See discussions, stats, and author profiles for this publication at: https://www.researchgate.

net/publication/377809257

Multiplicity analysis of positive weak solutions in a quasi-linear Dirichlet

problem inspired by Kirchhoff-type phenomena

Article in The International Journal of Nonlinear Analysis and Applications (IJNAA) · January 2024
DOI: 10.22075/ijnaa.2023.32433.4823

CITATIONS READS

0 54

2 authors:

Ahmed Ahmed Mohamed Saad Bouh Elemine Vall

Sidi Mohamed Ben Abdellah University Université de Nouakchott
18 PUBLICATIONS 55 CITATIONS 22 PUBLICATIONS 60 CITATIONS

SEE PROFILE SEE PROFILE

All content following this page was uploaded by Mohamed Saad Bouh Elemine Vall on 30 January 2024.

The user has requested enhancement of the downloaded file.

Int. J. Nonlinear Anal. Appl. In Press, 1–11
ISSN: 2008-6822 (electronic)
http://dx.doi.org/10.22075/ijnaa.2023.32433.4823

Multiplicity analysis of positive weak solutions in a quasi-linear

Dirichlet problem inspired by Kirchhoff-type phenomena
Ahmed AHMEDa,∗, Mohamed Saad Bouh Elemine Vallb

a Mathematicsand Computer Sciences Department, Research Unit Geometry, Algebra, Analysis and Applications, Faculty of Science and
Technology, University of Nouakchott, Nouakchott, Mauritania
b Department of Industrial Engineering and Applied Mathematics, Professional University Institute, University of Nouakchott, Nouakchott,
Mauritania

(Communicated by Abdolrahman Razani)

Abstract

The main focus of this paper lies in investigating the existence of infinitely many positive weak solutions for the
following elliptic-Kirchhoff equation with Dirichlet boundary condition

N Z   
 X
 1 ∂u pi (x) ∂ ∂u pi (x)−2 ∂u
− Mi dx = f (x, u) in Ω,
 i=1 Ω pi (x) ∂xi ∂xi ∂xi ∂xi
u=0 on ∂Ω.


The methodology adopted revolves around the technical approach utilizing the direct variational method within the
framework of anisotropic variable exponent Sobolev spaces.
Keywords: Nonlinear elliptic equations, Variational methods applied on PDEs, Positive solutions to PDEs
2020 MSC: Primary 35A15; Secondary 35H30, 35J60, 35B09

1 Introduction
Over the past years, differential equations have been a focal point of research, owing to their extensive practical
implications and widespread use in numerous fields.
Proposed by Kirchhoff [20], the Kirchhoff differential equations offer an extension to D’Alembert’s wave equation,
accommodating the effects of string length changes during vibrations
!
Z L 2
∂2u P0 E ∂u ∂2u
ρ 2 − + dx = 0, (1.1)
∂t h 2L 0 ∂x ∂x2

in this context, L denotes the length of the chord, h represents the area of the cross-section, E stands for the Young’s
modulus of the material, denotes the density, and P0 corresponds to the initial tension. The Kirchhoff equation (1.1)

∗ Corresponding author
Email addresses: ahmedmath2001@gmail.com (Ahmed AHMED ), saad2012bouh@gmail.com (Mohamed Saad Bouh ELEMINE VALL )

Received: November 2023 Accepted: December 2023

2 Ahmed, Elemine Vall

L Z 2
P0 ∂u E
exhibits a unique characteristic in its inclusion of a non-local coefficient + h dx, which is dependent on
2L
0 ∂x
Z L 2
E ∂u 2
the average 2L of the kinetic energy 12 ∂u∂x within the interval [0, L]. Thus, the equation loses its property
0 ∂x
of being a point-wise identity. See also [6, 19, 32] for related topics.
In recent times, numerous mathematicians, physicists, and engineers have shown a keen interest in anisotropic
variable exponent Sobolev spaces. The motivation behind this stems from the crucial role these spaces play in
modeling real-world phenomena, including electrorheological and image restoration, magnetorheological fluids, and
elastic materials, (look at, for example [5, 8, 10, 30, 33, 34, 35, 36]).
In the present paper we study the existence of positive solutions of the nonhomogeneous anisotropic Kirchhoff
problem  N Z   
− X M
 1 ∂u pi (x) ∂ ∂u pi (x)−2 ∂u
i dx = f (x, u) in Ω,
i=1 Ω pi (x) ∂xi ∂xi ∂xi ∂xi (1.2)

u=0 on ∂Ω,


where Ω ⊂ RN , (N > 3) represents a bounded domain with a smooth boundary ∂Ω, and pi , i = 1, . . . N are continuous
functions. Additionally, for each i = 1, . . . N , Mi and f are continuous functions which satisfies some conditions detailed
in Section 3.
The differential operator
N 
pi (x)−2

X ∂ ∂u ∂u
∆p⃗(·) u = , (1.3)
i=1
∂x i ∂xi ∂xi

involved in problem (1.2) is an anisotropic variable exponent p⃗(·)-Laplace operator which represents an extension of
the operator  
∆p(·) u = div |∇u|p(x)−2 ∇u . (1.4)

The p(·)-Laplacian operator, obtained by setting each pi (x) to be equal to p(x) for i = 1, . . . N , serves as a natural
extension of the isotropic p-Laplacian operator

∆p u = div |∇u|p−2 ∇u ,


(1.5)

where p > 1 denotes a real constant. In the classical Sobolev spaces, F. J. S. A. Corrêa, R. G. Nascimento [12] have
established the existence of solutions for problem (1.2) in this particular case p-Kirchhoff-type equation, for additional
results, refer to [22, 28, 29, 31].
In the Sobolev variable exponent setting, G. Dai and D. Liu [13] has analyzed the problem (1.2) in the context of
p(x)-Kirchhoff-type equation, see also [2, 9, 11, 14, 18] for related topics.
The investigation of problem (1.2) in anisotropic variable exponent Sobolev spaces has been previously addressed
by other researchers (see [7, 16, 25]). However, our study stands apart due to the entirely distinct hypotheses adopted,
which subsequently lead to different and novel findings.
The shift from a variable exponent to an anisotropic variable exponent inevitably introduces fresh complexities. To
tackle these challenges, we adopt a combined approach, utilizing traditional methodologies alongside modern techniques
specifically designed for handling problems of anisotropic nature with variable exponents. The organization of this
paper is as follows: In Section 2, we provide an introduction to anisotropic variable exponent Sobolev spaces, laying
the necessary groundwork for the subsequent analysis. Section 3 is dedicated to presenting the assumptions under
which our problem yields positive solutions, accompanied by an illustrative example.

2 Preliminary
Let Ω denote a smooth bounded domain in RN , where we introduce the following definitions:

C+ (Ω) = p ∈ C(Ω) such that 1 < p− ≤ p+ < ∞ ,



where
p− = ess inf p(x) : x ∈ Ω and p+ = ess sup p(x) : x ∈ Ω .
 
Multiplicity analysis of positive weak solutions ... 3

For any p ∈ C+ (Ω), we introduce the Lebesgue space with variable exponent Lp(·) (Ω), which encompasses all
measurable functions u : Ω 7−→ R such that the convex modular
Z
ρp(·) (u) := |u|p(x) dx,
Ω

remains finite. Consequently, we define the norm

n u o
∥u∥Lp(·) (Ω) = ∥u∥p(·) = inf λ > 0 : ρp(·) ≤1 ,
λ
as the Luxemburg norm in Lp(·) (Ω). As a separable Banach space, (Lp(·) (Ω), | · |p(·) ) exhibits desirable properties.
′
Additionally, the space Lp(·) (Ω) is uniformly convex, making it reflexive, and its dual space is isomorphic to Lp (·) (Ω),
1 1
with + = 1. Lastly, we arrive at the following Hölder-type inequality.
p(·) p′ (·)
Z  1 1  ′
u(x)v(x)dx ≤ − + − ′ ∥u∥p(·) ∥v∥p′ (·) for all u ∈ Lp(·) (Ω) and v ∈ Lp (·) (Ω). (2.1)
Ω p (p )

The modular ρp(·) of the space Lp(·) (Ω) assumes a crucial role in handling the generalized Lebesgue spaces. The
ensuing result is presented:

Proposition 2.1. (See [15]). Considering un , u ∈ Lp(·) (Ω), with p+ < +∞, we observe the subsequent properties:
− +
1. If ∥u∥p(·) > 1, then ∥u∥pp(·) < ρp(·) (u) < ∥u∥pp(·) .
+ −
2. For ∥u∥p(·) < 1, we have ∥u∥pp(·) < ρp(·) (u) < ∥u∥pp(·) .

3. The condition ∥u∥p(·) < 1 (respectively = 1; > 1) is equivalent to ρp(·) (u) < 1 (respectively = 1; > 1).
4. When ∥un ∥p(·) → 0 (respectively → +∞), it implies ρp(·) (un ) → 0 (respectively → +∞).
 
u
5. Lastly, we have ρp(·) = 1.
∥u∥p(·)

1,p(·)
The definition of W0 (Ω) involves taking the closure of C0∞ (Ω) in W 1,p(·) (Ω), and

N p(x)
for p(x) < N,

p∗ (x) = N − p(x)
∞ for p(x) ≥ N.


Proposition 2.2. (See [15]).

1,p(·)
(i) For 1 < p− ≤ p+ < ∞, both W 1,p(·) (Ω) and W0 (Ω) are separable and reflexive Banach spaces.
1,p(·)
(ii) If q(x) ∈ C+ (Ω̄) and q(x) < p∗ (x) holds true for each x ∈ Ω, then the embedding W0 (Ω) ,→,→ Lq(·) (Ω) is
continuous and compact.

We now introduce the anisotropic Sobolev space with variable exponent, which serves as the foundation for studying
our main problem. Consider N variable exponents p1 (·), . . . , pN (·) belonging to C+ (Ω). We use the notation
n o ∂u
p⃗(·) = p1 (·), . . . , pN (·) and Di u = for i = 1, . . . , N,
∂xi

and we set it for all x ∈ Ω,

n o n o
pM (·) = max p1 (·), . . . , pN (·) and pm (·) = min p1 (·), ..., pN (·) .
4 Ahmed, Elemine Vall

The following notations are introduced:

n o n o n o
p = min p− 1 , p−
2 , . . . , p −
N , p+
= max p − −
1 , p2 , . . . , p −
N , p = max p+
1 , p+
2 , . . . , p +
N , (2.2)

and
N n o
p∗ = N
, p,∞ = max p∗ , p+ . (2.3)
X 1
−1
p−
i=1 i

In the context of this paper, we make the assumption that

N
X 1
> 1. (2.4)
p−
i=1 i

The definition of the anisotropic variable exponent Sobolev space W 1,⃗p(·) (Ω) is as follows:
n o
W 1,⃗p(·) (Ω) = Di u ∈ Lpi (·) (Ω), i = 1, 2, . . . , N ,

equipped with the norm

N
X
∥u∥W 1,⃗p(·) (Ω) = ∥u∥1,⃗p(·) = ∥Di u∥Lpi (·) (Ω) , (2.5)
i=1

1,⃗
p(·)
Furthermore, W0 (Ω) is defined as the closure of C0∞ (Ω) in W 1,⃗p(·) (Ω) under the norm (2.5). The dual space
′
of W 0 1,⃗
p(·)
(Ω) is denoted as W −1,⃗p(·) (Ω), where p⃗′ (x) = p′0 (x), . . . , p′N (x), satisfying p′ 1(x) + pi1(x) = 1 (see [26, 27])
i
1,⃗
p(·)

for the constant exponent case). The reflexivity of the Banach space W0 (Ω), |u|1,⃗p(·) has been established in
[24]. For a more comprehensive treatment of anisotropic variable exponent Sobolev spaces, researchers may delve into
[1, 3, 17, 21, 24].

Proposition 2.3. (See [4, 23]). The bounded domain Ω ⊂ RN , with a smooth boundary and N > 3, satisfies relation
(2.4).

1. Considering any q ∈ C+(Ω) satisfying the condition 1 < q(x) < p, ∞, for all x ∈ Ω, then
1,⃗
p(·)
W0 (Ω) ,→,→ Lq(·) (Ω).

2. Assume that p > N then

1,⃗
p(·)
W0 (Ω) ,→,→ C 0 (Ω).

3 Fundamental assumptions and main results

For the entirety of this paper, we make the assumption that the following set of conditions is satisfied:
Assume that f : Ω × R 7−→ R is Carathéodory functions satisfying the following condition

(H1) There exists a constant τ > 0 such that supt∈[0,τ ] f (·, t) ∈ L∞ (Ω).

(H2) Suppose that (an )n and (bn )n be two positive sequences such that
Z an Z t
0 < an < bn , lim bn = 0, and f (x, s)ds = sup f (x, s)ds for almost all x ∈ Ω and n ∈ N.
n→∞ 0 t∈[an ,bn ] 0

(H3) There is a sequence (ϑn )n , which is a subset of the interval [0, bn ], such that
Z ϑn
ess inf f (x, s)ds > 0.
Ω 0
Multiplicity analysis of positive weak solutions ... 5

For the function Mi , i = 1, . . . N , we set forth the subsequent assumptions.

(H4) Mi is a differentiable on R+ and there is positive constant m such that

Mi (t) ≥ m for all t ≥ 0.
1,⃗
p(·)
Functionals are defined for any u ∈ W0 (Ω) as follows:
N Z
pi (x)
 Z
X 1 ∂u
Φ(u) = Mi
c dx − F (x, u)dx, (3.1)
i=1
p
Ω i (x) ∂xi Ω
Z t Z t
where F (x, t) = f (x, s)ds and M
ci (t) = Mi (s)ds.
0 0

1,⃗
p(·)
Definition 3.1. For any measurable function u ∈ W0 (Ω) to be considered a weak solution of the elliptic problem
1,⃗
p(·)
(1.2), it must satisfy the condition that, for all v ∈ W0 (Ω),
N Z
pi (x)
Z X N pi (x)−2
Z
X 1 ∂u ∂u ∂u ∂v
Mi dx dx = f (x, u)v(x)dx. (3.2)
i=1 Ω pi (x) ∂xi Ω i=1 ∂xi ∂xi ∂xi Ω

1,⃗
p(·) 1,⃗
p(·)
It is easy to see that Φ ∈ C 1 (W0 (Ω), R) (see [7, 25]), and the function u ∈ W0 (Ω) is deemed a weak solution of
(1.2) if and only if it corresponds to a critical point of the functional Φ.
Considering our assumptions on f , we can find positive constants k and τ such that |f (·, t)| ≤ k for every 0 ≤ τ ≤ t
and almost every x ∈ Ω. Without any loss of generality, we can suppose that bn ≤ τ for every n ∈ N. Let’s proceed by
defining 
0
 if t ≤ 0,
g(·, t) = f (·, t) if 0 < t ≤ τ, (3.3)

f (·, τ ) if t > τ.


Hence, we have
|g(·, t)| ≤ k, (3.4)
for almost every x ∈ Ω and every t ∈ R. Next, we take into account the following problem
 N Z   
− X M
 1 ∂u pi (x) ∂ ∂u pi (x)−2 ∂u
i dx = g(x, u) in Ω,
i=1 Ω pi (x) ∂xi ∂xi ∂xi ∂xi (3.5)

u=0 on ∂Ω.


We can identify the weak solutions of (3.5) as the critical points of the functional
N Z  Z
X 1 ∂u pi (x)
Ψ(u) = Mi
c dx − G(x, u)dx, (3.6)
i=1 Ω pi (x) ∂xi Ω
Z t
1,⃗
p(·)
where G(x, t) = g(x, s)ds. By (3.4), it is clear that Ψ is well defined and Gâteaux differentiable in W0 (Ω) (see
0
[7, 25]). For every fixed n ∈ N, we define
n o
1,⃗
p(·)
Kn (u) = u ∈ W0 (Ω) : 0 < u(x) ≤ bn a.e. Ω . (3.7)

Having established the necessary groundwork, we can now present the main findings of this paper.

Theorem 3.2. Assume assumptions (H1)-(H4) hold true and f (·, 0) = 0. Then, there exists a sequence (un )n ⊂
1,⃗
p(·)
W0 (Ω) of positive, homoclinic weak solutions of (1.2) such that
lim Ψ(un ) = 0 and lim ∥un ∥1,⃗p(·) = 0. (3.8)
n→+∞ n→+∞

Theorem 3.3. To enhance the organization and clarity, we divided the proof into three steps.
6 Ahmed, Elemine Vall

Step 1 :. Auxiliary lemmas.

Lemma 3.4. Assume assumptions (H1), (3.4) and (H4) are satisfied. Then, the functionals Ψ is weakly lower
semi-continuous.
1,⃗
p(·) 1,⃗
p(·)
Proof . For each i = 0, . . . , N and any u ∈ W0 (Ω), we can define the functionals Ji and H : W0 (Ω) −→ R as
follows: Z
1 ∂u pi (x) ∂u
Ji = dx, where = u,
Ω pi (x) ∂xi ∂x0
Z
H(u) = − G(x, u)dx.
Ω
1,⃗
p(·)
Claim 1: Consider a sequence (un )n with the property that un ⇀ u in W0 (Ω). As Ji is convex, for every n, we
obtain
Ji (u) ≤ Ji (un ) + ⟨Ji′ (u), u − un ⟩.
Taking the limit as n → ∞ in the above inequality, we observe that Ji is sequentially weakly lower semi-continuous. As
a result, we obtain:
Z X N Z X N
1 ∂u pi (x) 1 ∂un pi (x)
dx ≤ lim inf dx. (3.9)
Ω i=1 pi (x) ∂xi n→+∞ Ω i=0 pi (x) ∂xi

By utilizing (3.9) and considering the continuity and monotonicity of M

ci , we obtain

N Z
pi (x)

X 1 ∂un
lim inf J(un ) = lim inf M
ci dx
n→+∞ n→+∞
i=0 Ω pi (x) ∂xi
N  Z pi (x)

X 1 ∂un
≥ M
ci lim inf dx
i=0
n→+∞ Ω pi (x) ∂xi
N Z
pi (x)

X 1 ∂u
≥ M
ci dx
i=0 Ω pi (x) ∂xi
≥ J(u). (3.10)

That is to say, J demonstrates sequential weak lower semi-continuity.

1,⃗
p(·)
Claim 2: H is sequentially weakly continuous. Let (un )n be a sequence in W0 (Ω) such that un ⇀ u weakly in
1,⃗
p(·)
W0 (Ω). So, by (3.4) and Proposition 2.3. Therefore, it is easy to show that lim H(un ) = H(u), and hence H is
n→∞
sequentially weakly lower semicontinuous. Similarly, just like we demonstrated for the mapping H, it is possible to
establish the sequential weak lower semi-continuity of Φ. Since Ψ = J − H, we complete the proof. □

Lemma 3.5. On Kn , the functional Ψ is boundedly below, and the infimum mn over Kn is attained at un ∈ Kn .

Proof . To begin with, considering any u ∈ Kn , we find that

N Z
pi (x)
 Z Z
X 1 ∂u
Ψ(u) = M
ci dx − G(x, u)dx ≥ − G(x, u)dx
i=1 Ω pi (x) ∂xi Ω Ω

≥ −kbn meas(Ω). (3.11)

In conclusion, we deduce that Ψ is bounded from below on Kn . It is apparent that Kn possesses the properties of
1,⃗
p(·)
convexity and closedness, thus establishing its weak closedness within W0 (Ω). Consider the sequence (un )n in Kn
1
such that Ψ(un ) lies between mn and mn + n for all n ∈ N, where mn = inf Kn Ψ. Next, if ∥un ∥1,⃗p(·) ≤ 1, our objective
Multiplicity analysis of positive weak solutions ... 7

is achieved; otherwise, we proceed with the following steps

N Z
pi (x)
 Z
X 1 ∂u
Ψ(u) = M
ci dx − G(x, u)dx
i=1 Ω pi (x) ∂xi Ω
Z 
1 ∂u pi (x)
N Z  dx Z
X
Ω pi (x) ∂xi
= Mi (s)ds − G(x, u)dx
i=1 0 Ω
N
!
p Z
mX ∂u
≥ −1 − G(x, u)dx
p i=1 ∂xi pi (·) Ω

m p mN
≥ ∥un ∥1,⃗p(·) − − kbn (measΩ).
pN p−1 pN p−1

Which yields
m p mN
∥un ∥1,⃗p(·) ≤ mn + 1 + + kbn |Ω|, (3.12)
pN p−1 pN p−1
1,⃗
p(·)
for all n ∈ N, thus (un )n is bounded in W0 (Ω) which is a reflexive space. Therefore, by considering a sub-sequence
denoted as (un )n , we observe weak convergence towards a specific element un ∈ Kn . This leads us to the conclusion
that Ψ(un ) = mn , utilizing the concept of weakly sequentially lower semi-continuity of Ψ. □

Step 2 :. A priori estimates.

We start this step by proving in the following result that the sequence (un )n is bounded almost everywhere.

Proposition 3.6. For all n ∈ N, we have 0 ≤ un (x) ≤ an a.e. x ∈ Ω.

Proof
 . Let Λn  = {x ∈ Ω : bn ≥ un (x) > an } and suppose that meas(Λn ) > 0. Define the function σn (t) =
min max(t, 0), an and set hn = σn (un ). It is clear that from the definition and the continuity of σn we get hn ∈ Kn .
As a consequence, we obtain that (
un (x) if x ∈ Ω\Λn ,
hn (x) = (3.13)
an if x ∈ Λn .
Then, we can write
Ψ(hn ) − Ψ(un )
N Z  Z Z hn N Z  Z Z un
X 1 ∂hn pi (x) X 1 ∂un pi (x)
= Mi
c dx − g(x, t)dtdx − Mi
c dx + g(x, t)dtdx
i=1 Ω pi (x) ∂xi Ω 0 i=1 Ω pi (x) ∂xi Ω 0
N Z ! N Z 
X 1 ∂hn pi (x) X 1 ∂hn pi (x)
= Mi
c dx + Mi
c dx
i=1 Ω\Λn pi (x) ∂xi i=1 Λn pi (x) ∂xi
Z Z hn Z Z hn
− g(x, t)dtdx − g(x, t)dtdx
Ω\Λn 0 Λn 0
N Z ! N Z 
X 1 ∂un pi (x) X 1 ∂un pi (x)
− M
ci dx − M
ci dx
i=1 Ω\Λn pi (x) ∂xi i=1 Λn pi (x) ∂xi
Z Z un Z Z un
+ g(x, t)dtdx + g(x, t)dtdx
Ω\Λn 0 Λn 0
N Z
pi (x)
 Z Z an Z Z un
X 1 ∂un
=− M
ci dx − g(x, t)dtdx + g(x, t)dtdx
i=1 Λn pi (x) ∂xi Λn 0 Λn 0
N Z
pi (x)
 Z Z hn
X 1 ∂un
=− M
ci dx − g(x, t)dtdx
i=1 Λn pi (x) ∂xi Λn un
N Z
pi (x)
 Z Z an
X 1 ∂un
=− M
ci dx − g(x, t)dtdx ≤ 0. (3.14)
i=1 Λn pi (x) ∂xi Λn un
8 Ahmed, Elemine Vall
Z Z an
Because g(x, t)dtdx ≥ 0. Hence, Ψ(hn ) ≥ Ψ(un ) = inf Kn Ψ, then every term should be zero. In particular,
Λn un

N Z pi (x)
Z
X 1 ∂un  
dx = G(x, an ) − G(x, un ) dx. (3.15)
i=1 Λn pi (x) ∂xi Λn

Therefore, meas(Λn ) = 0, which means 0 ≤ un (x) ≤ an almost every where x ∈ Ω. □ Next we show that the sequence
(un )n formed of weak solutions of problem (3.5) as mentioned in the following result.

1,⃗
p(·)
Proposition 3.7. The terms of (un )n are local minimum points of Ψ in W0 (Ω).
Z u
Proof . Set Γn = {x ∈ Ω : bn ≥ u(x) > an }. So, we have g(x, t)dt = 0 for any x ∈ Ω\Γn . In the other hand, if
σn (u)
x ∈ Γn , then one has the following three cases.
Z u
(a) If u(x) < 0, then g(x, t)dt = 0.
σn (u)
Z u
(b) If an < u(x) ≤ bn , then by (H2), g(x, t)dt ≤ 0.
σn (u)
Z u Z u Z u  
(c) If bn < u(x), then g(x, t)dt = g(x, t)dt ≤ kdt = k u(x) − an , by (3.4).
σn (u) an an

Fix a real p,∞ such that p,∞ > q(x) + 1 > p for every x ∈ Ω, then the following constant is finite

k(µ − an )
λ = sup .
µ≥bn (µ − an )q(x)+1
Z u
Then, for almost every where x ∈ Ω, we have g(x, t)dt ≤ λ|(u(x) − σn (u(x)))|q(x)+1 . Then, since when p ≤ N ,
σn (u)
the space is W 1,⃗p(·) (Ω) compactly embedded in Lq(·)+1 (Ω) and continuously embedded in C 0 (Ω) elsewhere, there is a
positive constant c such that
Z Z u
q(x)+1
g(x, t)dtdx ≤ cq(x)+1 λ∥(u − σn (u))∥1,⃗p(·) . (3.16)
Ω σn (u)

Therefore, we can write

N Z  Z Z u
X 1 ∂u pi (x)
Ψ(u) − Ψ(σn (u)) = M
ci dx − g(x, t)dtdx
i=1 Ω pi (x) ∂xi Ω 0
N Z  Z Z σn (u)
X 1 ∂σn (u) pi (x)
− Mi
c dx + g(x, t)dtdx
i=1 Ω pi (x) ∂xi Ω 0
N Z  Z Z u
X 1 ∂u pi (x)
= M
ci dx − g(x, t)dtdx
i=1 Γn pi (x) ∂xi Γn σn (u)
N Z  Z Z u
X 1 ∂u ∂σn (u) pi (x)
= M
ci − dx − g(x, t)dtdx
i=1 Ω pi (x) ∂xi ∂xi Ω σn (u)
N Z 
X 1 ∂u ∂σn (u) pi (x) q(x)+1
≥ Mi
c − dx − λcq(x)+1 ∥(u − σn (u))∥1,⃗p(·) (by (3.16))
i=1 Ω pi (x) ∂x i ∂x i
m q(x)+1
≥ ∥(u − σn (u))∥p1,⃗p(·) − λcq(x)+1 ∥(u − σn (u))∥1,⃗p(·) , (3.17)
pN p−1
Multiplicity analysis of positive weak solutions ... 9

Since σn (u) ∈ Kn , we have Ψ(σn (u)) ≥ Ψ(un ) and preserving the generality of our analysis, let’s assume that
∥(u − σn (u))∥1,⃗p(·) ≤ 1 cause we need small values of ∥(u − σn (u))∥1,⃗p(·) . Then
m p q(x)+1 q(x)+1
Ψ(u) ≥ Ψ(un ) + p−1 ∥(u − σn (u))∥1,⃗p(·) − λc ∥(u − σn (u))∥1,⃗p(·)
p(N )
 
m q(x)+1−p
≥ Ψ(un ) + − λc q(x)+1
∥(u − σn (u))∥1,⃗p(·) ∥(u − σn (u))∥p1,⃗p(·) . (3.18)
p(N )p−1

1,⃗
p(·)
The continuity of σn allows us to choose a positive value δ > 0 such that, for any u ∈ W0 (Ω), the condition
q(x)+1−p m
∥(u − σn (u))∥1,⃗p(·) < δ, ∥(u − σn (u))∥1,⃗p(·) ≤ , (3.19)
p(N )p−1 λcq(x)+1
this implies that un is a local minimum of Ψ. □

Proposition 3.8. The sequence (mn )n is strictly negative and converges to zero.

Proof . In view of condition (H3), we have ϑn ∈ Kn

Z Z ϑn
mn ≤ Ψ(ϑn ) = − f (x, t)dtdx < 0. (3.20)
Ω 0

To prove that lim mn = 0 it is sufficient to observe that for every n ∈ N and u ∈ Kn , we have
n→+∞

0 > mn = Ψ(un ) ≥ −kbn |Ω|. (3.21)

Since (bn )n converges to zero, we conclude the required result. □

Step 3 :. Proof of Theorem 3.2. Since the terms of (un )n are local minima of Ψ, they are weak solutions of (1.2). In
virtue of Proposition 3.6, we have 0 ≤ un (x) ≤ an for almost every where x ∈ Ω and since (an )n converges to zero. An
infinite number of distinct sequences (un )n can be found such that lim ∥un ∥L∞ (Ω) = 0. Furthermore, we have
n→+∞

N Z
pi (x)
 Z
X 1 ∂un
mn = Ψ(un ) ≥ M
ci dx − G(x, un )dx
i=1 Ω pi (x) ∂xi Ω
N Z
pi (x)

X 1 ∂un
≥ M
ci dx − kbn |Ω|. (3.22)
i=1 Ω pi (x) ∂xi

Thus, if ∥un ∥1,⃗p(·) ≤ 1, we have

m
∥un ∥p1,⃗p(·) ≤ mn + kbn |Ω| −→ 0. (3.23)
pN p−1

Thus, limn→+∞ ∥u∥1,⃗p(·) = 0, which completes our proof.

Now, we present an example to illustrate the main results.

θ
Example 3.9. We define Mi (t) = (1 + t) i for i = 1, . . . N , where θi > 0. It is worth noting that Mi (t) ≥ 1 for all
t ≥ 0, which directly verifies the condition stated in (H4). Let
(
(1 + |x|2 )(p + 2)tp+1 sin t1p − p(1 + |x|2 )t cos t1p



if t > 0,
f (x, t) = (3.24)
0 otherwise.

It is easy to compute directly that

(
(1 + |x|2 )tp+2 sin 1


tp if t > 0,
F (x, t) = (3.25)
0 otherwise.
10 Ahmed, Elemine Vall

We shall now consider the following nonlinear perturbed Kirchhoff problem

 N  Z θi  
 X 1 ∂u pi (x) ∂ ∂u pi (x)−2 ∂u
 − 1+ dx


Ω pi (x) ∂xi ∂xi ∂xi ∂x
i=1
2 p+1 1

2 1

i (3.26)


 = (1 + |x| )(p + 2)t sin tp − p(1 + |x| )t cos t p in Ω,
u = 0 on ∂Ω.


Let (an )n , (bn )n , and (ϑn )n be three positive sequences satisfying the conditions:
  p1   p1   p1
1 1 1
an = , bn = 3π and ϑn = π , (3.27)
2nπ + 2π 2nπ + 2
4nπ + 2

for every n ∈ N. Then one easily deduces

Z an Z t
f (x, s)ds = sup f (x, s)ds,
0 t∈[an ,bn ] 0

and F (x, ϑn ) > 0. So conditions (H2) and (H3) have been verified. Having satisfied all the assumptions of Theorem
3.2, we can affirm the existence of a sequence of positive, homoclinic weak solutions (un )n in W 1,⃗p(x) (Ω) for the problem
(3.26).

References
[1] A. Ahmed, H. Hjiaj, and A. Touzani, Existence of infinitely many weak solutions for a Neumann elliptic equations
involving the p⃗(·)-Laplacian operator, Rend. Circ. Mat. Palermo (2) 64 (2015), no. 3, 459–473.
[2] M. Allaoui and A. Ourraoui, Existence results for a class of p(x)-Kirchhoff problem with a singular weight, Mediterr.
J. Math. 13 (2016), no. 2, 677–686.
[3] C.O. Alves and A. El Hamidi, Existence of solution for an anisotropic equation with critical exponent, Nonlinear
Anal. TMA. 4 (2005), 611–624.
[4] A.E. Amrouss and A. El Mahraoui, Infinitely many solutions for anisotropic elliptic equations with variable
exponent, Proyec. J. Math. 40 (2021), no. 5, 1071–1096.
[5] S.N. Antontsev and J.F. Rodrigues, On stationary thermorheological viscous flows, Ann. Univer. Ferrara. 52
(2006), no. 1, 19–36.
[6] A. Arosio and S. Panizzi, On the well-posedness of the Kirchhoff string, Trans. Amer. Math. Soc. 348 (1996),
no. 1, 305–330.
[7] M. Avci, R.A. Mashiyev, and B. Cekic, Solutions of an anisotropic nonlocal problem involving variable exponent,
Adv. Nonlinear Anal. 2 (2013), no. 3, 325–338.
[8] M.M. Boureanu, A. Matei, and M. Sofonea, Non-linear problems with p(·)-growth conditions and applications to
anti-plane contact models, Adv. Nonlinear Stud. 14 (2014), no. 2, 295–313.
[9] F. Cammaroto and L. Vilasi, Multiple solutions for a Kirchhoff-type problem involving the p(x)-Laplacian operator,
Nonlinear Anal. 74 (2011), no. 5, 1841–1852.
[10] Y. Chen, S. Levine, and R. Rao, Variable exponent, linear growth functionals in image restoration, SIAM J. Appl.
Math. 66 (2006), no. 4, 1383–1406.
[11] N.T. Chung, Multiplicity results for a class of p(x)-Kirchhoff type equations with combined nonlinearities, Electron.
J. Qual. Theory Differ. Equ. 42 (2012), 1–13.
[12] F.J.S.A. Corrêa and R.G. Nascimento, On a nonlocal elliptic system of p-Kirchhoff-type under Neumann boundary
condition, Math. Model. Comput. 49 (2009), no. 3-4, 598–604.
[13] G. Dai and D. Liu, Infinitely many positive solutions for a p(x)-Kirchhoff-type equation, J. Math. Anal. Appl. 359
(2009), no. 2, 704–710.
Multiplicity analysis of positive weak solutions ... 11

[14] G. Dai and R. Hao, Existence of solutions for a p(x)-Kirchhoff-type equation, J. Math. Anal. Appl. 359 (2009),
no. 1, 275–284.
[15] L. Diening, P. Harjulehto, P. Hästö, and M. Růz̆ic̆ka, Lebesgue and Sobolev Spaces with Variable Exponents, Lecture
Notes in Mathematics, vol. 2017, Springer-Verlag, Heidelberg, 2011.
[16] G.C.G. dos Santos, J.R.S. Silva, S.C.Q. Arruda, and L.S. Tavares, Existence and multiplicity results for critical
anisotropic Kirchhoff-type problems with nonlocal nonlinearities, Complex Var. Elliptic Equ. 67 (2022), no. 4,
822–842.
[17] X.L. Fan, Anisotropic variable exponent Sobolev spaces and p⃗(·)-Laplacian equations, Complex Var. Elliptic Equ.
56 (2011), no. 7-9, 623–642.
[18] X.L. Fan, On nonlocal p(x)-Laplacian Dirichlet problems, Nonlinear Anal. 72 (2010), no. 7-8, 3314–3323.
[19] J.R. Graef, S. Heidarkhani, and L. Kong, A variational approach to a Kirchhoff-type problem involving two
parameters, Results Math. 63 (2013), no. 3-4, 877–889.
[20] G. Kirchhoff, Mechanik, Teubner, Leipzig, Germany, 1883.
[21] B. Kone, S. Ouaro, and S. Traore, Weak solutions for anisotropic nonlinear elliptic equations with variable
exponents, Electron. J. Differ. Equ. 144 (2009), 1–11.
[22] D. Liu, On a p-Kirchhof equation via fountain theorem and dual fountain theorem, Nonlinear Anal. 72 (2010), no.
1, 302-308.
[23] M. Mihăilescu and G. Morosanu, Existence and multiplicity of solutions for an anisotropic elliptic problem involving
variable exponent growth conditions, Appl. Anal. 89 (2010), no. 2, 257–271.
[24] M. Mihăilescu, P. Pucci, and V. Rădulescu, Eigenvalue problems for anisotropic quasilinear elliptic equations with
variable exponent, J. Math. Anal. Appl. 340 (2008), no. 1, 687–698.
[25] A. Ourraoui, Multiplicity of solutions for p⃗(·)-Laplacian elliptic Kirchhoff type equations, Appl. Math. E-Notes. 20
(2020), 124–132.
[26] J. Rákosnı́k, Some remarks to anisotropic Sobolev spaces I, Beiträge Anal. 13 (1979), 55-68.
[27] J. Rákosnı́k, Some remarks to anisotropic Sobolev spaces II, Beiträge Anal. 15 (1981), 127–140.
[28] A. Razani and G. M. Figueiredo, Degenerated and competing anisotropic (p, q)-Laplacians with weights, Appl.
Anal. 102 (2023), no. 16, 4471–4488.
[29] A. Razani and G.M. Figueiredo, A positive solution for an anisotropic (p, q)-Laplacian, Discrete Contin. Dyn. Syst.
Ser. S 16 (2023), no. 6, 1629–1643.
[30] A. Razani and G.M. Figueiredo, Existence of infinitely many solutions for an anisotropic equation using genus
theory, Math. Meth. Appl. Sci. 45 (2022), no. 12, 7591-7606
[31] A. Razani, Nonstandard competing anisotropic (p, q)-Laplacians with convolution, Bound. Value Probl. 2022
(2022), 87.
[32] B. Ricceri, On an elliptic Kirchhoff-type problem depending on two parameters, J. Glob. Optim. 46 (2010), no. 4,
543–549.
[33] M. Růz̆ic̆ka, Electrorheological Fluids: Modeling and Mathematical Theory, Lecture Notes in Mathematics, Springer,
Berlin. 1748, 2000.
[34] T. Soltani and A. Razani, Solutions for an anisotropic elliptic problem involving nonlinear terms, Q. Math. (2023),
1-20. https://doi.org/10.2989/16073606.2023.2190045
[35] R. Stanway, J.L. Sproston, and A.K. El-Wahed, Applications of electro-rheological fluids in vibration control: A
survey, Smart Mater. Struct. 5 (1996), no. 4, 464-482.
[36] V.V. Zhikov, Averaging of functionals in the calculus of variations and elasticity, Math. USSR-Izvestiya 29 (1987),
no. 1, 33–66.

View publication stats