Home Mathematics Global existence, asymptotic behavior, and finite time blow up of solutions for a class of generalized thermoelastic system with p-Laplacian
Article Open Access

Global existence, asymptotic behavior, and finite time blow up of solutions for a class of generalized thermoelastic system with p-Laplacian

  • Yuxuan Chen EMAIL logo
Published/Copyright: November 15, 2025

Abstract

In this work, we examine the initial boundary value problem for the thermoelastic system with p -Laplacian, subject to a novel nonlinearity condition within a bounded domain. First, we introduce the blow-up solution for the problem under consideration for the negative initial energy. By introducing a set of potential wells, we construct invariant sets of solutions for the thermoelastic system with p -Laplacian. Under subcritical and critical initial energy scenarios, we derive the global existence and asymptotic behavior of weak solutions, as well as blow-up phenomena occurring within a finite time. Finally, we provide the lower and upper bounds of the blow-up time for the thermoelastic system with p -Laplacian.

MSC 2020: 35A01; 35D30; 35B40

1 Introduction

This article is concerned with the dynamical behavior of solutions for the following initial boundary value problem of the generalized thermoelastic system with p -Laplacian, which arises in the modeling of nonlinear elastic heat conducting body:

(1.1) u t t Δ p u + θ = f ( u ) , x Ω , t > 0 , θ t Δ θ = u t , x Ω , t > 0 , u ( x , t ) Ω = 0 , θ ( x , t ) Ω = 0 , t > 0 , u ( x , 0 ) = u 0 ( x ) , u t ( x , 0 ) = u 1 ( x ) , x Ω , θ ( x , 0 ) = θ 0 ( x ) , x Ω ,

where u = u ( x , t ) and θ = θ ( x , t ) denote the elongation of a plate and the temperature difference to the equilibrium state, respectively, and Δ p is the classical p -Laplacian operator.

Δ p u div ( u p 2 u ) , 2 p < n .

The initial data u 0 W 0 1 , p ( Ω ) , u 1 L 2 ( Ω ) , θ 0 H 0 1 ( Ω ) , the domain Ω R n is bounded, and its boundary Ω is smooth.

The nonlinearity f ( u ) satisfies the following assumption:

  1. f C 1 , f ( 0 ) = f ( 0 ) = 0 ;

  2. f ( u ) is monotone and convex for u > 0 , concave for u < 0 ;

  3. For p < α γ < n p n p , we have

    (1.2) u f ( u ) γ F ( u ) σ

    and

    (1.3) α F ( u ) u f ( u ) + β u p + α σ ,

    where F ( u ) = 0 u f ( s ) d s , σ > 0 and 0 < β λ 1 , p ( α p ) p , and λ 1 , p is the first eigenvalue of p -Laplacian.

Thermoelasticity has always been an extremely important subject in both theoretical and practical fields. As Dafermos mentioned [8], the theory of thermoelasticity is an appropriate model to explain the attenuation of the amplitude of free vibrations in certain elastic bodies. From a physical perspective, the theory of thermoelasticity is a combination of the theory of elasticity and the theory of heat conduction [15]. It involves the effect of heat on the deformation of an elastic medium, as well as the reverse effect of deformation on the thermal state of the medium. Thermal stresses arise when the time rate of change of the heat sources in the medium or the time rate of change of the thermal boundary conditions on the medium is compared with the characteristics of structural vibrations. In this case, the solution for the temperature and stress fields should be derived from the coupled thermoelastic equations [10]. From an application standpoint, thermoelasticity involves the interaction between the thermal and mechanical fields. In these fields, the phenomenon of thermomechanical coupling is extremely common in design and performance improvement, and often brings about many interesting phenomena. In fact, many new structures and devices often involve complex elastic materials and require operation in highly nonlinear states. For example, beams, plates, alloys, and composite materials are typical examples of such media. Therefore, the application needs of these materials in various fields, such as thermal protection systems, mechanical component design, heat treatment processes, and piezoelectric materials, have attracted widespread attention [23]. In classical thermodynamics, extensive research on nonlinear elasticity and thermoelasticity models can be found in the literature. This includes reviews, expositions, general models, and applications aimed at studying the effects of nonlinear coupling, such as those in [4,5,12,46,48].

Since the pioneering work of Dafermos [8] on linear thermoelasfticity, significant progress has been made on the mathematical aspect of thermoelasticity. In the beginning, people mainly considered the dynamical problems of classical thermoelastic systems [33], the 1D linear model of which is given as follows:

(1.4) ρ u t t a u x x + b θ x = 0 , k θ t a θ x x + b u x t = 0 ,

where ρ denotes the mass density, a the elasticity coefficient, b the stress-temperature, and k the heat conductivity. The functions u and θ are the displacement of the solid elastic material and the temperature difference. The study of existence of solutions to systems of thermoelasticity dates back to 1960s; however, the initial work was related to the linear simplification (linear thermoelasticity is addressed in detail shortly after). In 1960s, Dafermos [8] discussed the existence of solution of the classical thermoelastic system and showed the asymptotic stability of the system under certain condition. Slemrod [45] showed the global existence of smooth solution for small data when the boundary is either traction free and at a constant temperature or rigidly clamped and thermally insulated. In the case of Dirichlet boundary condition for which the boundary is rigidly clamped and held at a constant temperature, Racke and Shibata [43] proved global existence of a smooth solution. For large data for a class of systems of thermoelastic type with nonlocal nonlinearities, Mũnoz Rivera and Racke [34] investigated smoothing properties of following coupled systems in an abstract setting:

u t t + M ( A u 2 , θ 2 ) A 2 u + N ( A u 2 , θ 2 ) A θ = 0 , R ( A u 2 , θ 2 ) θ t + Q ( A u 2 , θ 2 ) A θ N ( A u 2 , θ 2 ) A u t = 0 ,

and obtained the existence of global solutions. Moreover, decay rates of Sobolev norms of the solutions as time tends to infinity are also investigated. Compared with the linear case, the study of existence of solution for nonlinear thermoelastic systems requires more sophisticated methods. We refer to the relevant work done by [31,32,37].

The asymptotic behavior of solutions to the thermoelastic equations as t + has always been a highly concerned issue in this field, attracting numerous scholars to conduct in-depth research. It is well known [16,22] that the one-dimensional linear thermoelastic system, associated with various types of boundary conditions, decays to zero exponentially. For one-dimensional models, it has been proven that the dissipation provided by the difference in temperature is strong enough to produce a uniform rate of decay of the solution, see Kim [22] and Marzocchi et al. [35]. However, the situation is different in two and three dimensions. In these cases, the displacement vector field has two or three degrees of freedom, while the difference in temperature, which produces the dissipative effect of the system, has only one degree of freedom. As a result, the dissipation is weaker than in the one-dimensional case, and a uniform rate of decay is not generally expected, as demonstrated by Henry [14]. For the multi-dimensional case, we have the pioneering work of Dafermos [8], who proved that the solution of the n -dimensional anisotropic thermoelastic material is asymptotically stable as time tends to infinity. However, the decay of the displacement is not to zero but to an undamped oscillation. On the other hand, the difference in temperature, as well as the divergence of the displacement, always tends to zero as time tends to infinity. The uniform rate of decay for the solution in two- or three-dimensional space was obtained by Jiang et al. [17] under special conditions, such as radial symmetry. Lebeau and Zuazua [25] showed that the decay rate is never uniform when the domain is convex. To address the lack of uniform decay in multi-dimensional cases, additional damping mechanisms are necessary. For the time-dependent system of linear thermoelasticity for isotropic bodies,

ρ ( x ) u t t μ ( x ) Δ u + μ ( x ) Δ ( Δ u ) Δ ( ( λ + 2 μ ) Δ u ) + 2 μ u μ [ u + ( u ) r ] + β ( γ ( x ) θ ) + α u t + q ( x ) u = f ( x , t ) , ρ ( x ) θ t ( k ( x ) θ ) + β γ ( x ) u t = g ( x , t ) ,

Pereira and Menzala [41] introduced a linear internal damping mechanism effective throughout the domain and established a uniform decay rate. Oliveira and Charão [38] improved upon the results in [41] by including a weak localized dissipative term effective only in a neighborhood of part of the boundary. They proved an exponential decay result when the damping term is linear and a polynomial decay result for a nonlinear damping term. For more studies focusing on the existence, regularity, and asymptotic behavior of solutions to the equations of nonlinear thermoelasticity, we refer the reader to books by Jiang and Racke [18] and Zheng [51].

Regarding the nonexistence results in the literature, there are numerous studies on this topic. These include the blowup rate of solutions to initial-boundary value problems for thermoelastic systems, in both one-dimensional [20,42] and multidimensional settings (see [1,36] and the references therein). Specifically, for thermoelasticity with second sound, Messaoudi [36] investigated the n -dimensional nonlinear thermoelastic system with a nonlinear source term competing with the damping factor,

u t t μ Δ u ( μ + λ ) div u + β θ = u p 2 u , θ t + γ div q + δ div u t = 0 , τ q t + q + κ θ = 0 .

He established a local existence result and demonstrated that solutions with negative energy blow up in finite time. Later, Qin and Rivera [42] focused on the Cauchy problem in nonautonomous nonlinear one-dimensional thermoelastic models that follow both Fourier’s law of heat flux and the theory proposed by Gurtin and Pipkin,

u t t = a u x + b θ x + d u x m u t + f ( u ) , c θ t = K θ x x + g θ x x + b u x t + p u x + q θ x .

They studied the blow-up phenomena of solutions in finite time when the initial energy is negative. Based on the results by Messaoudi [36], it has been shown that for a particular type of semilinear thermoelastic equations, weak solutions collapse in finite time. Almost at the same time, Kirane and Tatar [20] examined the Cauchy problem for a nonlinear system arising in thermoelasticity and proved that its solution develops singularities in finite time, depending on the size and regularity of the initial data. Their work is distinguished from previous studies by relaxing the requirements on the initial data, allowing for a slightly more general and nonautonomous forcing term, and permitting the insertion of gradient terms in both equations of the system. This result improves upon Messaoudi’s work by accommodating a more general set of initial data. To provide threshold critical exponents depending on the space dimension n , Boutefnouchet et al. [2] studied the nonexistence of global solutions to the Cauchy problem for systems of parabolic-hyperbolic or hyperbolic thermoelasticity equations posed in R N . For power nonlinearities, they demonstrated the nonexistence of global solutions for three systems of thermoelasticity using the nonlinear capacity method. Recently, Ding and Zhou [9] studied the thermoelastic system with p -Laplacian in a bounded domain. By using the potential well theory, they discussed the properties of global existence and finite time blow-up for the solutions, and give the lower and upper bounds of blow-up time to the blow-up solutions. For more details, one can refer the readers to [26,44,47,49] and the references therein.

The study of the long-time dynamic behavior of nonlinear thermoelastic systems is of great significance. In classical models, the hyperbolic elastic system and the parabolic heat conduction model are usually studied separately. However, in the fields of mechanics, physics, and engineering, many string vibration problems are often accompanied by thermal effects, which in turn influence the vibrational behavior, thereby forming an interesting coupled hyperbolic–parabolic system. This coupled system is not only significant in physical reality but also highly valuable in its mathematical properties. As a coupled hyperbolic-parabolic system, it possesses both the typical characteristics of hyperbolic systems and the features of parabolic systems, which makes its study highly attractive. One key question is: In such a coupled system, which characteristic, the hyperbolic or the parabolic, dominates? It turns out that the answer to this question depends on various factors, such as the spatial dimension, initial conditions, nonlinear indices, and the specific type of problem being studied [11,13,30,39]. When studying the relationship between initial conditions and the dynamic behavior of solutions in thermoelastic problems, there are certain particularities. The dissipative effect in thermoelastic problems can be caused not only by the system’s inherent damping but also by the mechanism of heat conduction. Therefore, a natural question arises: Is the dissipative effect caused by heat conduction strong enough to prevent the explosive growth of solutions under large data or at least under small data conditions? To answer this question, it is necessary to deeply explore the dependence of the long-time dynamic behavior of solutions of nonlinear thermoelastic systems on the initial data. Many previous studies have focused particularly on the exponential decay of solutions to thermoelastic systems, that is, the asymptotic behavior of solutions decaying to the equilibrium point (exponentially or polynomially) as time tends to infinity. These studies are undoubtedly of great significance. However, the focus of this article is different from that of previous literature. We are concerned with the dependence of the solutions of thermoelastic systems on the initial data. Specifically, we will discuss the different dynamic behaviors of solutions caused by changes in the magnitude of the initial data and systematically characterize the features of the initial data.

The goal of this article is to establish the result for the global existence of a weak solution and the blow-up phenomena in a finite time frame for the nonlinear system (1.1) in the context of initial energy states, as well as the long-time behavior of global solutions. Specifically, inspired by [6], we consider a class of generalized nonlinear sources that depend on the characteristics of the domain, i.e.,

(1.5) α 0 u f ( s ) d s u f ( u ) + β u p + α σ , u > 0 .

These nonlinear sources are particularly important in the study of p -Laplacian equations because they are closely related to the domain and eigenvalues. Moreover, this type of nonlinear source can cover a wider range of nonlinear types. In other words, it is important to note that the condition (1.5) for p = 2 on the nonlinearity f ( u ) includes the following cases:

(1.6) ( 2 + ε ) F ( u ) u f ( u ) ,

(1.7) ( 2 + ε ) F ( u ) u f ( u ) + σ ,

(1.8) ( 2 + ε ) F ( u ) u f ( u ) + β u 2 + σ ,

where 0 < β ε λ 1 2 , σ > 0 and F ( u ) = 0 u f ( s ) d s . Clearly, condition (1.8) can encompass conditions (1.6) and (1.7), where the constant β in condition (1.8) is dependent on the eigenvalues of the domain. If the first eigenvalue is arbitrarily small, then condition (1.8) approaches condition (1.7). Our proof strategy heavily hinges on two advanced methods in the variational analysis of nonlinear evolution equations: we combine the potential well method for proving global existence and nonexistence results in thermoelastic systems [9] with a variational approach to hyperbolic-parabolic partial differential equations [51], which allows us to consider the dissipation effect induced by internal damping through heat conduction. Indeed, the potential well method, introduced by Sattinger and Payne [40], can effectively establish the dependence of the long-time dynamic behavior of solutions on the initial data. So far, it can be said that the applications of the potential well method to the study of various nonlinear evolution equations are still relevant. In particular, research on generalized nonlinear sources is still ongoing. This is confirmed by the recently published papers devoted to this direction, for example, see [27,29] and the references therein.

We end this section with a discussion of the organization of the article. Section 2 introduces some preparatory lemmas, defines the family of stable and unstable sets, and provides a local well-posedness result. In Section 3, we discuss the finite time blowup of solutions under negative initial energy. Section 4 is devoted to presenting a sufficient condition for the global existence and finite time blowup of weak solutions under subcritical initial energy, including the decay of global solutions and the lower and upper bound estimates for the blowup time. In Section 5, we extend the results obtained for subcritical initial energy to the critical initial energy case.

2 Preliminaries

Throughout this article, we denote the standard Lebesgue space L q ( Ω ) and the Sobolev spaces W 0 m , q ( Ω ) ( 1 m < , 1 q ) with their usual scaler products and norms. In particular, we use H 0 m ( Ω ) = W 0 m , 2 ( Ω ) . As usual, ( , ) denotes L 2 -inner product, , denotes the dual paring between W 0 1 , p ( Ω ) and its dual W 1 , p ( Ω ) , and , p denotes the dual paring between L p ( Ω ) and its dual L p ( Ω ) , where p = p p 1 .

Definition 2.1

Let f ( u ) satisfy ( i ) ( i i i ) and u 0 W 0 1 , p ( Ω ) , u 1 L 2 ( Ω ) , θ 0 H 0 1 ( Ω ) . Suppose that u , θ : Ω × [ 0 , T ) R are two functions. The function ( u , θ ) is called a weak solution of system (1.1) on Ω × ( 0 , T ) , if u L ( 0 , T ; W 0 1 , p ( Ω ) ) , u t L ( 0 , T ; L 2 ( Ω ) ) and θ L 2 ( 0 , T ; H 0 1 ( Ω ) ) satisfies

(2.1) ( u t , ψ ) + 0 t u p 2 u , ψ p d τ + 0 t ( θ , ψ ) d τ = 0 t ( f ( u ) , ψ ) d τ + ( u 1 , ψ ) ,

(2.2) ( θ , φ ) + 0 t ( θ , φ ) d τ = 0 t ( u t , φ ) d τ + ( θ 0 , φ )

for a.e. t ( 0 , T ) and any ψ W 0 1 , p ( Ω ) , φ H 0 1 ( Ω ) .

Furthermore, the total energy of system (1.1) is denoted as follows:

(2.3) E ( t ) E ( u , u t , θ ) = 1 2 u t L 2 ( Ω ) 2 + 1 2 θ L 2 ( Ω ) 2 + 1 p u L p ( Ω ) p Ω ( F ( u ) σ ) d x .

Lemma 2.2

Let f ( u ) satisfy ( i ) ( i i i ) , the weak solution ( u , θ ) of system (1.1) satisfies

(2.4) E ( t ) + 0 t θ ( τ ) L 2 ( Ω ) 2 d τ = E ( 0 ) , 0 t < T .

Proof

Multiplying the equation (1.1) 1 and (1.1) 2 by u t and θ , respectively, then we have

(2.5) u t t , u t + u p 2 u , u t p + ( θ , u t ) = ( f ( u ) , u t )

and

(2.6) ( θ t , θ ) + ( θ , θ ) = ( u t , θ ) .

Differentiating and then integrating the energy functional E ( t ) , one can easy to verify that

(2.7) 0 t E ( τ ) d τ = 1 2 0 t Ω d d τ ( u t ( τ ) 2 + θ ( τ ) 2 ) d x d τ + 1 p 0 t Ω d d τ u ( τ ) p d x d τ 0 t Ω d d τ ( F ( u ( τ ) ) σ ) d x d τ = 0 t ( u τ τ ( τ ) , u τ ( τ ) ) + ( θ τ ( τ ) , θ ( τ ) ) + ( u ( τ ) p 2 u ( τ ) , ( u ( τ ) ) τ ) d τ 0 t ( F u ( u ( τ ) ) u τ ( τ ) ) d τ .

By substituting (2.6) into (2.5), then (2.7) turns into

0 t E ( τ ) d τ = 0 t [ u τ τ ( τ ) Δ p u ( τ ) f ( u ) ] ( u ( τ ) ) τ + ( θ ( τ ) , θ τ ( τ ) ) d τ = 0 t ( θ ( τ ) , θ ( τ ) ) d τ ,

which yields

E ( t ) + 0 t θ ( τ ) L 2 ( Ω ) 2 d τ = E ( 0 ) .

For the sake of completeness, we state here a local existence and uniqueness theorem for the nonlinear thermoelastic system (1.1). Its proof can be obtained by following a similar argument to the method mentioned in studies by Chen [3], Messaoudi [36], and Dafermos [8].

Proposition 2.3

(Local existence and uniqueness [3,36]) Let f ( u ) satisfy ( i ) ( i i i ) . Assume that 2 p < n , p < α γ < n p n p . Then given any u 0 W 0 1 , p ( Ω ) , u 1 L 2 ( Ω ) , θ 0 H 0 1 ( Ω ) , system (1.1) has a unique local solution satisfying u L ( 0 , T ; W 0 1 , p ( Ω ) ) , u t L ( 0 , T ; L 2 ( Ω ) ) and θ L 2 ( 0 , T ; H 0 1 ( Ω ) ) for T small enough.

Now, let’s introduce the potential energy functional J ( u ) and the Nehari functional I ( u ) related to problem (1.1) as follows:

J ( u ) = 1 p u L p ( Ω ) p Ω ( F ( u ) σ ) d x

and

I ( u ) = u L p ( Ω ) p Ω u f ( u ) d x ,

where u W 0 1 , p ( Ω ) , 2 p < n . The depth of potential well d is defined by

d inf u N J ( u ) ,

where the Nehari manifold N is defined by

N = { u W 0 1 , p ( Ω ) \ { 0 } I ( u ) = 0 } .

In addition, the depth of potential well d can be understood as a ground state energy, which is a positive constant by Remark 2.11.

Moreover, the Nehari manifold N helps us to divide the following two sets:

N + { u W 0 1 , p ( Ω ) \ { 0 } I ( u ) > 0 } and N { u W 0 1 , p ( Ω ) \ { 0 } I ( u ) < 0 } .

Lemma 2.4

Suppose that f ( u ) satisfy ( i ) ( i i i ) , then we obtain

  1. Q f ( Q ) ( p 1 ) f ( Q ) 0 holds for Q 0 , Q W 0 1 , p ( Ω ) ;

  2. F ( u ) σ A u γ for some A > 0 and any u R ;

  3. F ( u ) σ B u λ for u 1 with B = F ( u ˜ ) σ u ˜ λ > 0 and λ = α 1 + β > p .

Proof

We can use the basic variational method to prove (i). The details are as follows, according to the proof of Lemma 2.1 in Payne-Sattinger’s paper [40], we known that a critical point of functional

J ( u ) = 1 p Ω u p d x Ω F ( u ) d x ,

is a solution of Euler-Lagrange equation ( p -Poisson equation)

Δ p Q = f ( Q ) ,

where the existence and uniqueness of the solution can be found in Chapter 2 of [28]. Given that the functional J ( u ) is concave due to the convexity of f ( u ) for u > 0 , we can investigate its behavior at nontrivial critical points. Specifically, at a nontrivial critical point w , the second variation of J ( u ) , denoted by i ( 0 ) , will fail to be positive definite. In virtue of the fact i ( 0 ) 0 , we intend to deduce that Q f ( Q ) ( p 1 ) f ( Q ) 0 . Denote

i ( τ ) = J ( Q + τ v ) = 1 p Ω ( Q + τ v ) p d x Ω F ( Q + τ v ) d x .

Next, we calculus the second variation of i ( τ ) . The first derivative yields

i ( τ ) = Ω ( Q + τ v ) p 2 ( Q + τ v ) v d x Ω f ( Q + τ v ) v d x

and

i ( τ ) = ( p 2 ) Ω ( Q + τ v ) p 4 ( Q + τ v ) v 2 d x + Ω ( Q + τ v ) p 2 v 2 d x Ω v 2 f ( Q + τ v ) d x .

Then the second variation of J ( ) at a critical point Q is

i ( 0 ) = ( p 2 ) Ω Q p 4 Q v 2 d x + Ω Q p 2 v 2 d x .

Choosing the test function v to be Q , we then arrive at

i ( 0 ) = ( p 1 ) Ω Q p d x Ω Q 2 f ( Q ) d x = Ω Q [ ( p 1 ) ( Δ p Q ) Q f ( Q ) ] d x = Ω Q [ Q f ( Q ) ( p 1 ) f ( Q ) ] d x .

Due to i ( 0 ) 0 , we conclude that Q f ( Q ) ( p 1 ) f ( Q ) 0 .

(ii) By assumption ( i ) ( i i i ) on f ( u ) , the growth condition u f ( u ) γ F ( u ) σ tells us that

(2.8) f ( u ) F ( u ) σ γ u for u 0 .

Fixed u ˜ 0 , we can integrate (2.8) from u ˜ to u and then to obtain that

u ˜ u f ( s ) F ( s ) σ d s u ˜ u γ s d s .

With the help of F ( u ) = f ( u ) , we arrive at

ln F ( u ) σ F ( u ˜ ) σ ln u γ u ˜ γ ,

which means that F ( u ) σ A u γ , where A = F ( u ˜ ) σ u ˜ γ .

(iii) From assumption ( i ) ( i i i ) on f ( u ) , we obtain α F ( u ) u f ( u ) + β u p + α σ . Then, for any u 1 , one can deduce that

α u f ( u ) F ( u ) σ + β u p 1 F ( u ) σ ( 1 + β ) f ( u ) F ( u ) σ ,

then we obtain

(2.9) f ( u ) F ( u ) σ λ u ,

where λ = α 1 + β . As the method we used on (ii), it is easy to deduce that F ( u ) σ B u λ , where B = F ( u ˜ ) σ u ˜ λ .□

Lemma 2.5

Suppose that f ( u ) satisfy ( i ) ( i i i ) , u W 0 1 , p ( Ω ) , u L p ( Ω ) 0 , and

(2.10) ϕ ( ε ) = 1 ε p 1 Ω u f ( ε u ) d x .

Then the function ϕ satisfies the following properties:

  1. ϕ ( ε ) is increasing on 0 < ε < .

  2. lim ε 0 ϕ ( ε ) = 0 and lim ε + ϕ ( ε ) = + .

Proof

Taking the derivative of ϕ ( ε ) with respect to ε , we obtain

d ϕ ( ε ) d ε = 1 ε p Ω ( ε u 2 f ( ε u ) ( p 1 ) u f ( ε u ) ) d x = 1 ε p Ω u ( ε u f ( ε u ) ( p 1 ) f ( ε u ) ) d x > 0 .

Using (i) in Lemma 2.4, we know that ϕ ( ε ) is increasing on 0 < ε < .

The property lim ε 0 ϕ ( ε ) = 0 comes from the fact that

ϕ ( ε ) 1 ε p Ω ε u f ( ε u ) d x γ A ε p Ω ε u γ d x = γ ε γ p A u L γ ( Ω ) γ .

Furthermore, by using the fact u f ( u ) 0 and (iii) in Lemma 2.4 as follows:

ϕ ( ε ) = 1 ε p Ω ε u f ( ε u ) d x 1 ε p Ω ε ε u f ( ε u ) d x 1 ε p Ω ε ( α F ( ε u ) β ε p u p α σ ) d x α B ε λ p Ω ε u λ d x β Ω ε u p d x ,

where Ω ε = { x Ω u ( x ) 1 ε } and lim ε + Ω ε u λ d x = u L λ ( Ω ) λ > 0 , we then obtain that lim ε + ϕ ( ε ) = + . Hence, we conclude the results.□

If the nonlinearity source f ( u ) satisfies conditions ( i ) ( i i i ) , it follows that all nontrivial critical points are a priori unstable equilibria for the system (1.1). The subsequent lemma will establish the local minimum of J , identify a critical point ε * , and confirm the positiveness of the depth of the potential well d .

Lemma 2.6

Suppose that f ( u ) satisfy ( i ) ( i i i ) , u W 0 1 , p ( Ω ) with u L p ( Ω ) 0 . Then we have

  1. lim ε 0 J ( ε u ) = 0 .

  2. d d ε J ( ε u ) = 1 ε I ( ε u ) .

  3. lim ε + J ( ε u ) = .

  4. There exists a unique ε * = ε * ( u ) > 0 to show that

    (2.11) d d ε J ( ε u ) ε = ε * = 0 .

  5. J ( ε u ) is increasing on 0 ε ε * , decreasing on ε * ε < and takes the maximum at ε = ε * .

  6. I ( ε u ) > 0 for 0 < ε < ε * , I ( ε u ) < 0 for ε * < ε < and I ( ε * u ) = 0 .

Proof

(i) According to F ( u ) σ A u γ in Lemma 2.4 and

J ( ε u ) = ε p p Ω u p d x Ω ( F ( ε u ) σ ) d x ,

we obtain the result immediately.

(ii) By a simple calculations, we can easily obtain

d d ε J ( ε u ) = ε p 1 Ω u p d x Ω u f ( ε u ) d x = 1 ε I ( ε u ) .

(iii) Let

J ( ε u ) = ε p p Ω u p d x Ω ( F ( ε u ) σ ) d x ε p p Ω u p d x B ε λ Ω ε u λ d x .

Since λ > p , we obtain J ( ε u ) as ε + . This guarantees that there exists a critical point ε * .

(iv) Suppose that there two roots of equation d J ( ε u ) d ε = 0 as ε 1 < ε 2 . Then we arrive at

ε 1 p 1 Ω u p d x Ω u f ( ε 1 u ) d x = 0

and

ε 2 p 1 Ω u p d x Ω u f ( ε 2 u ) d x = 0 .

Removing the term Ω u p d x from the aforementioned equations yields

Ω u f ( ε 2 u ) ε 2 p 1 f ( ε 1 u ) ε 1 p 1 d x = 0 .

This expression can be rewritten by the following substitution Q = ε 1 u and ε = ε 2 ε 1 > 1 as follows:

(2.12) 1 ε p 1 Ω Q f ( ε Q ) d x = Ω Q f ( Q ) d x .

One can easily verify that equation (2.12) does not hold. Thus, we obtain the uniqueness of critical point.

Finally, according to Lemma 2.5, we deduce the case (v)–(vi), and

d d ε J ( ε u ) = 1 ε I ( ε u ) = ε p 1 Ω u p d x ϕ ( ε ) .

To introduce the family of potential wells G δ and δ , we begin to denote

I δ ( u ) = δ Ω u p d x Ω u f ( u ) d x , δ > 0

and

d ( δ ) inf { J ( u ) u W 0 1 , p ( Ω ) \ { 0 } , I δ ( u ) = 0 } .

Lemma 2.7

(Sobolev inequality) Let Ω be a bounded, open subset of R n , suppose Ω is C 1 . Assume that 1 < p < n , p * = n p n p , and f W 1 , p ( Ω ) . Then we have f L p * ( Ω ) , with the estimate f L p * ( Ω ) C 1 f W 1 , p ( Ω ) , where the constant C 1 depending only on p, n, and Ω .

Lemma 2.8

Suppose that f ( u ) satisfy ( i ) ( i i i ) , u W 0 1 , p ( Ω ) and

r ( δ ) = δ a C 2 γ 1 γ p and a = sup u f ( u ) u γ ,

where C 2 is the optimal embedding constant for W 0 1 , p ( Ω ) L γ ( Ω ) . One can deduce that

  1. if 0 < u L p ( Ω ) < r ( δ ) , then I δ ( u ) > 0 . In particular, for δ = 1 , if 0 < u L p ( Ω ) < r ( 1 ) , then I ( u ) > 0 .

  2. if I δ ( u ) < 0 , then u L p ( Ω ) > r ( δ ) . In particular, for δ = 1 , if I ( u ) < 0 , then u L p ( Ω ) > r ( 1 ) .

  3. if I δ ( u ) = 0 , then u L p ( Ω ) r ( δ ) . In particular, for δ = 1 , if I ( u ) = 0 , then u L p ( Ω ) r ( 1 ) .

Proof

  1. According to Lemma 2.7, we obtain from 0 < u L p ( Ω ) < r ( δ ) that

    Ω u f ( u ) d x Ω u f ( u ) d x a Ω u γ d x = a u L γ ( Ω ) γ a C 2 γ u L p ( Ω ) γ p u L p ( Ω ) p < δ u L p ( Ω ) p ,

    which means that I δ ( u ) > 0 .

  2. Since I δ ( u ) < 0 , we arrive at

    δ u L p ( Ω ) p < Ω u f ( u ) d x a u L γ ( Ω ) γ a C 2 γ u L p ( Ω ) γ p u L p ( Ω ) p ,

    which yields u L p ( Ω ) > r ( δ ) .

  3. According to I δ ( u ) = 0 , we obtain

    δ u L p ( Ω ) p = Ω u f ( u ) d x a u L γ ( Ω ) γ a C 2 γ u L p ( Ω ) γ p u L p ( Ω ) p ,

    which means that u L p ( Ω ) r ( δ ) .□

Lemma 2.9

(Theorem 1.1, [19]) There exists λ 1 , p > 0 and 0 < Q W 0 1 , p ( Ω ) in Ω R n such that

(2.13) ( Q ( x ) p 2 Q ( x ) ) + λ 1 , p Q ( x ) = 0 , x Ω , Q ( x ) = 0 , x Ω ,

where λ 1 , p is given by

λ 1 , p inf v W 0 1 , p ( Ω ) \ { 0 } Ω v p d x Ω v p d x > 0 .

In addition, one can recall that λ 1 , p denotes the first eigenvalue of the p-Laplacian operator, with Q being an associated eigenfunction.

Next lemma describes the properties of d ( δ ) for a family of potential wells.

Lemma 2.10

Suppose that f ( u ) satisfy ( i ) ( i i i ) . Then

  1. d ( δ ) > a ( δ ) r p ( δ ) for 0 < δ < α p β λ 1 , p , where a ( δ ) = 1 p δ α β λ 1 , p α .

  2. lim δ 0 d ( δ ) = 0 and there exists a unique b, α p β λ 1 , p b α p such that d ( b ) = 0 and d ( δ ) > 0 for 0 < δ < b .

  3. d ( δ ) is strictly increasing on 0 δ 1 , strictly decreasing on 1 δ < and takes the maximum d = d ( 1 ) at δ = 1 .

Proof

  1. If I δ ( u ) = 0 and u L p ( Ω ) 0 , then we obtain u L p ( Ω ) r ( δ ) via Lemma 2.8. By using I δ ( u ) = 0 , Lemmas 2.9 and 2.8, one can deduce that

    J ( u ) = 1 p u L p ( Ω ) p Ω ( F ( u ) σ ) d x 1 p u L p ( Ω ) p 1 α Ω u f ( u ) d x β α u L p ( Ω ) p 1 p δ α β λ 1 , p α u L p ( Ω ) p a ( δ ) r p ( δ ) ,

    where 0 < δ < α p β λ 1 , p . In addition, according to the assumption on α , β , we know that α p β λ 1 , p > 1 . Lemma 2.5

  2. tells us that for u W 0 1 , p ( Ω ) , u L p ( Ω ) 0 and each δ > 0 , one can denote a unique ε ( δ ) = ϕ 1 ( δ u L p ( Ω ) p ) to lead to

    (2.14) ε p ϕ ( ε ) = Ω ε u f ( ε u ) d x = δ ( ε u ) L p ( Ω ) p .

    Thus, for I δ ( ε u ) = 0 , we obtain

    lim δ 0 ε ( δ ) = 0 and lim δ + ε ( δ ) = + .

    From Lemma 2.6, one can deduce that

    lim δ 0 J ( ε u ) = lim ε 0 J ( ε u ) = 0 and lim δ 0 d ( δ ) = 0

    and

    lim δ + J ( ε u ) = lim ε + J ( ε u ) = and lim δ + d ( δ ) = .

    Obviously, we deduce the existence of d ( b ) = 0 for b α p β λ 1 , p and d ( δ ) > 0 for 0 < δ < b by using the part (i) and above expressions. In addition, applying the facts that u f ( u ) γ F ( u ) σ and I δ ( u ) = 0 , it is easy to obtain the upper bound of b γ p . Then we have

    J ( u ) = 1 p u L p ( Ω ) p Ω ( F ( u ) σ ) d x 1 p u L p ( Ω ) p γ 1 Ω u f ( u ) d x = 1 p δ γ u L p ( Ω ) p > 0 , for δ > γ p .

  3. Indeed, we only need to demonstrate that d ( δ ) < d ( δ ) for any 0 < δ < δ < 1 or 1 < δ < δ < b . For any given δ and δ satisfying 0 < δ < δ < 1 or 1 < δ < δ < ξ and any u W 0 1 , p ( Ω ) with I δ ( u ) = 0 and u L p ( Ω ) 0 , we now show that there exists a v W 0 1 , p ( Ω ) and a constant ε ( δ , δ ) > 0 such that

    J ( v ) > J ( u ) c ( δ , δ ) ,

    where v satisfies I δ ( v ) = 0 and v L p ( Ω ) 0 . To begin, we denote ε ( δ ) for u as (2.14), then I δ ( ε ( δ ) u ) = 0 and ε ( δ ) = 1 . Let g ( ε ) = J ( ε u ) , then we have

    d d ε g ( ε ) = 1 ε ( ε u ) L p ( Ω ) p Ω ε u f ( ε u ) d x = ( 1 δ ) ε p 1 u L p ( Ω ) p ,

    where one can noticed the fact that δ u L p ( Ω ) p = Ω u f ( u ) d x . Next, we pick v = ε ( δ ) u , then I δ ( v ) = 0 and v L p ( Ω ) 0 . For 0 < δ < δ < 1 , it yields

    J ( u ) J ( v ) = J ( ε ( δ ) u ) J ( ε ( δ ) u ) = g ( 1 ) g ( ε ( δ ) ) = ( 1 δ ) ε ( δ ) u L p ( Ω ) p ( 1 ε ( δ ) ) > ( 1 δ ) ε ( δ ) r p ( δ ) ( 1 ε ( δ ) ) c ( δ , δ ) .

    For 1 < δ < δ < b , we deduce

    J ( u ) J ( v ) = J ( ε ( δ ) u ) J ( ε ( δ ) u ) = g ( 1 ) g ( ε ( δ ) ) > ( δ 1 ) r p ( δ ) ε ( δ ) ( ε ( δ ) 1 ) .

Remark 2.11

According to Lemma 2.10, it is easily to verify that the depth of potential well is a positive constant, i.e.,

d 1 p δ α β λ 1 , p α r p ( 1 ) > 0 ,

where r ( 1 ) is defined in Lemma 2.8.

Lemma 2.12

Suppose f ( u ) satisfy ( i ) ( i i i ) and 0 < δ < α p β λ 1 , p . Then we conclude

  1. if J ( u ) d ( δ ) and I δ ( u ) > 0 , then

    0 < u L p ( Ω ) p < d ( δ ) a ( δ ) .

    In particular, for δ = 1 , if J ( u ) d and I ( u ) > 0 , then

    0 < u L p ( Ω ) p < d a ( 1 ) .

  2. if J ( u ) d ( δ ) and I δ ( u ) = 0 , then

    0 < u L p ( Ω ) p < d ( δ ) a ( δ ) .

    In particular, for δ = 1 , if J ( u ) d and I ( u ) = 0 , then

    0 < u L p ( Ω ) p < d a ( 1 ) .

  3. if J ( u ) d ( δ ) and u L p ( Ω ) p > d ( δ ) a ( δ ) , then I δ ( u ) < 0 . In particular, for δ = 1 , if J ( u ) d and u L p ( Ω ) p > d a ( 1 ) , then I ( u ) < 0 .

Proof

The case (i) follows from

J ( u ) = 1 p u L p ( Ω ) p Ω ( F ( u ) σ ) d x 1 p u L p ( Ω ) p 1 α Ω u f ( u ) d x β α Ω u p d x > 1 p δ α β λ 1 , p α u L p ( Ω ) p = a ( δ ) u L p ( Ω ) p .

The proof of cases (ii) and (iii) follows from a similar argument.□

Then a family of potential wells can be defined for 0 < δ < b as follows:

G δ { u W 0 1 , p ( Ω ) I δ ( u ) > 0 , J ( u ) < d ( δ ) } { 0 } , δ { u W 0 1 , p ( Ω ) I δ ( u ) < 0 , J ( u ) < d ( δ ) } .

In the case δ = 1 , we have

G { u W 0 1 , p ( Ω ) I ( u ) > 0 , J ( u ) < d } { 0 } , { u W 0 1 , p ( Ω ) I ( u ) < 0 , J ( u ) < d } .

From the definition of G δ , δ and Lemma 2.10, we derive the following properties:

Lemma 2.13

If 0 < δ < δ 1 and 1 δ < δ < b , then G δ G δ and δ δ , respectively.

Lemma 2.14

Suppose f ( u ) satisfy ( i ) ( i i i ) and u W 0 1 , p ( Ω ) satisfying 0 < J ( u ) < d . If δ 1 < δ 2 are two roots of the equation J ( u ) = d ( δ ) , then the sign of I δ ( u ) is unchangeable for δ 1 < δ < δ 2 .

Proof

From J ( u ) > 0 , it is easy to verify that u L p ( Ω ) 0 . Arguing by contradiction. we assume that the sign of I δ ( u ) is changeable for δ 1 < δ < δ 2 , then there exists a first δ * ( δ 1 , δ 2 ) leading to I δ * ( u ) = 0 . Then, in virtue of the definition of d ( δ ) , we know that J ( u ) d ( δ * ) , which contradicts

J ( u ) = d ( δ 1 ) = d ( δ 2 ) < d ( δ * ) .

Therefore, the assumption does not hold, and the original proposition is valid.□

3 Finite time blow up of the solution for system (1.1) with negative initial energy E ( 0 ) < 0

In this section, we discuss the finite time blow up of the solution to the thermoelastic system (1.1) with p -Laplacian under the negative initial energy level E ( 0 ) < 0 .

Lemma 3.1

[24] Suppose that Φ ( t ) C 2 ( [ 0 , ) ) is a positive function satisfying the following inequality

Φ ( t ) Φ ( t ) ( 1 + γ ) ( Φ ( t ) ) 2 2 C 1 Φ ( t ) Φ ( t ) C 2 Φ 2 ( t ) ,

where C 1 , C 2 0 and γ > 0 are constants. If

C 1 + C 2 > 0 , Φ ( 0 ) > 0 , Φ ( 0 ) + γ 2 γ 1 Φ ( 0 ) > 0 ,

then we have lim t t * Φ ( t ) = , where

t * t 2 * = 1 2 C 1 2 + γ C 2 log γ 1 Φ ( 0 ) + γ Φ ( 0 ) γ 2 Φ ( 0 ) + γ Φ ( 0 ) ,

and γ 1 = C 1 2 + γ C 2 C 1 , γ 2 = C 1 2 + γ C 2 C 1 .

Theorem 3.2

Let Ω be a bounded domain of R n with smooth boundary Ω . Suppose f ( u ) satisfy ( i ) ( i i i ) . If u 0 W 0 1 , p ( Ω ) , u 1 L 2 ( Ω ) and θ 0 H 0 1 ( Ω ) satisfy E ( 0 ) < 0 and ε 4 ( u 1 , u 0 ) > u 0 L 2 ( Ω ) 2 , ε > 0 , then there cannot exist a solution u of system (1.1) existing for all times T * , i.e., blow up in a finite time in the sense that

(3.1) lim t T * u ( t ) L 2 ( Ω ) 2 = + ,

where the maximal existence time T * satisfies the following estimate:

(3.2) 0 < T * ε 1 2 log ε ( u 1 , u 0 ) + u 0 L 2 ( Ω ) 2 ε ( u 1 , u 0 ) u 0 L 2 ( Ω ) 2 .

Proof

Assume that ( u , θ ) = ( u ( t ) , θ ( t ) ) is any weak solution of system (1.1) for t [ 0 , T ) with E ( 0 ) < 0 and ε 4 ( u 1 , u 0 ) > u 0 L 2 ( Ω ) 2 , ε > 0 , and T be the maximum existence time. Now we prove that u blows up in a finite time. Arguing by contradiction, we suppose T = and set

(3.3) Ψ ( t ) u ( t ) L 2 ( Ω ) 2 , t 0 ,

with the constant M 0 > 0 . Then we obtain

Ψ ( t ) = 2 ( u ( t ) , u t ( t ) ) , t > 0 .

By employing the condition (iii), Lemma 2.9 and 0 < β λ 1 , p ( α p ) p , we estimate the second derivative of Ψ ( t ) with respect to time

Ψ ( t ) = 2 u ( t ) , u t t ( t ) + 2 u t ( t ) L 2 ( Ω ) 2 = 2 Ω u ( t ) ( Δ p u ( t ) + u ( t ) f ( u ( t ) ) θ ( t ) ) d x + 2 u t ( t ) L 2 ( Ω ) 2 = 2 Ω u ( t ) p d x + 2 Ω u ( t ) f ( u ( t ) ) d x 2 Ω u ( t ) θ ( t ) d x + 2 u t ( t ) L 2 ( Ω ) 2 2 Ω u ( t ) p d x + 2 Ω ( α F ( u ( t ) ) β u p ( t ) α σ ) d x 2 Ω u ( t ) θ ( t ) d x + 2 u t ( t ) L 2 ( Ω ) 2 = 2 α 1 p Ω u ( t ) p d x Ω ( F ( u ( t ) ) σ ) d x + 2 α p 2 Ω u ( t ) p d x 2 β Ω u p ( t ) d x 2 Ω u ( t ) θ ( t ) d x + 2 u t ( t ) L 2 ( Ω ) 2 2 α 1 p Ω u ( t ) p d x Ω ( F ( u ( t ) ) σ ) d x + 2 α p 2 Ω u ( t ) p d x 2 β Ω u p ( t ) d x 1 α u ( t ) L 2 ( Ω ) 2 α θ ( t ) L 2 ( Ω ) 2 + 2 u t ( t ) L 2 ( Ω ) 2 2 α E ( t ) + 2 λ 1 , p ( α p ) p β Ω u p ( t ) d x 1 α u ( t ) L 2 ( Ω ) 2 + ( 2 + α ) u t ( t ) L 2 ( Ω ) 2 2 α E ( t ) 1 α u ( t ) L 2 ( Ω ) 2 .

Therefore, Ψ ( t ) can be rewritten as follows:

Ψ ( t ) 2 α E ( 0 ) + 2 α 0 t θ ( τ ) L 2 ( Ω ) 2 d τ 1 α u ( t ) L 2 ( Ω ) 2 2 α E ( 0 ) 1 α u ( t ) L 2 ( Ω ) 2 .

Then by taking ε > 0 and using Lemma 2.2, we deduce that

Ψ ( t ) Ψ ( t ) 4 + ε 4 ( Ψ ( t ) ) 2 Ψ ( t ) ( 2 α E ( 0 ) 1 α u ( t ) L 2 ( Ω ) 2 ) ( 4 + ε ) Ψ ( t ) u t ( t ) L 2 ( Ω ) 2 Ψ ( t ) 2 α E ( 0 ) 1 α u ( t ) L 2 ( Ω ) 2 ( 4 + ε ) u t ( t ) L 2 ( Ω ) 2 Ψ ( t ) ( 4 + ε ) θ ( t ) L 2 ( Ω ) 2 + 2 ( 4 + ε ) ( J ( u ) E ( 0 ) ) 2 α E ( 0 ) 1 α u ( t ) L 2 ( Ω ) 2 Ψ ( t ) 2 ( 4 + ε ) ( J ( u ) E ( 0 ) ) 2 α E ( 0 ) 1 α u ( t ) L 2 ( Ω ) 2 .

Using the assumption condition that E ( 0 ) < 0 , together with the decay property of the energy in Lemma 2.2, it follows that

E ( 0 ) E ( t ) = 1 2 u t ( t ) L 2 ( Ω ) 2 + 1 2 θ ( t ) L 2 ( Ω ) 2 + J ( u ) J ( u ) ,

thus we deduce

(3.4) Ψ ( t ) Ψ ( t ) 1 + ε 4 ( Ψ ( t ) ) 2 ( Ψ ( t ) ) 2 .

Taking C 1 = 0 , C 2 = 1 , γ = ε 4 , then we know that the inequality in Lemma 3.1 is satisfied. Moreover, in virtue of ε 4 ( u 1 , u 0 ) > u 0 L 2 ( Ω ) 2 , we have Ψ ( 0 ) + 4 ε Ψ ( 0 ) > 0 .

Hence, by Lemma 3.1, we see that the expression (3.4) for t 0 implies Ψ ( t ) for t t * < , which contradicts T = . Then u blows up in finite time. Moreover, similar to the aforementioned steps and using Lemma 3.1, one can easily to obtain

T ε 1 2 log ε ( u 1 , u 0 ) + u 0 L 2 ( Ω ) 2 ε ( u 1 , u 0 ) u 0 L 2 ( Ω ) 2 .

That completes the proof.□

4 Long-time behavior of the solution for system (1.1) with subcritical initial energy E ( 0 ) < d

In this section, we will discuss the initial data partition problem under the subcritical initial energy level. We will prove that when the initial energy is below the depth of the potential well E ( 0 ) < d , the solutions starting from any initial data have and only have two types of dynamical behaviors, namely, asymptotically tending to the zero equilibrium solution or blowing up in finite time. This provides the threshold conditions for the partition of initial data under the subcritical energy level. Specifically, we first use the Nehari flow to complete the invariance of the two sets inside and outside the potential well with respect to the initial data. Subsequently, we prove that when the initial data are located in the set N + , the solution exists globally and provide energy decay estimates. When the initial value is located in the set N , the solution will blow up in finite time. Finally, we give the upper and lower bounds estimates for the blow-up time of the blow-up solution.

4.1 Invariant sets of solutions

Theorem 4.1

Let f ( u ) satisfy ( i ) ( i i i ) , u 0 W 0 1 , p ( Ω ) , u 1 L 2 ( Ω ) and θ 0 H 0 1 ( Ω ) . Assume that 0 < e < d , then the equation d ( δ ) = e has two roots δ 1 < δ 2 , and we formulate the following properties,

  1. All solutions of problem (1) with E ( 0 ) = e belong to G δ for δ 1 < δ < δ 2 , provided I ( u 0 ) > 0 .

  2. All solutions of problem (1) with E ( 0 ) = e belong to δ for δ 1 < δ < δ 2 , provided I ( u 0 ) < 0 .

Proof

For case (i), we assume that ( u , θ ) is the weak solution of system (1.1) with E ( 0 ) = e and I ( u 0 ) > 0 . If I ( u 0 ) > 0 , then one can deduce that

(4.1) 0 t θ ( τ ) L 2 ( Ω ) 2 d τ + E ( t ) E ( 0 ) = d ( δ 1 ) = d ( δ 2 ) < d ( δ ) , δ 1 < δ < δ 2 .

According to (4.1) and the definition of d ( δ ) , it follows that I δ ( u 0 ) > 0 and E ( 0 ) < d ( δ ) , i.e, u 0 G δ for 0 < δ < b . Next, we claim that u ( t ) G δ for δ 1 < δ < δ 2 and 0 < t < T , where T is the maximal existence time of solution u . Arguing by contradiction, there exists a first time t 0 ( 0 , T ) such that u ( t 0 ) G δ 0 for some δ 0 ( δ 1 , δ 2 ) , and

I δ 0 ( u ( t 0 ) ) = 0 or J ( u ( t 0 ) ) = d ( δ 0 ) .

In virtue of (4.1), it easily verifies that J ( u ( t 0 ) ) = d ( δ 0 ) is impossible. If I δ 0 ( u ( t 0 ) ) = 0 , then by the definition of d ( δ 0 ) , we obtain that J ( u ( t 0 ) ) d ( δ 0 ) , which contradicts with (4.1).

For case (ii), we again employ that ( u , θ ) is the weak solution of system (1.1) with

E ( 0 ) = e and I ( u 0 ) < 0 .

As in the previous case, u 0 δ can be rebuilt by using Lemma 2.14 and (4.1).

The proof of u ( t ) δ for δ ( δ 1 , δ 2 ) follows from arguing by contradiction. Suppose that t 0 ( 0 , T ) is the first time such that u ( t ) δ for t [ 0 , t 0 ) and u ( t 0 ) G δ 0 , which means,

I δ 0 ( u ( t 0 ) ) = 0 or J ( u ( t 0 ) ) = d ( δ 0 )

for some δ 0 ( δ 1 , δ 2 ) . From (4.1), it follows that J ( u ( t 0 ) ) d ( δ 0 ) . If I δ 0 ( u ( t 0 ) ) = 0 , then I δ 0 ( u ( t ) ) < 0 for t ( 0 , t 0 ) and Lemma 2.8 yields

u ( t ) L p ( Ω ) > r ( δ 0 ) and u ( t 0 ) L p ( Ω ) r ( δ 0 ) .

Hence, by the definition of d ( δ 0 ) , we have J ( u ( t 0 ) ) d ( δ 0 ) , which contradicts (4.1).□

Now we construct a global weak solution for system (1.1) by the Galerkin approximation method. We now begin with the definition of the Krasnoselskii genus [21]. We denote by A the class of all closed subsets A E \ { 0 } that are symmetric with respect to the origin, that means, ϕ A implies ϕ A , where E is a Banach space.

Definition 4.2

Let A A . The Krasnoselskii genus ϖ ( A ) of A is defined as being the least positive integer m such that there is an odd mapping μ C ( A , R m ) such that μ ( x ) 0 for all x A . If such a m does not exist, we set ϖ ( A ) = . Furthermore, by definition, ϖ ( ) = 0 .

Lemma 4.3

([50], Theorem 3.1.1) Let 2 p < n and p < α γ < n p n p hold and T ( 0 , ) be fixed. Then the embedding

{ φ φ L 2 ( 0 , T ; W 0 1 , p ( Ω ) ) , φ t L 2 ( 0 , T ; L 2 ( Ω ) ) } L 2 ( 0 , T ; L γ ( Ω ) )

is compact.

Lemma 4.4

([50], Lemma 3.1.7 and Remark 3.1.4) Let B be a reflexive Banach space and 0 < T < . Suppose 1 < q < , φ L q ( 0 , T ; B ) , and the sequence { φ m } m = 1 L q ( 0 , T ; B ) satisfies, as m ,

φ m φ weakly i n L q ( 0 , T ; B ) , φ m t φ t weakly i n L q ( 0 , T ; B ) .

Then φ m ( 0 ) φ ( 0 ) weakly in B .

Lemma 4.5

Let y ( t ) : R + R + be a nonincreasing function, and assume that there is a constant A > 0 such that

(4.2) s + y ( t ) d t A y ( s ) , 0 s < + ,

then y ( t ) y ( 0 ) e 1 t A , for all t 0 .

4.2 Global existence of solutions

Theorem 4.6

(Global existence for E ( 0 ) < d ) Let f ( u ) satisfy ( i ) ( i i i ) , u 0 W 0 1 , p ( Ω ) , u 1 L 2 ( Ω ) , and θ 0 H 0 1 ( Ω ) . Assume that u 0 G . The initial-boundary value problem (1.1) admits a global weak solution u L ( 0 , ; W 0 1 , p ( Ω ) ) along with u t L 2 ( 0 , ; L 2 ( Ω ) ) and θ L 2 ( 0 , ; H 0 1 ( Ω ) ) . Moreover, the global weak solution have the following exponential asymptotic behavior:

(4.3) 1 2 u t ( t ) L 2 ( Ω ) 2 + 1 2 θ ( t ) L 2 ( Ω ) 2 + 0 t θ ( τ ) L 2 ( Ω ) 2 d τ + 1 p 1 α β λ 1 , p α u ( t ) L p ( Ω ) p E ( 0 ) e c t , t 0 ,

where c is a positive constant.

Proof

We divide this proof into two steps. Step 1: Global existence of solutions.

For j = 1 , 2 , , we let

ϒ j { Y { u L p ( Ω ) : u L p ( Ω ) = 1 } : Y is symmetric, compact and ϖ ( Y ) j } ,

where ϖ ( ) is the Krasnoselskii genus denoted by Definition 4.2. According to [7], we deduce

a j = inf Y ϒ j sup u Y u L p ( Ω ) p , j = 1 , 2 , ,

are the eigenvalues of the p -Laplacian. The operator Δ p : W 0 1 , p ( Ω ) W 1 , p ( Ω ) defined by

Δ p u , v = u p 2 u , v p = Ω u p 2 u v d x , u , v W 0 1 , p ( Ω ) ,

is a hemicontinuous, coercive, and monotone operator on W 0 1 , p ( Ω ) . According to Minty-Browder, one can deduce that there exists a basis { μ j } j = 1 of W 0 1 , p ( Ω ) satisfies the following eigenvalue problem:

Δ p μ j = a j μ j , x Ω , μ j = 0 , x Ω .

Moreover, it is easy to verify that there is a basis { ν j } j = 1 of H 0 1 ( Ω ) such that ν j is an eigenfunction of the Laplacian operator corresponding to the eigenvalue b j :

Δ ν j = b j ν j , x Ω , ν j = 0 , x Ω .

Hence, we can choose { μ j } j = 1 and { ν j } j = 1 as the Galerkin basis for Δ p and Δ in W 0 1 , p ( Ω ) and H 0 1 ( Ω ) . Then we construct the approximate solutions u m ( t ) and θ m ( t ) for problem (1.1):

(4.4) u m = j = 1 m g j m ( t ) μ j ( x ) , θ m = j = 1 m f j m ( t ) ν j ( x ) , m = 1, 2 , ,

which satisfy, for j = 1, 2 , , m ,

(4.5) u m t t , μ j + u m p 2 u m , μ j p + ( θ m , μ j ) = ( f ( u m ) , μ j ) , ( θ m t , ν j ) + ( θ m , ν j ) = ( u m t , ν j )

with initial conditions u m ( 0 ) = u 0 m , u m t ( 0 ) = u 1 m , and θ m ( 0 ) = θ 0 m , where u 0 m , u 1 m , and θ 0 m are chosen in span { μ 1 , , μ m } and span { ν 1 , , ν m } such that

(4.6) u 0 m u 0 in W 0 1 , p ( Ω ) , as m ,

(4.7) u 1 m u 1 in L 2 ( Ω ) , as m ,

(4.8) θ 0 m θ 0 in H 0 1 ( Ω ) , as m .

According to Carathéodory’s theorem, we find that (4.5) lead to a system of ordinary differential equations in the variable t that has a local solution ( u m ( t ) , θ m ( t ) ) on [ 0 , T m ] .

Next, multiplying (4.5) by g j m ( t ) and f j m ( t ) , respectively, and summing for j from 1 to m , we reach

(4.9) u m t t , u m t + u m p 2 u m , u m t p + ( θ m , u m t ) = ( f ( u m ) , u m t ) ,

(4.10) ( θ m t , θ m ) + ( θ m , θ m ) = ( u m t , θ m ) .

Substituting (4.10) into (4.9), we obtain d d t E m ( t ) = θ m ( t ) L 2 ( Ω ) 2 , i.e.,

(4.11) E m ( t ) + 0 t θ m ( τ ) L 2 ( Ω ) 2 d τ = E m ( 0 ) ,

where

E m ( t ) = 1 2 u m t ( t ) L 2 ( Ω ) 2 + 1 2 θ m ( t ) L 2 ( Ω ) 2 + J ( u m ( t ) ) ,

and J ( ) is the function defined in (1.5).

From E ( 0 ) < d , (4.6)–(4.8), we see that E m ( 0 ) < d for sufficiently large m . Thus, we obtain from (3.8) that

(4.12) 1 2 u m t ( t ) L 2 ( Ω ) 2 + 1 2 θ m ( t ) L 2 ( Ω ) 2 + J ( u m ( t ) ) + 0 t θ m ( t ) L 2 ( Ω ) 2 d τ < d , 0 t T m ,

holds for sufficiently large m . Since I ( u 0 ) > 0 , we have u 0 G . Then it follows from (4.6) that u m ( 0 ) G for sufficiently large m . Next, we plan to prove that u m ( t ) G for any t [ 0 , T m ] and sufficiently large m . Arguing by contradiction, there exist a first time t 0 ( 0 , T m ] and a sufficiently large m leading to I ( u m ( t 0 ) ) = 0 and u m ( t 0 ) 0 , then we deduce that u m ( t 0 ) N . So we obtain J ( u m ( t 0 ) ) d , which contradicts (4.12). Thus, u m ( t ) W for any t [ 0 , T m ] and sufficiently large m , which yields I ( u m ( t ) ) 0 for any t [ 0 , T m ] and sufficiently large m .

Since (4.12), we know that I ( u m ( t ) ) 0 for sufficiently large m and from (4.12) and

J ( u m ) = 1 p u m L p ( Ω ) p Ω ( F ( u m ) σ ) d x 1 p u m L p ( Ω ) p 1 α Ω u m f ( u m ) d x β α Ω u m p d x 1 p 1 α β λ 1 , p α u m L p ( Ω ) p + 1 α I ( u m ) ,

we establish

(4.13) 1 2 u m t ( t ) L 2 ( Ω ) 2 + 1 2 θ m ( t ) L 2 ( Ω ) 2 + 0 t θ m ( τ ) L 2 ( Ω ) 2 d τ + 1 p 1 α β λ 1 , p α u m ( t ) L p ( Ω ) p < d , t [ 0 , )

for sufficiently large k . Thus, for 0 t < , (4.13) tells us the following estimates:

(4.14) Ω u m p 2 u m p p 1 d x = u m L p ( Ω ) p < 1 p 1 α β λ 1 , p α 1 d ,

(4.15) u m L γ ( Ω ) γ C 2 γ u m L p ( Ω ) γ < C 2 γ 1 p 1 α β λ 1 , p α γ p d γ p ,

(4.16) 0 t θ m L 2 ( Ω ) 2 d τ < d ,

(4.17) u m t L 2 ( Ω ) 2 2 d ,

(4.18) θ m L 2 ( Ω ) 2 2 d ,

and

(4.19) f ( u m ) L γ γ 1 ( Ω ) γ γ 1 Ω ( γ A u m γ 1 ) γ γ 1 d x = γ γ γ 1 A γ γ 1 u m L γ ( Ω ) γ < γ γ γ 1 A γ γ 1 C 2 γ 1 p 1 α β λ 1 , p α γ p d γ p ,

where C 2 is the optimal constant of the Sobolev embedding W 0 1 , p ( Ω ) L γ ( Ω ) .

Having in mind the aforementioned energy estimates, one can deduce that the sequence { u m } m = 1 is bounded in L ( 0 , ; W 0 1 , p ( Ω ) ) , { u m t } m = 1 is bounded in L ( 0 , ; L 2 ( Ω ) ) , and further { θ m } m = 1 is bounded in L ( 0 , ; H 0 1 ( Ω ) ) .

Moreover, according to (4.14)–(4.18) for each T ˜ > 0 , there exist u , θ and the subsequences of { u m } m = 1 and { θ m } m = 1 , still denoted by { u m } m = 1 and { θ m } m = 1 , respectively, such that (as m )

(4.20) u m t u t weakly star in L ( 0 , T ˜ ; L 2 ( Ω ) ) ,

(4.21) u m t u t weakly in L 2 ( 0 , T ˜ ; L 2 ( Ω ) ) ,

(4.22) θ m θ weakly star in L ( 0 , T ˜ ; L 2 ( Ω ) ) ,

(4.23) θ m θ weakly in L 2 ( 0 , T ˜ ; H 0 1 ( Ω ) ) ,

(4.24) u m u weakly star in L ( 0 , T ˜ ; W 0 1 , p ( Ω ) ) ,

(4.25) u m u weakly in L 2 ( 0 , T ˜ ; W 0 1 , p ( Ω ) ) .

By (4.21), (4.25), and Lemma 4.3, we obtain

(4.26) u m u strongly in L 2 ( 0 , T ˜ ; L γ ( Ω ) ) .

Then

u m ( x , t ) u ( x , t ) for a.e. ( x , t ) Ω × ( 0 , T ˜ ) ,

which, combining with (4.14) and (4.19), yields

(4.27) u m p 2 u m u p 2 u weakly star in L ( 0 , T ˜ ; L p p 1 ( Ω ) ) ,

(4.28) u m p 2 u m u p 2 u weakly in L 2 ( 0 , T ˜ ; L p p 1 ( Ω ) ) ,

(4.29) f ( u m ) f ( u ) weakly star in L ( 0 , T ˜ ; L γ γ 1 ( Ω ) ) ,

(4.30) f ( u m ) f ( u ) weakly in L 2 ( 0 , T ˜ ; L γ γ 1 ( Ω ) ) .

Next, we integrate the two equations of (4.5) over [ 0 , t ] , then, for j = 1 , , m , it follows that

( u m t , μ j ) + 0 t u m p 2 u m , μ j p d τ + 0 t ( θ m , μ j ) d τ = 0 t ( f ( u m ) , μ j ) d τ + ( u m t ( 0 ) , μ j ) ,

(4.31) ( θ m , ν j ) + 0 t ( θ m , ν j ) d τ = 0 t ( u m τ , ν j ) d τ + ( θ m ( 0 ) , ν j ) .

By (4.21) and (4.23), we obtain that, for each T ˜ > 0 ,

0 T ˜ ( u m t , μ j ) d t 0 T ˜ ( u t , μ j ) d t , 0 T ˜ ( θ m , ν j ) d t 0 T ˜ ( θ , ν j ) d t , as m ,

which tells us that, for a.e. t [ 0 , ) ,

( u m t , μ j ) ( u t , μ j ) , ( θ m , ν j ) ( θ , ν j ) , as m .

Therefore, making m in (3.25), and from (4.21), (4.23), (4.25), (4.28), and (4.30), one can deduce that, for a.e. t [ 0 , ) ,

( u t , μ j ) + 0 t u p 2 u , μ j p d τ + 0 t ( θ , μ j ) d τ = 0 t ( f ( u ) , μ j ) d τ + ( u t ( 0 ) , μ j ) , ( θ , ν j ) + 0 t ( θ , ν j ) d τ = 0 t ( u τ , ν j ) d τ + ( θ ( 0 ) , ν j ) , j = 1, 2 , .

Since { μ j } j = 1 and { ν j } j = 1 are dense in W 0 1 , p ( Ω ) and H 0 1 ( Ω ) respectively, then we obtain

(4.32) ( u t , ψ ) + 0 t u p 2 u , ψ p d τ + 0 t ( θ , ψ ) d τ = 0 t ( f ( u ) , ψ ) d τ + ( u t ( 0 ) , ψ ) ,

( θ , ϕ ) + 0 t ( θ , ϕ ) d τ = 0 t ( u τ , ϕ ) d τ + ( θ ( 0 ) , ϕ ) ,

hold for any ψ W 0 1 , p ( Ω ) , ϕ H 0 1 ( Ω ) and a.e. t ( 0 , ) .

Now, for sufficiently small ε > 0 , we define

(4.33) S { h ( t ) ψ ( x ) h ( t ) L 2 ( 0 , ε ) , ψ ( x ) W 0 1 , p ( Ω ) } ,

then S is dense in L 2 ( 0 , ε ; W 0 1 , p ( Ω ) ) . Now, we take

H = j = 1 m α j μ j α j R , m N .

Since { μ j } j = 1 is a basis of W 0 1 , p ( Ω ) , for each ψ W 0 1 , p ( Ω ) , there exist a sequence { ψ j } j = 1 H leading to lim j ψ j = ψ in W 0 1 , p ( Ω ) . By using the definition of H , we show that the first equation of (4.5) is true by replacing μ j with ψ j , which means,

(4.34) u m t t , ψ j + u m p 2 u m , ψ j p + ( θ m , ψ j ) = ( f ( u m ) , ψ j ) .

Note lim j ψ j = ψ in W 0 1 , p ( Ω ) , making j in (4.34), one can deduce

(4.35) u m t t , ψ + u m p 2 u m , ψ p + ( θ m , ψ ) = ( f ( u m ) , ψ ) .

By multiplying the both sides of (4.35) by h ( t ) for a.e. t 0 , we arrive at

u m t t , h ( t ) ψ + u m p 2 u m , h ( t ) ψ p + ( θ m , h ( t ) ψ ) = ( f ( u m ) , h ( t ) ψ ) .

Therefore, employing (4.14)–(4.18), and Hölder’s inequality, it is easy to obtain

(4.36) 0 ε u m τ τ , h ( τ ) ψ d τ 0 ε u m p 2 u m , h ( τ ) ψ p d τ + 0 ε ( θ m , h ( τ ) ψ ) d τ + 0 ε ( u m f ( u m ) , h ( τ ) ψ ) d τ 0 ε u m L p ( Ω ) 2 p 2 d τ 1 2 0 ε h ( τ ) ψ L p ( Ω ) 2 d τ 1 2 + 0 ε θ m L 2 ( Ω ) 2 d τ 1 2 0 ε h ( τ ) ψ L 2 ( Ω ) 2 d τ 1 2 + 0 ε u m L γ ( Ω ) 2 ( γ 1 ) d τ 1 2 0 ε h ( τ ) ψ L γ ( Ω ) 2 d τ 1 2 ε 1 2 p d γ γ p p 1 p 0 ε h 2 ( τ ) d τ 1 2 ψ L p ( Ω ) + ( 2 d ε ) 1 2 0 ε h 2 ( τ ) d τ 1 2 ψ L 2 ( Ω ) + ε 1 2 C d γ 1 γ 0 ε h 2 ( τ ) d τ 1 2 ψ L γ ( Ω ) .

Since W 0 1 , p ( Ω ) L 2 ( Ω ) and W 0 1 , p ( Ω ) L γ ( Ω ) , then by (4.36), there exists a constant C 3 > 0 such that

(4.37) 0 ε u m τ τ , h ( τ ) ψ d τ C 3 ε 1 2 p d γ γ p p 1 p + ( 2 d ε ) 1 2 + ε 1 2 C d γ 1 γ 0 ε h 2 ( τ ) d τ 1 2 ψ L p ( Ω ) .

Since S is dense in L 2 ( 0 , ε ; W 0 1 , p ( Ω ) ) , where S is the set defined in (4.33), by (4.37), it follows that

u m t t L 2 ( 0 , ε ; W 1 , p ( Ω ) ) C 3 ε 1 2 p d γ γ p p 1 p + ( 2 d ε ) 1 2 + ε 1 2 C d γ 1 γ ,

i.e., u m t t is bounded in L 2 ( 0 , ε ; W 1 , p ( Ω ) ) and then there exists a subsequence of { u m t t } m = 1 denoted by { u m t t } m = 1 again such that

(4.38) u m t t u t t weakly in L 2 ( 0 , ε ; W 1 , p ( Ω ) ) as m .

Below, for sufficiently small ε > 0 , we denote

(4.39) Q { h ( t ) ϕ ( x ) h ( t ) L 2 ( 0 , ε ) , ϕ ( x ) H 0 1 ( Ω ) } ,

hence, Q is dense in L 2 ( 0 , ε ; H 0 1 ( Ω ) ) . Furthermore, we take

K = j = 1 m α j ν j α j R , m N .

Since { ν j } j = 1 is a basis of H 0 1 ( Ω ) , for any ϕ H 0 1 ( Ω ) , there exist a sequence { ϕ j } j = 1 K such that lim j ϕ j = ϕ in H 0 1 ( Ω ) . By the definition of K , it is easy to see that the second equation of (4.5) is true by replacing ν j with ϕ j , i.e.,

( θ m t , ϕ j ) + ( θ m , ϕ j ) = ( u m t , ϕ j ) .

Note lim j ϕ j = ϕ in H 0 1 ( Ω ) , letting j in the above equality, we reach

(4.40) ( θ m t , ϕ ) + ( θ m , ϕ ) = ( u m t , ϕ ) .

By multiplying the both sides of (4.40) by h ( t ) , we show that

( θ m t , h ( t ) ϕ ) + ( θ m , h ( t ) ϕ ) = ( u m t , h ( t ) ϕ ) .

Therefore, from (4.16), (4.17), and Hölder’s inequality, it follows that

0 ε θ m τ , h ( τ ) ϕ * d τ 0 ε ( θ m , h ( τ ) ϕ ) d τ + 0 ε ( u m τ , h ( τ ) ϕ ) d τ 0 ε θ m L 2 ( Ω ) 2 d τ 1 2 0 ε h ( τ ) ϕ L 2 ( Ω ) 2 d τ 1 2 + 0 ε u m τ L 2 ( Ω ) 2 d τ 1 2 0 ε h ( τ ) ϕ L 2 ( Ω ) 2 d τ 1 2 d 1 2 0 ε h 2 ( τ ) d τ 1 2 ϕ L 2 ( Ω ) + ( 2 d ε ) 1 2 0 ε h 2 ( τ ) d τ 1 2 ϕ L 2 ( Ω ) ,

where , * denotes the dual pairing between H 0 1 ( Ω ) and H 1 ( Ω ) .

Due to the embedding H 0 1 ( Ω ) L 2 ( Ω ) , then by the aforementioned inequality, there exists a C 4 > 0 such that

0 ε θ m τ , h ( τ ) ϕ * d τ C 4 ( d 1 2 + ( 2 d ε ) 1 2 ) 0 ε h 2 ( τ ) d τ 1 2 ϕ L 2 ( Ω ) .

By using the fact that Q is dense in L 2 ( 0 , ε ; H 0 1 ( Ω ) ) , where Q is defined in (4.39), by the aforementioned inequality, we obtain

θ m t L 2 ( 0 , ε ; H 1 ( Ω ) ) C 4 ( d 1 2 + ( 2 d ε ) 1 2 ) ,

i.e., θ m t is bounded in L 2 ( 0 , ε ; H 1 ( Ω ) ) and then there exists a subsequence of { θ m t } m = 1 denoted by { θ m t } m = 1 again such that

(4.41) θ m t θ t weakly in L 2 ( 0 , ε ; H 1 ( Ω ) ) as m .

In view of (4.38), (4.41), (4.21), (4.23), and (4.25), by using Lemma 4.4, we have u m ( 0 ) u ( 0 ) weakly in L 2 ( Ω ) , and u m t ( 0 ) u t ( 0 ) weakly in W 1 , p ( Ω ) , and θ m ( 0 ) θ ( 0 ) weakly in H 1 ( Ω ) .Then by (4.4) and (4.6)–(4.8), we obtain u ( 0 ) = u 0 W 0 1 , p ( Ω ) , u t ( 0 ) = u 1 L 2 ( Ω ) and θ ( 0 ) = θ 0 H 0 1 ( Ω ) . Hence, it follows from (4.32) that

( u t , ψ ) + 0 t u p 2 u , ψ p d τ + 0 t ( θ , ψ ) d τ = 0 t ( f ( u ) , ψ ) d τ + ( u 1 , ψ ) , ( θ , ϕ ) + 0 t ( θ , ϕ ) d τ = 0 t ( u τ , ϕ ) d τ + ( θ 0 , ϕ ) .

hold for a.e. t ( 0 , ) and each ψ W 0 1 , p ( Ω ) , ϕ H 0 1 ( Ω ) .

To the end, we claim that (2.4) holds for a.e. t ( 0 , ) . By arbitrariness of T ˜ ( 0 , ) , (4.26), and (4.30), we verify that for a.e. t ( 0 , )

(4.42) u m u strongly in L γ γ 1 ( Ω ) as m ,

(4.43) Ω ϕ ( f ( u m ) f ( u ) ) d x 0 as m for any ϕ L γ γ 1 ( Ω ) .

Note (1.3), (4.24), and arbitrariness of T ˜ ( 0 , ) , we obtain u ( t ) W 0 1 , p ( Ω ) for a.e. fixed t ( 0 , ) , which is continuously embedding in L γ γ 1 ( Ω ) . Thus, using (4.43), we obtain, for a.e. t ( 0 , )

(4.44) Ω u ( f ( u m ) f ( u ) ) d x 0 as m .

Therefore, by (4.19), (4.42), (4.44), and Hölder’s inequality, one can obtain, for a.e. t ( 0 , ) ,

(4.45) Ω u m f ( u m ) d x Ω u f ( u ) d x Ω ( u m u ) f ( u m ) d x + Ω u ( f ( u m ) f ( u ) ) d x C 5 γ γ 1 u m u L γ γ 1 ( Ω ) + Ω u ( f ( u m ) f ( u ) ) d x 0

as m , where C 5 is a general positive constant. From the convergence of u m , u m t , θ m , and the definition of E ( t ) , we conclude that

1 2 u t ( t ) L 2 ( Ω ) 2 + 1 2 θ ( t ) L 2 ( Ω ) 2 + 1 p u ( t ) L p ( Ω ) p + 0 t θ ( t ) L 2 ( Ω ) 2 d τ liminf m 1 2 u m t ( t ) L 2 ( Ω ) 2 + 1 2 θ m ( t ) L 2 ( Ω ) 2 + 1 p u m ( t ) L p ( Ω ) p + 0 t θ m ( t ) L 2 ( Ω ) 2 d τ liminf m E m ( t ) + γ γ γ 1 A γ γ 1 u m ( t ) L γ ( Ω ) γ + 0 t θ m ( t ) L 2 ( Ω ) 2 d τ = lim m ( E m ( 0 ) + γ γ γ 1 A γ γ 1 u m ( t ) L γ ( Ω ) γ ) = E ( 0 ) + γ γ γ 1 A γ γ 1 u ( t ) L γ ( Ω ) γ .

Step 2: Exponential asymptotic behavior of solutions.

By using Theorem 4.1, we have u ( t ) N + for any t 0 and then I δ ( u ( t ) ) > 0 . Due to Lemma 2.6, one can infer from I ( u ) 0 that there is a constant ε * > 1 such that I δ ( ε * u ( t ) ) = 0 . Combining this fact with the following equality

I δ ( ε * u ( t ) ) = δ ( ε * ) p u ( t ) L p ( Ω ) p Ω ε * u ( t ) f ( ε * u ( t ) ) d x = δ ( ( ε * ) p ( ε * ) γ ) u ( t ) L p ( Ω ) p + ( ε * ) γ I δ ( u ( t ) ) ,

we obtain that

(4.46) I δ ( u ( t ) ) = δ ( 1 ( ε * ) p γ ) u ( t ) L p ( Ω ) p .

Next, we estimate the constant ε * . By variational characterization of d ( δ ) , we know

(4.47) d ( δ ) J ( ε * u ( t ) ) = ( ε * ) p p u ( t ) L p ( Ω ) p Ω ( F ( ε * u ( t ) ) σ ) d x ( ε * ) p p u ( t ) L p ( Ω ) p 1 γ Ω ε * u ( t ) f ( ε * u ( t ) ) d x 1 γ I δ ( ε * u ( t ) ) + ( ε * ) p 1 p δ p γ u ( t ) L p ( Ω ) p ( ε * ) p ( γ δ ) γ p u ( t ) L p ( Ω ) p .

On the other hand, by the nonincreasing property of energy functional E ( t ) in Lemma 2.2, we have

(4.48) E ( 0 ) E ( t ) J ( u ( t ) ) = 1 p u ( t ) L p ( Ω ) p Ω ( F ( u ( t ) ) σ ) d x 1 p u ( t ) L p ( Ω ) p 1 α Ω u ( t ) f ( u ( t ) ) d x β α Ω u ( t ) p d x 1 p δ α β λ 1 , p α u ( t ) L p ( Ω ) p + 1 α I δ ( u ( t ) ) 1 p δ α β λ 1 , p α u ( t ) L p ( Ω ) p .

Together with (4.47) and (4.48), it follows from Theorem 4.1 and E ( 0 ) < d that

(4.49) ε * d ( δ ) E ( 0 ) 1 p > 1 .

Thus, in virtue of (4.46) and (4.49), we claim that

(4.50) I δ ( u ( t ) ) 1 d ( δ ) E ( 0 ) p γ p δ u ( t ) L p ( Ω ) p .

Moreover, as a consequence, it follows from the definition of I δ ( u ) that

(4.51) Ω u ( t ) f ( u ( t ) ) d x = δ u ( t ) L p ( Ω ) p I δ ( u ( t ) ) δ u ( t ) L p ( Ω ) p 1 d ( δ ) E ( 0 ) p γ p δ u ( t ) L p ( Ω ) p = δ d ( δ ) E ( 0 ) p γ p u ( t ) L p ( Ω ) p .

To prove the asymptotic behavior of solutions, we now denote the following Lyapunov functional:

(4.52) L ( t ) = K 1 E ( t ) + K 2 ( u t ( t ) , u ( t ) ) , t 0 ,

where K 1 and K 2 are constants determined later. By a simple calculations, we obtain

(4.53) d d t ( u t ( t ) , u ( t ) ) = u t ( t ) L 2 ( Ω ) 2 + u t t ( t ) , u ( t ) = u t ( t ) L 2 ( Ω ) 2 u ( t ) L p ( Ω ) p + Ω u ( t ) f ( u ( t ) ) d x ( θ ( t ) , u ( t ) ) .

With the help of the Hölder and ε -Young inequalities, we have

(4.54) ( θ , u ) 1 2 ε ˜ θ L 2 ( Ω ) 2 + ε ˜ 2 u L 2 ( Ω ) 2 ,

for any ε ˜ > 0 . Hence, together with (4.53) and (4.54), one has

(4.55) d d t ( u t ( t ) , u ( t ) ) u t ( t ) L 2 ( Ω ) 2 u ( t ) L p ( Ω ) p + Ω u ( t ) f ( u ( t ) ) d x + ε ˜ 2 u ( t ) L 2 ( Ω ) 2 + 1 2 ε ˜ θ ( t ) L 2 ( Ω ) 2 .

According to (4.51), we deduce from (4.55) that

(4.56) d d t ( u t ( t ) , u ( t ) ) u t ( t ) L 2 ( Ω ) 2 1 δ d ( δ ) E ( 0 ) p γ p u ( t ) L p ( Ω ) p + ε ˜ 2 u ( t ) L 2 ( Ω ) 2 + 1 2 ε ˜ θ ( t ) L 2 ( Ω ) 2 .

Next, we plan to give the estimate for L ( t ) as follows. By the definition of L ( t ) in (4.52), the energy inequality (2.4), we show that

(4.57) L ( t ) = K 1 E ( t ) + K 2 d d t ( u t ( t ) , u ( t ) ) K 1 θ ( t ) L 2 ( Ω ) 2 + K 2 u t ( t ) L 2 ( Ω ) 2 K 2 1 δ d ( δ ) E ( 0 ) p γ p u ( t ) L p ( Ω ) p + K 2 ε ˜ 2 u ( t ) L 2 ( Ω ) 2 + K 2 2 ε ˜ θ ( t ) L 2 ( Ω ) 2 .

In virtue of the Poincaré inequality v C v , for any v H 0 1 ( Ω ) , we know

(4.58) L ( t ) K 2 u t ( t ) L 2 ( Ω ) 2 K 2 1 δ d ( δ ) E ( 0 ) p γ p u ( t ) L p ( Ω ) p + K 2 ε ˜ 2 u ( t ) L 2 ( Ω ) 2 + K 2 2 ε ˜ K 1 C 2 θ ( t ) L 2 ( Ω ) 2 .

Recalling I ( u ) > 0 for all t 0 and Lemma 2.4, one can deduce that

(4.59) u L p ( Ω ) p > Ω u f ( u ) d x B u λ λ .

Thus, taking 0 < η < 1 and using the embedding inequality u L 2 ( Ω ) C 6 u L λ ( Ω ) , (4.58) becomes

(4.60) L ( t ) K 2 u t ( t ) L 2 ( Ω ) 2 K 2 1 η δ d ( δ ) E ( 0 ) p γ p u ( t ) L p ( Ω ) p η K 2 u ( t ) L p ( Ω ) p + K 2 ε ˜ 2 u ( t ) L 2 ( Ω ) 2 + K 2 2 ε ˜ K 1 C 2 θ ( t ) L 2 ( Ω ) 2 K 2 u t ( t ) L 2 ( Ω ) 2 K 2 1 η δ d ( δ ) E ( 0 ) p γ p u ( t ) L p ( Ω ) p η K 2 B u ( t ) L λ ( Ω ) λ + K 2 ε ˜ 2 u ( t ) L 2 ( Ω ) 2 + K 2 2 ε ˜ K 1 C 2 θ ( t ) L 2 ( Ω ) 2 K 2 u t ( t ) L 2 ( Ω ) 2 K 2 1 η δ d ( δ ) E ( 0 ) p γ p u ( t ) L p ( Ω ) p η K 2 B C 6 λ u ( t ) L 2 ( Ω ) λ + K 2 ε ˜ 2 u ( t ) L 2 ( Ω ) 2 + K 2 2 ε ˜ K 1 C 2 θ ( t ) L 2 ( Ω ) 2 = K 2 u t ( t ) L 2 ( Ω ) 2 K 2 1 η δ d ( δ ) E ( 0 ) p γ p u ( t ) L p ( Ω ) p + K 2 ε ˜ 2 η K 2 B C 6 λ u ( t ) L 2 ( Ω ) λ 2 u ( t ) L 2 ( Ω ) 2 + K 2 2 ε ˜ K 1 C 2 θ ( t ) L 2 ( Ω ) 2 .

In virtue of (4.15), it follows that u ( t ) L 2 ( Ω ) < C d is bounded for t > 0 , then we can choose a enough small ε ˜ to ensure η ε ˜ C d λ 2 < 1 and pick the parameter K 2 enough small such that

ε K 2 ˜ 1 2 B u L 2 ( Ω ) λ 2 C 6 λ C d λ 2 = ε K 2 ˜ 1 B ϱ C 6 λ 0, 0 < ϱ < 1

and

K 2 2 ε ˜ K 1 C 2 < 0 .

Therefore, (4.60) turns into

(4.61) L ( t ) K 2 u t ( t ) L 2 ( Ω ) 2 K 2 1 η δ d ( δ ) E ( 0 ) p γ p u ( t ) L p ( Ω ) p K 1 C 2 K 2 2 ε ˜ θ ( t ) L 2 ( Ω ) 2 .

According to the definition of E ( t ) and I ( u ) > 0 for all t 0 , we know that

(4.62) E ( t ) 1 2 u t ( t ) L 2 ( Ω ) 2 + 1 2 θ ( t ) L 2 ( Ω ) 2 + 1 p 1 α β λ 1 , p α u ( t ) L p ( Ω ) p + 1 α I ( u ) 1 2 u t ( t ) L 2 ( Ω ) 2 + 1 α I ( u )

and

(4.63) E ( t ) 1 2 u t ( t ) L 2 ( Ω ) 2 + 1 2 θ ( t ) L 2 ( Ω ) 2 + 1 p 1 α β λ 1 , p α u ( t ) L p ( Ω ) p + 1 α I ( u ) 1 2 θ ( t ) L 2 ( Ω ) 2 + 1 p 1 α β λ 1 , p α u ( t ) L p ( Ω ) p + 1 α I ( u ) .

Thus, it follows from (4.61), (4.62), and (4.63) that

(4.64) L ( t ) K 2 u t ( t ) L 2 ( Ω ) 2 + 1 α I ( u ) K 2 1 η δ d ( δ ) E ( 0 ) p γ p u ( t ) L p ( Ω ) p K 1 C 2 K 2 2 ε ˜ θ ( t ) L 2 ( Ω ) 2 1 α I ( u ) K 2 + K 3 2 E ( t ) K 4 E ( t ) ,

where K 3 and K 4 are constants that can ensure the inequality holds after magnification.

Now, we show the equivalent of L ( t ) and E ( t ) . By using the Cauchy-Schwartz inequality and Poincaré inequality, it follows from (4.50) that

(4.65) ( u t ( t ) , u ( t ) ) 1 2 u t ( t ) L 2 ( Ω ) 2 + C 2 2 u ( t ) L 2 ( Ω ) 2 1 + p C 2 δ 1 1 d ( δ ) E ( 0 ) p γ p 1 E ( t ) ,

where C is the optimal constant in the embedding H 0 1 ( Ω ) L 2 ( Ω ) . Therefore, we obtain

L ( t ) K 1 E ( t ) K 2 ( u t ( t ) , u ( t ) ) K 2 1 + p C 2 δ 1 1 d ( δ ) E ( 0 ) p γ p 1 E ( t ) K 2 K E ( t ) .

Then, we obtain by taking K > K 1 K 2 that

( K 2 K K 1 ) E ( t ) L ( t ) ( K 2 K + K 1 ) E ( t ) .

Finally, it follows from (4.64) that there exist a positive constant c such that

L ( t ) c E ( t ) , t t * ,

which yields (4.3) immediately.□

4.3 Finite time blow up of solutions

Lemma 4.7

Suppose f ( u ) satisfy ( i ) ( i i i ) and u W 0 1 , p ( Ω ) satisfy I ( u ) < 0 . Then there must exist a λ * ( 0 , 1 ) such that I ( λ * u ) = 0 . In addition, there holds

(4.66) I ( u ) < α ( J ( u ) d ) .

Proof

Denote φ ( λ ) λ γ p u L γ ( Ω ) γ + 1 for each λ ( 0 , ) . Then for any λ > 0 , it follows that

(4.67) I ( λ u ) = λ p u L p ( Ω ) p λ 1 p Ω u f ( λ u ) d x λ p ( u L p ( Ω ) p λ γ p u L γ ( Ω ) γ ) = λ p ( u L p ( Ω ) p φ ( λ ) ) .

Since I ( u ) < 0 , together with (4.67) and Lemma 2.8-(ii), we obtain

(4.68) φ ( 1 ) > u L p ( Ω ) p > ( r ( 1 ) ) p > 0 .

In addition, it easily verify that lim λ 0 + φ ( λ ) = 0 , which, combining with (4.68), yields that there exists a λ * ( 0,1 ) such that φ ( λ * ) = u L p ( Ω ) p and I ( λ * u ) = 0 .

Below, we give (4.66). Taking g ( λ ) α J ( λ u ) I ( λ u ) for λ > 0 , by a simple calculation, we achieve

g ( λ ) = α J ( λ u ) I ( λ u ) = α p λ u L p ( Ω ) p α Ω ( F ( λ u ) σ ) d x I ( λ u ) α p λ u L p ( Ω ) p Ω λ u f ( λ u ) d x β Ω ( λ u ) p d x I ( λ u ) λ p α p 1 β λ 1 , p u L p ( Ω ) p .

Hence, by Lemma 2.8-(ii), it follows that

g ( λ ) = α p p β λ 1 , p λ p 1 u L p ( Ω ) p > α p p β λ 1 , p λ p 1 ( r ( 1 ) ) p > 0 ,

Finally, we obtain that g ( λ ) is strictly increasing for any λ > 0 , recalling the fact that 0 < λ * < 1 , we know g ( 1 ) > g ( λ * ) , i.e.,

α J ( u ) I ( u ) > α J ( λ * u ) I ( λ * u ) = α J ( λ * u ) α d ,

which yields (4.66).□

Theorem 4.8

Let f ( u ) satisfy ( i ) ( i i i ) , u 0 W 0 1 , p ( Ω ) , u 1 L 2 ( Ω ) and θ 0 H 0 1 ( Ω ) . Suppose that u 0 and α 2 ( u 1 , u 0 ) > u 0 L 2 ( Ω ) 2 . Then the solution of system (1.1) blows up in a finite time, i.e., there exists a time T * > 0 such that

(4.69) lim t T * u ( t ) L 2 ( Ω ) 2 = + .

Furthermore, the upper bound of blowup time T is estimated by

T 1 α 2 log α 2 ( u 1 , u 0 ) + u 0 L 2 ( Ω ) 2 α 2 ( u 1 , u 0 ) u 0 L 2 ( Ω ) 2 .

And the lower bound of blowup time T is estimated in the following two cases:

  1. if p < γ n p 2 ( n p ) with 2 < p < n < 2 p 2 2 p p 2 or 1 < γ n n 2 with n > p = 2 , then T is estimated by

    (4.70) T u 0 L γ ( Ω ) γ 1 P 1 s 2 γ p + γ A s + P 2 d s ,

    where P 1 and P 2 is defined in (4.78).

  2. if max { p , n p 2 ( n p ) } < γ < 2 n p p 2 2 ( n p ) with max { 2 , 4 n n + 4 } p < n , then T is estimated by

    (4.71) T u 0 L ρ ( Ω ) ρ 1 A 1 s μ + A 2 s θ + A 3 d s ,

    where ρ = n p 2 ( n p ) , the constants A 1 , A 2 , A 3 and the exponents μ , θ are defined in (4.86).

Proof

We divide this proof into two steps.

Step 1: Blow up in finite time and upper bound estimate of the blowup time.

Let u = u ( t ) , t [ 0 , T ) be any weak solution of problem (1.1) with E ( 0 ) < d and I ( u 0 ) < 0 , and T be the maximum existence time. First, we obtain u ( t ) N for all t [ 0 , T ) by Theorem 4.1-(ii). Now we prove that u blows up in finite time. By contradiction, we let T = and recall the auxiliary function Ψ ( t ) u ( t ) L 2 ( Ω ) 2 denoted in (3.3) for t [ 0 , T ) , then we have

(4.72) Ψ ( t ) = 2 ( u t ( t ) , u ( t ) ) and ( Ψ ( t ) ) 2 4 u ( t ) L 2 ( Ω ) 2 u t ( t ) L 2 ( Ω ) 2 = 4 Ψ ( t ) u t ( t ) L 2 ( Ω ) 2 .

By the definition of I ( u ) , we conclude that

(4.73) Ψ ( t ) = 2 u t ( t ) L 2 ( Ω ) 2 + 2 u t t ( t ) , u ( t ) = 2 u t ( t ) L 2 ( Ω ) 2 2 I ( u ) 2 ( u ( t ) , θ ( t ) ) .

Then by Hölder’s inequality, the definition of E ( t ) in (2.3), (2.4), and Lemma 4.7, one has

Ψ ( t ) Ψ ( t ) 2 + α 4 ( Ψ ( t ) ) 2 Ψ ( t ) ( 2 u t ( t ) L 2 ( Ω ) 2 2 I ( u ) 2 ( u ( t ) , θ ( t ) ) ) ( 2 + α ) Ψ ( t ) u t ( t ) L 2 ( Ω ) 2 Ψ ( t ) ( α u t ( t ) L 2 ( Ω ) 2 2 I ( u ) u ( t ) L 2 ( Ω ) 2 θ ( t ) L 2 ( Ω ) 2 ) Ψ ( t ) ( ( α 1 ) θ ( t ) L 2 ( Ω ) 2 + 2 α ( J ( u ) E ( 0 ) ) 2 I ( u ) u ( t ) L 2 ( Ω ) 2 ) > Ψ ( t ) ( 2 α ( J ( u ) d ) 2 I ( u ) ) ( Ψ ( t ) ) 2 > ( Ψ ( t ) ) 2 .

Take C 1 = 0 , C 2 = 1 , γ = α 2 4 , γ 1 = α 2 2 , and γ 2 = α 2 2 , then we know that Ψ ( 0 ) = u 0 L 2 ( Ω ) 2 > 0 , C 1 + C 2 > 0 . By α 2 ( u 1 , u 0 ) > u 0 L 2 ( Ω ) 2 , we have Ψ ( 0 ) + γ 2 γ 1 Ψ ( 0 ) > 0 . Therefore, by using Lemma 3.1, we obtain Ψ ( t ) for t t * < , which contradicts T = . That means that u blows up in a finite time. In addition, as we did in the aforementioned steps, using Lemma 3.1, we arrive at

T 1 α 2 log α 2 ( u 1 , u 0 ) + u 0 L 2 ( Ω ) 2 α 2 ( u 1 , u 0 ) u 0 L 2 ( Ω ) 2 .

Step 2: Lower bound estimate of the blow-up time.

First, by Step 1, we know that the weak solutions of problem (1.1) will blow up in a finite time, i.e., lim t T u ( t ) L 2 ( Ω ) 2 = . Then it follows from L γ ( Ω ) L 2 ( Ω ) that lim t T u ( t ) L γ ( Ω ) γ = . Using (2.3) and (2.4), we reach

(4.74) 1 2 u t L 2 ( Ω ) 2 + 1 2 θ L 2 ( Ω ) 2 + 1 p u L p ( Ω ) p E ( 0 ) + Ω ( F ( u ) σ ) d x E ( 0 ) + A u L γ ( Ω ) γ E ( 0 ) + A u L γ ( Ω ) γ .

In the following, we divide the remainder proof into two cases.

Case 1: p < γ n p 2 ( n p ) with 2 < p < n < 2 p 2 2 p p 2 or 1 < γ n n 2 with n > p = 2 .

We define the auxiliary function ϕ ( t ) u ( t ) L γ ( Ω ) γ for t > 0 , then by using the Hölder and Young inequality, we deduce

(4.75) ϕ ( t ) = γ Ω u ( t ) γ 2 u ( t ) u t ( t ) d x γ 2 ( u ( t ) L 2 ( γ 1 ) ( Ω ) 2 ( γ 1 ) + u t ( t ) L 2 ( Ω ) 2 ) .

Since 2 ( γ 1 ) > p , using the inequality a s + b s ( a + b ) s ( a , b 0 , s 1 ) , it is easy to see that

(4.76) u L 2 ( γ 1 ) ( Ω ) 2 ( γ 1 ) C 3 2 ( γ 1 ) u L p ( Ω ) 2 ( γ 1 ) C 3 2 ( γ 1 ) u L p ( Ω ) 2 ( γ 1 ) + u L γ ( Ω ) 2 γ ( γ 1 ) p C 3 2 ( γ 1 ) ( u L p ( Ω ) p + u L γ ( Ω ) γ ) 2 ( γ 1 ) p ,

where C 3 is the optimal embedding constant of W 0 1 , p ( Ω ) L 2 ( γ 1 ) ( Ω ) .

Combining (4.74)–(4.76) yield

(4.77) ϕ ( t ) γ C 3 2 ( γ 1 ) 2 ( ϕ ( t ) + u ( t ) L p ( Ω ) p ) 2 ( γ 1 ) p + γ E ( 0 ) + γ A ϕ ( t ) γ C 3 2 ( γ 1 ) 2 ( p E ( 0 ) + ( p A + 1 ) ϕ ( t ) ) 2 ( γ 1 ) p + γ E ( 0 ) + γ A ϕ ( t ) 2 2 ( γ 1 ) p 2 γ C 3 2 ( γ 1 ) ( p 2 ( γ 1 ) p E ( 0 ) 2 ( γ 1 ) p + ( p A + 1 ) 2 ( γ 1 ) p ϕ 2 ( γ 1 ) p ( t ) ) + γ E ( 0 ) + γ A ϕ ( t ) = P 1 ϕ 2 ( γ 1 ) p ( t ) + γ A ϕ ( t ) + P 2 ,

where

(4.78) P 1 = 2 2 ( γ 1 ) p 2 γ C 3 2 ( γ 1 ) ( p A + 1 ) 2 ( γ 1 ) p , P 2 = 2 2 ( γ 1 ) p 2 γ C 3 2 ( γ 1 ) p 2 ( γ 1 ) p E ( 0 ) 2 ( γ 1 ) p + γ E ( 0 ) .

Integrating (4.77) from 0 to t , one has

ϕ ( 0 ) ϕ ( t ) 1 P 1 z 2 r p + γ A z + P 2 d z t ,

which, combining with lim t T u ( t ) L γ ( Ω ) γ = , implies (5.9).

Case 2: max { p , n p 2 ( n p ) } < γ < 2 n p p 2 2 ( n p ) with max { 2 , 4 n n + 4 } p < n .

We know the embedding W 0 1 , p ( Ω ) L 2 γ ( Ω ) does not satisfy in this case. Thus, we need to re-estimate the lower bound of the bow-up time by employing the interpolation inequality as follows. To begin, we denote

(4.79) ψ ( t ) u ( t ) L ρ ( Ω ) ρ with ρ = n p 2 ( n p ) .

Since p 4 n n + 4 , then it follows from L ρ ( Ω ) L 2 ( Ω ) that lim t T u ( t ) L ρ ( Ω ) ρ = . By the Schwartz inequality, one can arrive at

(4.80) ψ ( t ) = ρ Ω u ( t ) ρ 2 u ( t ) u t ( t ) d x ρ 2 ( u ( t ) L 2 ρ 2 ( Ω ) 2 ρ 2 + u t ( t ) L 2 ( Ω ) 2 ) .

Since 2 ( ρ 1 ) = n p 2 n + 2 p n p < n p n p 2 * , in virtue of the embedding W 0 1 , p ( Ω ) L 2 * ( Ω ) , we deduce that

(4.81) u L 2 ρ 2 ( Ω ) 2 ρ 2 C 4 2 α 2 u L p ( Ω ) 2 α 2 ,

where C 4 is the optimal embedding constant of W 0 1 , p ( Ω ) L 2 α 2 ( Ω ) .

Combining (4.74), (4.80), and (4.81), we obtain

(4.82) ψ ( t ) ρ C 4 2 ( ρ 1 ) 2 ( p E ( 0 ) + p A u ( t ) L γ ( Ω ) γ ) 2 ρ 2 p + ρ E ( 0 ) + ρ A u ( t ) L γ ( Ω ) γ .

In virtue of the interpolation inequality, we arrive at

(4.83) u L γ ( Ω ) γ u L ρ ( Ω ) γ κ u L 2 * ( Ω ) γ ( 1 κ ) C 1 γ ( 1 κ ) u L ρ ( Ω ) γ κ u L p ( Ω ) γ ( 1 κ ) ,

where C 1 is the optimal constant of the Sobolev embedding W 0 1 , p ( Ω ) L 2 * ( Ω ) and

1 γ = κ ρ + 1 κ 2 * , κ = 2 * γ γ ( 0, 1 ) .

Using the Young inequality and (4.74), it follows that

u L γ ( Ω ) γ ( p γ ( 1 κ ) ) C 1 p γ ( 1 κ ) p γ ( 1 κ ) p u L ρ ( Ω ) p κ γ p γ ( 1 κ ) + γ ( 1 κ ) p u L p ( Ω ) p L ( u L ρ ( Ω ) ρ ) ϑ + γ ( 1 κ ) E ( 0 ) + γ ( 1 κ ) A u L γ ( Ω ) γ ,

where

L = ( p γ ( 1 κ ) ) C 1 γ ( 1 κ ) p γ ( 1 κ ) p , ϑ = p κ γ p γ ( 1 κ ) .

This implies that

(4.84) u L γ ( Ω ) γ L κ ( u L ρ ( Ω ) ρ ) ϑ + γ ( 1 κ ) κ E ( 0 ) = L κ ψ ϑ ( t ) + γ ( 1 κ ) κ E ( 0 ) .

Let C * = max { 2 2 ρ 2 2 p p , 1 2 } , then we obtain from (4.82) and (4.84) that

(4.85) ψ ( t ) ρ C 4 2 ( ρ 1 ) 2 p E ( 0 ) + p L κ γ ψ ϑ ( t ) + p ( 1 κ ) κ E ( 0 ) 2 ρ 2 p + ρ E ( 0 ) + L ρ κ γ ψ ϑ ( t ) + ρ ( 1 κ ) κ E ( 0 ) ρ C 4 2 ( ρ 1 ) C * p κ 2 ρ 2 p E ( 0 ) 2 ρ 2 p + ρ C 4 2 ρ 2 C * p L κ γ 2 ρ 2 p ψ μ ( t ) + ρ κ E ( 0 ) + L ρ κ γ ψ ϑ ( t ) = A 1 ψ μ ( t ) + A 2 ψ ϑ ( t ) + A 3 ,

where

(4.86) μ = ϑ ( 2 ρ 2 ) p , A 1 = α C 4 2 ρ 2 C * p L κ γ 2 ρ 2 p , A 2 = L ρ κ γ , A 3 = ρ C 4 2 ρ 2 C * p κ 2 ρ 2 p E ( 0 ) 2 ρ 2 p + ρ κ E ( 0 ) , ϑ = κ γ p γ ( 1 κ ) .

By integrating (4.85) from 0 to t , one has

ψ ( 0 ) ψ ( t ) 1 A 1 z μ + A 2 z ϑ + A 3 d z t ,

which, together with lim t T u ( t ) L ρ ( Ω ) ρ = , implies (4.69).□

5 Long-time behavior of the solution for system (1.1) with critical initial energy E ( 0 ) = d

5.1 Global existence of solutions

In this subsection, we will employ the scaling method to extend the conclusions obtained under low initial energy levels to the critical initial energy level. This primarily includes the global existence of solutions, their asymptotic behavior, and the dependence of finite-time blow-up and blow-up time estimates on the initial data.

Theorem 5.1

(Global existence for E ( 0 ) = d ) Let f ( u ) satisfy ( i ) ( i i i ) , u 0 W 0 1 , p ( Ω ) , u 1 L 2 ( Ω ) and θ 0 H 0 1 ( Ω ) . Assume that E ( 0 ) = d and I ( u 0 ) 0 . The initial-boundary value problem (1.1) admits a global weak solution u L ( 0 , ; W 0 1 , p ( Ω ) ) along with u t L 2 ( 0 , ; L 2 ( Ω ) ) and θ L 2 ( 0 , ; H 0 1 ( Ω ) ) . Moreover, the global solution have the following exponential asymptotic behavior

(5.1) 1 2 u t ( t ) L 2 ( Ω ) 2 + 1 2 θ ( t ) L 2 ( Ω ) 2 + 0 t θ ( t ) L 2 ( Ω ) 2 d τ + 1 p 1 α β λ 1 , p α u ( t ) L p ( Ω ) p E ( 0 ) e c t , t t * > 0 ,

where c is a positive constant.

Proof

We divide this proof into two steps. Step 1: Global existence of solutions.

For the case u 0 L p ( Ω ) 0 , we suppose that ρ m = 1 1 m and u 0 m = ρ m u 0 , m 2 . System (1.1) can be rewritten as follows:

(5.2) u t t Δ p u + θ = f ( u ) , x Ω , t > 0 , θ t Δ θ = u t , x Ω , t > 0 , u ( x , t ) Ω = 0 , θ ( x , t ) Ω = 0 , t > 0 , u ( x , 0 ) = u 0 m ( x ) , u t ( x , 0 ) = u 1 ( x ) , x Ω , θ ( x , 0 ) = θ 0 ( x ) , x Ω .

According to I ( u 0 ) 0 and Lemma 2.6, one can infer that ε * = ε * ( u 0 ) 1 . Thus, we know I ( u 0 m ) > 0 . By the relationship between J ( u ) and I ( u ) , we obtain

J ( u 0 ) > J ( u 0 m ) = 1 p u 0 m L p ( Ω ) p Ω ( F ( u 0 m ) σ ) d x 1 p u 0 m L p ( Ω ) p 1 α Ω u 0 m f ( u 0 m ) d x β α Ω u 0 m p d x > 1 p δ α β λ 1 , p α u 0 m L p ( Ω ) p + 1 α I ( u ) 1 p δ α β λ 1 , p α u 0 m L p ( Ω ) p > 0 .

Hence, it follows that

0 < E m ( 0 ) = 1 2 u 1 L 2 ( Ω ) 2 + 1 2 θ 0 L 2 ( Ω ) 2 + J ( u 0 m ) < 1 2 u 1 L 2 ( Ω ) 2 + 1 2 θ 0 L 2 ( Ω ) 2 + J ( u 0 ) = E ( 0 ) d ,

which tells us that u 0 m N + . Furthermore, for each m , since Theorem 4.6, there exists a global weak solution u m , θ m of system (5.2) such that u L ( 0 , ; W 0 1 , p ( Ω ) ) along with u t L 2 ( 0 , ; L 2 ( Ω ) ) and θ L 2 ( 0 , ; H 0 1 ( Ω ) ) , and

(5.3) ( u t , ψ ) + 0 t u p 2 u , ψ p d τ + 0 t ( θ , ψ ) d τ = 0 t ( f ( u ) , ψ ) d τ + ( u t ( 0 ) , ψ ) , ( θ , ϕ ) + 0 t ( θ , ϕ ) d τ = 0 t ( u τ , ϕ ) d τ + ( θ ( 0 ) , ϕ ) ,

for any ψ W 0 1 , p ( Ω ) , ϕ H 0 1 ( Ω ) and a.e. t ( 0 , ) .

In addition, with the help of (2.4), we know

(5.4) E m ( t ) E m ( 0 ) < d ,

which, combining with the same argument as we did in Theorem 4.1, we deduce that u m ( t ) N + .

For the case u 0 p = 0 , we obtain J ( u 0 ) = 0 by I ( u 0 ) 0 . Therefore, it is easy to verify that

E ( 0 ) = 1 2 u 1 L 2 ( Ω ) 2 + 1 2 θ 0 L 2 ( Ω ) 2 + J ( u 0 ) = 1 2 u 1 L 2 ( Ω ) 2 + 1 2 θ 0 L 2 ( Ω ) 2 = d .

Let ρ m = 1 1 m and u 1 m = ρ m u 1 , θ 0 m = ρ m u 0 , m 2 . We consider the following problem:

(5.5) u t t Δ p u + θ = f ( u ) , x Ω , t > 0 , θ t Δ θ = u t , x Ω , t > 0 , u ( x , t ) Ω = 0 , θ ( x , t ) Ω = 0 , t > 0 , u ( x , 0 ) = u 0 ( x ) , u t ( x , 0 ) = u 1 m ( x ) , x Ω , θ ( x , 0 ) = θ 0 m ( x ) , x Ω .

Obviously, it follows that

(5.6) 0 < E m ( 0 ) = 1 2 u 1 m L 2 ( Ω ) 2 + 1 2 θ 0 m L 2 ( Ω ) 2 + J ( u 0 ) = 1 2 ρ m u 1 L 2 ( Ω ) 2 + 1 2 ρ m θ 0 L 2 ( Ω ) 2 < 1 2 u 1 L 2 ( Ω ) 2 + 1 2 θ 0 L 2 ( Ω ) 2 = d .

In virtue of (5.6) and Theorem 4.6, we conclude that there exists a global weak solution u m , θ m of the system (5.5) such that u L ( 0 , ; W 0 1 , p ( Ω ) ) along with u t L 2 ( 0 , ; L 2 ( Ω ) ) and θ L 2 ( 0 , ; H 0 1 ( Ω ) ) . The remainder of the proof for Theorem 5.1 is the same as those of Theorem 4.6. Hence, we omit them.

Step 2: Exponential asymptotic behavior of solutions.

Recall that the existence of a global weak solution of problem (1.1) is given in Theorem 5.1 with E ( 0 ) = d and I ( u 0 ) > 0 , which implies I ( u ) 0 for 0 t < . Now we consider the following two cases:

  1. Assume that I ( u ) > 0 for 0 t < . Picking any t 1 > 0 and letting

    d 1 = d 0 t 1 θ ( τ ) L 2 ( Ω ) 2 d τ ,

    then by using the fact that

    E ( t ) E ( 0 ) 0 t 1 θ ( τ ) L 2 ( Ω ) 2 d τ = d 1 < d ,

    and u ( t ) N + δ for δ 1 < δ < δ 2 and t 1 t < , where δ 1 < δ 2 are two roots of equation d ( δ ) = d 1 , we obtain I δ ( u ) 0 for t t 1 . The rest of proof is same as we did in Step 2 of Theorem 4.6. By using the similar Lyapunov functional denoted in (4.52), we can also obtain

    L ( t ) c E ( t ) , t t 1 ,

    which yields (5.1) immediately.

  2. Assume that there exists t 1 > 0 such that I ( u ( t 1 ) ) = 0 and I ( u ( t ) ) > 0 for 0 t < t 1 . If θ ( t ) L 2 ( Ω ) 2 0 for t > 0 , then we know 0 t θ ( τ ) L 2 ( Ω ) 2 d τ > 0 , which implies that

    E ( t 1 ) = d 0 t 1 θ ( τ ) L 2 ( Ω ) 2 d τ < d .

    However, it is easy to verify from I ( u ( t 1 ) ) = 0 that

    E ( t 1 ) = 1 2 u t ( t 1 ) L 2 ( Ω ) 2 + 1 2 θ ( t 1 ) L 2 ( Ω ) 2 + J ( u ( t 1 ) ) = 1 2 u t ( t 1 ) L 2 ( Ω ) 2 + 1 2 θ ( t 1 ) L 2 ( Ω ) 2 + d < d ,

    which means that u t ( t 1 ) L 2 ( Ω ) 2 + θ ( t 1 ) L 2 ( Ω ) 2 < 0 . This is clearly a contradiction. If θ ( t ) L 2 ( Ω ) 2 = 0 , then we deduce that 0 t θ ( τ ) L 2 ( Ω ) 2 d τ 0 . Similar to the previous discussion, we can obtain through repeated verification that

    E ( t 1 ) = 1 2 u t ( t 1 ) L 2 ( Ω ) 2 + 1 2 θ ( t 1 ) L 2 ( Ω ) 2 + J ( u ( t 1 ) ) = 1 2 u t ( t 1 ) L 2 ( Ω ) 2 + 1 2 θ ( t 1 ) L 2 ( Ω ) 2 + d d ,

    which tells us that u t ( t 1 ) L 2 ( Ω ) = 0 for t > 0 . By using the triangle inequality, for any Banach space valued differentiable function v , we have

    d d t v ( t ) d d t v ( t ) .

    Therefore, u t ( t 1 ) L 2 ( Ω ) = 0 implies that u ( t 1 ) L 2 ( Ω ) = c for t > 0 . However, it contradicts with I ( u ( t 1 ) ) = 0 . Hence, the assumption does not hold, and this case does not exist.□

5.2 Finite time blowup of solutions

Theorem 5.2

Let f ( u ) satisfy ( i ) ( i i i ) , u 0 W 0 1 , p ( Ω ) , u 1 L 2 ( Ω ) and θ 0 H 0 1 ( Ω ) . Suppose that E ( 0 ) = d , I ( u 0 ) < 0 and α 2 ( u 1 , u 0 ) > u 0 L 2 ( Ω ) 2 . Then the nontrivial solution of the initial-boundary value problem (1.1) blows up in a finite time, i.e., there exists time T > t ¯ > 0 such that

(5.7) lim t T 0 t u ( τ ) L γ ( Ω ) γ d τ = + .

Furthermore, the upper bound of blowup time T is estimated by

T 1 α 2 log α 2 ( u 1 , u 0 ) + u 0 L 2 ( Ω ) 2 α 2 ( u 1 , u 0 ) u 0 L 2 ( Ω ) 2 .

And the lower bound of blowup time T is estimated in the following two cases:

  1. if p < γ n p 2 ( n p ) with 2 < p < n < 2 p 2 2 p p 2 or 1 < γ n n 2 with n > p = 2 , then T is estimated by

    (5.8) T u 0 L γ ( Ω ) γ 1 P 1 s 2 γ p + γ A s + P 2 d s ,

    where P 1 and P 2 is defined in (4.78).

  2. if max { p , n p 2 ( n p ) } < γ < 2 n p p 2 2 ( n p ) with max { 2 , 4 n n + 4 } p < n , then T is estimated by

    (5.9) T u 0 L ρ ( Ω ) ρ 1 A 1 s μ + A 2 s θ + A 3 d s ,

    where ρ = n p 2 ( n p ) , the constants A 1 , A 2 , A 3 and the exponents μ , θ are defined in (4.86).

Proof

Let ( u , θ ) be a local solution to problem (1.1). We prove that if ( u , θ ) is a nontrivial solution, then the solution blows up in a finite time, i.e., T < . Indeed, we first obtain from Lemma 2.2 that there exists a time t ¯ > 0 such that

E ( t ¯ ) = d 0 t ¯ θ ( τ ) L 2 ( Ω ) 2 d τ < d

holds. Then we know from Theorem 4.1-(ii) that I ( u ( t ) ) < 0 for all t > t ¯ . On the other hand, recalling the auxiliary function Ψ ( t ) u ( t ) L 2 ( Ω ) 2 denoted in (3.3) for t [ t ¯ , T ) , then we also have (4.72) and (4.73). Then by the Hölder and Young inequality, the definition of E ( t ) in (2.3), (2.4), and Lemma 4.7, it follows:

Ψ ( t ) Ψ ( t ) 2 + α 4 ( Ψ ( t ) ) 2 Ψ ( t ) ( 2 u t ( t ) L 2 ( Ω ) 2 2 I ( u ) 2 ( u ( t ) , θ ( t ) ) ) ( 2 + α ) Ψ ( t ) u t ( t ) L 2 ( Ω ) 2 Ψ ( t ) ( α u t ( t ) L 2 ( Ω ) 2 2 I ( u ) u ( t ) L 2 ( Ω ) 2 θ ( t ) L 2 ( Ω ) 2 ) Ψ ( t ) ( ( α 1 ) θ ( t ) L 2 ( Ω ) 2 + 2 α ( J ( u ) d ) 2 I ( u ) u ( t ) L 2 ( Ω ) 2 ) > Ψ ( t ) ( 2 α ( J ( u ) d ) 2 I ( u ) ) ( Ψ ( t ) ) 2 > ( Ψ ( t ) ) 2 .

Obviously, the proof for the remaining part is the same as the process we argue in Theorem 4.8, we hence omit it and obtain the results in the case E ( 0 ) = d immediately.□

Acknowledgment

The author would like to thank the anonymous referees for their valuable comments and suggestions that help to improve and clarify the article greatly.

  1. Funding information: The research was supported by the Fundamental Research Funds in Heilongjiang Natural Science Funds for Distinguished Young Scholar (JQ2023A005).

  2. Author contributions: The sole author is responsible for the entire work and has read and approved the final manuscript.

  3. Conflict of interest: The author states no conflict of interest.

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

References

[1] B. Ahmad, A. Alsaedi, and M. Kirane, Nonexistence results for the Cauchy problem of time fractional nonlinear systems of thermo-elasticity, Math. Meth. Appl. Sci. 40 (2017), 4272–4279, https://doi.org/10.1002/mma.4303. Search in Google Scholar

[2] M. Boutefnouchet, H. Erjaee, M. Kirane, and M. Qafsaoui, Nonexistence results for the Cauchy problem for some fractional nonlinear systems of thermo-elasticity type, Z. Angew. Math. Mech. 96 (2016), 1119–1128, https://doi.org/10.1002/zamm.201500134. Search in Google Scholar

[3] W. H. Chen, Cauchy problem for thermoelastic plate equations with different damping mechanisms, Commun. Math. Sci. 18 (2020), no. 2, 429–457, https://doi.org/10.4310/CMS.2020.v18.n2.a7. Search in Google Scholar

[4] S. H. Chen, J. B. Han, R. Z. Xu, C. Yang, and M. N. Zhang, Well-posedness of damped Kirchhoff-type wave equation with fractional Laplacian, Adv. Nonlinear Stud. 25 (2025), 673–718, https://doi.org/10.1515/ans-2023-0172. Search in Google Scholar

[5] Y. Chu, B. Wen, and L. B. Cheng, Existence and blow up for viscoelastic hyperbolic equations with variable exponents, Commun. Anal. Mech. 16 (2024), 717–737, https://doi.org/10.3934/cam.2024032. Search in Google Scholar

[6] S. Y. Chung and M. J. Choi, A new condition for the concavity method of blow-up solutions to p-Laplacian parabolic equations, J. Differential Equations 265 (2018), 6384–6399, https://doi.org/10.1016/j.jde.2018.07.03. Search in Google Scholar

[7] C. V. Coffman, Lyusternik-Schnirelman theory and eigenvalue problems for monotone potential operators, J. Funct. Analysis 14 (1973), 237–252, https://doi.org/10.1016/0022-1236(73)90070-0. Search in Google Scholar

[8] C. M. Dafermos, On the existence and the asymptotic stability of solutions to the equations of linear thermoelasticity, Arch. Rational Mech. Anal. 29 (1968), 241–271, https://doi.org/10.1007/BF00276727. Search in Google Scholar

[9] H. Ding and J. Zhou, Global existence and blow-up for a thermoelastic system with p-Laplacian, Appl. Anal. 101 (2022), 6690–6708, https://doi.org/10.1080/00036811.2021.1941906. Search in Google Scholar

[10] A. D. Drozdov and V. B. Kolmanovskii, Stability in Viscoelasticity, North-Holland, Amsterdam, 1994. Search in Google Scholar

[11] J. B. Han, K. Y. Wang, R. Z. Xu, and C. Yang, Global quantitative stability of wave equations with strong and weak dampings, J. Differential Equations, 390 (2024), 228–344, https://doi.org/10.1016/j.jde.2024.01.033. Search in Google Scholar

[12] J. B. Han, R. Z. Xu, and Y. B. Yang, Asymptotic behavior and finite time blow up for damped fourth order nonlinear evolution equation, Asymptot. Anal. 122 (2021), 349–369, https://doi.org/10.3233/ASY-201621. Search in Google Scholar

[13] J. B. Han, R. Z. Xu, and C. Yang, Continuous dependence on initial data and high energy blowup time estimate for porous elastic system, Commun. Anal. Mech. 15 (2023), 214–244, https://doi.org/10.3934/cam.2023012. Search in Google Scholar

[14] D. B. Henry, Non-decay of thermoelastic vibrations in dimension ≥ 3, Topics in Analysis, Chapter 2. Publ. Secc. Mat., Univ. Autònoma Barcelona. 31 (1987), 29–84, https://doi.org/10.5565/PUBLMAT_31187_03. Search in Google Scholar

[15] J. Ignaczak and M. Ostoja-Starzewski, Thermoelasticity with Finite Wave Speeds, Oxford mathematical Monographs. Oxford University Press, New York, 2010, https://doi.org/10.1093/acprof:oso/9780199541645.001.0001. Search in Google Scholar

[16] T. Irmscher and R. Racke, Sharp decay rates in parabolic and hyperbolic thermoelasticity, IMA J. Appl. Math. 71 (2006), 459–478, https://doi.org/10.1093/imamat/hxh110. Search in Google Scholar

[17] S. Jiang, R. J. E. Munoz, and R. Racke, Asymptotic stability and global existence in thermoelasticity with symmetry, Quart. Appl. Math. 56 (1998), no. 2, 259–275, https://doi.org/10.1090/qam/1622566. Search in Google Scholar

[18] S. Jiang and R. Racke, Evolution equations in thermoelasticity, π Monographs Surveys Pure Appl. Math. 112, Chapman & Hall / CRC, Boca Raton, 2000, https://doi.org/10.1201/9781482285789. Search in Google Scholar

[19] B. Kawohl and P. Lindqvist, Positive eigenfunctions for the p-Laplace operator revisited, Analysis 26 (2006), 545–550, https://doi.org/10.1524/anly.2006.26.4.545. Search in Google Scholar

[20] M. Kirane and N. Tatar, A nonexistence result to a Cauchy problem in nonlinear one-dimensional thermoelasticity, J. Math. Anal. Appl. 254 (2001), 71–86, https://doi.org/10.1006/jmaa.2000.7203. Search in Google Scholar

[21] M. A. Krasnosel’skii, Topological methods in the theory of nonlinear integral equations, The Macmillan Co., NewYork, 1964. Translated by A. H. Armstrong; translation edited by J. Burlak, A Pergamon Press Book. Search in Google Scholar

[22] J. V. Kim, On the energy decay of a linear thermoelastic bar and plate, SIAM J. Math. Anal. 23 (1992), 889–899, https://doi.org/10.1137/0523047. Search in Google Scholar

[23] J. E. Lagnese, G. Leugering, and E. J. P. G. Schmidt, Modeling, analysis and control of dynamic elastic multi-link structures, Systems & Control: Foundations & Applications, Birkhäuser, Boston, Inc., Boston, MA, 1994, https://doi.org/10.1007/978-1-4612-0273-8. Search in Google Scholar

[24] H. A. Levine, Instability and nonexistence of global solutions to nonlinear wave equations of the form Putt=Au+F(u), Trans. Amer. Math. Soc. 192 (1974), 1–21, https://doi.org/10.2307/1996814. Search in Google Scholar

[25] G. Lebeau and E. Zuazua, Sur la décroissance non uniforme de l’énergie dans le systéeme de la thermoelasticité linéaire, C. R. Acad. Sci. Paris. Sér. I Math. 324 (1997), 409–415, https://doi.org/10.1016/S0764-4442(97)80077-8. Search in Google Scholar

[26] W. Lian, V. D. Radulescu, R. Z. Xu, Y. B. Yang, and N. Zhao, Global well-posedness for a class of fourth-order nonlinear strongly damped wave equations, Adv. Calc. Var. 14 (2019), 589–611, https://doi.org/10.7494/OpMath.2019.39.2.297. Search in Google Scholar

[27] Q. Lin and R. Xu, Global well-posedness of the variable-order fractional wave equation with variable exponent nonlinearity, J. Lond. Math. Soc. 111 (2025), e70091. https://doi.org/10.1112/jlms.70091. Search in Google Scholar

[28] J. L. Lions, Quelques Méthodes de Résolution des Problémes aux Limites non Linéaires, Etudes mathematiques, Dunod-Gauthier-Villars, 1969 (French). Search in Google Scholar

[29] Y. B. Luo, R. Z. Xu, and C. Yang, Global well-posedness for a class of semilinear hyperbolic equations with singular potentials on manifolds with conical singularities, Calc. Var. PDE 61 (2022), 1–47, https://doi.org/10.1007/s00526-022-02316-2. Search in Google Scholar

[30] W. Lian, R. Z. Xu, and Y. Niu, Global well-posedness of coupled parabolic systems, Sci. China Math. 63 (2020), 321–356, https://doi.org/10.1007/s11425-017-9280-x. Search in Google Scholar

[31] Y. Liu, X. Long, and L. Zhang, Long-time dynamics for a coupled system modeling the oscillations of suspension bridges, Commun. Anal. Mech. 17 (2025), 15–40, https://doi.org/10.3934/cam.2025002. Search in Google Scholar

[32] Y. Liu, B. Moon, V. D. Radulescu, R. Z. Xu, and C. Yang, Qualitative properties of solution to a viscoelastic Kirchhoff-like plate equation, J. Math. Phys. 64 (2023), 051511, https://doi.org/10.1063/5.0149240. Search in Google Scholar

[33] H. W. Lord and Y. Shulman, A generalized dynamical theory of thermoelasticity, J. Mech. Phys. Solids 15 (1967), 299–309, https://doi.org/10.1016/0022-5096(67)90024-5. Search in Google Scholar

[34] J. E. Mũnoz Rivera and R. Racke, Large solutions and smoothing properties for nonlinear thermoelastic systems, J. Differential Equations 127 (1996), 454–483, https://doi.org/10.1006/jdeq.1996.0078. Search in Google Scholar

[35] A. Marzocchi, J. E. M. Rivera, and M. G. Naso, Asymptotic behavior and exponential stability for a transmission problem in thermoelasticity, Math. Meth. Appl. Sci. 25 (2002), no 11, 955–980, https://doi.org/10.1002/mma.323. Search in Google Scholar

[36] S. A. Messaoudi, Local existence and blow up in thermoelasticity with second sound, Commun. Partial Differ. Equ. 26 (2002), no. 8, 1681–1693, https://doi.org/10.1081/PDE-120005852. Search in Google Scholar

[37] J. E. Mũnoz Rivera and R. Racke, Smoothing properties, decay, and global existence of solutions to nonlinear coupled systems of thermoelastic type, SIAM J. Math. Anal. 26 (1995), 1547–1563, https://doi.org/10.1137/S0036142993255058. Search in Google Scholar

[38] J. C. Oliveira and R. C. Charao, Stabilization of a locally damped thermoelastic system, Comput. Appl. Math. 27 (2008), no. 3, 319–357, https://doi.org/10.1007/s40314-008-5169-9. Search in Google Scholar

[39] T. T. Pang, K. Y. Wang, R. Z. Xu, and C. Yang, Global existence and non-existence for nonlinear wave equation with variable-exponent nonlinear weak damping term and power-type nonlinearity, Commun. Math. Sci. 23 (2025), 1933–1982, https://doi.org/10.4310/CMS.250802032704. Search in Google Scholar

[40] L. E. Payne and D. H. Sattinger, Saddle points and instability of nonlinear hyperbolic equations, Israel J. Math. 22 (1975), 273–303, https://doi.org/10.1007/BF02761595. Search in Google Scholar

[41] D. C. Pereira and G. P. Menzala, Exponential stability in linear thermoelasticity: the inhomogeneous case, Appl. Anal. 44 (1992), 21–36, https://doi.org/10.1080/00036819208840066. Search in Google Scholar

[42] Y. Qin and J. M. Rivera, Blow-up of solutions to the Cauchy problem in nonlinear one-dimensional thermoelasticity, J. Math. Anal. Appl. 292 (2004), 160–193, https://doi.org/10.1016/j.jmaa.2003.11.050. Search in Google Scholar

[43] R. Racke and Y. Shibata, Global smooth solutions and asymptotic stability in one dimensional thermoelasticity Arch. Ration. Mech. Anal. 116 (1991), 1–34, https://doi.org/10.1007/BF00375601. Search in Google Scholar

[44] C. A. Raposo, J. O. Ribeiro, and A. P. Cattai, Global solution for a thermoelastic system with p-Laplacian, Appl. Math. Lett. 86 (2018), 119–125, https://doi.org/10.1016/j.aml.2018.06.029. Search in Google Scholar

[45] M. Slemrod, Global existence, uniqueness and asymptotic stability of classical smooth solutions in one-dimensional nonlinear thermoelasticity, Arch. Ration. Mech. Anal. 76 (1981), 97–133, https://doi.org/10.1007/BF00251248. Search in Google Scholar

[46] E. H. G. Tavares, M. A. J. Silva, Y. N. Li, V. Narciso, and Z. Yang, Dynamics of a thermoelastic Balakrishnan-Taylor beam model with fractional operators, Appl. Math. Opt. 89 (2024), 17, https://doi.org/10.1007/s00245-023-10086-2. Search in Google Scholar

[47] R. Z. Xu and J. Su, Global existence and finite time blow-up for a class of semilinear pseudo-parabolic equations, J. Funct. Anal. 264 (2013), 2732–2763, https://doi.org/10.1016/j.jfa.2013.03.010. Search in Google Scholar

[48] C. Yang, V. D. Radulescu, R. Z. Xu, and M. Y. Zhang, Global well-posedness analysis for the nonlinear extensible beam equations in a class of modified Woinowsky-Krieger models, Adv. Nonlinear Stud. 22 (2022), 436–468, https://doi.org/10.1515/ans-2022-0024. Search in Google Scholar

[49] Y. B. Yang and R. Z. Xu, Nonlinear wave equation with both strongly and weakly damped terms: supercritical initial energy finite time blow up, Commun. Pure Appl. Anal. 18 (2019), 1351–1358, https://doi.org/10.3934/CPAA.2019065. Search in Google Scholar

[50] S. M. Zheng, Nonlinear evolution equations, volume 133 of Chapman & Hall/CRC Monographs and Surveys in Pure and Applied Mathematics. Chapman & Hall/CRC, Boca Raton, FL, 2004, https://doi.org/10.1201/9780203492222. Search in Google Scholar

[51] S. Zheng, Nonlinear Parabolic Equations and Hyperbolic-Parabolic Coupled Systems, Pitman Monographs Surveys Pure Applied Mathematics, vol. 76. Longman, Harlow, 1995, https://doi.org/10.1201/9780429154225. Search in Google Scholar

Received: 2025-07-02
Revised: 2025-09-17
Accepted: 2025-10-13
Published Online: 2025-11-15

© 2025 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. Research Articles
  2. Incompressible limit for the compressible viscoelastic fluids in critical space
  3. Concentrating solutions for double critical fractional Schrödinger-Poisson system with p-Laplacian in ℝ3
  4. Intervals of bifurcation points for semilinear elliptic problems
  5. On multiplicity of solutions to nonlinear Dirac equation with local super-quadratic growth
  6. Entire radial bounded solutions for Leray-Lions equations of (p, q)-type
  7. Combinatorial pth Calabi flows for total geodesic curvatures in spherical background geometry
  8. Sharp upper bounds for the capacity in the hyperbolic and Euclidean spaces
  9. Positive solutions for asymptotically linear Schrödinger equation on hyperbolic space
  10. Existence and non-degeneracy of the normalized spike solutions to the fractional Schrödinger equations
  11. Existence results for non-coercive problems
  12. Ground states for Schrödinger-Poisson system with zero mass and the Coulomb critical exponent
  13. Geometric characterization of generalized Hajłasz-Sobolev embedding domains
  14. Subharmonic solutions of first-order Hamiltonian systems
  15. Sharp asymptotic expansions of entire large solutions to a class of k-Hessian equations with weights
  16. Stability of pyramidal traveling fronts in time-periodic reaction–diffusion equations with degenerate monostable and ignition nonlinearities
  17. Ground-state solutions for fractional Kirchhoff-Choquard equations with critical growth
  18. Existence results of variable exponent double-phase multivalued elliptic inequalities with logarithmic perturbation and convections
  19. Homoclinic solutions in periodic partial difference equations
  20. Classical solution for compressible Navier-Stokes-Korteweg equations with zero sound speed
  21. Properties of minimizers for L2-subcritical Kirchhoff energy functionals
  22. Asymptotic behavior of global mild solutions to the Keller-Segel-Navier-Stokes system in Lorentz spaces
  23. Blow-up solutions to the Hartree-Fock type Schrödinger equation with the critical rotational speed
  24. Qualitative properties of solutions for dual fractional parabolic equations involving nonlocal Monge-Ampère operator
  25. Regularity for double-phase functionals with nearly linear growth and two modulating coefficients
  26. Uniform boundedness and compactness for the commutator of an extension of Riesz transform on stratified Lie groups
  27. Normalized solutions to nonlinear Schrödinger equations with mixed fractional Laplacians
  28. Existence of positive radial solutions of general quasilinear elliptic systems
  29. Low Mach number and non-resistive limit of magnetohydrodynamic equations with large temperature variations in general bounded domains
  30. Sharp viscous shock waves for relaxation model with degeneracy
  31. Bifurcation and multiplicity results for critical problems involving the p-Grushin operator
  32. Asymptotic behavior of solutions of a free boundary model with seasonal succession and impulsive harvesting
  33. Blowing-up solutions concentrated along minimal submanifolds for some supercritical Hamiltonian systems on Riemannian manifolds
  34. Stability of rarefaction wave for relaxed compressible Navier-Stokes equations with density-dependent viscosity
  35. Singularity for the macroscopic production model with Chaplygin gas
  36. Global strong solution of compressible flow with spherically symmetric data and density-dependent viscosities
  37. Global dynamics of population-toxicant models with nonlocal dispersals
  38. α-Mean curvature flow of non-compact complete convex hypersurfaces and the evolution of level sets
  39. High-energy solutions for coupled Schrödinger systems with critical growth and lack of compactness
  40. On the structure and lifespan of smooth solutions for the two-dimensional hyperbolic geometric flow equation
  41. Well-posedness for physical vacuum free boundary problem of compressible Euler equations with time-dependent damping
  42. On the existence of solutions of infinite systems of Volterra-Hammerstein-Stieltjes integral equations
  43. Remark on the analyticity of the fractional Fokker-Planck equation
  44. Continuous dependence on initial data for damped fourth-order wave equation with strain term
  45. Unilateral problems for quasilinear operators with fractional Riesz gradients
  46. Boundedness of solutions to quasilinear elliptic systems
  47. Existence of positive solutions for critical p-Laplacian equation with critical Neumann boundary condition
  48. Non-local diffusion and pulse intervention in a faecal-oral model with moving infected fronts
  49. Nonsmooth analysis of doubly nonlinear second-order evolution equations with nonconvex energy functionals
  50. Qualitative properties of solutions to the viscoelastic beam equation with damping and logarithmic nonlinear source terms
  51. Shape of extremal functions for weighted Sobolev-type inequalities
  52. One-dimensional boundary blow up problem with a nonlocal term
  53. Doubling measure and regularity to K-quasiminimizers of double-phase energy
  54. General solutions of weakly delayed discrete systems in 3D
  55. Global well-posedness of a nonlinear Boussinesq-fluid-structure interaction system with large initial data
  56. Optimal large time behavior of the 3D rate type viscoelastic fluids
  57. Local well-posedness for the two-component Benjamin-Ono equation
  58. Self-similar blow-up solutions of the four-dimensional Schrödinger-Wave system
  59. Existence and stability of traveling waves in a Keller-Segel system with nonlinear stimulation
  60. Existence and multiplicity of solutions for a class of superlinear p-Laplacian equations
  61. On the global large regular solutions of the 1D degenerate compressible Navier-Stokes equations
  62. Normal forms of piecewise-smooth monodromic systems
  63. Fractional Dirichlet problems with singular and non-locally convective reaction
  64. Sharp forced waves of degenerate diffusion equations in shifting environments
  65. Global boundedness and stability of a quasilinear two-species chemotaxis-competition model with nonlinear sensitivities
  66. Non-existence, radial symmetry, monotonicity, and Liouville theorem of master equations with fractional p-Laplacian
  67. Global existence, asymptotic behavior, and finite time blow up of solutions for a class of generalized thermoelastic system with p-Laplacian
  68. Formation of singularities for a linearly degenerate hyperbolic system arising in magnetohydrodynamics
  69. Linear stability and bifurcation analysis for a free boundary problem arising in a double-layered tumor model
  70. Hopf's lemma, asymptotic radial symmetry, and monotonicity of solutions to the logarithmic Laplacian parabolic system
  71. Generalized quasi-linear fractional Wentzell problems
  72. Existence, symmetry, and regularity of ground states of a nonlinear Choquard equation in the hyperbolic space
  73. Normalized solutions for NLS equations with general nonlinearity on compact metric graphs
  74. Boundedness and global stability in a predator-prey chemotaxis system with indirect pursuit-evasion interaction and nonlocal kinetics
  75. Review Article
  76. Existence and stability of contact discontinuities to piston problems
Downloaded on 29.1.2026 from https://www.degruyterbrill.com/document/doi/10.1515/anona-2025-0127/html
Scroll to top button