, using Theorem 1, with sufficiently small w k < r 0 1 and p > 1 very close to 1 so that 2 ( 1 + ε ) p r 0 < 4 π , we obtain

(3.12) limsup k + Ω f ( x , h ( w k ) ) h ( w k ) w k d x < + .

Now using Lemma 6- ( h 8 ) and (3.12), for some large N ( 1 ) N , we deduce

Ω { x : h ( w k ) ( x ) > N } f ( x , h ( w k ) ) d x 1 N Ω { x : h ( w k ) ( x ) > N } f ( x , h ( w k ) ) h ( w k ) d x 2 N Ω { x : h ( w k ) ( x ) > N } f ( x , h ( w k ) ) h ( w k ) w k d x = O 1 N .

The last relation, together with the Lebesgue dominated convergence theorem, implies that

(3.13) Ω f ( x , h ( w k ) ) d x = Ω { x : h ( w k ) ( x ) N } f ( x , h ( w k ) ) d x + Ω { x : h ( w k ) ( x ) > N } f ( x , h ( w k ) ) d x = Ω { x : h ( w k ) ( x ) N } f ( x , h ( w k ) ) d x + O 1 N Ω f ( x , h ( w 0 ) ) d x , as  k +  and  N + .

Since by ( f 4 ) and (3.13), F ( x , h ( w k ) ) M 1 ( 1 + f ( x , h ( w k ) ) ) L 1 ( Ω ) , for all k N , using the Lebesgue dominated convergence theorem, we obtain

(3.14) Ω F ( x , h ( w k ) ) d x Ω F ( x , h ( w 0 ) ) d x as k + .

Now from the weak lower semi-continuity of the norm, we have

r 0 liminf k + w k w 0 ,

which yields that w 0 B r 0 . Therefore,

J λ 0 ( w 0 ) m 0 .

Furthermore, recalling (3.11) and (3.14), we obtain

m 0 = lim k + J λ 0 ( w k ) J λ 0 ( w 0 ) m 0 .

Thus,

J λ 0 ( w 0 ) = m 0 < 0 .

This yields that w 0 ( 0 ) is a local minimizer of J λ 0 in H 0 1 ( Ω ) .

Next, we claim that w 0 is a weak solution to problem (2.7). Note that, for any ϕ 0 , ϕ H 0 1 ( Ω ) ,

(3.15) liminf t 0 + J λ 0 ( w 0 + t ϕ ) J λ 0 ( w 0 ) t 0 .

It can be derived from the last expression that Δ w 0 0 in Ω in the weak sense, and hence, by the strong maximum principle, w 0 > 0 in Ω . Furthermore, by employing Fatou’s lemma in (3.15), we infer that

(3.16) Ω w 0 ψ d x λ Ω α ( x ) h ( w 0 ) q h ( w 0 ) ψ d x + λ Ω f ( x , h ( w 0 ) ) h ( w 0 ) ψ d x for all ψ H 0 1 ( Ω ) , ψ 0 .

Now for any ϕ H 0 1 ( Ω ) and ε > 0 , taking ψ = ( w 0 + ε ϕ ) + as a test function in (3.15) and dividing it by ε > 0 , we obtain

Ω w 0 ϕ d x + ε Ω ϕ 2 d x λ ε Ω α ( x ) ( h ( w 0 + ε ϕ ) 1 q h ( w 0 ) 1 q ) d x + λ ε Ω ( F ( x , h ( w 0 + ε ϕ ) ) F ( x , h ( w 0 ) ) ) d x .

Letting the limit ε 0 + in the last expression, we deduce

Ω w 0 ϕ d x λ Ω α ( x ) h ( w 0 ) q h ( w 0 ) ϕ d x + λ Ω f ( x , h ( w 0 ) ) h ( w 0 ) ϕ d x .

Again, by taking ϕ in place of ϕ , we obtain the reverse inequality in the last relation. Therefore, w 0 is a weak solution to problem (2.7).

Next, we discuss the case 1 q < 3 . For that, we consider the following problem:

(3.17) Δ w = λ α ( x ) ( h ( w ) + ε ) q h ( w ) + λ f ( x , h ( w ) ) h ( w ) in Ω , w > 0 in Ω , w = 0 on Ω ,

where 0 < ε < 1 is sufficiently small. We show the existence of solution to (3.17) by constructing a sub-solution and a super-solution to (3.17). Let W be a solution to

Δ W = 1 in Ω ; W = 0 on Ω .

Then by the maximum principle, W > 0 in Ω and by the standard elliptic regularity theory, W C 1 ( Ω ¯ ) and hence, W is bounded on Ω ¯ . Set

w ¯ w λ ̲ + M W ,

where M > 0 is a sufficiently large real constant and w λ ̲ is the solution to (3.1). Therefore,

(3.18) Δ w ¯ = λ α ( x ) h ( w λ ̲ ) q h ( w λ ̲ ) + M .

So, using Lemma 6- ( h 3 ) , ( h 6 ) combining with ( f 2 ) and the continuity of f , for any ε > 0 , there exists some constant c ( ε , Ω ) , such that

(3.19) λ f ( x , h ( w λ ̲ + M W ) ) h ( w λ ̲ + M W ) < λ c exp ( ( 1 + ε ) h ( w λ ̲ + M W ) 4 ) < λ c exp ( 2 ( 1 + ε ) ( w λ ̲ + M W ) 2 ) < λ c exp ( 2 ( 1 + ε ) ( w λ ̲ + M W ) 2 ) = λ C ( M ) < M for sufficiently small λ > 0 .

Plugging (3.19) in (3.18), we infer that w ¯ is a super-solution to (3.17). Now let us set

w ̲ m W

for some sufficiently small constant m = m ( λ ) > 0 such that w ̲ < 1 . Therefore,

Δ w ̲ = m .

We claim that w ̲ is a sub-solution to (3.17). To prove the claim, it is enough to show

(3.20) m < λ α 0 h ( m W ) ( h ( m W ) + ε ) q .

In virtue of Lemma 6- ( h 4 ) , ( h 5 ) , ( h 8 ) , we deduce

λ α 0 m ( h ( m W ) + ε ) q h ( m W ) λ α 0 m ( h ( m W ) + ε ) q h ( m W ) 2 m W λ α 0 m ( m W + 1 ) q h ( 1 ) 2 λ α 0 m q + 1 ( W + 1 m ) q h ( 1 ) 2 > 1 ,

for sufficiently small 0 < m < 1 . This establishes the claim.

Now for λ > 0 sufficiently small, choosing M sufficiently large and m sufficiently small so that M m , from (3.19) and (3.20), we obtain 0 < w ̲ w ¯ . Hence, there is a solution w ̲ w ε w ¯ to (3.17) in H 0 1 ( Ω ) . Since w ̲ , w ¯ do not depend on ε ,

w ε ( x ) w λ ( x ) pointwise in Ω as ε 0 + .

Next, we will show that w λ is a weak solution to (2.7). From Lemma 6- ( h 4 ) , ( h 5 ) , ( h 8 ) , we deduce

(3.21) α ( x ) ( h ( w ε ) + ε ) q h ( w ε ) w ε α ( x ) ( h ( w ε ) + ε ) q h ( w ε ) α ( x ) h ( w ̲ ) 1 q α w ̲ 1 q , if 0 q 1 , α h ( 1 ) 1 q w ̲ 1 q , if 1 q 3 .

On the other hand, by ( f 1 ) and Lemma 6- ( h 10 ) , it follows that f ( x , h ( s ) ) h ( s ) = f ( x , h ( s ) ) h ( s ) 3 . h ( s ) 3 h ( s ) s . s is increasing in s > 0 . By using this together with Lemma 6- ( h 4 ) , we find

(3.22) f ( x , h ( w ε ) ) h ( w ε ) w ε f ( x , h ( w ¯ ) ) h ( w ¯ ) .

Testing (3.17) against the test function w ε and then using (3.24), (3.25), we obtain

Ω w ε 2 d x = λ Ω α ( x ) ( h ( w ε ) + ε ) q h ( w ε ) w ε d x + λ Ω f ( x , h ( w ε ) ) h ( w ε ) w ε d x λ C ( α ) Ω w ̲ 1 q d x + λ Ω f ( x , h ( w ¯ ) ) h ( w ¯ ) d x < + .

Thus, { w ε } is bounded in H 0 1 ( Ω ) . So, up to some sub-sequence, w ε w λ in H 0 1 ( Ω ) as ε 0 . Again, testing (3.17) against any ϕ H 0 1 ( Ω ) , we deduce

(3.23) Ω w ε ϕ d x = λ Ω α ( x ) ( h ( w ε ) + ε ) q h ( w ε ) ϕ d x + λ Ω f ( x , h ( w ε ) ) h ( w ε ) ϕ d x .

For ϕ 0 , by Lemma 6- ( h 13 ) , ( h 5 ) , ( h 8 ) ,

(3.24) α ( x ) ( h ( w ε ) + ε ) q h ( w ε ) ϕ α ( x ) h ( w ̲ ) q h ( w ̲ ) ϕ α ϕ h ( w ̲ ) q C h ( 1 ) q w ̲ q .

Again for ϕ 0 , using the fact that f ( x , h ( s ) ) h ( s ) is increasing in s > 0 , we have

(3.25) f ( x , h ( w ε ) ) h ( w ε ) ϕ ϕ f ( x , h ( w ¯ ) ) h ( w ¯ ) .

Then, (3.23), (3.24), and (3.25) combining with the Lebesgue dominated convergence theorem yield that

(3.26) Ω w λ ϕ d x = λ Ω α ( x ) h ( w λ ) q h ( w λ ) ϕ d x + λ Ω f ( x , h ( w λ ) ) h ( w λ ) ϕ d x as  ε 0 .

Since for any ϕ H 0 1 ( Ω ) , ϕ = ϕ + ϕ , the relation in (3.26) holds for all ϕ H 0 1 ( Ω ) . Thus, w λ is a solution to (2.7). This completes the proof of the lemma.□

In the next lemma, we aim to discuss the regularity result for the solutions of (2.7).

Lemma 14

Assume that ( f 1 ) ( f 4 ) and ( α 1 ) hold. Let h be defined as in (2.2). If w H 0 1 ( Ω ) is any weak solution to (2.7) for λ ( 0 , Λ * ] , then w L ( Ω ) C φ q + ( Ω ) .

Proof

Let w H 0 1 ( Ω ) be a weak solution to (2.7). First, in spirit of [17, Lemma A.4], we show that w is in L ( Ω ) . For that, let us define a C 1 cut-off function ψ : R [ 0 , 1 ] as follows:

ψ ( s ) = 0 if s 0 , 1 if s 1 ,

with ψ ( s ) 0 . Now for any ε > 0 , define

ψ ε ( s ) = ψ s 1 ε for s R .

Note that ( ψ ε w ) = ( ψ ε w ) w . Hence, ψ ε w H 0 1 ( Ω ) . Let v C c ( Ω ) with v 0 . Now using ϕ ( ψ ε w ) v as a test function in (2.8), we obtain

(3.27) Ω w ( ψ ε w ) v d x λ Ω α ( x ) h ( w ) q h ( w ) ( ψ ε w ) v d x λ Ω f ( x , h ( w ) ( x ) ) h ( w ) ( ψ ε w ) v d x = 0 .

Since

w ( ψ ε w ) v = w 2 ( ψ w ) v + ( w v ) ( ψ ε w ) ,

from (3.27), we deduce

Ω ( w v ) ( ψ ε w ) d x λ Ω α ( x ) h ( w ) q h ( w ) ( ψ ε w ) v d x + λ Ω f ( x , h ( w ) ( x ) ) h ( w ) ( ψ ε w ) v d x .

In the last relation, letting ε 0 + and using Lemma 6- ( h 13 ) , ( h 3 ) , we obtain

Ω ( w 1 ) + v d x λ Ω { x : w ( x ) > 1 } α ( x ) h ( 1 ) q h ( 1 ) v d x + λ Ω { x : w ( x ) > 1 } f ( x , h ( w ) ) h ( w ) v d x C + λ Ω f ( x , h ( w ) ) v d x .

Now following the arguments as in [20, Lemma 10, Theorem C] combining with Theorem 1, from the last relation, we infer that ( w 1 ) + L ( Ω ) . Hence, w L ( Ω ) .

Now we claim that w λ ̲ w a.e. in Ω . Suppose the claim is not true. Then using ( w λ ̲ w ) + as the test function in Δ ( w λ ̲ w ) λ α ( x ) ( h ( w λ ̲ ) q h ( w λ ̲ ) h ( w ) q h ( w ) ) in Ω and recalling Lemma 6- ( h 13 ) , we deduce

0 Ω ( w λ ̲ w ) + 2 d x λ Ω α ( x ) ( h ( w λ ̲ ) q h ( w λ ̲ ) h ( w ) q h ( w ) ) ( w λ ̲ w ) + d x 0 .

Hence, the claim holds. Next, let z λ be a solution to the problem,

(3.28) Δ z λ = λ ( α ( x ) ) h ( z λ ) q h ( z λ ) + g ( w ) exp ( 2 w 2 ) in Ω , z λ > 0 in Ω , z λ = 0 on Ω .

Then, arguing similarly as above, we obtain w z λ . Now, it can be checked that the results in Theorem 3.1 hold for the problem (3.28). Therefore, z λ is unique and w λ ̲ w z λ . So, in the light of Theorem 3.1- ( i ) , there exist two positive constants C 1 ( λ , q ) 1 , and C 2 ( λ , q ) such that

(3.29) C 1 ( q , λ ) δ ( x ) w C 2 ( q , λ ) δ ( x ) if 0 < q < 1 ,

(3.30) C 1 ( q , λ ) δ ( x ) 2 q + 1 w C 2 ( q , λ ) δ ( x ) 2 q + 1 if 1 < q < 3 .

Moreover, if q = 1 , there exists a constant C ( λ ) > 0 , and for any ε > 0 small enough, there exists a constant C ε ( λ ) > 0 such that

(3.31) C ( λ ) δ ( x ) w C ε ( λ ) δ ( x ) 1 ε .

Finally, combining (3.29), (3.30), and (3.31) and recalling the standard elliptic regularity theory, it follows that w C φ q + ( Ω ) .

Remark 15

Using Lemma 14, and adapting the proof of Corollary 1.1 in [2] (also see [17]), one can show that if w H 0 1 ( Ω ) is any weak solution to (2.7), then w C ( Ω ¯ ) . Moreover, when 0 < q < 1 , w C 1 ( Ω ¯ ) .

The following lemma basically ensures the non-existence of solution to (2.7) for λ > Λ * .

Lemma 16

Let the conditions in Theorem 4hold and let h be defined as in (2.2). Then, 0 < Λ * < + .

Proof

From Lemma 13, we can infer that Λ * > 0 . Thus, we are left to show that Λ * < + . Suppose this is not true. Then, there exists a sequence { λ k } Q such that λ k + as k + . For s > 0 , let us define the function

N λ ( s ) λ [ h ( s ) q + f ( x , h ( s ) ) ] h ( s ) s .

We claim that there exist k 0 N sufficiently large and β = β ( λ k 0 ) > 0 such that for all s > 0 ,

(3.32) N λ k 0 ( s ) = λ k 0 [ h ( s ) q + f ( x , h ( s ) ) ] h ( s ) s > β > λ ˜ 1 , Ω ( ϱ ) ,

where λ ˜ 1 , Ω ( ϱ ) is the first eigenvalue of the problem (1.6) with ϱ ( x ) = min { 1 , α ( x ) } . Indeed, for any arbitrary k N and for s [ 1 n , n ] , n N , let us consider N λ k ( s ) . Since N λ k is a continuous function, there exists s n s n , k [ 1 n , n ] such that

N λ k ( s n ) N λ k ( s ) for all s [ 1 n , n ] .

Now we show that, up to some sub-sequence, s n s 0 , k ( 0 , + ) as n + . If not, we have either s n 0 or s n + as n + . For both the cases, by applying Lemma 6- ( h 6 ) , ( h 8 ) , we obtain

lim n + N λ k ( s ) lim n + N λ k ( s n ) = + .

That is, N λ k + for all s ( 0 , + ) and for all k N , which is absurd. Hence, s n s 0 , k ( 0 , + ) as n + and

(3.33) N λ k ( s ) λ k [ h ( s 0 , k ) q + f ( x , h ( s 0 , k ) ) ] h ( s 0 , k ) s 0 , k for all s > 0 .

Arguing in a similar manner as in (3.33), it can be deduced that s k s 0 , k s 0 ( 0 , + ) , up to some sub-sequence, as k + . Using this fact, from (3.33), we obtain (3.32). Thus, the claim follows.

Since λ k 0 Q , for λ = λ k 0 , let w λ k 0 be a solution to (2.7). So, it follows that

Δ w λ k 0 β ϱ ( x ) w λ k 0 Δ w λ k 0 N λ k 0 ϱ ( x ) w λ k 0 = Δ w λ k 0 ϱ ( x ) w λ k 0 λ k 0 [ h ( w λ k 0 ) q + f ( x , h ( w λ k 0 ) ) ] h ( w λ k 0 ) w λ k 0 Δ w λ k 0 λ k 0 [ α ( x ) h ( w λ k 0 ) q f ( x , h ( w λ k 0 ) ) h ( w λ k 0 ) ] = 0 .

This implies that Δ w λ k 0 β ϱ ( x ) w λ k 0 > 0 in Ω , which in view of strong maximum principle yields that w λ k 0 > 0 in Ω . Now by applying Picone’s identity for φ 1 , Ω and w λ k 0 , we derive

0 Ω φ 1 , Ω 2 d x Ω φ 1 , Ω 2 w λ k 0 w λ k 0 d x Ω φ 1 , Ω 2 d x Ω β ϱ ( x ) φ 1 , Ω 2 d x = ( λ ˜ 1 , Ω ( ϱ ) β ) Ω ϱ ( x ) φ 1 , Ω 2 d x .

Therefore, λ ˜ 1 , Ω ( ϱ ) β , which contradicts (3.32). Thus, the proof of the lemma follows.□

In the next result, by using a sub-super solution technique, we show the existence of at least one solution to (2.7).

Proposition 17

Let the conditions in Theorem 4be satisfied and let h be defined as in (2.2). Then for each λ ( 0 , Λ * ) , (2.7) admits a nontrivial solution in H 0 1 ( Ω ) C φ q + ( Ω ) .

Proof

Let λ ( 0 , Λ * ) and λ ( λ , Λ * ) . Then from the definition of Λ * and Lemma 13, one can see that w λ H 0 1 ( Ω ) forms a weak solution to (2.7) for λ = λ . Let w λ ̲ be as in Theorem 11. Then

(3.34) Δ w λ ̲ = λ α ( x ) h ( w λ ̲ ) q h ( w λ ̲ ) λ α ( x ) h ( w λ ̲ ) q h ( w λ ̲ ) + λ f ( x , h ( w λ ̲ ) ) h ( w λ ̲ ) , x Ω .

Thus, w λ ̲ is a weak sub-solution to (2.7). Therefore, w λ and w ̲ λ satisfy the following:

Δ w λ λ α ( x ) h ( w λ ) q h ( w λ ) in Ω , Δ w λ ̲ λ α ( x ) h ( w λ ̲ ) q h ( w λ ̲ ) in Ω .

Hence, Lemma 10 yields that w λ ̲ w λ . Now we consider the closed convex subset Y λ of H 0 1 ( Ω ) as follows:

(3.35) Y λ { w H 0 1 ( Ω ) : w λ ̲ w w λ } .

Let { w k } Y λ be such that w k w 0 in H 0 1 ( Ω ) as k + . Then, up to a sub-sequence, w k ( x ) w 0 ( x ) pointwise a.e. in Ω . Since w λ is a solution of (2.7), by Lemma 14, w λ C φ q + . Now for 1 < q < 3 , using (3.29) and Lemma 6- ( h 5 ) , we obtain

(3.36) α ( x ) h ( w k ) 1 q α ( x ) w k 1 q α ( x ) w λ 1 q α ( C 2 ( λ , q ) ) 1 q δ 1 q L 1 ( Ω ) .

Next, for q = 1 , by Lemma 6- ( h 5 ) and (3.31), for any sufficiently small ε > 0

(3.37) α ( x ) log ( h ( w k ) ) α ( x ) log ( w k ) α ( x ) log ( w λ ) α ( x ) w λ α C ε ( λ ) δ 1 ε L 1 ( Ω ) .

For 1 < q < 3 , using h is increasing, (3.3) with c 1 ( q , λ ) > 0 small enough such that c 1 δ 2 1 + q < 1 and Lemma 6- ( h 8 ) , we obtain

(3.38) α ( x ) h ( w k ) 1 q α ( x ) h ( w λ ̲ ) 1 q α ( x ) h ( c 1 δ 2 1 + q ) 1 q α c 1 1 q h ( 1 ) 1 q δ 2 ( 1 q ) 1 + q L 1 ( Ω ) ,

since 2 ( 1 q ) 1 + q > 1 . Furthermore, using ( f 2 ) in combination with Lemma 6- ( h 6 ) and Theorem 1, we deduce

(3.39) F ( x , h ( w k ) ) < C exp ( ( 1 + ε ) h ( w k ) 4 ) C exp ( 2 ( 1 + ε ) w λ 2 ) L 1 ( Ω ) .

Therefore, by the Lebesgue dominated convergence theorem,

Ω α ( x ) h ( w k ) 1 q d x Ω α ( x ) h ( w 0 ) 1 q d x , if q 1 ; Ω α ( x ) log ( h ( w k ) ) d x Ω α ( x ) log ( h ( w 0 ) ) d x , if q = 1 ; Ω F ( x , h ( w k ) ) d x Ω F ( x , h ( w 0 ) ) d x .

Using the last three limits and the weak lower semicontinuity property of the norm, it follows that J λ is weakly lower semicontinuous on Y λ . Since Y λ is weakly sequentially closed subset of H 0 1 ( Ω ) , there exists a w λ Y λ such that

(3.40) inf w Y λ J λ ( w ) = J λ ( w λ ) .

Now we show that w λ is a weak solution to (2.7).

For φ H 0 1 ( Ω ) and ε > 0 small enough, we define

v ε min { w λ , max { w λ ̲ , w λ + ε φ } } = w λ + ε φ φ ε + φ ε Y λ ,

where φ ε max { 0 , w λ + ε φ w λ } and φ ε max { 0 , w λ ̲ w λ ε φ } . By construction, v ε Y λ and φ ε , φ ε H 0 1 ( Ω ) . Since w λ + t ( v ε w ) Y λ , for each 0 < t < 1 , using (3.40), Lemma 8 and mean value theorem, we obtain

0 lim t 0 + J λ ( w λ + t ( v ε w λ ) ) J λ ( w λ ) t = Ω w λ ( v ε w λ ) d x λ lim t 0 + Ω α ( x ) h ( w λ + θ t ( v ε w λ ) ) q h ( w λ + θ t ( v ε w λ ) ) ( v ε w λ ) d x λ Ω f ( x , h ( w λ ) ) h ( w λ ) ( v ε w λ ) d x

for some 0 < θ < 1 . From the definition of φ ε , φ ε , we obtain that v ε w λ H 0 1 ( Ω ) , which yields that

( h ( w λ ̲ ) q h ( w λ ̲ ) ) ( v ε w λ ) L 1 ( Ω ) .

Moreover, using Lemma 6- ( h 13 ) , we obtain

h ( w λ + θ t ( v ε w λ ) ) q h ( w λ + θ t ( v ε w λ ) ) ( v ε w λ ) ( h ( w λ ̲ ) q h ( w λ ̲ ) ) ( v ε w λ )

for all t ( 0 , 1 ) . From the last relation, it follows that

(3.41) Ω w λ φ d x λ Ω α ( x ) h ( w λ ) q h ( w λ ) φ d x λ Ω f ( x , h ( w λ ) ) h ( w λ ) φ d x 1 ε ( E ε E ε ) ,

where

E ε Ω w λ φ ε d x λ Ω α ( x ) h ( w λ ) q h ( w λ ) φ ε d x λ Ω f ( x , h ( w λ ) ) h ( w λ ) φ ε d x ; E ε Ω w λ φ ε d x λ Ω α ( x ) h ( w λ ) q h ( w λ ) φ ε d x λ Ω f ( x , h ( w λ ) ) h ( w λ ) φ ε d x .

We define the set Ω ε { x Ω : ( w λ + ε φ ) ( x ) w λ ( x ) > w λ ( x ) } so that Ω ε 0 as ε 0 + . Next, using the fact that w λ is a super-solution to (2.7) together with Lemma 6- ( h 13 ) , we estimate the following:

1 ε E ε = 1 ε Ω ( w λ w λ ) φ ε d x + Ω w λ φ ε d x λ Ω ( α ( x ) h ( w λ ) q + f ( x , h ( w λ ) ) ) h ( w λ ) φ ε d x 1 ε Ω ε ( w λ w λ ) 2 d x + Ω ε ( w λ w λ ) φ d x + λ ε Ω ε α ( x ) × ( h ( w λ ) q h ( w λ ) h ( w λ ) q h ( w λ ) ) φ ε d x + λ ε Ω ε ( f ( x , h ( w λ ) ) h ( w λ ) f ( x , h ( w λ ) ) h ( w λ ) ) φ ε d x Ω ε ( w λ w λ ) φ d x λ Ω ε α ( x ) ( h ( w λ ) q h ( w λ ) h ( w λ ) q h ( w λ ) ) φ d x λ Ω ε f ( x , h ( w λ ) ) h ( w λ ) f ( x , h ( w λ ) ) h ( w λ ) φ d x = o ( 1 ) as ε 0 + .

Arguing similarly, we have

1 ε E ε o ( 1 ) as ε 0 + .

Thus, from (3.41), we obtain

Ω w λ φ d x λ Ω α ( x ) h ( w λ ) q h ( w λ ) φ d x λ Ω f ( x , h ( w λ ) ) h ( w λ ) φ d x o ( 1 ) as ε 0 +

for all φ H 0 1 ( Ω ) . Considering φ in place of φ and following the similar arguments as mentioned earlier, we infer that w λ is a weak solution to (2.7). Moreover, from the construction of w λ and Lemma 14, it follows that w λ C φ q + . This concludes the proof of the proposition.□

Lemma 18

Assume that the conditions in Theorem 4hold and let h be defined as in (2.2). Let λ ( 0 , Λ * ) . Then any weak solution to (2.7) obtained in Proposition 17is a local minimizer for the functional J λ .

Proof

We prove this lemma for the case q 1 . For q = 1 , the proof follows in a similar fashion.

Now suppose the statement of the lemma does not hold. So, let us assume that w λ is not a local minimum of J λ , where w λ is a solution to (2.7) obtained in Lemma 17. Then there exists a sequence { w k } H 0 1 ( Ω ) such that

(3.42) w k w λ 0 as k + and J λ ( w k ) < J λ ( w λ ) .

Next, we define w ̲ w λ ̲ and w ¯ w λ as a sub-solution and a super-solution to (2.7), respectively, as defined in the proof of Proposition 17. Furthermore, we define

v k max { w ̲ , min { w k , w ¯ } } = w ̲ , if w k < w ̲ , w k , if w ̲ w k w ¯ , w ¯ , if w k > w ¯ , u k ̲ ( w k w ̲ ) , u k ¯ ( w k w ¯ ) + , S ̲ k supp ( u k ̲ ) , S ¯ k supp ( u k ¯ ) .

Then, w k = v k u k ̲ + u k ¯ and v k Y λ , where the set Y λ is defined in (3.35). Then we can express J λ ( w k ) as

(3.43) J λ ( w k ) = J λ ( v k ) + A k + B k ,

where

(3.44) A k 1 2 S k ¯ ( w k 2 w ¯ 2 ) d x λ 1 q S k ¯ α ( x ) ( h ( w k ) 1 q h ( w ¯ ) 1 q ) d x λ S k ¯ ( F ( x , h ( w k ) ) F ( x , h ( w ¯ ) ) ) d x ,

(3.45) B k 1 2 S k ̲ ( w k 2 w ̲ 2 ) d x λ 1 q S k ̲ α ( x ) ( h ( w k ) 1 q h ( w ̲ ) 1 q ) d x λ S k ̲ ( F ( x , h ( w k ) ) F ( x , h ( w ̲ ) ) ) d x .

From Proposition 17, we obtain J λ ( w k ) J λ ( w λ ) + A k + B k . We intend to show that A k , B k 0 for large k , as we will see later

(3.46) lim k + S k ¯ = 0 and lim k + S k ̲ = 0 .

Let us prove (3.46). For any b > 0 , let us define the set

Ω b { x : δ ( x ) > b } .

Now recalling the proof of Proposition 17, we have

w ¯ w λ > w ̲ w λ ̲ > 0 .

On the other hand, by ( f 1 ) and Lemma 6, we have f ( x , h ( s ) ) h ( s ) is increasing in s > 0 . Therefore, combining the above facts together with Lemma 6- ( h 3 ) , ( h 12 ) and using the mean value theorem and (3.2) and (3.3), for x Ω b 2 , we obtain

Δ ( w ¯ w λ ) λ α ( x ) [ h ( w ¯ ) q h ( w ¯ ) h ( w λ ) q h ( w λ ) ] = λ α ( x ) [ q ( h ( w * ) ) 2 h ( w * ) q 1 + h ( w * ) q h ( w * ) ] ( w ¯ w λ ) , for some w * ( w λ , w ¯ ) λ α ( x ) [ q h ( w * ) q 1 2 h ( w * ) q ] ( w ¯ w λ ) λ α ( x ) [ q h ( w λ ̲ ) q 1 2 h ( w λ ̲ ) q ] ( w ¯ w λ ) λ α ( x ) [ q h ( c 1 ( q , λ ) ( δ ( x ) ) a ) q 1 + 2 h ( c 1 ( q , λ ) ( δ ( x ) ) a ) q ] ( w ¯ w λ ) λ α ( x ) q h c 1 ( q , λ ) b 2 a q 1 + 2 h c 1 ( q , λ ) b 2 a q ( w ¯ w λ ) ,

where a = 1 , if 0 < q < 1 and a = 2 q + 1 , if 1 < q < 3 . By applying [7, Theorem 3], from the last relation, we infer that for any b > 0 , there exists a constant C * > 0 such that

w ¯ w λ C * b 2 > 0 in Ω b .

Given ε > 0 , choose b > 0 such that Ω \ Ω b < ε 2 . Since we assumed that w k w λ in H 0 1 ( Ω ) , for sufficiently large k N , we obtain

S k ¯ Ω \ Ω b + S k ¯ Ω b ε 2 + S k ¯ Ω b w k w λ w ¯ w λ d x ε 2 + 4 C * 2 b 2 S k ¯ Ω b ( w k w λ ) 2 d x < ε + C w k w λ 2 .

This yields that S k ¯ 0 as k + . In a similar fashion as mentioned earlier, considering w ̲ w λ , we obtain S k ̲ 0 as k + . Therefore, as k + ,

(3.47) v k ¯ 2 = S k ¯ ( w k w ¯ ) 2 d x 2 w k w λ 2 + S k ¯ ( w λ w ¯ ) 2 d x 0 .

Similarly, v k ̲ 2 0 as k + . Since w ¯ is a super-solution to (2.7), by using Lemma 6- ( h 9 ) and mean value theorem, from (3.44), we obtain

(3.48) A k = 1 2 S k ¯ ( ( w ¯ + v k ¯ ) 2 w ¯ 2 ) d x λ 1 q S k ¯ α ( x ) ( h ( w ¯ + v k ¯ ) 1 q h ( w ¯ ) 1 q ) d x λ S k ¯ ( F ( x , h ( w ¯ + v k ¯ ) ) F ( x , h ( w ¯ ) ) ) d x , = 1 2 v k ¯ 2 + S k ¯ w ¯ v k ¯ d x λ 1 q S k ¯ α ( x ) h ( w ¯ + θ v k ¯ ) q h ( w ¯ + θ v k ¯ ) v k ¯ d x λ S k ¯ f ( x , h ( w ¯ + θ v k ¯ ) ) h ( w ¯ + θ v k ¯ ) v k ¯ d x , θ ( 0 , 1 ) 1 2 v k ¯ 2 + λ S k ¯ α ( x ) h ( w ¯ ) q h ( w ¯ ) v k ¯ d x + λ S k ¯ f ( x , h ( w ¯ ) ) h ( w ¯ ) v k ¯ d x λ 1 q S k ¯ α ( x ) h ( w ¯ + θ λ v k ¯ ) q h ( w ¯ + θ v k ¯ ) v k ¯ d x λ S k ¯ f ( x , h ( w ¯ + θ v k ¯ ) ) h ( w ¯ + θ v k ¯ ) v k ¯ d x 1 2 v k ¯ 2 + λ S k ¯ f ( x , h ( w ¯ ) ) h ( w ¯ ) v k ¯ d x λ S k ¯ f ( x , h ( w ¯ + θ v k ¯ ) ) h ( w ¯ + θ v k ¯ ) v k ¯ d x = 1 2 v k ¯ 2 + λ θ S k ¯ ( f ( x , h ( w ¯ + θ ˜ v k ¯ ) ) ( h ( w ¯ + θ ˜ v k ¯ ) ) 2 + f ( x , h ( w ¯ + θ ˜ v k ¯ ) ) h ( w ¯ + θ ˜ v k ¯ ) ) v k ¯ 2 d x , θ ˜ ( 0 , 1 ) = 1 2 v k ¯ 2 + λ θ S k ¯ ( f ( x , h ( w ¯ + θ ˜ v k ¯ ) ) ( h ( w ¯ + θ ˜ v k ¯ ) ) 2 2 f ( x , h ( w ¯ + θ ˜ v k ¯ ) ) h ( w ¯ + θ ˜ v k ¯ ) ( h ( w ¯ + θ ˜ v k ¯ ) ) 4 ) v k ¯ 2 d x .

Now by the definition of the function f , we have

f ( x , h ( s ) ) = ( g ( x , h ( s ) ) + 4 h ( s ) 3 g ( x , h ( s ) ) ) exp ( h ( s ) 4 ) 4 h ( s ) 3 f ( x , h ( s ) ) .

By using this combining with Lemma 6- ( h 13 ) , ( h 12 ) , ( h 3 ) , ( h 6 ) and Hölder’s inequality, from (3.48), we deduce

A k 1 2 v k ¯ 2 + λ θ S k ¯ ( 4 f ( x , h ( w ¯ + θ ˜ v k ¯ ) ) h ( w ¯ + θ ˜ v k ¯ ) 3 ( h ( w ¯ + θ ˜ v k ¯ ) ) 2 2 f ( x , h ( w ¯ + θ ˜ v k ¯ ) ) h ( w ¯ + θ ˜ v k ¯ ) ( h ( w ¯ + θ ˜ v k ¯ ) ) 4 ) v k ¯ 2 d x 1 2 v k ¯ 2 λ θ S k ¯ ( 4 f ( x , h ( w ¯ + θ ˜ v k ¯ ) ) h ( w ¯ + θ ˜ v k ¯ ) 3 ( h ( w ¯ + θ ˜ v k ¯ ) ) 3 2 f ( x , h ( w ¯ + θ ˜ v k ¯ ) ) h ( w ¯ + θ ˜ v k ¯ ) h ( w ¯ + θ ˜ v k ¯ ) ) v k ¯ 2 d x 1 2 v k ¯ 2 λ 2 θ S k ¯ f ( x , h ( w ¯ + θ ˜ v k ¯ ) ) v k ¯ 2 d x 1 2 v k ¯ 2 λ C S k ¯ exp ( ( 1 + ε ) h ( w ¯ + θ ˜ v k ¯ ) 4 ) v k ¯ 2 d x , by ( f 2 ) for ε > 0 1 2 v k ¯ 2 λ C S k ¯ exp ( 2 ( 1 + ε ) ( w ¯ + θ ˜ v k ¯ ) 2 ) v k ¯ 2 d x 1 2 v k ¯ 2 λ C S k ¯ exp ( 4 ( 1 + ε ) ( w ¯ + θ ˜ v k ¯ ) 2 ) d x 1 2 S k ¯ v k ¯ 4 d x 1 2 1 2 v k ¯ 2 λ C S k ¯ v k ¯ 2 0 , for large k N ,

where in the last line, we used Theorem 1, (3.47), and (3.46). Similarly, from (3.45), we can show that B k 0 for sufficiently large k N . Thus, from (3.43), for large k N , it yields that

J λ ( w k ) J λ ( w λ ) ,

which is a contradiction to (3.42). Hence, w λ is a local minimum of J λ over H 0 1 ( Ω ) .□

The next result is a consequence of Lemma 18, which yields that at the threshold level of λ , i.e., for λ = Λ * , we have a weak solution to (2.7).

Proposition 19

Let the conditions in Theorem 4be satisfied and let h be defined as in (2.2). Then for λ = Λ * , (2.7) admits a weak solution in H 0 1 ( Ω ) C φ q + ( Ω ) .

Proof

From the definition of Λ * , there exists a increasing sequence { λ k } Q such that λ k Λ * as k + . Hence, by Proposition 17, { w λ k } H 0 1 ( Ω ) C φ q + ( Ω ) Y λ k is a sequence of positive weak solutions to (2.7) with λ = λ k , where Y λ k is defined as in (3.35). Therefore,

(3.49) Ω w λ k 2 d x λ k Ω α ( x ) h ( w λ k ) q h ( w λ k ) w λ k d x λ k Ω f ( x , h ( w λ k ) ) h ( w λ k ) w λ k d x = 0 .

For λ = λ k , let w λ k ̲ denote the unique solution to (3.1), which is a sub-solution to (2.7) as in (3.34). So, using Lemma 6- ( h 4 ) , we obtain

(3.50) Ω w λ k ̲ 2 d x = λ k Ω α ( x ) h ( w λ k ̲ ) q h ( w λ k ̲ ) w λ k ̲ d x λ k Ω α ( x ) h ( w λ k ̲ ) 1 q d x .

Now Lemma 18 yields that w λ k is a local minimizer of J λ k for each k N . So,

(3.51) J λ k ( w λ k ) = min w Y λ k J λ k ( w ) J λ k ( w λ k ̲ ) 1 2 Ω w λ k ̲ 2 d x λ k 1 q Ω α ( x ) h ( w λ k ̲ ) 1 q d x if q 1 ; 1 2 Ω w λ k ̲ 2 d x λ k Ω α ( x ) log h ( w λ k ̲ ) d x if q = 1 .

Plugging (3.50) in (3.51), we obtain

(3.52) J λ k ( w λ k ) β k λ k 1 2 1 1 q Ω α ( x ) h ( w λ k ̲ ) 1 q d x if q 1 ; λ k 2 Ω α ( x ) d x λ k Ω α ( x ) log h ( w λ k ̲ ) d x if q = 1 .

Thus, for all 0 < q < 3 , since 0 < λ 1 λ 2 λ k Λ * , from (3.2) and (3.3) and Theorem 11- ( i v ) , we infer that

(3.53) sup k β k < + .

Case-I: 0 < q < 1 .

Now (3.49) and (3.52) combining with Lemma 6- ( h 4 ) and ( f 3 ) imply that

(3.54) β k + λ k 1 1 q 1 2 Ω α ( x ) h ( w λ k ) 1 q d x λ k 2 Ω f ( x , h ( w λ k ) ) h ( w λ k ) w λ k d x Ω F ( x , h ( w λ k ) ) d x λ k 1 4 1 τ Ω f ( x , h ( w λ k ) ) h ( w λ k ) d x .

Plugging (3.54) and (3.53) in (3.49) and using Lemma 6- ( h 4 ) , ( h 5 ) together with the fact that τ > 4 , the Sobolev inequality, and the Hölder inequality, we obtain

(3.55) w λ k 2 C 1 + C 2 Λ * Ω α ( x ) h ( w λ k ) 1 q d x C 1 + C 2 Λ * α Ω w λ k 1 q d x C 1 + C 3 w λ k 1 q .

Case-II: q = 1 .

Again using (3.49), (3.52), Lemma 6- ( h 4 ) and ( f 3 ) combining with the inequality log h ( s ) log s < s , for s > 0 , we obtain

(3.56) β k + λ k Ω α ( x ) d x λ k 2 Ω α ( x ) h ( w λ k ) 1 h ( w λ k ) w λ k d x + λ k 2 Ω f ( x , h ( w λ k ) ) h ( w λ k ) w λ k d x Ω F ( x , h ( w λ k ) ) d x λ k 4 Ω α ( x ) d x + λ k 1 4 1 τ Ω f ( x , h ( w λ k ) ) h ( w λ k ) d x .

This gives that

(3.57) β k + 3 4 λ k Ω α ( x ) λ k 1 4 1 τ Ω f ( x , h ( w λ k ) ) h ( w λ k ) d x .

Using (3.57) and (3.53), from (3.49), we deduce

(3.58) w λ k 2 C 4 ( β k + Λ * α Ω ) < + .

Case-III: 1 < q < 3 .

From (3.54), it follows that

λ k 1 4 1 τ Ω f ( x , h ( w λ k ) ) h ( w λ k ) d x β k + λ k 1 q 1 + 1 2 Ω α ( x ) h ( w λ k ) 1 q d x .

Employing the last relation in (3.49) and using Theorem 11- ( i v ) , Lemma 6- ( h 8 ) and (3.3) for λ 2 with c 1 ( q , λ 2 ) > 0 small enough such that c 1 ( q , λ 2 ) δ 2 1 + q < 1 , we obtain

(3.59) w λ k 2 C 5 + C 6 λ k Ω h ( w λ k ) 1 q d x C 5 + C 6 λ k Ω h ( w λ k ̲ ) 1 q d x C 5 + C 6 Λ * Ω h ( w λ 2 ̲ ) 1 q d x C 5 + C 6 Λ * Ω h c 1 ( q , λ 2 ) δ 2 1 + q 1 q d x C 5 + C 6 Λ * h ( 1 ) 1 q ( c 1 ( q , λ 2 ) ) 1 q Ω δ 2 ( 1 q ) 1 + q d x < + ,

since 2 ( 1 q ) 1 + q > 1 .

Thus, each of the expressions in (3.55), (3.59), (3.58) from Cases I, II, III, respectively, yields that

(3.60) limsup k + w λ k < + .

Therefore, up to a sub-sequence, there exists w Λ * H 0 1 ( Ω ) such that w λ k w Λ * weakly in H 0 1 ( Ω ) and w λ k ( x ) w Λ * ( x ) a.e. in Ω as k + . Also, by the construction, it follows that w λ k w λ k ̲ w λ 1 ̲ . So,

w Λ * ( x ) = lim k + w λ k ( x ) > w λ 1 ̲ ( x ) > 0 a.e. in Ω .

Thus, by the Lebesgue dominated convergence theorem, for any ϕ C c ( Ω ) , we have

(3.61) Ω α ( x ) h ( w λ k ) q h ( w λ k ) ϕ d x Ω α ( x ) h ( w Λ * ) q h ( w Λ * ) ϕ d x as k + .

Next, we show that for any ϕ C c ( Ω ) ,

(3.62) Ω f ( x , h ( w λ k ) ) h ( w λ k ) ϕ d x Ω f ( x , h ( w Λ * ) ) h ( w Λ * ) ϕ d x as k + .

To prove (3.62), first observe that for 1 < q < 3 , using Lemma 6- ( h 4 ) , and arguing similarly as in (3.38), α ( x ) h ( w λ k ) q h ( w λ k ) w λ k α h ( w λ k ) 1 q α h ( w λ 1 ̲ ) 1 q L 1 ( Ω ) . Thus, by the Lebesgue dominated convergence theorem, we have

(3.63) Ω α ( x ) h ( w λ k ) q h ( w λ k ) w λ k d x Ω α ( x ) h ( w Λ * ) q h ( w Λ * ) w Λ * d x as k + .

Furthermore, for 0 < q 1 , (3.63) follows similarly as in (3.11). Hence, from (3.49), (3.60) and (3.63), we obtain

limsup k + Ω f ( x , h ( w λ k ) ) h ( w λ k ) w λ k d x < + .

Then repeating a similar argument as in (3.13), we obtain (3.62). Gathering (3.62) and (3.63), we finally infer that w Λ * H 0 1 ( Ω ) is a positive weak solution to (2.7). Moreover, in light of Lemma 14, we have w Λ * H 0 1 ( Ω ) C φ q + ( Ω ) . Hence, the proof of the proposition is complete.□

Proof of Theorem 4

Combining Lemmas 13, 14, 16, and 18 along with Propositions 17 and 19, we infer that w λ H 0 1 ( Ω ) C φ q + ( Ω ) is a weak solution to (2.7). Now by Lemma 6- ( h 1 ) , we have h is a C function and Lemma 6- ( h 8 ) , ( h 11 ) ensure that h ( s ) behaves like s when s is close to 0. Therefore, we can conclude that h ( w λ ) H 0 1 ( Ω ) C φ q + ( Ω ) forms a weak solution to problem ( P * ).□

4 Proof of Theorem 5: Multiplicity result

This section is dedicated toward establishing the existence of second solution of (2.7) using the mountain pass lemma in combination with Ekeland variational principle. Let us define the set

T { w H 0 1 ( Ω ) : w w λ a.e. in Ω } .

Since by Lemma 18, w λ is a local minimizer for J λ , it follows that J λ ( w ) J λ ( w λ ) whenever w λ w σ 0 , for some small constant σ 0 > 0 . Then, one of the following cases holds:

  1. (Zero Altitude): inf { J λ ( w ) w T , w w λ = σ } = J λ ( w λ ) for all σ ( 0 , σ 0 ) .

  2. (Mountain Pass): There exists a σ 1 ( 0 , σ 0 ) such that inf { J λ ( w ) w T , w w λ = σ 1 } > J λ ( w λ ) .

Now for the case ( Z A ) , inspired by [19] and [2], we prove the existence of second weak solution to (2.7) in the following result.

Proposition 20

Let the conditions in Theorem 5hold and let λ ( 0 , Λ * ) . Suppose that ( Z A ) holds. Then for all σ ( 0 , σ 0 ) , (2.7) admits a second solution v λ H 0 1 ( Ω ) C φ q + ( Ω ) such that v λ w λ and v λ w λ = σ .

Proof

Let us fix σ ( 0 , σ 0 ) and r > 0 such that σ r > 0 and σ + r < σ 0 . Define the set

A { w T : 0 < σ r w w λ σ + r } .

Clearly A is closed in H 0 1 ( Ω ) and by ( Z A ) , inf w A J λ ( w ) = J λ ( w λ ) . So, for any minimizing sequence { w k } A satisfying w k w λ = σ , by Ekeland variational principle, we obtain another sequence { v k } A such that

(4.1) J λ ( v k ) J λ ( w k ) J λ ( w λ ) + 1 k w k v k 1 k J λ ( v k ) J λ ( v ) + 1 k v v k for all v A .

For z T , we can choose ε > 0 small enough so that v k + ε ( z v k ) A . So, from (4.1), we obtain

J λ ( v k + ε ( z v k ) ) J λ ( v k ) ε 1 k z v k .

Letting ε 0 + , using the fact that v k w λ for each k N and following the similar arguments as in (2.5) and (2.6) for the singular term in the last relation, we obtain

(4.2) 1 k z v k Ω v k ( z v k ) d x λ Ω α ( x ) h ( v k ) q h ( v k ) ( z v k ) d x λ Ω f ( x , h ( v k ) ) h ( v k ) ( z v k ) d x

for all z T . Since { v k } is a bounded sequence in H 0 1 ( Ω ) , there exists v λ H 0 1 ( Ω ) such that, up to a sub-sequence, v k v λ weakly in H 0 1 ( Ω ) and pointwise a.e. in Ω as k + . Since v k w λ for each k , v λ w λ a.e. in Ω .

Claim: v λ is a weak solution to (2.7).

For ϕ H 0 1 ( Ω ) and ε > 0 , we set

ψ k , ε ( v k + ε ϕ w λ ) , ψ ε ( v λ + ε ϕ w λ ) .

Clearly, ψ k , ε H 0 1 ( Ω ) . This gives that ( v k + ε ϕ + ψ k , ε ) T . Taking z = v k + ε ϕ + ψ k , ε in (4.2), we deduce

(4.3) 1 k ( ε ϕ + ψ k , ε ) Ω v k ( ε ϕ + ψ k , ε ) λ Ω α ( x ) h ( v k ) q h ( v k ) ( ε ϕ + ψ k , ε ) d x λ Ω f ( x , h ( v k ) ) h ( v k ) ( ε ϕ + ψ k , ε ) d x .

Note that ψ k , ε w λ + ε ϕ . Hence, using the Sobolev embedding and the Lebesgue dominated convergence theorem, as k + , ψ k , ε ψ ε in L p ( Ω ) , p [ 1 , + ) ; ψ k , ε ψ ε weakly in H 0 1 ( Ω ) and ψ k , ε ( x ) ψ ε ( x ) pointwise a.e. in Ω . Now by using Lemma 6- ( h 13 ) , we obtain

α ( x ) h ( v k ) q h ( v k ) ( ε ϕ + ψ k , ε ) α h ( v λ ) q h ( v λ ) ( v λ + 2 ε ϕ ) .

Furthermore, by using the fact that both h and f ( x , ) are non-decreasing functions, we obtain

f ( x , h ( v k ) ) h ( v k ) ( ε ϕ + ψ k , ε ) f ( x , h ( w λ + ε ϕ ) ) ( w λ + ε ϕ ) .

Hence, by employing the Lebesgue dominated convergence theorem, as k + , we infer

(4.4) Ω α ( x ) h ( v k ) q h ( v k ) ( ε ϕ + ψ k , ε ) d x Ω α ( x ) h ( v λ ) q h ( v λ ) ( ε ϕ + ψ ε ) d x ,

(4.5) Ω f ( x , h ( v k ) ) h ( v k ) ( ε ϕ + ψ k , ε ) d x Ω f ( x , h ( v λ ) ) h ( v λ ) ( ε ϕ + ψ ε ) d x .

We define the sets

Ω k , ε supp ψ k , ε , Ω ε supp ψ ε and Ω 0 { x Ω : v λ ( x ) w λ ( x ) } .

Then,

(4.6) Ω ε \ Ω 0 0 as ε 0 ;

(4.7) Ω k , ε \ Ω ε + Ω ε \ Ω k , ε 0 as k + .

Therefore, as k + ,

(4.8) Ω v k ψ k , ε d x = Ω ε v k ψ ε d x Ω k , ε ( v k v λ ) 2 d x + Ω k , ε v λ ( v λ v k ) d x + o ( 1 ) Ω v k ψ ε d x + Ω ε v λ ( v λ v k ) d x + o ( 1 ) = Ω v k ψ ε d x + o ( 1 ) .

Combining (4.3), (4.4), (4.5), (4.6), (4.7), and (4.8), letting k + and using Lemma 6- ( h 13 ) , we obtain

(4.9) Ω v λ ϕ d x λ Ω α ( x ) h ( v λ ) q h ( v λ ) ϕ d x λ Ω f ( x , h ( v λ ) ) h ( v λ ) ϕ d x 1 ε Ω v λ ψ ε d x λ Ω α ( x ) h ( v λ ) q h ( v λ ) ψ ε d x λ Ω f ( x , h ( v λ ) ) h ( v λ ) ψ ε d x = 1 ε Ω ( v λ w λ ) ψ ε d x + λ Ω α ( x ) [ h ( v λ ) q h ( v λ ) h ( w λ ) q h ( w λ ) ] ψ ε d x + λ Ω [ f ( x , h ( v λ ) ) h ( v λ ) f ( x , h ( w λ ) ) h ( w λ ) ] ψ ε d x = 1 ε Ω ε ( v λ w λ ) ( w λ v λ ε ϕ ) d x + λ Ω ε α ( x ) [ h ( v λ ) q h ( v λ ) h ( w λ ) q h ( w λ ) ] ( w λ v λ ε ϕ ) d x + λ Ω ε [ f ( x , h ( v λ ) ) h ( v λ ) f ( x , h ( w λ ) ) h ( w λ ) ] ( w λ v λ ε ϕ ) d x Ω ε ( v λ w λ ) ϕ d x λ Ω ε α ( x ) [ h ( v λ ) q h ( v λ ) h ( w λ ) q h ( w λ ) ] ϕ d x λ Ω ε [ f ( x , h ( v λ ) ) h ( v λ ) f ( x , h ( w λ ) ) h ( w λ ) ] ϕ d x + λ ε Ω ε [ f ( x , h ( v λ ) ) h ( v λ ) f ( x , h ( w λ ) ) h ( w λ ) ] ( w λ v λ ) d x .

Since v λ w λ , using Lemma 6- ( h 9 ) and the mean value theorem, if x Ω ε ,

(4.10) [ f ( x , h ( v λ ) ) h ( v λ ) f ( x , h ( w λ ) ) h ( w λ ) ] ( w λ v λ ) d x ( v λ w λ ) 2 [ f ( x , h ( ξ λ ) ) ( h ( ξ λ ) ) 2 + f ( x , h ( ξ λ ) ) h ( ξ λ ) ] , ξ λ ( w λ , v λ ) ε 2 f ( x , h ( ξ λ ) ) ϕ 2 .

Plugging (4.10) into (4.9), letting ε 0 + and using (4.6), we obtain

Ω v λ ϕ d x λ Ω α ( x ) h ( v λ ) q h ( v λ ) ϕ d x λ Ω f ( x , h ( v λ ) ) h ( v λ ) ϕ d x o ( 1 ) λ ε Ω ε f ( x , h ( ξ λ ) ) ϕ 2 d x = o ( 1 ) .

Considering ϕ in place of ϕ and arguing similarly as mentioned earlier, we obtain the reverse inequality in the last relation. Therefore,

Ω v λ ϕ d x λ Ω α ( x ) h ( v λ ) q h ( v λ ) ϕ d x λ Ω f ( x , h ( v λ ) ) h ( v λ ) ϕ d x = 0

for all ϕ H 0 1 ( Ω ) . So, v λ H 0 1 ( Ω ) is a weak solution to (2.7), and thus, the claim is proved. Moreover, by Lemma 14, v λ C φ q + ( Ω ) .

Now we show that v λ w λ . For that, it is enough to prove that

(4.11) v k v λ in H 0 1 ( Ω ) as k + .

By applying the Brézis-Lieb lemma, we obtain

v k 2 v k v λ 2 = v λ 2 + o ( 1 ) .

By putting z = v λ in (4.2) and using the fact that v k v λ in T as k + , we obtain that

(4.12) Ω ( v k v λ ) 2 d x o ( 1 ) λ Ω α ( x ) h ( v k ) q h ( v k ) ( v λ v k ) d x λ Ω f ( x , h ( v k ) ) h ( v k ) ( v λ v k ) d x .

Let us denote u k v k w λ v k w λ . Now recalling ( f 2 ) and (3.6) for any ε > 0 , Lemma 6- ( h 3 ) , ( h 4 ) , ( h 5 ) and using Hölder’s inequality, for some large N 1 , we deduce

(4.13) Ω { x : v k ( x ) > N } f ( x , h ( v k ) ) h ( v k ) v k d x C Ω { x : v k ( x ) > N } exp ( ( 1 + ε ) h ( v k ) 4 ) h ( v k ) v k d x C Ω { x : v k ( x ) > N } exp ( 2 ( 1 + ε ) v k 2 ) v k d x C Ω { x : v k ( x ) > N } exp ( 3 ( 1 + ε ) v k 2 ) d x = C Ω { x : v k ( x ) > N } exp ( ( 1 + ε ) v k 2 ) exp ( 4 ( 1 + ε ) ( v k w λ + w λ ) 2 ) d x C exp ( ( 1 + ε ) N 2 ) Ω { x : v k ( x ) > N } exp ( 8 ( 1 + ε ) [ ( v k w λ ) 2 + w λ 2 ] ) d x C exp ( ( 1 + ε ) N 2 ) Ω { x : v k ( x ) > N } exp ( 8 ( 1 + ε ) u k 2 v k w λ 2 ) exp ( 8 ( 1 + ε ) w λ 2 ) d x C exp ( ( 1 + ε ) N 2 ) Ω { x : v k ( x ) > N } exp ( 8 p ( 1 + ε ) u k 2 v k w λ 2 ) d x 1 p Ω exp ( 8 p ( 1 + ε ) w λ 2 ) d x 1 p C exp ( ( 1 + ε ) N 2 ) Ω { x : v k ( x ) > N } exp ( 8 p ( 1 + ε ) u k 2 ( σ + r ) 2 ) d x 1 p Ω exp ( 8 p ( 1 + ε ) w λ 2 ) d x 1 p .

Now choosing p > 1 and σ 0 > 0 appropriately small so that 8 ( 1 + ε ) p ( σ + r ) < 8 ( 1 + ε ) p σ 0 < 4 π and then applying Theorem 1, from (4.13), we derive

(4.14) Ω { x : v k ( x ) > N } f ( x , h ( v k ) ) h ( v k ) v k d x = O ( exp ( ( 1 + ε ) N 2 ) ) .

Since f ( x , h ( v k ) ) h ( v k ) v k f ( x , h ( v λ ) ) h ( v λ ) v λ pointwise a.e. in Ω as k + , using (4.14) and applying the Lebesgue dominated convergence theorem, it follows that

(4.15) Ω f ( x , h ( v k ) ) h ( v k ) v k d x = Ω { x : v k N } f ( x , h ( v k ) ) h ( v k ) v k d x + Ω { x : v k > N } f ( x , h ( v k ) ) h ( v k ) v k d x = Ω { x : v k N } f ( x , h ( v k ) ) h ( v k ) v k d x + O ( exp ( ( 1 + ε ) N 2 ) ) Ω f ( x , h ( v λ ) ) h ( v λ ) v λ d x as k + and N + .

In a similar way, we also have Ω f ( x , h ( v k ) ) h ( v k ) v λ d x Ω f ( x , h ( v λ ) ) h ( v λ ) v λ d x as k + . This, together with (4.15), yields that

(4.16) Ω f ( x , h ( v k ) ) h ( v k ) ( v k v λ ) d x 0 as k + .

Next, we show that

(4.17) Ω α ( x ) h ( v k ) q h ( v k ) ( v k v λ ) d x 0 as k + .

By the construction, we have v k , v λ w λ w λ ̲ . Next, we consider the three cases separately as following:

Case I: 0 < q < 1 . Using Lemma 6- ( h 3 ) , ( h 8 ) , (3.2) with c 1 ( q , λ ) > 0 small enough such that c 1 ( q , λ ) δ < 1 and (3.29) for v λ , we obtain

(4.18) α ( x ) h ( v k ) q h ( v k ) v λ α ( x ) h ( w λ ̲ ) q v λ α ( x ) h ( c 1 ( q , λ ) δ ) q C 2 ( q , λ ) δ α C ( q , λ ) h ( 1 ) q δ 1 q L 1 ( Ω ) .

Moreover, by Lemma 6- ( h 4 ) , ( h 5 ) , it follows that

(4.19) α ( x ) h ( v k ) q h ( v k ) v k α ( x ) h ( v k ) 1 q α v k 1 q L 1 ( Ω )

due to the Sobolev inequality and the Hölder inequality. Now using (3.10), Ω v k 1 q d x Ω v λ 1 q d x as k + . So, by the Lebesgue dominated convergence theorem, as k + , combining (4.18) and (4.19), we obtain (4.17).

Case II: q = 1 . Again using Lemma 6- ( h 13 ) , ( h 3 ) , ( h 8 ) , (3.4) with c ( λ ) > 0 small enough such that c ( λ ) δ < 1 and (3.31) for v λ with any small 0 < ε < 1 ,

(4.20) h ( v k ) 1 h ( v k ) v λ h ( w λ ̲ ) 1 h ( w λ ̲ ) v λ h ( c ( λ ) δ ) C ε ( λ ) δ 1 ε c ( λ ) C ε ( λ ) h ( 1 ) 1 δ 1 δ 1 ε = C ( λ , ε ) δ ε L 1 ( Ω ) .

From this and the hypothesis ( α 1 ) , (4.17) follows, thanks to the Lebesgue dominated convergence theorem.

Case III: 1 < q < 3 . In view of (3.3) with c 1 ( q , λ ) > 0 small enough such that c 1 ( q , λ ) δ 2 1 + q < 1 , recalling (3.30) for v λ and using Lemma 6- ( h 13 ) , ( h 3 ) , ( h 8 ) , it yields that

(4.21) h ( v k ) q h ( v k ) v λ h ( w λ ̲ ) q h ( w λ ̲ ) v λ h c 1 δ 2 1 + q q v λ ( c 1 h ( 1 ) ) q δ 2 q 1 + q C 2 ( q , λ ) δ 2 1 + q C ( q , λ ) δ 2 ( 1 q ) 1 + q L 1 ( Ω )

since 2 ( 1 q ) 1 + q > 1 . Again, repeating a similar argument as in (3.38), it follows that

(4.22) h ( v k ) q h ( v k ) v k h ( w λ ̲ ) 1 q L 1 ( Ω ) .

Thus, (4.21) and (4.22), combining with the hypothesis ( α 1 ) and the Lebesgue dominated convergence theorem, yield (4.17).

Therefore, taking into account (4.16) and (4.17), from (4.12), we finally obtain (4.11). This completes the proof of the proposition.□

Now we prove the existence of second solution in the case where ( M P ) occurs.

Now for k N , we define the Moser function k : Ω R as follows:

k ( x ) = 1 2 π ( log k ) 1 2 if 0 x 1 k , log 1 x ( log k ) 1 2 if 1 k x 1 , 0 if x 1 .

Then, k H 0 1 ( Ω ) and supp k B 1 . Moreover, we define the function

k ( x ) k x x 0 .

We choose x 0 and is such a way so that supp k Ω . Note that k = 1 .

Lemma 21

Let the conditions in Theorem 5and ( M P ) hold. Then

  1. J λ ( w λ + t k ) as t + uniformly for k large.

  2. sup t 0 J λ ( w λ + t k ) < J λ ( w λ ) + π for large k .

Proof

We prove this lemma for the case q 1 . When q = 1 , the proof follows similarly.

(i) By ( f 2 ) , there exist two positive constants C 1 , C 2 such that

F ( x , s ) C 1 exp ( h ( s ) 4 ) C 2

for all x Ω , s 0 . By using this combining with Lemma 6- ( h 8 ) for large t 1 and the Hölder’s inequality, we obtain

J λ ( w λ + t k ) = 1 2 Ω ( w λ + t k ) 2 d x λ 1 q Ω α ( x ) h ( w λ + t k ) 1 q d x λ Ω F ( x , h ( w λ + t k ) ) d x 1 2 w λ 2 + t w λ k + 1 2 t 2 k 2 + C 2 λ Ω C 1 λ Ω exp ( h ( w λ + t k ) 4 ) t 2 + C 3 ( λ ) t λ C 1 B k ( x 0 ) exp ( h ( 1 ) 4 ( w λ + t k ) 2 ) d x t 2 + C 3 ( λ ) t C 4 ( λ ) B k ( x 0 ) exp t 2 h ( 1 ) 4 2 π log k d x = t 2 + C 3 ( λ ) t C 4 ( λ ) k h ( 1 ) 4 2 π t 2 2 uniformly in  k  as  t + .

(ii). Suppose ( i i ) does not hold. So, there exists a sub-sequence of { k } , still denoted by { k } , such that

(4.23) max t 0 J λ ( w λ + t k ) J λ ( w λ ) + π .

Then,

(4.24) J λ ( w λ + t k ) = 1 2 w λ 2 + t Ω w λ k d x + t 2 2 λ 1 q Ω α ( x ) h ( w λ + t k ) 1 q d x λ Ω F ( x , h ( w λ + t k ) ) d x = J λ ( w λ ) + λ 1 q Ω α ( x ) h ( w λ ) 1 q d x + λ Ω F ( x , h ( w λ ) ) d x + t λ Ω α ( x ) h ( w λ ) q h ( w λ ) k d x + t λ Ω f ( x , h ( w λ ) ) h ( w λ ) k d x + t 2 2 λ 1 q Ω α ( x ) h ( w λ + t k ) 1 q d x λ Ω F ( x , h ( w λ + t k ) ) d x .

From ( i ) and (4.23), it follows that there exists t k ( 0 , + ) and L 0 > 0 satisfying t k L 0 , such that

(4.25) max t 0 J λ ( w λ + t k ) = J λ ( w λ + t k k ) J λ ( w λ ) + π .

Now (4.24) and (4.25) imply that

λ 1 q Ω α ( x ) h ( w λ ) 1 q d x + λ Ω F ( x , h ( w λ ) ) d x + t k λ Ω α ( x ) h ( w λ ) q h ( w λ ) k d x + t k λ Ω f ( x , h ( w λ ) ) h ( w λ ) k d x + t k 2 2 λ 1 q Ω α ( x ) h ( w λ + t k k ) 1 q d x λ Ω F ( x , h ( w λ + t k k ) ) d x π .

Using mean value theorem and Lemma 6- ( h 9 ) , we deduce

(4.26) λ 1 q Ω α ( x ) h ( w λ ) 1 q d x Ω α ( x ) h ( w λ + t k k ) 1 q d x + t k λ Ω α ( x ) h ( w λ ) q h ( w λ ) k d x = λ t k Ω α ( x ) h ( w λ + t k θ k k ) q h ( w λ + t k θ k k ) k d x Ω α ( x ) h ( w λ ) q h ( w λ ) k d x = λ t k 2 θ k Ω α ( x ) k 2 [ q h ( w λ + t k ξ k k ) q 1 h ( w λ + t k ξ k k ) + h ( w λ + t k ξ k k ) q h ( w λ + t k ξ k k ) ] = λ t k 2 θ k Ω α ( x ) k 2 [ q h ( w λ + t k ξ k k ) q 1 h ( w λ + t k ξ k k ) + 2 h ( w λ + t k ξ k k ) 1 q ( h ( w λ + t k ξ k k ) ) 4 ] ,

where θ k , ξ k ( 0 , 1 ) . Now by Lemma 6- ( h 3 ) , ( h 12 ) ,

2 h ( s ) 1 q ( h ( s ) ) 4 = 2 h ( s ) 1 q ( h ( s ) h ( s ) ) 2 ( h ( s ) ) 2 h ( s ) 1 q h ( s ) .

Plugging the last relation in (4.26) and using Lemma 6- ( h 13 ) , ( h 3 ) , ( h 8 ) together with the fact that w λ + t k ξ k k w λ ̲ > 0 in B k ( x 0 ) , we obtain

(4.27) λ 1 q Ω α ( x ) h ( w λ ) 1 q d x Ω α ( x ) h ( w λ + t k k ) 1 q d x + t k λ Ω α ( x ) h ( w λ ) q h ( w λ ) k d x λ t k 2 θ k ( q + 1 ) Ω α ( x ) k 2 h ( w λ + t k ξ k k ) q 1 h ( w λ + t k ξ k k ) d x λ t k 2 θ k ( q + 1 ) α Ω k 2 h ( w λ ̲ ) q 1 h ( w λ ̲ ) d x λ t k 2 θ k ( q + 1 ) α ( h ( 1 ) ) 1 q Ω k 2 w λ ̲ q 1 d x = t k 2 O 1 log k .

On the other hand, again using the mean value theorem and Lemma 6- ( h 9 ) , ( h 12 ) , ( h 3 ) , we infer

(4.28) t k Ω f ( x , h ( w λ ) ) h ( w λ ) k d x + Ω F ( x , h ( w λ ) ) d x Ω F ( x , h ( w λ + t k k ) ) d x = t k Ω f ( x , h ( w λ ) ) h ( w λ ) k d x Ω f ( x , h ( w λ + t k θ ˜ k k ) ) h ( w λ + t k θ ˜ k k ) k d x = t k 2 Ω θ ˜ k k 2 [ f ( x , h ( w λ + t k ξ ˜ k k ) ) ( h ( w λ + t k ξ ˜ k k ) ) 2 + f ( x , h ( w λ + t k ξ ˜ k k ) ) h ( w λ + t k ξ ˜ k k ) ] d x = t k 2 θ ˜ k Ω k 2 [ f ( x , h ( w λ + t k ξ ˜ k k ) ) ( h ( w λ + t k ξ ˜ k k ) ) 2 + 2 f ( x , h ( w λ + 2 t k ξ ˜ k k ) ) h ( w λ + t k ξ ˜ k k ) ( h ( w λ + t k ξ ˜ k k ) ) 4 ] d x t k 2 θ ˜ k Ω k 2 [ f ( x , h ( w λ + t k ξ ˜ k k ) ) ( h ( w λ + t k ξ ˜ k k ) ) 2 + 2 f ( x , h ( w λ + 2 t k ξ ˜ k k ) ) h ( w λ + t k ξ ˜ k k ) h ( w λ + t k ξ ˜ k k ) ] d x t k 2 θ ˜ k Ω k 2 [ ( 2 M 0 ) f ( x , h ( w λ + t k ξ ˜ k k ) ) + L ] h ( w λ + t k ξ ˜ k k ) d x [ by (1.5) ] L t k 2 θ ˜ k Ω k 2 d x = t k 2 O 1 log k ,

where θ ˜ k , ξ ˜ k ( 0 , 1 ) . Now gathering (4.23), (4.27), and (4.28), it follows that t k 2 2 + O 1 log k t k 2 π . That is,

(4.29) t k 2 2 π O 1 log k .

In virtue of the relation d d t J λ ( w λ + t k ) t = t k = 0 , we have

(4.30) t k 2 + t k Ω w λ k d x λ Ω α ( x ) h ( w λ + t k k ) q h ( w λ + t k k ) t k k d x = λ Ω f ( x , h ( w λ + t k k ) ) h ( w λ + t k k ) t k k d x .

Now we estimate the right-hand side in (4.30).

By using the fact that w λ C > 0 on B k ( x 0 ) combining with ( f 1 ) and Lemma 6- ( h 4 ) , we obtain

(4.31) Ω f ( x , h ( w λ + t k k ) ) h ( w λ + t k k ) t k k d x B k ( x 0 ) f ( x , h ( C + t k k ) ) h ( C + t k k ) t k k d x 1 2 B k ( x 0 ) f ( x , h ( C + t k k ) ) h ( C + t k k ) t k k C + t k k d x .

By ( f 1 ) , since g ( x , s ) is non-decreasing in s , we obtain

(4.32) lim s + s f ( x , s ) exp ( s 4 ) = lim s + s g ( x , s ) = + .

Therefore, for any b > 0 , there exists some constant M b > 0 such that

(4.33) f ( x , h ( C + t k k ) ) h ( C + t k k ) > b exp ( h ( C + t k k ) 4 ) for all t k > M b .

On the other hand, by Lemma 6- ( h 7 ) , for any ε > 0 , there exists m ε > 0 such that

(4.34) h ( C + t k k ) 4 > ( C + t k k ) 2 ( 2 ε ) for all t k m ε .

Plugging (4.33) and (4.34) in (4.31) and using (4.29), for t k > max { M b , m ε } , we obtain

(4.35) B k ( x 0 ) f ( x , h ( w λ + t k k ) ) h ( w λ + t k k ) t k k d x b 2 B k ( x 0 ) exp ( h ( C + t k k ) 4 ) t k k C + t k k d x b 2 B k ( x 0 ) exp ( ( C + t k k ) 2 ( 2 ε ) ) t k k C + t k k d x b 2 t k k ( x 0 ) C + t k k ( x 0 ) B k ( x 0 ) exp ( ( C + t k k ) 2 ( 2 ε ) ) d x = b 2 t k k ( x 0 ) C + t k k ( x 0 ) exp ( [ C 2 + 2 C t k k ( x 0 ) + t k 2 M k ( x 0 ) 2 ] ( 2 ε ) ) B k ( x 0 ) = b 2 t k k ( x 0 ) C + t k k ( x 0 ) C 0 exp ( 2 C t k k ( x 0 ) + t k 2 M k ( x 0 ) 2 ) B k ( x 0 ) as ε 0 + = b 2 1 1 + C t k k ( x 0 ) C 0 π 2 exp 2 C t k k ( x 0 ) + 2 log k t k 2 2 π 1 b 2 1 1 + C 2 π t k log k C 1 π 2 exp ( 2 C t k k ( x 0 ) ) .

Now incorporating (4.35) and (4.31) in (4.30) and using Hölder’s inequality, we obtain

+ > C 2 L 0 2 + w λ L 0 t k 2 + w λ t k b 2 λ 1 + C 2 π t k log k C 1 π 2 exp ( 2 C t k k ( x 0 ) ) .

Since { t k } is bounded away from 0, for some C 3 > 0 , we have t k k ( x 0 ) > C 3 ( log k ) 1 2 + as k + . Hence, letting k + in the last relation, it yields that

C 2 b 2 λ C 1 π 2 exp ( 2 C C 3 log k ) + .

This is absurd since, C 2 < + . Hence, ( i i ) follows. This completes the proof of the lemma.□

Next, we recall the following result due to Lions [24].

Theorem 22

Let { u k : u k = 1 } be a sequence in H 0 1 ( Ω ) converging weakly in H 0 1 ( Ω ) to a non-zero function u . Then, for every 0 < p < ( 1 u ) 1 ,

sup k Ω exp ( 4 π p u k 2 ) d x < + .

Proposition 23

Let the conditions in Theorem 5hold and let λ ( 0 , Λ * ) . Suppose that ( M P ) holds. Then (2.7) admits a second solution v λ H 0 1 ( Ω ) C φ q + ( Ω ) satisfying v λ w λ and v λ w λ = σ 1 .

Proof

First, we define the complete metric space

Γ { η C ( [ 0 , 1 ] , T ) : η ( 0 ) = w λ , η ( 1 ) w λ > σ 1 , J λ ( η ( 1 ) ) < J λ ( w λ ) }

with the metric

d ( η , η ˜ ) = max t [ 0 , 1 ] { η ( t ) η ˜ ( t ) } for all η , η ˜ Γ .

Using Lemma 21- ( i ) , we have Γ . Let us set

γ 0 = inf η Γ max t [ 0 , 1 ] J λ ( η ( t ) ) .

Then, in light of Lemma 21- ( i i ) and the condition ( M P ) , we obtain

J λ ( w λ ) < γ 0 J λ ( w λ ) + π .

Set Φ ( η ) max t [ 0 , 1 ] J λ ( η ( t ) ) for η Γ . Then employing Ekeland’s variational principle to the functional Φ on Γ , we obtain a sequence { η k } Γ such that

Φ ( η k ) < γ 0 + 1 k and Φ ( η k ) < Φ ( η ) + 1 k Φ ( η ) Φ ( η k ) Γ for all η Γ .

Again, applying Ekeland’s variational principle to J λ on Γ and arguing exactly similar to [4, Lemma 3.5], one can show that there exists a sequence { v k } T such that

(4.36) J λ ( v k ) γ 0 as k + Ω v k ( w v k ) d x λ Ω α ( x ) h ( v k ) q h ( v k ) ( w v k ) d x λ Ω f ( x , h ( v k ) ) h ( v k ) ( w v k ) d x C k ( 1 + w ) for all w T .

Now choosing w = 2 v k in (4.36), we obtain

(4.37) Ω v k 2 d x λ Ω α ( x ) h ( v k ) q h ( v k ) v k d x λ Ω f ( x , h ( v k ) ) h ( v k ) v k d x C k ( 1 + 2 v k ) .

In virtue of Lemma 6- ( h 4 ) , from (4.37), we obtain

(4.38) Ω v k 2 d x λ 2 Ω α ( x ) h ( v k ) 1 q d x λ 2 Ω f ( x , h ( v k ) ) h ( v k ) d x C k ( 1 + 2 v k ) .

We claim that

(4.39) limsup k + v k < + .

For 0 < q < 1 , using (4.36), (4.38), ( f 3 ) , Lemma 6- ( h 5 ) and the Sobolev embedding, as k + , it follows that

(4.40) γ 0 + 4 C τ v k + o ( 1 ) 1 2 2 τ v k 2 λ 1 1 q 1 τ Ω α ( x ) h ( v k ) 1 q d x λ τ Ω [ τ F ( x , h ( v k ) ) f ( x , h ( v k ) ) h ( v k ) ] d x 1 2 2 τ v k 2 λ C ( q , Ω ) 1 1 q 1 τ α v k 1 q d x .

Since τ > 4 and 2 > 1 q > 0 , the last relation yields that { v k } is bounded in H 0 1 ( Ω ) , and hence, (4.39) holds.

Next, for q = 1 , using (4.40) and following the arguments in similar way as in (3.56) and (3.57), we obtain (4.39).

Finally, for the case 1 < q < 3 , again using (4.40) and the exact arguments in (3.59) used for estimating the singular term, as k + , we obtain

1 2 2 τ v k 2 ( 1 + C ) v k + γ 0 + o ( 1 ) ,

which gives (4.39). Therefore, there exists a v λ H 0 1 ( Ω ) such that v k v λ weakly in H 0 1 ( Ω ) and pointwise a.e. in Ω as k + . Arguing similarly as in the proof of Proposition 20, we infer that v λ is a weak solution to (2.7) and v λ H 0 1 ( Ω ) C φ q + ( Ω ) .

Claim: v λ w λ .

To establish this claim, it is sufficient to prove that v k v λ in H 0 1 ( Ω ) as k + . For that, we need to establish

(4.41) Ω f ( x , h ( v k ) ) h ( v k ) v k d x Ω f ( x , h ( v λ ) ) h ( v λ ) v λ d x as k + .

Now borrowing the similar arguments as in (4.17), we have

(4.42) Ω α ( x ) h ( v k ) 1 q d x Ω α ( x ) h ( v λ ) 1 q d x as k + .

Next, we show that

(4.43) Ω F ( x , h ( v k ) ) d x Ω F ( x , h ( v λ ) ) d x as k + .

Since { v k } is bounded in H 0 1 ( Ω ) , from (4.38), it follows that

(4.44) limsup k + Ω f ( x , h ( v k ) ) h ( v k ) v k d x < + .

Now using (4.44) and Lemma 6- ( h 8 ) , for some large N 1 , we obtain

Ω { x : h ( v k ) ( x ) > N } f ( x , h ( v k ) ) d x 1 N Ω { x : h ( v k ) ( x ) > N } f ( x , h ( v k ) ) h ( v k ) d x 2 N Ω { x : h ( v k ) ( x ) > N } 2 f ( x , h ( v k ) ) h ( v k ) v k = O 1 N .

The last relation, combining with the Lebesgue dominated convergence theorem, yields

(4.45) Ω f ( x , h ( v k ) ) d x = Ω { x : h ( v k ) ( x ) N } f ( x , h ( v k ) ) d x + Ω { x : h ( v k ) ( x ) > N } f ( x , h ( v k ) ) d x = Ω { x : h ( v k ) ( x ) N } f ( x , h ( v k ) ) d x + O 1 N Ω f ( x , h ( v λ ) ) d x ,

as k + and N + . Since by ( f 4 ) , F ( x , h ( v k ) ) M 1 ( 1 + f ( x , h ( v k ) ) ) for all k N , using (4.45) and the Lebesgue dominated convergence theorem, (4.43) follows. Therefore, using (4.42) and (4.43) with the weak lower semicontinuity property of the norm, we derive

(4.46) J λ ( v λ ) liminf k + J λ ( v k ) .

Supposing the contrary, let us assume that { v k } does not converge in H 0 1 ( Ω ) strongly. Then (4.42), (4.43), and (4.46) imply that J λ ( v λ ) < γ 0 . By the hypothesis, J λ ( w λ ) J λ ( v λ ) . So, for sufficiently small ε > 0 , by Lemma 21, we have

(4.47) ( γ 0 J λ ( v λ ) ) ( 1 + ε ) ( max t [ 0 , 1 ] J ( w λ + t k ) J λ ( w λ ) ) ( 1 + ε ) < π .

Set

ζ 0 lim k Ω α ( x ) h ( v k ) 1 q d x + λ Ω F ( x , h ( v k ) ) d x .

Then,

(4.48) lim k + v k 2 = 2 lim k + J λ ( v k ) + λ Ω α ( x ) h ( v k ) 1 q d x + λ Ω F ( x , h ( v k ) ) d x = 2 ( γ 0 + ζ 0 ) .

Taking into account (4.47) and (4.48), we deduce

( 1 + ε ) v k 2 < 2 π ( γ 0 + ζ 0 ) γ 0 J λ ( v λ ) = 2 π ( γ 0 + ζ 0 ) γ 0 + ζ 0 1 2 v λ 2 = 2 π 1 1 2 v λ 2 γ 0 + ζ 0 1 .

Hence, we choose r > 0 such that ( 1 + ε ) 2 π v k 2 < p r 2 π < ( 1 u λ 2 ) 1 , where u λ v λ ( 2 ( γ 0 + ζ 0 ) ) 1 2 . Also, note that u k v k v k v λ ( 2 ( γ 0 + ζ 0 ) ) 1 2 = u λ weakly in H 0 1 ( Ω ) as k + . Therefore, using Theorem 22, for any ε > 0 , we obtain

sup k Ω exp ( 4 π p u k 2 ) d x < + ,

which together with the fact that p > ( 1 + ε ) 2 π v k 2 yields

(4.49) sup k Ω exp ( 2 ( 1 + ε ) v k 2 ) d x < + .

Now using Lemma 6- ( h 6 ) , ( f 2 ) and (4.49), we obtain

(4.50) Ω f ( x , h ( v k ) ) h ( v k ) v k d x = Ω { x : v k ( x ) > N > 1 } f ( x , h ( v k ) ) h ( v k ) v k d x + Ω { x : v k ( x ) N } f ( x , h ( v k ) ) h ( v k ) v k d x = O Ω { x : v k ( x ) > N } exp 1 + ε 2 h ( v k ) 4 d x + Ω { x : v k ( x ) N } f ( x , h ( v k ) ) h ( v k ) v k d x = O Ω { x : v k ( x ) > N } exp 2 1 + ε 2 v k 2 d x + Ω { x : v k ( x ) N } f ( x , h ( v k ) ) h ( v k ) v k d x = O exp ( ε N 2 ) Ω { x : v k ( x ) > N } exp ( 2 ( 1 + ε ) v k 2 ) d x + Ω { x : v k ( x ) N } f ( x , h ( v k ) ) h ( v k ) v k d x = O ( exp ( ε N 2 ) ) + Ω { x : v k ( x ) N } f ( x , h ( v k ) ) h ( v k ) v k d x Ω f ( x , h ( v λ ) ) h ( v λ ) v λ d x , as k + and N + ,

thanks to the Lebesgue dominated convergence theorem. Thus, we obtain (4.41). Hence, taking into account (4.42) and (4.41) and arguing similarly as in the proof of Proposition 20, we can infer that v k v λ strongly in H 0 1 ( Ω ) as k + and v λ w λ . Hence, the claim is verified. Thus, the proof of the proposition follows.□

Proof of Theorem 5

From Proposition 23, it follows that v λ H 0 1 ( Ω ) C φ q + ( Ω ) is a weak solution to (2.7). Moreover, v λ w λ , where w λ is another weak solution to (2.7). Now by Lemma 6- ( h 1 ) , we have h which is a C function and Lemma 6- ( h 8 ) , ( h 11 ) ensure that h ( s ) behaves like s when s is close to 0. Therefore, h ( v λ ) H 0 1 ( Ω ) C φ q + ( Ω ) forms a weak solution to problem ( P * ) and h ( v λ ) h ( w λ ) , where h ( w λ ) is another solution to problem ( P * ) obtained in Theorem 4. This concludes the proof of Theorem 5.□

Acknowledgement

The authors are thankful to the editor and the reviewers for their valuable suggestions.

  1. Funding information: Reshmi Biswas is funded by Fondecyt Postdocotal Project No. 3230657, Sarika Goyal is funded by SERB under the grant SPG/2022/002068, K. Sreenadh is funded by Janaki & K. A. Iyer Chair Professor grant.

  2. Author contributions: Reshmi Biswas: writing – original draft preparation, Sarika Goyal: writing – original draft preparation, K. Sreenadh – writing, reviewing and editing.

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

  4. Data availability statement: No data were used for the research described in the article.

References

[1] A. Adimurthi, Existence of positive solutions of the semilinear Dirichlet problem with critical growth for the n-Laplacian, Ann. Scuola Norm. Sup. Pisa Cl. Sci. 17 (1990), 393–413. Search in Google Scholar

[2] A. Adimurthi and J. Giacomoni, Multiplicity of positive solutions for singular and critical elliptic problem in R2, Commun. Contemp. Math. 8 (2006), 621–656. 10.1142/S0219199706002222Search in Google Scholar

[3] A. Ambrosetti, H. Brézis, and G. Cerami, Combined effect of concave and convex nonlinearities in some elliptic problems, J. Funct. Anal. 122 (1994), 519–543. 10.1006/jfan.1994.1078Search in Google Scholar

[4] M. Badiale and G. Tarantello, Existence and multiplicity for elliptic problems with critical growths and discontinuous nonlinearities, Nonlinear Anal. 29 (1997), 639–677. 10.1016/S0362-546X(96)00071-5Search in Google Scholar

[5] K. Bal, P. Garain, I. Mandal, and K. Sreenadh, Multiplicity result to a singular quasilinear Schrödinger equation, J. Math. Anal. Appl. 497 (2021), 124904. 10.1016/j.jmaa.2020.124904Search in Google Scholar

[6] F. Bass and N. N. Nasanov, Nonlinear electromagnetic-spin waves, Phys. Rep. 189 (1990), 165–223. 10.1016/0370-1573(90)90093-HSearch in Google Scholar

[7] H. Brézis and L. Nirenberg, H1 versus C1 local minimizers, C. R. Acad. Sci. Paris Sér. I Math. 317 (1983), no. 5, 465–472. Search in Google Scholar

[8] M. Colin and L. Jeanjean, Solutions for a quasilinear Schrödinger equation: A dual approach, Nonlinear Anal. 56 (2004), 213–226. 10.1016/j.na.2003.09.008Search in Google Scholar

[9] M. G. Crandall, P. H. Rabinowitz, and L. Tartar, On a Dirichlet problem with a singular nonlinearity, Comm. Partial Differential Equations 2 (1977), 193–222. 10.1080/03605307708820029Search in Google Scholar

[10] D. G. de Figueiredo, O. H. Miyagaki, and B. Ruf, Elliptic equations in R2 with nonlinearities in the critical growth range, Calc. Var. Partial Differential Equations 3 (1995), no. 2, 139–153. 10.1007/BF01205003Search in Google Scholar

[11] J. M. do Ó, O. H. Miyagaki, and S. H. M. Soares, Soliton solutions for quasilinear Schrödinger equations: The critical exponential case, Nonlinear Anal. 67 (2007), 3357–3372. 10.1016/j.na.2006.10.018Search in Google Scholar

[12] J. M. do Ó and U. Severo, Solitary waves for a class of quasilinear Schrödinger equations in dimension two, Calc. Var. Partial Differential Equations 38 (2010), 275–315. 10.1007/s00526-009-0286-6Search in Google Scholar

[13] J. M. do Ó, O. H. Miyagaki, and S. H. M. Soares, Soliton solutions for quasilinear Schrödinger equations with critical growth, J. Differential Equations 248 (2010), 722–744. 10.1016/j.jde.2009.11.030Search in Google Scholar

[14] R. Dhanya, S Prashanth, K. Sreenadh, and S. Tiwari, Critical growth elliptic problem in R2 with singular discontinuous nonlinearities, Adv. Differ. Equ. 19 (2014), 409–440. 10.57262/ade/1396558057Search in Google Scholar

[15] R. Dhanya, S. Prashanth, S. Tiwari, and K. Sreenadh, Elliptic problems in RN with critical and singular discontinuous nonlinearities, Complex Var. Elliptic Equ. 61 (2016), 1668–1688. 10.1080/17476933.2016.1198787Search in Google Scholar

[16] P. Drabek, A. Kufner, and F. Nicolosi, Quasilinear Elliptic Equations with Degenerations and Singularities. Walter de Gruyter, Berlin, 2011. Search in Google Scholar

[17] J. Giacomoni and K. Saoudi, Multiplicity of positive solutions for a singular and critical problem, Nonlinear Anal. 71 (2009), 4060–4077. 10.1016/j.na.2009.02.087Search in Google Scholar

[18] R. W. Hasse, A general method for the solution of nonlinear soliton and kink Schrödinger equation, Z. Phys. B. 37 (1980), 83–87. 10.1007/BF01325508Search in Google Scholar

[19] Y. Haitao, Multiplicity and asymptotic behaviour of positive solutions for a singular semilinear elliptic problem, J. Differential Equations 189 (2003), 487–512. 10.1016/S0022-0396(02)00098-0Search in Google Scholar

[20] N. Hirano, C. Saccon, and N. Shioji, Existence of multiple positive solutions for singular elliptic problems with concave and convex nonlinearities, Adv. Differential Equations 9 (2004), 197–220. 10.57262/ade/1355867973Search in Google Scholar

[21] Y. He and G. B. Li, Concentrating soliton solutions for quasilinear Schrödinger equations involving critical Soblev exponents, Discrete Contin. Dyn. Syst. 36 (2016), 73–762. 10.3934/dcds.2016.36.731Search in Google Scholar

[22] L. Jeanjean, T. J. Luo, and Z. Q. Wang, Multiple normalized solutions for quasi-linear Schrödinger equations, J. Differential Equations 259 (2015), 3894–3928. 10.1016/j.jde.2015.05.008Search in Google Scholar

[23] S. Kurihara, Large-amplitude quasi-solitons in superfluids films, J. Phys. Soc. Japan 50 (1981), 3262–3267. 10.1143/JPSJ.50.3262Search in Google Scholar

[24] P. L. Lions, The concentration compactness principle in the calculus of variations part-I, Rev. Mat. Iberoamericana 1 (1985), 185–201. 10.4171/rmi/6Search in Google Scholar

[25] Q. Liu, X. Q. Liu, and Z. Q. Wang, Multiple sign-changing solutions for quasilinear elliptic equations via perturbation method, Comm. Partial Differential Equations 39 (2014), 2216–2239. 10.1080/03605302.2014.942738Search in Google Scholar

[26] J. Q. Liu, Y. Q. Wang, and Z. Q. Wang, Solutions for quasilinear Schrödinger equations via the Nehari method, Comm. Partial Differential Equations 29 (2004), 879–901. 10.1081/PDE-120037335Search in Google Scholar

[27] J. Q. Liu and Z. Q. Wang, Soliton solutions for quasilinear Schrödinger equations, I, Proc. Amer. Math. Soc. 131 (2003), 441–448. 10.1090/S0002-9939-02-06783-7Search in Google Scholar

[28] X. Q. Liu, J. Q. Liu, and Z. Q. Wang, Quasilinear elliptic equations with critical growth via perturbation method, J. Differential Equations 254 (2013), 102–124. 10.1016/j.jde.2012.09.006Search in Google Scholar

[29] J. Moser, A sharp form of an inequality by N. Trudinger, Indiana Univ. Math. J. 20 (1971), 1077–1092. 10.1512/iumj.1971.20.20101Search in Google Scholar

[30] J. M. B. do Ó and A. Moameni, Solutions for singular quasilinear Schrödinger equations with one parameter, Commun. Pure Appl. Anal. 9 (2010), no. 4, 1011–1023. 10.3934/cpaa.2010.9.1011Search in Google Scholar

[31] M. Poppenberg, K. Schmitt, and Z. Q. Wang, On the existence of soliton solutions to quasilinear Schrödinger equations, Calc. Var. Partial Differential Equations 14 (2002), 329–344. 10.1007/s005260100105Search in Google Scholar

[32] B. Ritchie, Relativistic self-focusing and channel formation in laser-plasma interactions, Phys. Rev. E. 50 (1994), 687–689. 10.1103/PhysRevE.50.R687Search in Google Scholar

[33] D. Ruiz and G. Siciliano, Existence of ground states for a modified nonlinear Schrödinger equation, Nonlinearity 23 (2010), 1221–1233. 10.1088/0951-7715/23/5/011Search in Google Scholar

[34] G. Santos, G. M. Figueiredo, and U. B. Severo, Multiple solution for a class of singular quasilinear problems, J. Math. Anal. Appl. 480 (2019), 123405. 10.1016/j.jmaa.2019.123405Search in Google Scholar

[35] C. A. Santos, M. Yang, and J. Zhou, Global multiplicity of solutions for a modified elliptic problem with singular terms, Nonlinearity 34 (2021), 7842–7871. 10.1088/1361-6544/ac2a50Search in Google Scholar

[36] U. Severo, Existence of weak solutions for quasilinear elliptic equatiuons involving the p-Laplacian, Electron. J. Differential Equations 2008 (2008), no. 56, 1–16. Search in Google Scholar

Received: 2023-06-17
Revised: 2024-03-25
Accepted: 2024-05-06
Published Online: 2024-06-08

© 2024 the author(s), 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. Existence of normalized peak solutions for a coupled nonlinear Schrödinger system
  3. Orbital stability of the trains of peaked solitary waves for the modified Camassa-Holm-Novikov equation
  4. Global boundedness in a two-dimensional chemotaxis system with nonlinear diffusion and singular sensitivity
  5. Existence and uniqueness of solution for a singular elliptic differential equation
  6. The bounded variation capacity and Sobolev-type inequalities on Dirichlet spaces
  7. Vanishing and blow-up solutions to a class of nonlinear complex differential equations near the singular point
  8. Existence, uniqueness, localization and minimization property of positive solutions for non-local problems involving discontinuous Kirchhoff functions
  9. Concentration phenomena for a fractional relativistic Schrödinger equation with critical growth
  10. Beyond the classical strong maximum principle: Sign-changing forcing term and flat solutions
  11. Monotonicity of solutions for parabolic equations involving nonlocal Monge-Ampère operator
  12. On a nonlinear Robin problem with an absorption term on the boundary and L1 data
  13. Multiple solutions for the quasilinear Choquard equation with Berestycki-Lions-type nonlinearities
  14. Gradient estimates for a class of higher-order elliptic equations of p-growth over a nonsmooth domain
  15. Global existence and decay estimates of the classical solution to the compressible Navier-Stokes-Smoluchowski equations in ℝ3
  16. Global existence and finite-time blowup for a mixed pseudo-parabolic r(x)-Laplacian equation
  17. Multiple positive solutions for a class of concave-convex Schrödinger-Poisson-Slater equations with critical exponent
  18. A minimization problem with free boundary for p-Laplacian weakly coupled system
  19. Multiplicity of semiclassical solutions for a class of nonlinear Hamiltonian elliptic system
  20. k-convex solutions for multiparameter Dirichlet systems with k-Hessian operator and Lane-Emden type nonlinearities
  21. Infinitely many solutions for Hamiltonian system with critical growth
  22. On optimal control in a nonlinear interface problem described by hemivariational inequalities
  23. Variational–hemivariational system for contaminant convection–reaction–diffusion model of recovered fracturing fluid
  24. Potential and monotone homeomorphisms in Banach spaces
  25. Critical fractional Schrödinger-Poisson systems with lower perturbations: the existence and concentration behavior of ground state solutions
  26. Supercritical Hénon-type equation with a forcing term
  27. Concentration of blow-up solutions for the Gross-Pitaveskii equation
  28. Mountain-pass-type solutions for Schrödinger equations in R2 with unbounded or vanishing potentials and critical exponential growth nonlinearities
  29. Regularity for critical fractional Choquard equation with singular potential and its applications
  30. Double phase anisotropic variational problems involving critical growth
  31. Normalized solutions of NLS equations with mixed nonlocal nonlinearities
  32. Nontrivial solutions for resonance quasilinear elliptic systems
  33. Standing waves for Choquard equation with noncritical rotation
  34. Low regularity conservation laws for Fokas-Lenells equation and Camassa-Holm equation
  35. Modified quasilinear equations with strongly singular and critical exponential nonlinearity
  36. Limit profiles and the existence of bound-states in exterior domains for fractional Choquard equations with critical exponent
  37. Nests of limit cycles in quadratic systems
  38. Regularity of minimizers for double phase functionals of borderline case with variable exponents
  39. Boundedness, existence and uniqueness results for coupled gradient dependent elliptic systems with nonlinear boundary condition
  40. Ground state solutions for magnetic Schrödinger equations with polynomial growth
  41. Nonoccurrence of Lavrentiev gap for a class of functionals with nonstandard growth
  42. Dynamics for wave equations connected in parallel with nonlinear localized damping
  43. Boussinesq's equation for water waves: Asymptotics in Sector I
  44. Optimal global second-order regularity and improved integrability for parabolic equations with variable growth
  45. Normalized solutions of Schrödinger equations involving Moser-Trudinger critical growth
  46. Normalized solutions for the double-phase problem with nonlocal reaction
  47. Normalized solutions for Sobolev critical fractional Schrödinger equation
  48. Well-posedness of Cauchy problem of fractional drift diffusion system in non-critical spaces with power-law nonlinearity
  49. Improved results on planar Klein-Gordon-Maxwell system with critical exponential growth
  50. Existence and multiplicity of solutions for a new p(x)-Kirchhoff equation
  51. Quasiconvex bulk and surface energies: C1,α regularity
  52. Time decay estimates of solutions to a two-phase flow model in the whole space
  53. Bounds for the sum of the first k-eigenvalues of Dirichlet problem with logarithmic order of Klein-Gordon operators
  54. The properties of a new fractional g-Laplacian Monge-Ampère operator and its applications
  55. A fractional profile decomposition and its application to Kirchhoff-type fractional problems with prescribed mass
  56. Cauchy problem for a non-Newtonian filtration equation with slowly decaying volumetric moisture content
  57. Multiplicity of normalized semi-classical states for a class of nonlinear Choquard equations
  58. The second Yamabe invariant with boundary
  59. Peaked solitary waves and shock waves of the Degasperis-Procesi-Kadomtsev-Petviashvili equation
  60. Normalized solutions for the Choquard equations with critical nonlinearities
  61. Boundedness and long-time behavior in a parabolic-elliptic system arising from biological transport networks
  62. Initial boundary value problem and exponential stability for the planar magnetohydrodynamics equations with temperature-dependent viscosity
  63. Nonexistence of mean curvature flow solitons with polynomial volume growth immersed in certain semi-Riemannian warped products
  64. Analysis of a vector-borne disease model with vector-bias mechanism in advective heterogeneous environment
  65. Normalized solutions for the Kirchhoff equation with combined nonlinearities in ℝ4
  66. Nonexistence of the compressible Euler equations with space-dependent damping in high dimensions
  67. Multiplicity of solutions for a nonhomogeneous quasilinear elliptic equation with concave-convex nonlinearities
  68. On semilinear inequalities involving the Dunkl Laplacian and an inverse-square potential outside a ball
  69. Besov regularity for the elliptic p-harmonic equations in the non-quadratic case
  70. Uniqueness and nondegeneracy of ground states for Δ u + u = ( I α u 2 ) u in R 3 when α is close to 2
  71. Existence of solutions for a class of quasilinear Schrödinger equations with Choquard-type nonlinearity
  72. Weighted Hardy-Adams inequality on unit ball of any even dimension
  73. Existence and regularity for a p-Laplacian problem in ℝN with singular, convective, and critical reaction
  74. Temporal periodic solutions of non-isentropic compressible Euler equations with geometric effects
  75. Nodal solutions for a zero-mass Chern-Simons-Schrödinger equation
  76. The modified quasi-boundary-value method for an ill-posed generalized elliptic problem
  77. Nonlocal heat equations with generalized fractional Laplacian
  78. Choquard equations with recurrent potentials
  79. Local and global well-posedness of the Maxwell-Bloch system of equations with inhomogeneous broadening
  80. The Riemann problem for two-layer shallow water equations with bottom topography
  81. Global stability and asymptotic profiles of a partially degenerate reaction diffusion Cholera model with asymptomatic individuals
  82. On Kirchhoff-Schrödinger-Poisson-type systems with singular and critical nonlinearity
  83. Energy-variational solutions for viscoelastic fluid models
  84. Stability on 3D Boussinesq system with mixed partial dissipation
  85. Existence and multiplicity results for non-autonomous second-order Hamiltonian systems
  86. Erratum
  87. Erratum to “Normalized solutions for the Kirchhoff equation with combined nonlinearities in ℝ4
Downloaded on 14.9.2025 from https://www.degruyterbrill.com/document/doi/10.1515/anona-2024-0019/html
Scroll to top button