Next Article in Journal
New Results Involving Riemann Zeta Function Using Its Distributional Representation
Previous Article in Journal
A New Adaptive Robust Sliding Mode Control Approach for Nonlinear Singular Fractional-Order Systems
 
 
Font Type:
Arial Georgia Verdana
Font Size:
Aa Aa Aa
Line Spacing:
Column Width:
Background:
Article

(p(x),q(x))-Kirchhoff-Type Problems Involving Logarithmic Nonlinearity with Variable Exponent and Convection Term

1
College of Science, Hohai University, Nanjing 210098, China
2
College of Science, Zhongyuan University of Technology, Zhengzhou 450007, China
3
College of Primary Education, Fuyang Preschool Teachers College, Fuyang 236015, China
*
Author to whom correspondence should be addressed.
Fractal Fract. 2022, 6(5), 255; https://doi.org/10.3390/fractalfract6050255
Submission received: 29 March 2022 / Revised: 30 April 2022 / Accepted: 3 May 2022 / Published: 6 May 2022

Abstract

:
In the present article, we study a class of Kirchhoff-type equations driven by the ( p ( x ) , q ( x ) ) -Laplacian. Due to the lack of a variational structure, ellipticity, and monotonicity, the well-known variational methods are not applicable. With the help of the Galerkin method and Brezis theorem, we obtain the existence of finite-dimensional approximate solutions and weak solutions. One of the main difficulties and innovations of the present article is that we consider competing ( p ( x ) , q ( x ) ) -Laplacian, convective terms, and logarithmic nonlinearity with variable exponents, another one is the weaker assumptions on nonlocal term M υ ( x ) and nonlinear term g.

1. Introduction

The purpose of the present article is to investigate the following ( p ( x ) , q ( x ) ) -Kirchhoff-type equations involving logarithmic nonlinearity and convection terms:
M p ( x ) δ p ( x ) ( η ) Δ p ( x ) η μ M q ( x ) δ q ( x ) ( η ) Δ q ( x ) η = λ | η | r ( x ) 2 η ln | η | + g ( x , η , η ) , in Ω , η | Ω = 0 ,
where r ( x ) C + ( Ω ) , μ , λ are real parameters, and Ω is an open bounded domain in R N with a smooth boundary.
Here, Δ γ ( x ) is a γ ( x ) -Laplace operator, defined by
Δ γ ( x ) η = d i v ( | η | γ ( x ) 2 η ) = i = 1 N | η | γ ( x ) 2 η x i , γ ( x ) p ( x ) , q ( x ) ,
for all x Ω and η C 0 ( R N ) , and denote
δ s ( x ) ( η ) = Ω 1 s ( x ) | η | s ( x ) d x , s ( x ) p ( x ) , q ( x ) .
From now on, we briefly state some major features of problem (1). One of the significant characteristics of the problem (1) is the presence of double non-local Kirchhoff terms, which were introduced in [1] as follows:
ρ 2 η ( x ) t 2 p 0 h + E 2 L 0 L η ( x ) t 2 d x 2 η ( x ) x 2 = 0 ,
where parameters ρ , p 0 , h , E , and L are real positive constants. Equation (4) is a nonlocal problem, which contains a nonlocal coefficient p 0 h + E 2 L 0 L η ( x ) t 2 d x , and has a wide range of applications and research in physical systems, such as non-homogeneous Kirchhoff-type equations in R N [2], nonlocal Kirchhoff equations of elliptic type [3], Kirchhoff–Schrödinger type equations [4], p ( x ) -Laplacian Dirichlet problem [5,6], Kirchhoff–Choquard equations involving variable-order [7,8], fractional p ( · ) -Kirchhoff type problem in R N [9], Kirchhoff-type equations involving the fractional p 1 ( x ) & p 2 ( x ) -Laplace operator [10], fractional p ( x , · ) -Kirchhoff-type problems in R N [11], and fractional Sobolev space and applications to nonlocal variational problems [12]. For more Kirchhoff-type problems, we also mention that [13] studied a class of Kirchhoff nonlocal fractional equations and obtained the existence of three solutions, Ref. [14] discussed a class of p-Kirchhoff equations via the fountain theorem and dual fountain theorem, and Ref. [15] researched the existence of non-negative solutions for a Kirchhoff type problem driven by a non-local integro-differential operator.
Let M i : R 0 + R + and p ( x ) , q ( x ) : R N ( 1 , + ) be continuous functions, which satisfy the following conditions:
H m :
There are some constants m υ ( x ) = m υ ( x ) ( ι ) > 0 ( υ ( x ) { p ( x ) , q ( x ) } ) for all ι > 0 such that
M υ ( x ) ( t ) m υ ( x ) , for any t > ι .
H p q :
The conditions that we impose on p ( x ) , q ( x ) are as follows:
1 < p : = inf x Ω ¯ p ( x ) p + : = sup x Ω ¯ p ( x ) < + ,
1 < q : = inf x Ω ¯ q ( x ) q + : = sup x Ω ¯ q ( x ) < + .
Another significant characteristic of the problem (1) is the presence of double operators, which comes from the following system
η t = div [ D η η ] + c ( x , η ) ,
where D η = | η | p 2 + | η | q 2 and c ( x , η ) is a polynomial of η . System (5) had a wide range of applications in the field of physics and related sciences, for example, on the stationary solutions of generalized reaction diffusion equations [16], elliptic problems with critical growth in R N [17], nontrivial solutions to nonlinear elliptic equation in R N [18], and fractional Choquard problems with variable order [19]. The function η in (5) describes a concentration, and the first term corresponds to the diffusion with a (generally non-constant) diffusion coefficient D η , whereas the second one is the reaction and relates to source and loss processes. Typically, in chemical and biological applications, the reaction term c ( x , η ) in (5) has a polynomial form with respect to the concentration η .
When M υ ( x ) = 1 ( υ ( x ) { p ( x ) , q ( x ) } ) and μ = 1 , Chung et al. in [20] devoted to the study of equations involving both p 1 ( x ) -Laplacian and p 2 ( x ) -Laplacian
( Δ ) p 1 ( · ) s η ( x ) + ( Δ ) p 2 ( · ) s η ( x ) + | η ( x ) | q ( x ) 2 η ( x ) = λ V 1 ( x ) | η ( x ) | r 1 ( x ) 2 η ( x ) λ V 2 ( x ) | η ( x ) | r 2 ( x ) 2 η ( x ) , x Ω , η ( x ) = 0 , x Ω ,
where p 1 , p 2 , q , r 1 , and r 2 are different continuous functions, while V 1 , V 2 are suitable weights. Equation (6) considered the local double Laplace operators, whose results differed from those of the single Laplace operator.
When M υ ( x ) = 1 ( υ ( x ) { p ( x ) , q ( x ) } ) and μ = 1 , we mention that Motreanu in [21] considered Dirichlet problems with competing operators
Δ p η ( x ) + Δ q η ( x ) = g ( x , η ( x ) , η ( x ) ) , in Ω , η ( x ) = 0 , on Ω ,
where Ω R N is a bounded domain. Equation (7) includes the sum Δ p + Δ q of the negative p-Laplacian Δ p and of the q-Laplacian Δ q , due to competition between Δ p and Δ q , and the operator Δ p + Δ q has a different behavior in comparison to the operator Δ p + Δ q . Moreover, the ellipticity and monotonicity property of the operator Δ p + Δ q are lost.
The third significant characteristic of the problem (1) is the presence of convection term g ( x , η , η ) , depending on the function η and on its gradient η , which makes the problem (1) non-variational, plays an important role in science and technology fields, and is widely used to describe physical phenomena. For example, due to convection and diffusion processes, particles or energy are converted and transferred inside physical systems. For the work related to this topic, we cite the interesting work [21,22,23,24] and their references.
The work in [25] focused on the p-Kirchhoff-type equations with gradient dependence in the reaction that is
M Ω | η ( x ) | p d x Δ p η ( x ) = g ( x , η ( x ) , η ( x ) ) , in Ω , η ( x ) = 0 , on Ω ,
where Ω R N is a bounded domain with a smooth boundary. The existence of solutions for the problem (8) was obtained by utilizing Galerkin’s approach.
One more reference on convection is Vetro [26], which was devoted to the study of the following p ( x ) -Kirchhoff-type equation:
Δ p ( x ) K η ( x ) = g ( x , η ( x ) , η ( x ) ) , in Ω , η | Ω = 0 .
The existence of weak solutions and generalized solutions for the problem (9) with gradient dependence was obtained via applying a topological method.
The nonlinearity g : Ω × R × R N R is a Carathéodory function, satisfying
H g 1 :
There exist some constants c < 1 , d > 0 and a function α [ 1 , p ) such that
g ( x , ϖ , ν ) ϖ c | ν | p ( x ) + d ( | ϖ | α ( x ) + 1 ) , for a . e . x Ω and all ( ϖ , ν ) R × R N .
H g 2 :
There exists a positive function ϕ ( x ) L p ( x ) ( Ω ) and some positive constants a and b such that
| g ( x , ϖ , ν ) | h ( x ) + a | ϖ | ϕ ( x ) + b | ν | ψ ( x ) p ( x ) , for a . e . x Ω and all ( ϖ , ν ) R × R N .
where ϕ ( x ) C ( Ω ¯ ) , ψ ( x ) C ( Ω ¯ ) such that 0 < ϕ ϕ + < p 1 , ψ p + < p 1 .
The last significant characteristic of the problem (1) is the presence of logarithmic nonlinearity. The interest in studying problems with logarithmic nonlinearity is motivated not only by the purpose of describing mathematical and physical phenomena but also by their application in realistic models. For instance, in the biological population, we use the function η ( x ) to represent the density of the population, and the logarithmic nonlinear term | η | r ( x ) 2 η ln | η | to denote external influencing factors.
Many scholars make efforts to investigate logarithmic nonlinearity, and, indeed, some important results were obtained; for example, see [27,28,29,30]. Peculiarly, Xiang et al. in [31] considered the following equation:
M ( [ η ] s , p p ) ( Δ ) p s η = h ( x ) | η | θ p 2 η ln | η | + λ | η | q 2 η , x Ω , η ( x ) = 0 , x R N \ Ω ,
where M ( [ η ] s , p p ) = [ η ] s , p p ( θ 1 ) and h ( x ) is a sign-changing function. The existence of least energy solutions (10) was obtained by utilizing the Nehari manifold method.
Until now, there have been few papers to handle the equations involving logarithmic nonlinearity with variable exponents. Recently, Boudjeriou in [32] studied the following initial value problem:
η t ( x ) Δ p ( x ) η ( x ) = | η ( x ) | s ( x ) 2 η ( x ) log ( | η ( x ) | ) , in Ω , t > 0 , η ( x ) = 0 , in Ω , t > 0 , η ( x , 0 ) = η 0 ( x ) , in Ω .
The weak solutions of Equation (11) were obtained under suitable conditions. Moreover, Zeng et al. in [33] were devoted to the study of equations with logarithmic nonlinearity and variable exponents by applying the logarithmic inequality.
Motivated by the previous and aforementioned cited works, there is no result for the Kirchhoff-type equations, which combine with variable exponents, competing ( p ( x ) , q ( x ) ) -Laplacian, logarithmic nonlinearity, and convection terms; therefore, we will investigate the existence of solutions for these kinds of equations, which are different from the work of [25,26,31,32]. Under weaker conditions on the nonlocal term M υ ( x ) and the nonlinearities g, we prove the existence of finite-dimensional approximate solutions by using the Galerkin method and obtain the existence of weak solutions with the help of the Brezis theorem. One of the main difficulties and innovations of the present article is that we consider competing ( p ( x ) , q ( x ) ) -Laplacian, convective term, and logarithmic nonlinearity with variable exponents; another one is the weaker assumptions on nonlocal term M υ ( x ) and nonlinear term g.
The present article is divided into six sections. Aside from Section 1, we have Section 2 given some preliminary notions and results about Lebesgue spaces and Sobolev spaces, and proved some technical lemmas. The finite-dimensional approximate solutions are obtained in Section 3. Section 4 discusses the existence of weak solutions by applying the Brezis theorem, and we give two examples of application of our theorems in Section 5 and present conclusions in Section 6.

2. Preliminary Results and Some Technical Lemmas

In this section, we briefly review some basic knowledge of generalized Lebesgue spaces and Sobolev spaces with variable exponents, and then give two technical lemmas.
For any real-valued function H defined on a domain Ω , we denote
C + ( Ω ¯ ) : = H ( x ) C ( Ω ¯ , R ) : 1 < H : = inf x Ω ¯ H ( x ) H ( x ) H + : = sup x Ω ¯ H ( x ) < + .
Letting ϑ ( x ) C + ( Ω ¯ ) , we define the generalized Lebesgue spaces with variable exponents as
L ϑ ( x ) ( Ω ) : = η : η is a measurable function and Ω | η | ϑ ( x ) d x < ,
provided with the Luxemburg norm
η ϑ ( x ) = η L ϑ ( x ) ( Ω ) : = inf χ > 0 : Ω η χ ϑ ( x ) d x 1 ;
then, ( L ϑ ( x ) ( Ω ) , · ϑ ( x ) ) is a separable and reflexive Banach spaces; see [34,35].
Lemma 1
(see [35]). Let ϑ ( x ) be the conjugate exponent of ϑ ˜ ( x ) C + ( Ω ¯ ) , that is,
1 ϑ ( x ) + 1 ϑ ˜ ( x ) = 1 , f o r a l l x Ω .
Assume that η L ϑ ( x ) ( Ω ) and ξ L ϑ ˜ ( x ) ( Ω ) ; then,
| Ω η ξ d x | 1 ϑ + 1 ϑ ˜ η ϑ ( x ) ξ ϑ ˜ ( x ) 2 η ϑ ( x ) ξ ϑ ˜ ( x ) .
Proposition 1
(see [36]). The modular of L ϑ ( x ) ( Ω ) , which is the mapping ρ ϑ ( x ) : L ϑ ( x ) ( Ω ) R , is defined by
ρ ϑ ( x ) ( η ) : = Ω | η | ϑ ( x ) d x .
Assume that η n , η L ϑ ( x ) ( Ω ) ; then, the following properties hold:
( 1 ) η ϑ ( x ) > 1 η ϑ ( x ) ϑ ρ ϑ ( x ) ( η ) η ϑ ( x ) ϑ + , ( 2 ) η ϑ ( x ) < 1 η ϑ ( x ) ϑ + ρ ϑ ( x ) ( η ) η ϑ ( x ) ϑ , ( 3 ) η ϑ ( x ) < 1 ( resp . = 1 , > 1 ) ρ ϑ ( x ) ( η ) < 1 ( resp . = 1 , > 1 ) , ( 4 ) η n ϑ ( x ) 0 ( resp . + ) ρ ϑ ( x ) ( η n ) 0 ( resp . + ) , ( 5 ) lim n | η n η | ϑ ( x ) = 0 lim n ρ ϑ ( x ) ( η n η ) = 0 .
Now, we consider the following generalized Sobolev spaces with variable exponents
W = W 1 , ϑ ( x ) ( Ω ) : = η L ϑ ( x ) ( Ω ) : | η | L ϑ ( x ) ( Ω ) ,
endowed with the norm
η W : = η ϑ ( x ) + η ϑ ( x ) ;
then, ( W , · W ) is a separable and reflexive Banach spaces, see [34].
Lemma 2
(see [34]). Assume that γ ( x ) C + ( Ω ¯ ) fulfills
1 < γ = min x Ω ¯ γ ( x ) γ ( x ) < ϑ * ( x ) = N ϑ ( x ) N ϑ ( x ) , f o r a n y x Ω ¯ .
Then, there exists C γ = C γ ( N , ϑ , γ , Ω ) > 0 such that
η γ ( x ) C γ η W ,
for any η W . Moreover, the embedding W L γ ( x ) ( Ω ) is compact.
Let W 0 denote the closure of C 0 ( Ω ) in W with respect to the norm η W 0 , which is the subspace of W. Thus, the spaces ( W 0 , · W 0 ) are also separable and reflexive Banach spaces.
Remark 1.
According to the Poincaré inequality, we know that η ϑ ( x ) and η W 0 are equivalent norms in W 0 . From now on, we work on W 0 and replace η W 0 by η ϑ ( x ) , that is,
η W 0 = η ϑ ( x ) , for all η W 0 .
Remark 2.
To simplify the presentation, we will denote the norm of W 0 by · instead of · W 0 . W 0 * denotes the dual space of W 0 .
Our technique of proof is based on Galerkin methods together with the fixed point theorem, whose proof may be found in Lions [37].
Lemma 3.
Let W 0 be a finite dimensional space with the norm · and let G : W 0 W 0 * be a continuous mapping. Assume that there is a constant R > 0 such that
G ( η ) , η 0 , for all η W 0 with η = R ,
then η W 0 exists with η R satisfying G ( η ) = 0 .
The following two Lemmas provide a useful growth estimate, related to logarithmic nonlinear terms, which play an important role during our proof process.
Lemma 4.
Assume that h ( x ) C + ( Ω ¯ ) ; then, we have the following estimate:
ln t 1 e h ( x ) t h ( x ) 1 e h t h ( x ) , for all t [ 1 , + ) .
Proof. 
Let h ( x ) C + ( Ω ¯ ) , and we construct the following function:
f ( t ) = ln t 1 e h ( x ) t h ( x ) , for all t [ 1 , + ) .
With respect to t, just by taking a simple derivative, we deduce
f ( t ) = 1 t 1 e t h ( x ) 1 , for all t [ 1 , + ) ,
and let f ( t ) = 0 ; then, t * = e h 1 ( x ) . It is obvious that t * is the unique maximum point of the function f ( t ) , so f ( t ) f ( t * ) = 0 for all t [ 1 , + ) . Therefore, based on the above discussion, we can obtain the stated conclusion. □
Lemma 5.
Assume that, for all η W 0 and h ( x ) , r ( x ) C + ( Ω ¯ ) , then the following inequality holds:
Ω | η | r ( x ) ln | η | d x C Ω 1 | Ω | + 1 e h max C h + + r + η h + + r + , C h + r η h + r ,
where C Ω 1 , C h + + r + , C h + r are some positive constants and h ( x ) + r ( x ) h + + r + < 2 p < p * ( x ) = N p ( x ) N p ( x ) .
Proof. 
Let Ω 1 = { x Ω : | η ( x ) | 1 } and Ω 2 = { x Ω : | η ( x ) | 1 } ; then,
Ω | η | r ( x ) ln | η | d x = Ω 1 | η | r ( x ) ln | η | d x + Ω 2 | η | r ( x ) ln | η | d x .
Since | η ( x ) | 1 , there exist M r 1 > 0 and M r 2 > 0 such that | η | r ( x ) < M r 1 and ln | η | < M r 2 . By a simple calculation, we obtain
Ω 1 | η | r ( x ) ln | η | d x < C Ω 1 | Ω | ,
where | Ω | denotes the Lebesgue measure of Ω and C Ω 1 > 0 . Using Lemma 4 with h ( x ) + r ( x ) h + + r + < p * ( x ) , we deduce
Ω 2 | η | r ( x ) ln | η | d x 1 e h Ω 2 | η | r ( x ) + h ( x ) d x 1 e h max η h ( x ) + r ( x ) h + + r + , η h ( x ) + r ( x ) h + r ,
in view of Lemma 2, and there exist some constants C h + + r + > 0 and C h + r > 0 such that
Ω 2 | η | r ( x ) ln | η | d x 1 e h max C h + + r + η h + + r + , C h + p η h + r .
It follows from (12) and (13) that
Ω | η | r ( x ) ln | η | d x C Ω 1 | Ω | + 1 e h max C h + + r + η h + + r + , C h + r η h + r .
This yields the stated conclusion. □

3. Finite Dimensional Approximate Solutions

Since W 0 is a reflexive and separable Banach space, see [34], and there exists an orthonormal basis { e 1 , . . . , e n , . . . } in W 0 , such that
W 0 = s p a n { e 1 , . . . , e n } ¯ .
Define X n = s p a n { e 1 , . . . , e n } , which means a sequence of vector X n subspaces of W 0 , satisfying
d i m ( X n ) < for all n 1 , X n X n + 1 for all n 1 , and n = 1 X n ¯ = W 0 .
It is known that X n and R N are isomorphic and, for η R N , we have a unique ξ X n by the identification
η Σ i = 1 N ξ i e i = ξ , η = | ξ | ,
where | · | is the Euclidian norm in R N .
Theorem 1.
Assume that conditions H m , H p q , and H g 1 are satisfied; then,
  • if 2 p > p + and p > α + , problem (1) admits a approximate solution for all μ 0 and λ 0 ,
  • if 2 p > p + , p > q + and p > α + , the problem (1) admits a approximate solution for all μ < 0 and λ 0 ,
  • if 2 q > p + and q > α + , problem (1) admits a approximate solution for all μ 0 and λ 0 ,
  • if 2 p > r + + h + and p > α + , problem (1) admits a approximate solution for all μ 0 and λ > 0 ,
  • if 2 p > r + + h + , p > q + and p > α + , problem (1) admits a approximate solution for all μ < 0 and λ > 0 ,
  • if 2 q > r + + h + and q > α + , problem (1) admits a approximate solution for all μ 0 and λ > 0 ,
that is, for all n 1 and φ X n , there exists η n X n such that
M p ( x ) δ p ( x ) ( η n ) η n , φ p ( x ) + μ M q ( x ) δ q ( x ) ( η n ) η n , φ q ( x ) = λ Ω | η n | r ( x ) 2 η n ln | η n | φ d x + Ω g ( x , η n , η n ) φ d x .
Proof. 
For all η X n , we consider the mapping G = ( G 1 , G 2 , . . . , G N ) : R N R by
G i = M p ( x ) δ p ( x ) ( η ) η , e i p ( x ) + μ M q ( x ) δ q ( x ) ( η ) η , e i q ( x ) λ Ω | η | r ( x ) 2 η ln | η | e i d x Ω g ( x , η , η ) e i d x .
The following work shows that, for each n 1 , problem (1) has an approximate solution η n in X n , namely
M p ( x ) δ p ( x ) ( η n ) η n , e i p ( x ) + μ M q ( x ) δ q ( x ) ( η n ) η n , e i q ( x ) = λ Ω | η n | r ( x ) 2 η n ln | η n | e i d x + Ω g ( x , η n , η n ) e i d x .
For η X n , we have
G , η = M p ( x ) δ p ( x ) ( η ) η , η p ( x ) + μ M q ( x ) δ q ( x ) ( η ) η , η q ( x ) Ω | η | r ( x ) ln | η | d x Ω g ( x , η , η ) η d x , 1 p + Ω | η | p ( x ) d x 2 + μ q + Ω | η | p ( x ) d x 2 λ Ω | η | r ( x ) ln | η | d x Ω g ( x , η , η ) η d x .
From H g 1 and Lemma 5, we have the following estimate:
G , η 1 p + Ω | η | p ( x ) d x 2 + μ q + Ω | η | q ( x ) d x 2 λ e h max C h + + r + η h + + r + , C h + r η h + r λ C Ω 1 | Ω | c Ω | η | p ( x ) d x d Ω ( | η | α ( x ) + 1 ) d x .
According to Remark 1 and Lemma 2, there exist some positive constants C α + and C α , such that
G , η 1 p + min η 2 p + , η 2 p + μ q + min η 2 q + , η 2 q λ e h max C h + + r + η h + + r + , C h + r η h + r d max C α + η α + , C α η α c max η p + , η p ( λ C Ω 1 + d ) | Ω | .
If η > 1 , then
G , η 1 p + η 2 p + μ q + min η 2 q + , η 2 q λ C h + + r + e h η h + + r + d C α η α + c η p + ( λ C Ω 1 + d ) | Ω | .
Combined with the above analysis, we deduce that
Case 1: Utilizing that 2 p > p + and p > α + with μ 0 and λ 0 , there exists a positive constant R, provided at a sufficiently large size, such that
G , η 1 p + η 2 p c η p + d C α η α + d | Ω | 0 ,
for all η X n , with η = R .
Case 2: Utilizing that 2 p > p + , p > q + and p > α + with μ < 0 and λ 0 , there exists a positive constant R, provided at a sufficiently large size, such that
G , η 1 p + η 2 p + μ q + η 2 q + c η p + d C α η α + d | Ω | 0 ,
for all η X n , with η = R .
Case 3: Utilizing that 2 q > p + and q > α + with μ 0 and λ 0 , there exists a positive constant R, provided at a sufficiently large size, such that
G , η μ q + η 2 q c η p + d C α η α + d | Ω | 0 ,
for all η X n , with η = R .
Case 4: Utilizing that 2 p > r + + h + and p > α + with μ 0 and λ > 0 , there exists a positive constant R, provided at a sufficiently large size, such that
G , η 1 p + η 2 p c η p + d C α η α + λ C h + + r + e h η h + + r + ( λ C Ω 1 + d ) | Ω | 0 ,
for all η X n , with η = R .
Case 5: Utilizing that 2 p > r + + h + , p > q + and p > α + with μ < 0 and λ > 0 , there exists a positive constant R, provided at a sufficiently large size, such that
G , η 1 p + η 2 p + μ q + η 2 q + c η p + d C α η α + λ C h + + r + e h η h + + r + ( λ C Ω 1 + d ) | Ω | 0 ,
for all η X n , with η = R .
Case 6: Utilizing that 2 q > r + + h + and q > α + with μ 0 and λ > 0 , there exists a positive constant R, provided at a sufficiently large size, such that
G , η μ q + η 2 q c η p + d C α η α + λ C h + + r + e h η h + + r + ( λ C Ω 1 + d ) | Ω | 0 ,
for all η X n , with η = R .
In the above six cases, G is continuous, so, in view of Lemma 3, problem (1) admits a approximate solution η n in X n W 0 with η n R .
Corollary 1.
Assume that the conditions of Theorem 1 are satisfied, then the sequence { η n } n 1 with η n X n constructed in Theorem 1 is bounded in W 0 .
Proof. 
If η n 1 for all n N , then the sequence { η n } n N is bounded in W 0 .
If η n > 1 for all n N , with η n in place of φ in (14), we have
M p ( x ) δ p ( x ) ( η n ) η n , η n p ( x ) + μ M q ( x ) δ q ( x ) ( η n ) η n , η n q ( x ) = λ Ω | η n | r ( x ) 2 η n ln | η n | η n d x + Ω g ( x , η n , η n ) η n d x .
On the basis of condition H g 1 and Lemma 5, it gives
1 p + Ω | η n | p ( x ) d x 2 + μ q + Ω | η n | q ( x ) d x 2 c Ω | η n | p ( x ) d x + d Ω ( | η n | α ( x ) + 1 ) d x + λ C Ω 1 | Ω | + λ e h max C h + + r + η n h + + r + , C h + r η n h + r .
Case 1: Recalling that 2 p > p + and p > α + with μ 0 and λ 0 , and, by Lemmas 1 and 2, we deduce
1 p + η n 2 p c η n p + + d C α + η n α + + d | Ω | .
Case 2: Recalling that 2 p > p + , p > q + and p > α + with μ < 0 and λ 0 , and by Lemmas 1 and 2, we deduce
1 p + η n 2 p μ q + η n 2 q + + c η n p + + d C α + η n α + + d | Ω | .
Case 3: Recalling that 2 q > p + and q > α + with μ 0 and λ 0 , and by Lemmas 1 and 2, we deduce
μ q + η n 2 q c η n p + + d C α η n α + + d | Ω | .
Case 4: Recalling that 2 p > r + + h + and p > α + with μ 0 and λ > 0 , and by Lemmas 1 and 2, we deduce
1 p + η n 2 p c η n p + + d C α η n α + + λ C h + + r + e h η n h + + r + + ( λ C Ω 1 + d ) | Ω | .
Case 5: Recalling that 2 p > r + + h + , p > q + and p > α + with μ < 0 and λ > 0 , and by Lemmas 1 and 2, we deduce
1 p + η n 2 p μ q + η n 2 q + + c η n p + + d C α η n α + + λ C h + + r + e h η n h + + r + + ( λ C Ω 1 + d ) | Ω | .
Case 6: Recalling that 2 p > r + + h + and q > α + with μ 0 and λ > 0 , and by Lemmas 1 and 2, we deduce
μ q + η n 2 q c η n p + d C α η n α + + λ C h + + r + e h η n h + + r + + ( λ C Ω 1 + d ) | Ω | .
In the above six cases, we conclude that the sequence { η n } n 1 is bounded in W 0 .

4. Existence of Weak Solutions

In this section, our interest is devoted to the existence of weak solutions for problem (1). The following are the main results of this section.
Theorem 2.
Assume that conditions H m , H p q , and H g 2 are satisfied, then, for all μ > 0 ,
  • if 2 p > ϕ + + 1 and 2 p > ( ψ p ) + , problem (1) admits at least one weak solution with λ 0 .
  • if 2 p > h + + r + , 2 p > ϕ + + 1 and 2 p > ( ψ p ) + , problem (1) admits at least one weak solution with λ > 0 .
  • if 2 q > ϕ + + 1 and 2 q > ( ψ p ) + , problem (1) admits at least one weak solution with λ 0 .
  • if 2 q > h + + r + , 2 q > ϕ + + 1 and 2 q > ( ψ p ) + , problem (1) admits at least one weak solution with λ > 0 .
Corollary 2.
Assume that the conditions of Theorem 2 are satisfied; then, the sequence { η n } n 1 with η n X n is bounded in W 0 .
Proof. 
The proof is similar to Corollary 1, which we omit. □
To prove Theorems 2, we use the Brezis theorem for pseudomonotone operators in the separable reflexive space (see (Theorem 27.A [38]). Let us define the operator T : W 0 W 0 * as
T η , φ = M p ( x ) δ p ( x ) ( η ) η , φ p ( x ) + μ M q ( x ) δ q ( x ) ( η ) η , φ q ( x ) λ Ω | η | r ( x ) 2 η ln | η | φ d x Ω g ( x , η , η ) φ d x ,
for all η , φ W 0 .
Lemma 6.
Assume that the conditions of Theorem 2 are satisfied; then, the operator T is bounded.
Proof. 
Let η W 0 be fixed and denote by Φ η the linear functional on W 0 , defined as
Φ η ( φ ) = Ω | η | υ ( x ) 2 η φ d x ,
for any φ W 0 and υ ( x ) { p ( x ) , q ( x ) } . By Hölder inequality,
| Φ η ( φ ) | η φ , for all η , φ W 0 .
Obviously, Φ η ( φ ) is bounded. From the hypothesis H m and Proposition 1, there exist some constants C υ 1 , C υ 2 > 0 such that
0 < C υ 1 M υ ( x ) δ υ ( x ) ( η ) C υ 2 ,
which, together with (16), there exists a constant C υ ( x ) > 0 such that
| M υ ( x ) δ υ ( x ) ( η ) η , φ υ ( x ) | C υ ( x ) .
In fact, by a simple calculation for the logarithmic nonlinear term, we deduce
Ω | η | r ( x ) 2 η ln | η | r + r + 1 d x = Ω 1 | η | r ( x ) 2 η ln | η | r + r + 1 d x + Ω 2 | η | r ( x ) 2 η ln | η | r + r + 1 d x C Ω 1 | Ω | + Ω 2 | η | r ( x ) 2 η ln | η | r + r + 1 d x .
Since r + < p * ( x ) , then, by using the continuous embedding L p * ( x ) ( Ω ) L r + ( Ω ) and combining Lemma 4, we deduce
Ω | η | r ( x ) 2 η ln | η | r + r + 1 d x C Ω 1 | Ω | + Ω | η | r + d x C Ω 1 | Ω | + C Ω 2 η p * ( x ) .
where C Ω 2 > 0 . Notice that the relation (18) implies that
| η | r ( x ) 1 ln | η | L r + r + 1 ( Ω ) C r + r + 1 ,
where C r + r + 1 > 0 . Using the Hölder inequality and taking into account the embeddings, for any φ W 0 with φ 1 ,
| Ω φ | η | r ( x ) 2 η ln | η | d x | φ L r + ( Ω ) | η | r ( x ) 2 η ln | η | L r + r + 1 ( Ω ) C r + r + 1 .
From hypothesis G 1 and Jensen’s inequality, for all η X , we have
Ω | g ( x , η , η ) | p ( x ) d x Ω | h ( x ) | + | a | η | ϕ ( x ) | + | b | η | ψ ( x ) p ( x ) | p ( x ) d x 3 ( q ) + 1 Ω | h ( x ) | p ( x ) d x + Ω | a | η | ϕ ( x ) | p ( x ) d x + Ω | b | η | ψ ( x ) p ( x ) | p ( x ) d x C p Ω | h ( x ) | p ( x ) d x + Ω | η | p ( x ) ϕ ( x ) d x + Ω | η | ψ ( x ) d x ,
where C p = 3 ( q ) + 1 max 1 , a ( p ) , a ( p ) + , b ( p ) , b ( p ) + . It follows from (20) and Proposition 1 that we have
Ω | g ( x , η , η ) | p ( x ) d x C p 3 + | h | p ( p ) + + | η | p ϕ ( p ϕ ) + + η ψ + C p 3 + | h | p ( p ) + + C p ϕ ( p ϕ ) + η p ϕ ( p ϕ ) + + η ψ + .
Hence, invoking Proposition 1, we infer
| g ( x , η , η ) | p 1 + C p [ 3 + | h | p ( p ) + + C p ϕ ( p ϕ ) + η p ϕ ( p ϕ ) + + η ψ + ] 1 ( p ) .
Utilizing Lemma 1and taking into account the embeddings, for all φ W 0 with φ 1 ,
| Ω g ( x , η , η ) φ d x | 2 | g ( x , η , η ) | p | φ | p 2 | g ( x , η , η ) | p .
Thus, it follows from these estimates (17), (19), and (22) that we easily determine the boundedness of T. □
Lemma 7.
Assume that the conditions of Theorem 2 are satisfied; then, the operator T is demicontinuous.
Proof. 
Assuming that η n η in W 0 , we show that T η n T η in W 0 * , that is,
M p ( x ) δ p ( x ) ( η n ) η n , φ p ( x ) + μ M q ( x ) δ q ( x ) ( η n ) η n , φ q ( x ) λ Ω | η n | r ( x ) 2 η n ln | η n | φ d x Ω g ( x , η n , η n ) φ d x M p ( x ) δ p ( x ) ( η ) η , φ p ( x ) + μ M q ( x ) δ q ( x ) ( η ) η , φ q ( x ) λ Ω | η | r ( x ) 2 η ln | η | φ d x Ω g ( x , η , η ) φ d x .
Since η n η in W 0 , up to a subsequence, we have
η n η and η n η , a . e . in Ω .
Thus, we have
| | η n | υ ( x ) 2 η n | p ( x ) η n ( 1 + η n p υ + 2 ) , υ { p ( x ) , q ( x ) } .
which imply that { | η n | υ ( x ) 2 η n } are bounded in L p ( Ω ) .
For υ ( x ) { p ( x ) , q ( x ) } , we obtain
| η n | r ( x ) 2 η n ln | η n | | η | r ( x ) 2 η ln | η | , a . e . in Ω , | η n | υ ( x ) 2 η n | η | υ ( x ) 2 η , a . e . in Ω , g ( x , η n , η n ) g ( x , η , η ) , a . e . in Ω .
Moreover, the boundedness of { η n } in W 0 and (22) imply that { g ( x , η n , η n ) } are bounded in L p ( Ω ) , and (19) implies that { | η n | r ( x ) 2 η n ln | η n | } are bounded in L r ( Ω ) . Thanks to (17), (24), and (26), combined with H m and Proposition 1, we obtain
M υ ( x ) δ υ ( x ) ( η n ) η n , φ υ ( x ) M υ ( x ) δ υ ( x ) ( η ) η , φ υ ( x ) .
Now, we show that the following conclusion holds:
Ω g ( x , η n , η n ) φ d x Ω g ( x , η , η ) φ d x .
Let g ( x , η n , η n ) , g ( x , η , η ) L p ( Ω ) , and
E ( N ) = x : | g ( x , η n , η n ) g ( x , η , η ) | 1 , for all n N .
Since meas ( E ( N ) ) meas ( Ω ) as N , and setting
F N = Ψ N L p ( x ) ( Ω ) : Ψ N 0 a . e . in Ω \ E ( N ) .
First, we prove that F N is dense in L p ( x ) ( Ω ) . Let f L p ( x ) ( Ω ) and
f N ( x ) = f ( x ) if   x ( E ( N ) ) , 0 if   x Ω \ ( E ( N ) ) .
Then,
ϱ p ( x ) f N ( x ) f ( x ) = E ( N ) | f N ( x ) f ( x ) | p ( x ) d x + Ω \ E ( N ) | f N ( x ) f ( x ) | p ( x ) d x = Ω \ E ( N ) | f ( x ) | p ( x ) d x = Ω | f ( x ) | p ( x ) χ Ω \ E ( N ) d x .
Taking Φ N = | f ( x ) | p ( x ) χ Ω \ E ( N ) for almost every x in Ω , we have
Φ N 0 a . e . in Ω and | Φ N | | f | p ( x ) .
Utilizing the dominated convergence theorem, we infer
ϱ p ( x ) ( f N ( x ) f ( x ) ) 0 as N ,
hence f N f in L p ( x ) ( Ω ) . Thus, F N is dense in L p ( x ) ( Ω ) .
Next, for all φ F N , let us show that
lim n Ω g ( x , η n , η n ) g ( x , η , η ) φ ( x ) d x = 0 .
Since φ 0 in Ω \ E ( N ) , it suffices to prove that
E ( N ) g ( x , η n , η n ) g ( x , η , η ) φ ( x ) d x 0 as n .
Let φ n = φ ( g ( x , η n , η n ) g ( x , η , η ) ) . Since | ( g ( x , η n , η n g ( x , η , η ) ) φ ( x ) | φ ( x ) a.e. in E ( N ) and φ n 0 a.e. in Ω , thanks to the dominated convergence theorem, we deduce φ n 0 in L 1 ( Ω ) , which implies that (29) holds.
It follows from the density of F N in L p ( x ) ( Ω ) that we deduce
lim n Ω g ( x , η n , η n ) φ ( x ) d x = lim n Ω g ( x , η , η ) φ ( x ) d x ,
for all φ F N , which implies that (28) holds.
Using the same discussion as above, one can conclude that
Ω | η n | r ( x ) 2 η n ln | η n | φ d x Ω | η | r ( x ) 2 η ln | η | φ d x .
As a result, it follows from (27), (28), and (30) that (23) holds, that is, the operator T is demicontinuous. □
Lemma 8.
Assume that the conditions of Theorem 2 are satisfied; then, for all μ > 0 , the operator T is coercive.
Proof. 
First, for all η W 0 , we note that
T η , η = M p ( x ) δ p ( x ) ( η ) η , η p ( x ) + μ M q ( x ) δ q ( x ) ( η ) η , η q ( x ) λ Ω | η | r ( x ) 2 η ln | η | η d x Ω g ( x , η , η ) η d x .
To estimate the first and second integral terms, we deduce
M p ( x ) δ p ( x ) ( η ) η , η p ( x ) + μ M q ( x ) δ q ( x ) ( η ) η , η q ( x ) 1 p + Ω | η | p ( x ) d x 2 + μ q + Ω | η | q ( x ) d x 2 1 p + min η 2 p + , η 2 p + μ q + min η 2 q + , η 2 q 1 p + η 2 p 1 + μ q + η 2 q 1 .
To estimate the third integral term, let Ω 1 = { x Ω : | η ( x ) | 1 } and Ω 2 = { x Ω : | η ( x ) | 1 } ; then,
Ω | η | r ( x ) ln | η | d x = Ω 1 | η | r ( x ) ln | η | d x + Ω 2 | η | r ( x ) ln | η | d x .
Using Lemma 4 with h ( x ) + r ( x ) h + + r + < p * ( x ) , we deduce
Ω 2 | η | r ( x ) ln | η | d x 1 e h Ω 2 | η | r ( x ) + h ( x ) d x 1 e h | η | h ( x ) + r ( x ) h + + r + + 1 ,
in view of Lemma 2, there exist some constants C h + + r + > 0 and C h + r > 0 such that
Ω 2 | η | r ( x ) ln | η | d x 1 e h C h + + r + η h + + r + + 1 .
This implies that
Ω | η | r ( x ) ln | η | d x C Ω 1 | Ω | + 1 e h C h + + r + η h + + r + + 1 ,
where | Ω | denotes the Lebesgue measure of Ω and C Ω 1 > 0 . This yields the stated conclusion.
To estimate the fourth integral term, we deduce from H g 2 , the Hölder-type inequality, and Proposition 1 that
Ω g ( x , η , η ) η d x Ω h ( x ) + a | η | ϕ ( x ) + b | η | ψ ( x ) p ( x ) η d x | h | p | η | p + a | η | ϕ ( x ) + 1 ϕ + + 1 + 1 + b η ( ψ p ) + + 1 | η | p C p | h | p η + a C ϕ ( x ) + 1 ϕ + + 1 η ϕ + + 1 + 1 + b C p η η ( ψ p ) + + 1 .
It follows (32), (33), and (34) that
T η , η 1 p + η 2 p 1 + μ q + η 2 q 1 λ C Ω 1 | Ω | λ e h C h + + r + η h + + r + + 1 C p | h | p η a C ϕ ( x ) + 1 ϕ + + 1 η ϕ + + 1 + 1 b C p η η ( ψ p ) + + 1 .
Case 1: Utilizing that 2 p > ϕ + + 1 and 2 p > ( ψ p ) + with λ 0 , for all η W 0 , such that
T η , η 1 p + η 2 p 1 C p | h | p η a C ϕ ( x ) + 1 ϕ + + 1 η ϕ + + 1 + 1 b C p η η ( ψ p ) + + 1 .
Case 2: Utilizing that 2 p > h + + r + , 2 p > ϕ + + 1 and 2 p > ( ψ p ) + with λ > 0 , for all η W 0 , such that
T η , η 1 p + η 2 p 1 λ C Ω 1 | Ω | λ e h C h + + r + η h + + r + + 1 C p | h | p η a C ϕ ( x ) + 1 ϕ + + 1 η ϕ + + 1 + 1 b C p η η ( ψ p ) + + 1 .
Case 3: Utilizing that 2 q > ϕ + + 1 and 2 q > ( ψ p ) + with λ 0 , for all η W 0 , such that
T η , η μ q + η 2 q 1 C p | h | p η a C ϕ ( x ) + 1 ϕ + + 1 η ϕ + + 1 + 1 b C p η η ( ψ p ) + + 1 .
Case 4: Utilizing that 2 q > h + + r + , 2 q > ϕ + + 1 and 2 q > ( ψ p ) + with λ > 0 , for all η W 0 , such that
T η , η μ q + η 2 q 1 λ C Ω 1 | Ω | λ e h C h + + r + η h + + r + + 1 C p | h | p η a C ϕ ( x ) + 1 ϕ + + 1 η ϕ + + 1 + 1 b C p η η ( ψ p ) + + 1 .
In the above four cases, we deduce the coerciveness of T from (35) as η .
Lemma 9.
Assume that the conditions of Theorem 2 are satisfied, then T is an ( S + ) -type operator.
Proof. 
Let { η n } W 0 be such that η n η in W 0 as n and
lim sup n T η n T η , η n η 0 .
First, note that
T η n , η n η = M p ( x ) δ p ( x ) ( η ) η , η n η p ( x ) + μ M q ( x ) δ q ( x ) ( η ) η , η n η q ( x ) λ Ω | η | r ( x ) 2 η ln | η | ( η n η ) d x Ω g ( x , η , η ) ( η n η ) d x .
Going if necessary up to a subsequence, we suppose there exists η W 0 such that
η n η , weakly in W 0 , η n η , strongly in L p ( x ) ( Ω ) , η n η , a . e . in Ω .
Indeed, by a simple calculation for the logarithmic nonlinear term, we deduce
Ω | η n | r ( x ) 2 η n ln | η n | r + r + 1 d x = Ω 1 | η n | r ( x ) 2 η n ln | η n | r + r + 1 d x + Ω 2 | η n | r ( x ) 2 η n ln | η n | r + r + 1 d x C Ω 1 | Ω | + Ω 2 | η n | r ( x ) 2 η n ln | η n | r + r + 1 d x .
Since r + < p * ( x ) , then, by using the continuous embedding L p * ( x ) ( Ω ) L r + ( Ω ) and combining Lemma 4, we deduce
Ω | η n | r ( x ) 2 η n ln | η n | r + r + 1 d x C Ω 1 | Ω | + Ω | η n | r + d x C Ω 1 | Ω | + C Ω 2 η n p * ( x ) .
where C Ω 2 > 0 . In conjunction with Hölder’s inequality, we obtain
Ω ( η n η ) | η n | r ( x ) 2 η n ln | η n | d x η n η L r + ( Ω ) | η n | r ( x ) 2 η n ln | η n | L r + r + 1 ( Ω ) .
Therefore, it follows from (37), (38) and (39) that
Ω ( η n η ) | η n | r ( x ) 2 η n ln | η n | d x 0 , as n .
In the same fashion, utilizing Lemma 1, we have
| Ω g ( x , η , η ) ( η n η ) d x | 2 | g ( x , η , η ) | p | ( η n η ) | p .
By the boundedness of { η n } W 0 and (37), we infer from the inequality above and the preceding estimate (21) that
| Ω g ( x , η , η ) ( η n η ) d x | 0 , as n .
If η n η in W 0 and lim sup n T η n T η , η n η 0 , as a consequence
lim n T η n , η n η = lim n T η n T η , η n η = 0 .
By (40), (41), (42), and H m , for υ { p ( x ) , q ( x ) } , as n , we deduce
Ω ( | η n | υ ( x ) 2 η n | η | υ ( x ) 2 η ) ( η n η ) d x 0 .
Using the following Simon inequalities
| u 1 u 2 | τ c τ [ ( | u 1 | τ 2 u 1 | u 2 | τ 2 u 2 ) ( u 1 u 2 ) ] τ 2 ( | u 1 | τ + | u 2 | τ ) 2 τ τ , 1 < τ < 2 , c ˜ τ ( | u 1 | τ 2 u 1 | u 2 | τ 2 u 2 ) ( u 1 u ) , τ 2 ,
for all u 1 , u 2 R N , where c τ and c ˜ τ are positive constants depending only on τ , we obtain
Ω | η n η | υ ( x ) d x Ω ( | η n | υ ( x ) 2 η n | η | υ ( x ) 2 η ) ( η n η ) d x .
Hence,
η n η 0 as n ,
that is, if η n η in W 0 and lim sup n T η n T η , η n η 0 , then η n η in W 0 . This shows the ( S + ) -property of T. □
Proof. 
From Section 2, evidently, we know that W 0 is a real, separable, and reflexive Banach spaces. Moreover, it follows from Lemmas 6–9 that the operator T satisfies all conditions of the Brezis theorem. Hence, invoking the Brezis theorem, we obtain that T η = 0 has at least one solution η in W 0 , i.e., problem (1) has at least one weak solution η . □

5. Examples

Now, we give two easy examples of application of our theorems. The first is when M p ( x ) ( t ) = a p + b p t , for all t 0 with a p > 0 , b p 0 and M q ( x ) ( t ) = a q + b q t , for all t 0 with a q > 0 , b q 0 . In this case, problem (1) reduces to the following form.
Example 1.
Consider the problem
a p + b p δ p ( x ) ( η ) ( Δ p ( x ) η ) + μ a q + b q δ q ( x ) ( η ) ( Δ q ( x ) η ) = λ | η | r ( x ) 2 η ln | η | + g ( x , η , η ) , in Ω , η | Ω = 0 ,
where Ω is an open bounded domain in R N with a smooth boundary.
It is clear that M p ( x ) ( t ) a p > 0 , for all t 0 and M q ( x ) ( t ) a q > 0 , for all t 0 . That is, the condition H m is satisfied. Thus, the results obtained in Theorems 1 and 2 stay true for problem (1). The problem and results are all new.
The second is when p ( x ) , q ( x ) , r ( x ) , α ( x ) are constant, that is, p ( x ) = p = c o n s t a n t ( 1 , + ) , q ( x ) = q = c o n s t a n t ( 1 , + ) , r ( x ) = r = c o n s t a n t ( 1 , + ) , α ( x ) = α = c o n s t a n t [ 1 , p ) and M p ( x ) ( t ) = ( a p + p b p t ) p 1 , for all t 0 with a p > 0 , b p 0 and M q ( x ) ( t ) = ( a q + q b q t ) q 1 , for all t 0 with a q > 0 , b q 0 . In this case, problem (1) becomes the following form.
Example 2.
Consider the problem
a p + b p Ω | η | p d x p 1 ( Δ p η ) + μ a q + b q Ω | η | q d x q 1 ( Δ q η ) = λ | η | r 2 η ln | η | + g ( x , η , η ) , in Ω , η | Ω = 0 ,
where Ω is an open bounded domain in R N with a smooth boundary.
The function g : Ω × R × R N R given by
g ( x , ϖ , ν ) = | ϖ | α 2 ϖ + ϖ 1 + ϖ 2 | ν | p 1 + γ ( x ) , for all ( x , ϖ , ν ) Ω × R × R N ,
with a constant α [ 1 , p ) , and some γ L ( Ω ) satisfies conditions H g 1 (see [21]). For p ( 1 , + ) , q ( 1 , + ) , the condition H p q is satisfied. It is clear that M p ( x ) ( t ) a p p 1 > 0 , for all t 0 and M q ( x ) ( t ) a q q 1 > 0 , for all t 0 . That is, condition H m is satisfied. Thus, the results obtained in Theorem 1 stay true for problem (2). The problem and results are also all new.

6. Conclusions

In this article, we study a kind of Kirchhoff-type elliptic problem, which combines with a variable exponent, competing ( p ( x ) , q ( x ) ) -Laplacian, logarithmic nonlinearity, and convection term. Due to the deficit of ellipticity, monotonicity, and variational structure, there are no available techniques to handle problem (1). A fundamental idea of the paper is to seek a solution to (1) as a limit of finite dimensional approximations. With the help of the Galerkin method and Brezis theorem, we obtain the existence of finite-dimensional approximate solutions and weak solutions, respectively. Our study extends previous results, such as from the elliptic problem with logarithmic nonlinearity or the convection term to ( p ( x ) , q ( x ) ) -Kirchhoff-type equations both logarithmic nonlinearity with variable exponents and convection terms. Finally, we consider that it will be a new field to study such problems (1) in fractional Sobolev spaces with variable exponents and in Sobolev spaces with variable exponents and variable fractional order.

Author Contributions

Each of the authors contributed to each part of this study equally, and all authors read and approved the final manuscript. All authors have read and agreed to the published version of the manuscript.

Funding

This work is supported by the Fundamental Research Funds for the Central Universities (B220203001), the Postgraduate Research & Practice Innovation Program of Jiangsu Province (KYCX21-0454), the Natural Science Foundation of Jiangsu Province (BK20180500), the National Key Research and Development Program of China (2018YFC1508100), the Special Soft Science Project of Technological Innovation in Hubei Province (2019ADC146), and the Natural Science Foundation of China (11701595).

Institutional Review Board Statement

Not applicable.

Informed Consent Statement

Not applicable.

Data Availability Statement

Data sharing does not apply to this article as no data sets were generated or analyzed during the current study.

Conflicts of Interest

The authors declare that they have no conflict of interest.

References

  1. Kirchhoff, G. Vorlesungen über Mathematische Physik; Band 1; Mechanik, B.G. Teubner: Leipzig, Germany, 1883. [Google Scholar]
  2. Ledesma, C. Multiplicity result for non-homogeneous fractional Schrödinger–Kirchhoff-type equations in RN. Adv. Nonlinear Anal. 2018, 7, 247–257. [Google Scholar] [CrossRef]
  3. Nyamoradi, N.; Chung, N. Existence of solutions to nonlocal Kirchhoff equations of elliptic type via genus theory. Electron. J. Differ. Equ. 2014, 86, 1–12. [Google Scholar]
  4. Lee, J.; Kim, J.M.; Kim, Y.H. Existence and multiplicity of solutions for Kirchhoff–Schrödinger type equations involving p(x)-Laplacian on the entire space RN. Nonlinear Anal. Real World Appl. 2019, 45, 620–649. [Google Scholar] [CrossRef]
  5. Fan, X.; Zhang, Q. Existence of solutions for p(x)-Laplacian Dirichlet problem. Nonliear Anal. 2003, 52, 1843–1852. [Google Scholar] [CrossRef]
  6. Kaufmann, U.; Rossi, J.; Vidal, R. Fractional Sobolev spaces with variable exponents and fractional p(x)-Laplacians. Electron. J. Qual. Theory Differ. Equ. 2017, 76, 1–10. [Google Scholar] [CrossRef]
  7. Biswas, R.; Tiwari, S. On a class of Kirchhoff–Choquard equations involving variable-order fractional p(·)-Laplacian and without Ambrosetti-Rabinowitz type condition. arXiv 2020, arXiv:2005.09221. [Google Scholar]
  8. Bahrouni, S.; Ounaies, H. Strauss and Lions type theorems for the fractional sobolev spaces with variable exponent and applications to nonlocal Kirchhoff–Choquard problem. Mediterr. J. Math. 2021, 18, 1–22. [Google Scholar] [CrossRef]
  9. Zuo, J.; Yang, L.; Liang, S. A variable-order fractional p(·)-Kirchhoff type problem in RN. Math. Methods Appl. Sci. 2021, 44, 3872–3889. [Google Scholar] [CrossRef]
  10. Zhang, J. Existence results for a kirchhoff-type equations involving the fractional p1(x)&p2(x)-Laplace operator. Collect. Math. 2022, 73, 271–293. [Google Scholar]
  11. Azroul, E.; Benkirane, A.; Shimi, M. Existence and multiplicity of solutions for fractional p(x,·)-Kirchhoff-type problems in RN. Appl. Anal. 2021, 100, 2029–2048. [Google Scholar] [CrossRef]
  12. Bahrouni, A.; Rǎdulescu, V. On a new fractional Sobolev space and applications to nonlocal variational problems with variable exponent. Discret. Contin. Dyn. Syst. Ser. S 2018, 11, 379–389. [Google Scholar] [CrossRef] [Green Version]
  13. Bisci, G.M.; Vilasi, L. On a fractional degenerate Kirchhoff-type problem. Commun. Contemp. Math. 2017, 19, 1550088. [Google Scholar] [CrossRef]
  14. Liu, D.C. On a p-Kirchhoff equation via fountain theorem and dual fountain theorem. Nonlinear Anal. 2010, 72, 302–308. [Google Scholar] [CrossRef]
  15. Fiscella, A.; Valdinoci, E. A critical Kirchhoff type problem involving a nonlocal operator. Nonlinear Anal. 2014, 94, 156–170. [Google Scholar] [CrossRef]
  16. Cherfils, L.; Il’yasov, V. On the stationary solutions of generalized reaction difusion equations with p&q-Laplacian. Commun. Pur. Appl. Anal. 2004, 4, 9–22. [Google Scholar]
  17. Figueiredo, G. Existence of positive solutions for a class of p&q elliptic problems with critical growth in RN. J. Math. Anal. Appl. 2011, 378, 507–518. [Google Scholar]
  18. Li, G.; Liang, X. The existence of nontrivial solutions to nonlinear elliptic equation of p&q-Laplacian type on RN. Nonlinear Anal. 2009, 71, 2316–2334. [Google Scholar]
  19. Zuo, J.; Fiscella, A.; Bahrouni, A. Existence and multiplicity results for p(·)&q(·) fractional Choquard problems with variable order. Complex Var. Elliptic Equ. 2022, 67, 500–516. [Google Scholar]
  20. Chung, N.; Toan, H. On a class of fractional Laplacian problems with variable exponents and indefinite weights. Collect. Math. 2020, 71, 223–237. [Google Scholar] [CrossRef]
  21. Motreanu, D. Quasilinear Dirichlet problems with competing operators and convection. Open Math. 2020, 18, 1510–1517. [Google Scholar] [CrossRef]
  22. Benboubker, M.; Azroul, E.; Barbara, A. Quasilinear elliptic problems with nonstandard growth. Electron. J. Differ. Equ. 2011, 62, 1–16. [Google Scholar]
  23. Fan, X.; Zhao, D. A class of De Giorgi type and Hölder continuity. Nonlinear Anal. 1999, 36, 295–318. [Google Scholar] [CrossRef]
  24. Ho, K.; Sim, I. A-priori bounds and existence for solutions of weighted elliptic equations with a convection term. Adv. Nonlinear Anal. 2017, 6, 427–445. [Google Scholar] [CrossRef]
  25. Ourraoui, A. On an elliptic equation of p-Kirchhoff type with convection term. C. R. Math. 2016, 354, 253–256. [Google Scholar] [CrossRef]
  26. Vetro, C. Variable exponent p(x)-Kirchhoff type problem with convection. J. Nonlinear Math. Phys. 2022, 506, 125721. [Google Scholar] [CrossRef]
  27. Li, Y.; Wang, D.; Zhang, J. Sign-changing solutions for a class of p-Laplacian Kirchhoff-type problem with logarithmic nonlinearity. AIMS Math. 2020, 5, 2100–2112. [Google Scholar] [CrossRef]
  28. Dai, P.; Mu, C.; Xu, G. Blow-up phenomena for a pseudo-parabolic equation with p-Laplacian and logarithmic nonlinearity terms. J. Math. Anal. Appl. 2019, 481, 123439. [Google Scholar] [CrossRef]
  29. Ardila, A.; Alex, H. Existence and stability of standing waves for nonlinear fractional Schrödinger equation with logarithmic nonlinearity. Nonlinear Anal. Theor. Methods Appl. 2017, 4, 52–64. [Google Scholar] [CrossRef] [Green Version]
  30. Truong, L. The nehari manifold for fractional p-Laplacian equation with logarithmic nonlinearity on whole space. Comput. Math. Appl. 2019, 78, 3931–3940. [Google Scholar] [CrossRef]
  31. Xiang, M.; Hu, D.; Yang, D. Least energy solutions for fractional Kirchhoff problems with logarithmic nonlinearity. Nonlinear Anal. 2020, 198, 111899. [Google Scholar] [CrossRef]
  32. Boudjeriou, T. On the diffusion p(x)-Laplacian with logarithmic nonlinearity. J. Elliptic Parabol.Equ. 2020, 6, 773–794. [Google Scholar] [CrossRef]
  33. Zeng, F.; Deng, Q.; Wang, D. Global Existence and Blow-Up for the Pseudo-parabolic p(x)-Laplacian Equation with Logarithmic Nonlinearity. J. Nonlinear Math. Phys. 2022, 29, 41–57. [Google Scholar] [CrossRef]
  34. Fan, X.; Zhao, D. On the spaces Lp(x)(Ω) and Wk,p(x)(Ω). J. Math. Anal. Appl. 2001, 263, 424–446. [Google Scholar] [CrossRef] [Green Version]
  35. Kováčik, O.; Rxaxkosník, J. On spaces Lp(x) and W1,p(x). Czechoslov. Math. J. 1991, 41, 592–618. [Google Scholar]
  36. Diening, L.; Harjulehto, P.; Hästö, P.; Ružička, M. Lebesgue and Sobolev Spaces with Variable Exponents; Lecture Notes in Mathematics 2017; Springer: Berlin/Heidelberg, Germany, 2011. [Google Scholar]
  37. Showalter, E. Monotone Operators in Banach Space and Nonlinear Partial Differential Equations; Mathematical Surveys and Monographs; American Mathematical Society: Providence, RI, USA, 1997; Volume 49. [Google Scholar]
  38. Zeidler, E. Nonlinear Functional Analysis and Its Applications. II/B: Nonlinear Monotone Operators; Springer: New York, NY, USA, 1990. [Google Scholar]
Publisher’s Note: MDPI stays neutral with regard to jurisdictional claims in published maps and institutional affiliations.

Share and Cite

MDPI and ACS Style

Bu, W.; An, T.; Qian, D.; Li, Y. (p(x),q(x))-Kirchhoff-Type Problems Involving Logarithmic Nonlinearity with Variable Exponent and Convection Term. Fractal Fract. 2022, 6, 255. https://doi.org/10.3390/fractalfract6050255

AMA Style

Bu W, An T, Qian D, Li Y. (p(x),q(x))-Kirchhoff-Type Problems Involving Logarithmic Nonlinearity with Variable Exponent and Convection Term. Fractal and Fractional. 2022; 6(5):255. https://doi.org/10.3390/fractalfract6050255

Chicago/Turabian Style

Bu, Weichun, Tianqing An, Deliang Qian, and Yingjie Li. 2022. "(p(x),q(x))-Kirchhoff-Type Problems Involving Logarithmic Nonlinearity with Variable Exponent and Convection Term" Fractal and Fractional 6, no. 5: 255. https://doi.org/10.3390/fractalfract6050255

APA Style

Bu, W., An, T., Qian, D., & Li, Y. (2022). (p(x),q(x))-Kirchhoff-Type Problems Involving Logarithmic Nonlinearity with Variable Exponent and Convection Term. Fractal and Fractional, 6(5), 255. https://doi.org/10.3390/fractalfract6050255

Article Metrics

Back to TopTop