Startseite Dynamics for wave equations connected in parallel with nonlinear localized damping
Artikel Open Access

Dynamics for wave equations connected in parallel with nonlinear localized damping

  • Yunlong Gao , Chunyou Sun EMAIL logo und Kaibin Zhang
Veröffentlicht/Copyright: 19. Juni 2024

Abstract

This study investigates the properties of solutions about one-dimensional wave equations connected in parallel under the effect of two nonlinear localized frictional damping mechanisms. First, under various growth conditions about the nonlinear dissipative effect, we try to establish the decay rate estimates by imposing minimal amount of support on the damping and provide some examples of exponential decay and polynomial decay. To achieve this, a proper observability inequality has been proposed and constructed based on some refined microlocal analysis. Then, the existence of a global attractor is proved when the damping terms are linearly bounded at infinity, a special weighting function has been used in this part, which eliminates undesirable terms of the higher order while contributing lower-order terms. Finally, we establish that the long-time behavior of solutions of the nonlinear system is completely determined by the dynamics of large finite number of functionals.

MSC 2010: 35B40; 35B41; 35L05; 35L20; 35Q74; 37L05

1 Introduction

The well-posedness, stability, blowup of solutions, and existence of attractors for evolution wave systems have attracted the attention of many scholars (e.g., [1,2,513,1927,2942,4650]). In this work, our main aim is to investigate the long-time dynamical behavior of the following system of coupled wave equations in parallel with nonlinear localized damping:

(1.1) ρ ( x ) u t t ( K ( x ) u x ) x + α ( u v ) + β 1 ( x ) g 1 ( u t ) + f 1 ( u , v ) = 0 , in ( 0 , L ) × ( 0 , ) , ρ ( x ) v t t ( K ( x ) v x ) x α ( u v ) + β 2 ( x ) g 2 ( v t ) + f 2 ( u , v ) = 0 , in ( 0 , L ) × ( 0 , ) ,

subject to initial-boundary conditions

(1.2) u ( 0 , t ) = u ( L , t ) = v ( 0 , t ) = v ( L , t ) = 0 , t > 0 , u ( x , 0 ) = u 0 ( x ) , u t ( x , 0 ) = u 1 ( x ) , v ( x , 0 ) = v 0 ( x ) , v t ( x , 0 ) = v 1 ( x ) , x ( 0 , L ) ,

where L > 0 is the length of the bounded region ( 0 , L ) , α > 0 is the spring coupling coefficient, ρ : ( 0 , L ) R + , K : ( 0 , L ) R + are C functions such that for all x ( 0 , L ) ,

(1.3) ρ 1 ρ ( x ) ρ 2 , K 1 K ( x ) K 2 ,

where ρ 1 , ρ 2 , K 1 , K 2 are positive constants. β i L ( 0 , L ) is the nonnegative function such that

(1.4) β i ( x ) β i ˜ > 0 , x I i , i = 1 , 2 ,

where I 1 , I 2 are open intervals contained in [ 0 , L ] satisfying ω I 1 I 2 . g 1 ( u t ) , g 2 ( v t ) represent the nonlinear damping terms and f 1 ( u , v ) , f 2 ( u , v ) are the nonlinear source terms.

There have been some main results regarding the study of linear and nonlinear parallel coupled wave equations. For example, Li and Wang [36] studied the conditions for the existence of partially synchronized solutions in a coupled dissipationless system of one-dimensional wave equations with different wave speeds within the framework of classical solutions. If the principal operator of System (1.1) in three-dimensional space is the Lamé operator, then the Lamé system is equivalent to three coupled wave equations. Yang et al. [50] studied the stability of solutions and the existence of global and exponential attractors for weakly damped Lamé systems with nonlinear time-varying delay. By considering different boundary conditions and wave propagation speeds, Najafi and Wang [39] proved the exponential stability of the linear system (1.1) which contained the linear full damping and without nonlinear source. Rajaram and Najafi [42] further proved that such system achieved an exact controllability result with Dirichlet boundary controls, without any restrictions on the coupling parameters. Moreover, the control time was independent of the coupling constants and was determined by geometric optics. Wu et al. [48] also investigated the exact controllability of the linear system (1.1) in the case where the principal part has variable coefficients. For the nonlinear system (1.1) and (1.2) with nonlinear full damping and source terms, Freitas et al. [26] had studied the singular limit dynamics of equations (1.1) and (1.2). To the best of our knowledge, Problems (1.1) and (1.2) with local damping have not been studied yet. For more details on the parallel connection wave equation, please refer to [40,41].

In recent years, scientists have shown a growing interest in studying the behavior of solutions to nonlinear models. Several of these models, such as Timoshenko system [7,8,21], Bresse system [11,38], porous elastic system [19,27], Lamé system [50], among others, are constituted by coupled nonlinear damped wave equations. These dissipative wave systems can be divided into fully (locally) dissipative systems (i.e., all the damping coefficients in the study area are strictly greater than zero) and partially (nonlocally) dissipative systems based on the damping coefficients. Generally, the study of fully dissipative systems is relatively simple, as we do not need to overly consider how the geometric characteristics of the study area affect dissipation. There are many achievements in this area, such as research on the stability, existence of attractors, and precise controllability of fully damped wave systems, we can refer to [21,26,27,2931,3942,4750] and other related studies.

In terms of physics, partially dissipative wave systems reflect the mechanism of localized energy loss in the medium. The energy of the wave will gradually decrease with the propagation distance but not completely disappear. This localized energy loss can reflect more better the energy attenuation characteristics of the medium in real systems. In terms of control and optimization, partially dissipative wave systems often have more applications. By designing control strategies, it is possible to adjust the region or frequency range of localized dissipation, thereby achieving goal control. Compared to fully dissipative wave systems, the study of partially dissipative systems is more challenging due to the potential requirement for geometric control conditions (GCCs) (e.g., [1,3,5]), and the scarcity of relevant literature, particularly for coupled systems. For example, Zuazua [22], Feireisl and Zuazua [23] investigated the exponential decay and existence of global attractors for semilinear single wave equation with locally distributed damping and nonlinear source term, and required to consider the case where the damping term grows linearly at infinity. For single wave equations with localized damping satisfying sublinear, or linear, or superlinear at infinity, their uniform decay rates had been studied in [9,16,17,35]. Note that these works all relied on the unique continuation property (UCP) of single wave equation proved in [33], which dictated whether wave equations that vanished in a subdomain must be identically null. For coupled wave systems with localized linear damping, the UCP can be derived from Holmgren’s uniqueness theorem, which is valid for equations with analytic coefficients [8,11,16]. However, the UCP for coupled wave systems with nonanalytic coefficients had been an open problem for a long time. Therefore, many previous studies on coupled wave equations with localized nonlinear damping and source terms treated the UCP as an a priori (without proof) condition (e.g., [7,20,24]). Recently, Cavalcanti et al. [10] proved the UCP of solutions to the general coupled wave equation system in hyper two-dimensional space in 2020. After that, Ma et al. [38] also obtained a new UCP for a system of one-dimensional wave equations through Carleman-type estimates. Combining the UCP with GCCs [1,3,5], the uniform stability and existence of attractors for various types of coupled hyperbolic equations (e.g., Klein-Gordon equation, Timoshenko system, Bress system, etc.) with localized damping had been studied in [2,6,7,10,25,38].

On the other hand, the coupled fully damped wave equations had shown some results in terms of stability under the influence of damping terms with different growth conditions at infinity (e.g., [19,27,29,30]). On the contrary, for the case of localized damping, the current results only consider the behavior of solutions with linear growth at infinity, and there are no stability results for other growth cases (i.e., sublinear, superlinear), [2,68,10,11,24,25,38,41]. Moreover, it is worth noting that Ferreira [24] also raised an open question about the mixed growth conditions, i.e., what is the stability of a coupled hyperbolic system with localized damping satisfying different growth conditions at infinity?

In this study, we will consider the uniform stability of solutions to Systems (1.1) and (1.2) under nine different combinations of growth conditions for nonlinear damping at infinity (Theorem 3.1). In addition, in contrast to the study of fully damped systems [26], we consider the existence of global attractors and determining functionals for the coupled system by imposing a minimal amount of support for the damping (Theorems 4.2 and 5.1).

In terms of stability analysis of Systems (1.1) and (1.2), we mainly face three challenges or innovations. First, we consider damping with different growth conditions at infinity, i.e., g 1 and g 2 can include sublinear, linear, superlinear cases, as shown in Table 1, which makes analysis and calculation more complex and difficult. Second, once we are dealing with locally distributed nonlinear damping β 1 ( x ) g 1 ( u t ) and β 2 ( x ) g 2 ( v t ) , the standard multiplier technique (e.g., [19,21]) fails as well because of the terms in the energy estimates that cannot be absorbed directly. Finally, we concern the generalized version of (1.1) with nonlinearities of the form f 1 ( u , v ) and f 2 ( u , v ) without growth restrictions. Hence, the reduction principle method proposed in [16], which reduces nonlinear stabilization to an observability inequality established for the corresponding linear problem, may not be a reliable approach. In order to overcome these difficulties, our approach is to prove the stabilizations (1.1) and (1.2) reduced themselves to prove an observability inequality, whose proof is now based on contradiction arguments in combination with microlocal analysis (e.g., [2,4,6,7,9,10,16,28]). On the other hand, in studying the infinite-dimensional dynamical system of the Systems (1.1) and (1.2), the main challenge is to prove that the dynamical system ( , S α ( t ) ) is gradient (Lemma 4.1) and asymptotically smooth (Lemma 4.3). To this end, we utilize the new UCP from Ma et al. [38] (Lemma 3.1) to verify that ( , S α ( t ) ) is gradient. However, during the proof of Lemma 4.3, we observe that the standard multipliers generate the terms of the energy level that cannot be absorbed directly, unlike in the case of a fully damping system, as demonstrated in [19,21,26,27]. To overcome this difficulty, some special weight functions are introduced, which eliminate undesirable terms of the higher order while contributing lower-order terms [8,11,32,38].

Table 1

( , ) represents the ordered array of damping g 1 and g 2 at infinity, where “L” means linear at infinity, “Sub” means sublinear at infinity, “Super” means superlinear at infinity. For example, “(L, Sub)” is interpreted as g 1 and g 2 are linear and sublinear at infinity, respectively.

Combination (L, L) (Sub, Sub) (Super, Super)
h ˜ ( s ) s s θ 1 2 θ 1 q 1 1 + s θ 2 2 θ 2 q 2 1 , ( θ 1 , θ 2 > 2 ) s θ 1 2 q 1 θ 1 q 1 1 + s θ 2 2 q 2 θ 2 q 2 1 , ( θ 1 > 2 q 1 , θ 2 > 2 q 2 )
Combination (L, Sub) (L, Super) (Sub, Super)
h ˜ ( s ) s + s θ 2 2 θ 2 q 2 1 , ( θ 2 > 2 ) s + s θ 2 2 q 2 θ 2 q 2 1 , ( θ 2 > 2 q 2 ) s θ 1 2 θ 1 q 1 1 + s θ 2 2 q 2 θ 2 q 2 1 , ( θ 1 > 2 , θ 2 > 2 q 2 )
Combination (Sub, L) (Super, L) (Super, Sub)
h ˜ ( s ) s θ 1 2 θ 1 q 1 1 + s , ( θ 1 > 2 ) s θ 1 2 q 1 θ 1 q 1 1 + s , ( θ 1 > 2 q 1 ) s θ 1 2 q 1 θ 1 q 1 1 + s θ 2 2 θ 2 q 2 1 , ( θ 1 > 2 q 1 , θ 2 > 2 )

The manuscript is organized as follows: In Section 2, we present the notations needed, we list the standing assumptions on the nonlinear terms and give well-posedness results. In Section 3, we prove the uniform decay rates of energy under the effect of two nonlinear localized frictional viscous mechanisms. In Section 4, we study the existence of global attractors. In Section 5, we give the construction of determining functionals. In Section 6, we state microlocal analysis background and abstract results of dynamical systems throughout the manuscript.

2 Preliminaries and well-posedness

In this section, we will present the notations, assumptions, and well-posedness of solutions for Problems (1.1) and (1.2).

2.1 Preliminaries

The relation a ( s ) b ( s ) will stand for a ( s ) C b ( s ) for some positive constant C independent of s . The relation a ( s ) b ( s ) will indicate that a ( s ) b ( s ) and b ( s ) a ( s ) . Next we begin with some notations on the standard space L p ( 0 , L ) ( 1 p + ) and Sobolev space H 0 1 ( 0 , L ) with dual space H 1 ( 0 , L ) . The L p norm is denoted by

u p if p 2 , u if p = 2 ,

and as usual ( , ) denotes L 2 -inner product. For the Sobolev space H 0 1 ( 0 , L ) , we have

u L π u x and u H 0 1 = u x .

Now, let us consider the Hilbert space

H 0 1 ( 0 , L ) × L 2 ( 0 , L ) × H 0 1 ( 0 , L ) × L 2 ( 0 , L ) ,

where inner product and norm on are defined as follows:

(2.1) U 1 , U 2 = 0 L [ ρ ( x ) ( ϕ 1 ϕ 2 + φ 1 φ 2 ) + K ( x ) ( u x 1 u x 2 + v x 1 v x 2 ) + α ( u 1 v 1 ) ( u 2 v 2 ) ] d x

and

(2.2) U i 2 = 0 L [ ρ ( x ) ( ϕ i 2 + φ i 2 ) + K ( x ) ( u x i 2 + v x i 2 ) + α ( u i v i ) 2 ] d x

for U i = ( u i , ϕ i , v i , φ i ) , i = 1 , 2 .

Finally, we consider the following some assumptions on g i and f i , i = 1 , 2 .

Assumption 2.1

For any i = 1 , 2 , we assume that

  1. g i : R R is continuous, monotone increasing, and satisfies

    (2.3) g i ( 0 ) = 0 , g i ( s ) s > 0 for s 0 ,

    and there exist constants m i , M i > 0 such that for all s 1 ,

    (2.4) m i s q i + 1 g i ( s ) s M i s q i + 1 , with q i 0 .

  2. f i C 1 ( R 2 ) and f i ( 0 , 0 ) = 0 , and there is a constant C f > 0 such that

    (2.5) f i ( u , v ) C f ( 1 + u p 1 + v p 1 ) , with p 1 .

  3. There exists a function F C 2 ( R 2 ) such that

    (2.6) F = ( f 1 , f 2 ) ,

    and there exists 0 β < K 1 π 2 2 L 2 such that, for all u , v R ,

    (2.7) F ( u , v ) β ( u 2 + v 2 )

    and

    (2.8) F ( u , v ) ( u , v ) F ( u , v ) β ( u 2 + v 2 ) .

Remark 2.1

By assumption (A1), if the order q i is bigger than, less than, or equal to 1, we say the map g i is superlinear, sublinear, or linearly bounded at infinity, respectively.

Remark 2.2

There is a large class of nonlinear force satisfying (A2) and (A3). For instance,

F ( u , v ) = a u + v p + 1 b u + v 2 + 2 c u v p + 1 2 , a , b , c > 0 .

2.2 Local and global well-posedness

Under the above assumptions and notations, we are able to state well-posedness of the Problems (1.1) and (1.2). To this end, denoting ϕ = u t , φ = v t , U = ( u , ϕ , v , φ ) , and U 0 = ( u 0 , u 1 , v 0 , v 1 ) , then Problems (1.1) and (1.2) can be rewritten as the following Cauchy problem in :

(2.9) d U ( t ) d t + A U ( t ) = ( U ( t ) ) , t > 0 , U ( 0 ) = U 0 ,

where the nonlinear unbounded operator A : D ( A ) is given by

(2.10) A U = ϕ 1 ρ ( x ) ( ( K ( x ) u x ) x + α ( u v ) + β 1 ( x ) g 1 ( ϕ ) ) φ 1 ρ ( x ) ( ( K ( x ) v x ) x α ( u v ) + β 2 ( x ) g 2 ( φ ) ) .

Here the domain of A is

D ( A ) = U ϕ , φ H 0 1 ( 0 , L ) , A U ( L 2 ( 0 , L ) ) 4 , β 1 ( x ) g 1 ( ϕ ) L 1 ( 0 , L ) H 1 ( 0 , L ) , and β 2 ( x ) g 2 ( φ ) L 1 ( 0 , L ) H 1 ( 0 , L ) .

The autonomous forcing terms are represented by a nonlinear function : defined by

(2.11) ( U ) = 0 1 ρ ( x ) f 1 ( u , v ) 0 1 ρ ( x ) f 2 ( u , v ) .

Next, we give the definition of weak and strong solutions for Problems (1.1) and (1.2).

Definition 2.1

A function U = ( u , u t , v , v t ) C ( [ 0 , T ) ; ) with U 0 = ( u 0 , u 1 , v 0 , v 1 ) is called a weak solution to (1.1) and (1.2), if the following identity is satisfied in the sense of distributions on [ 0 , T )

(2.12) d d t ( ρ ( x ) u t , ς ) + ( K ( x ) u x , ς x ) + α ( u v , ς ) + ( β 1 ( x ) g 1 ( u t ) , ς ) + ( f 1 ( u , v ) , ς ) = 0 ,

(2.13) d d t ( ρ ( x ) v t , ξ ) + ( K ( x ) v x , ξ x ) α ( u v , ξ ) + ( β 2 ( x ) g 2 ( v t ) , ξ ) + ( f 2 ( u , v ) , ξ ) = 0 ,

for all ς H 0 1 ( 0 , L ) and ξ H 0 1 ( 0 , L ) , and for some T > 0 . If a weak solution satisfies further

U C ( [ 0 , T ) ; D ( A ) ) C 1 ( [ 0 , T ) ; ) ,

then it is called a strong solution.

The well-posedness of Problems (1.1) and (1.2) is given in the following theorem.

Theorem 2.1

(Local and global in time) Suppose that (A1) and (A2) hold in Assumption 2.1.

  1. If q 1 = q 2 = 1 and U 0 , Problems (1.1) and (1.2) have a weak solution satisfying U C ( [ 0 , T max ) ; ) . If U 0 D ( A ) , Problems (1.1) and (1.2) have a strong solution satisfying U W 1 , ( [ 0 , T max ) ; ) , where T max is maximal existence time of solution ( u , v ) .

  2. (Energy identity) The following energy identity holds true:

    (2.14) U ( t ) + s t 0 L ( β 1 ( x ) g 1 ( u t ) u t + β 2 ( x ) g 2 ( v t ) v t ) d x d t = U ( s ) ,

    where 0 s < t < T max and the total energy U ( t ) is defined as

    U ( t ) 1 2 U 2 + 0 L F ( u , v ) d x .

  3. (Continuous dependence) The weak solution U C ( [ 0 , T max ) ; ) depends continuously on the initial data U 0 . That is, if U i ( t ) = ( u i , u t i , v i , v t i ) , i = 1 , 2 , are two solutions corresponding to initial data U 0 i ( t ) = ( u 0 i , u 1 i , v 0 i , v 1 i ) to Problems (1.1) and (1.2), then there exists a constant C ( U 0 1 , U 0 2 ) > 0 such that

    (2.15) U 1 ( t ) U 2 ( t ) e C T U 0 1 U 0 2 , 0 t T < T max .

    In particular, solution is unique.

  4. Further, if the assumption (A3) holds, then T max = + .

Proof

By employing the standard arguments of nonlinear semigroup theory, as outlined in [43] and Theorem 7.2 from [13], we can prove (i)–(iii) proceed with our proof following the methodology presented in the proofs of Theorems 2.4 and 2.5 in [27] (also referring to [9,16,30]). In order to prove (iv), we can use contradiction arguments. Indeed, let us suppose that T max < + . Using the assumption (2.3), (A3), and the energy identity (2.14), we obtain

(2.16) d d t U ( t ) = 0 L ( β 1 ( x ) g 1 ( u t ) u t + β 2 ( x ) g 2 ( v t ) v t ) d x 0 , t [ 0 , T max ) ,

in which U ( t ) is not-increasing. Thus, U ( t ) U ( 0 ) , for all t [ 0 , T max ) . On the other hand, from Assumption (2.7) and definition of U ( t ) yield

(2.17) 1 2 β L 2 π 2 K 1 U 2 U ( t ) , t [ 0 , T max ) ,

that is,

U 2 2 π 2 K 1 K 1 π 2 2 β L 2 U ( 0 ) , t [ 0 , T max ) .

Therefore, it follows that lim t T max U 2 < , together with Theorem 7.2 in [9], allows us to deduce T max = + immediately.□

Remark 2.3

If the inequalities (2.7) and (2.8) in (A3) are replaced by

F ( u , v ) β ( u 2 + v 2 )

and

F ( u , v ) F ( u , v ) ( u , v ) β ( u 2 + v 2 ) ,

then we can establish a potential well framework proposed by Sattinger [44] to prove that (iv) in Theorem 2.1 also holds with small initial energy. So under these conditions, we can also study the uniform stability of the third part [27,2931,47].

3 Asymptotic stability

In dissipative dynamical systems, the uniform stability of energy plays a significant role in control and optimization, analyzing long-term behavior, model validation, and more, especially in cases of polynomial and exponential decay. Therefore, in this section, we will investigate the uniform decay rate of energy for equations (1.1) and (1.2) by establishing an observability inequality.

3.1 Observability inequality

Before proving the observability inequality, we still need to make some assumptions. Due to the presence of localized damping and nonconstant coefficients ρ ( x ) , K ( x ) in Problem (1.1), we need to assume the following GCC to study stability. It is well-known that it is necessary and sufficient for stabilization and control of the linear wave equation, [1,3,5].

Assumption 3.1

ω geometrically controls the area ( 0 , L ) , that is, there exists T 0 > 0 , such that every geodesic of the metric G ( x ) , where G ( x ) = K ( x ) ρ ( x ) 1 , traveling with speed 1 and issued at t = 0 , intercepts ω in a time t < T 0 .

Remark 3.1

When we do not have any control on the geodesics of the metric G ( x ) = K ( x ) ρ ( x ) 1 , we have to assume damping everywhere on ( 0 , L ) , satisfying the following assumptions:

  1. For all x ( 0 , L ) , β i ( x ) > 0 , i = 1 , 2 .

  2. For all geodesic t I x ( t ) ( 0 , L ) of the metric G ( x ) = K ( x ) ρ ( x ) 1 , with 0 I , there exists t 0 such that β i ( x ( t ) ) > 0 , i = 1 , 2 .

If G = I d , the condition (a) above implies that (b) holds since the geodesics are straight. On the other hand, for a general metric condition, (a) can be verified without (b) holding, for instance, G admits a trapped bicharacteristics, Figure 1 in [7].

One of the ingredients for obtaining stability of wave equations with localized damping is a UCP. So we make the following lemma:

Lemma 3.1

Suppose that Assumptions (1.3) and (1.4) hold and K ( x ) ρ ( x ) is a definite constant. Then, for T > 0 large enough, any weak solution

( u , v ) C ( ( 0 , T ) ; ( L 2 ( 0 , L ) ) 2 ) C 1 ( ( 0 , T ) ; ( H 1 ( 0 , L ) ) 2 )

of problem

(3.1) ρ ( x ) u t t ( K ( x ) u x ) x + V 1 ( x , t ) u = V 3 ( x , t ) v , i n ( 0 , L ) × ( 0 , T ) , ρ ( x ) v t t ( K ( x ) v x ) x + V 2 ( x , t ) v = V 4 ( x , t ) u , i n ( 0 , L ) × ( 0 , T ) , u = v = 0 , i n ω × ( 0 , T ) ,

where V i ( x , t ) L 2 ( ( 0 , T ) ; L 2 ( 0 , L ) ) , i = 1 , 2 , 3 , 4 , and initial-boundary condition is (1.2), must vanish all over [ 0 , L ] × [ 0 , T ] .

Proof

Its proof can be obtained from Theorem 3.2 in [38].□

Remark 3.2

The lemma is called UCP. In the case of a single equation, these kinds of results were studied by Koch and Tataru [33], but for coupled systems, it remained an open problem before the year 2020. Recently, inspired by Dos Santos Ferreira [18] and Koch and Tataru [33], Cavalcanti et al. ([6], Proposition 2.3 and [10], Proposition 3.6) and Faria and Souza Franco ([20], Proposition 1.2) gave an example where the UCP holds at least locally, i.e., for the system

(3.2) P 1 ( x , D ) u + V 1 ( x , t ) u = V 3 ( x , t ) v , in Ω × ( 0 , T ) , P 2 ( x , D ) v + V 2 ( x , t ) v = V 4 ( x , t ) u , in Ω × ( 0 , T ) ,

where Ω R n , n > 2 , P 1 ( x , D ) , and P 2 ( x , D ) are second-order differential operators of real principal type with C coefficients, and V i ( x , t ) , i = 1 , 2 , 3 , 4 are elements of L ( ( 0 , T ) ; L n + 1 2 ( 0 , L ) ) . Afterwards Ma et al. [38] also gave a new UCP for one-dimensional coupled wave equations. The word “new” is reflected in the fact that there is no need to consider the analyticity of system coefficients (Charles et al. [11] had used a UCP derived from Holmgren uniqueness theorem [16], which is only valid for equations with analytic coefficients).

Now, we state and prove the observability inequality lemma for Problems (1.1) and (1.2).

Lemma 3.2

Under the assumptions of Assumptions 2.1, 3.1, and Lemma 3.1, let R > 0 be given. Then, for T > 0 large enough, there exists a constant C = C ( R ) > 0 such that every solution of (1.1) and (1.2) with ( u 0 , u 1 , v 0 , v 1 ) D ( A ) (specifically, if q 1 = q 2 = 1 , with ( u 0 , u 1 , v 0 , v 1 ) ) away from zero and U ( 0 ) R satisfies

(3.3) T U ( T ) C 0 T 0 L { β 1 ( x ) [ u t 2 ( t ) + g 1 2 ( u t ( t ) ) ] + β 2 ( x ) [ v t 2 ( t ) + g 2 2 ( v t ( t ) ) ] } d x d t .

Proof

If q 1 = q 2 = 1 and ( u 0 , u 1 , v 0 , v 1 ) , it is enough to show that (3.3) holds for regular solutions, and to then use a density argument. So, assuming ( u 0 , u 1 , v 0 , v 1 ) D ( A ) , we first assert that if ( u 0 , u 1 , v 0 , v 1 ) is far from zero, then U ( 0 ) > 0 . Indeed, we only need to consider the steady-state problem of the Systems (1.1) and (1.2), i.e.,

(3.4) ( K ( x ) u x ) x + α ( u v ) + f 1 ( u , v ) = 0 , in ( 0 , L ) , ( K ( x ) v x ) x α ( u v ) + f 2 ( u , v ) = 0 , in ( 0 , L ) , u ( 0 ) = u ( L ) = v ( 0 ) = v ( L ) = 0 .

If we multiply (3.4) 1 by u and (3.4) 2 by v , integrate over ( 0 , L ) and utilize (2.7) and (2.8), we have

K 1 π 2 L 2 2 β ( u 2 + v 2 ) 2 E U ( t ) = 0 .

This implies that u = v = 0 , i.e., this assertion holds.

Next our proof relies on contradiction arguments. If Lemma 3.2 does not hold, then there exists T > T 0 such that for all C > 0 , there exists a solution U C ( u C , u t C , v C , v t C ) for (1.1) and (1.2) verifying 0 < U C ( 0 ) R and such that U C violates (3.3). In particular, for each m N , we obtain the existence of a solution U m = ( u m , u t m , v m , v t m ) for Problems (1.1) and (1.2) verifying 0 < U m ( 0 ) R , whose corresponding solution satisfies the reverse inequality

U m ( T ) m T 0 T 0 L { β 1 ( x ) [ ( u t m ) 2 + ( g 1 ( u t m ) ) 2 ] + β 2 ( x ) [ ( v t m ) 2 + ( g 2 ( v t m ) ) 2 ] } d x d t .

By (2.16), it implies that U m ( T ) U m ( 0 ) . Then, we can obtain a sequence { U m } m N of solutions to Problems (1.1) and (1.2) such that

lim m + U m ( 0 ) 0 T 0 L { β 1 ( x ) [ ( u t m ) 2 + ( g 1 ( u t m ) ) 2 ] + β 2 ( x ) [ ( v t m ) 2 + ( g 2 ( v t m ) ) 2 ] } d x d t = + ,

which is equivalent to

(3.5) lim m + 0 T 0 L { β 1 ( x ) [ ( u t m ) 2 + ( g 1 ( u t m ) ) 2 ] + β 2 ( x ) [ ( v t m ) 2 + ( g 2 ( v t m ) ) 2 ] } d x d t U m ( 0 ) = 0 .

Once U m ( 0 ) R , for all m N , we can infer from (3.5) that

(3.6) lim m + 0 T 0 L { β 1 ( x ) [ ( u t m ) 2 + ( g 1 ( u t m ) ) 2 ] + β 2 ( x ) [ ( v t m ) 2 + ( g 2 ( v t m ) ) 2 ] } d x d t = 0 .

Due to the lengthy proof process, we will prove the lemma in the following four steps:

Step 1. Assert ( u m , v m ) ( 0 , 0 ) strongly in L ( 0 , T ; ( L q ( 0 , L ) ) 2 ) , as m + .

On the one hand, since U m ( t ) U m ( 0 ) R , for all t > 0 , we deduce there exists a subsequence of { U m } m N (without changing notation), such that

(3.7) ( u m , v m ) ( u , v ) weakly-star in L ( 0 , T ; ( H 0 1 ( 0 , L ) ) 2 ) , as m + ,

(3.8) ( u t m , v t m ) ( u t , v t ) weakly-star in L ( 0 , T ; ( L 2 ( 0 , L ) ) 2 ) , as m + ,

and from Aubin-Lions theorem ([45], Corollary 4), we obtain

(3.9) ( u m , v m ) ( u , v ) strongly in L ( 0 , T ; ( L q ( 0 , L ) ) 2 ) , as m + , for all q [ 2 , + ) ,

which yields that, as m + ,

(3.10) ( f 1 ( u m , v m ) , f 2 ( u m , v m ) ) ( f 1 ( u , v ) , f 2 ( u , v ) ) strongly in L 2 ( 0 , T ; ( L 2 ( 0 , L ) ) 2 ) .

Indeed, by (2.5), (3.9), Hölder’s inequality, and the embedding H 0 1 ( 0 , L ) L 4 p 4 ( 0 , L ) , we obtain

(3.11) 0 L f 1 ( u m , v m ) f 1 ( u , v ) 2 d x 0 L f 1 ( θ ˆ ) ( u m , v m ) + ( 1 θ ˆ ) ( u , v ) 2 ( u m , v m ) ( u , v ) 2 d x 0 L ( 1 + u m 2 p 2 + v m 2 p 2 + u 2 p 2 + v 2 p 2 ) ( u m , v m ) ( u , v ) 2 d x ( 1 + u x m 2 p 2 + v x m 2 p 2 + u x 2 p 2 + v x 2 p 2 ) ( u m u 4 2 + v m v 4 2 ) u m u 4 2 + v m v 4 2 0 , as m + ,

where 0 θ ˆ 1 . And analogously, we have 0 L f 2 ( u m , v m ) f 2 ( u , v ) 2 d x 0 , as m + .

At this point in the proof we will divide it into two cases: Case (a): u 0 and v 0 , Case (b): ( u , v ) = ( 0 , 0 ) . We note that these are the only two cases we should consider. In other words, it means that the situation where u = 0 , v 0 and u 0 , v = 0 will not occur. Indeed, if u = 0 then necessarily v = 0 . Because if we take the following sequence of problems into account:

(3.12) ρ ( x ) u t t m ( K ( x ) u x m ) x + α ( u m v m ) + β 1 ( x ) g 1 ( u t m ) + f 1 ( u m , v m ) = 0 , in ( 0 , L ) × ( 0 , T ) , ρ ( x ) v t t m ( K ( x ) v x m ) x α ( u m v m ) + β 2 ( x ) g 2 ( v t m ) + f 2 ( u m , v m ) = 0 , in ( 0 , L ) × ( 0 , T ) , u m ( 0 , t ) = u m ( L , t ) = v m ( 0 , t ) = v m ( L , t ) = 0 , t ( 0 , T ) , u m ( x , 0 ) = u 0 m ( x ) , u t m ( x , 0 ) = u 1 m ( x ) , x ( 0 , L ) , v m ( x , 0 ) = v 0 m ( x ) , v t m ( x , 0 ) = v 1 m ( x ) , x ( 0 , L ) ,

passing to the limit in (3.12) and taking (3.6)–(3.11) into account, we obtain

(3.13) α v + f 1 ( 0 , v ) = 0 , in ( 0 , L ) × ( 0 , T ) , ρ ( x ) v t t ( K ( x ) v x ) x + α v + f 2 ( 0 , v ) = 0 , in ( 0 , L ) × ( 0 , T ) , v ( 0 , t ) = v ( L , t ) = 0 , t ( 0 , T ) , v t = 0 , in ω × ( 0 , T ) .

Thus, for z = v t from (3.13) we have that

(3.14) ρ ( x ) z t t ( K ( x ) z x ) x + ( v f 1 ( 0 , v ) + v f 2 ( 0 , v ) ) z = 0 , in ( 0 , L ) × ( 0 , T ) , z ( 0 , t ) = z ( L , t ) = 0 , t ( 0 , T ) , z = 0 , in ω × ( 0 , T ) ,

which implies that z = v t 0 a.e. in ( 0 , L ) × ( 0 , T ) . Indeed, using Assumption (2.5) and v L ( 0 , T ; H 0 1 ( 0 , L ) ) , we obtain v f 1 ( 0 , v ) + v f 2 ( 0 , v ) L 2 ( 0 , T ; L 2 ( 0 , L ) ) . By combining the unique continuity property of a single wave equation, it can be proved. Therefore, (3.13) is rewritten as

(3.15) ( K ( x ) v x ) x + α v + f 2 ( 0 , v ) = 0 , in ( 0 , L ) × ( 0 , T ) , v ( 0 , t ) = v ( L , t ) = 0 , t ( 0 , T ) .

Multiplying (3.15) by v , integrating the resultant equations over ( 0 , L ) , and observing (3.15) 2 , we obtain that

(3.16) π 2 K 1 L 2 2 β v 2 0 L K ( x ) v x 2 d x + 0 L ( α v + f 2 ( 0 , v ) ) v d x = 0 .

Thus, we obtain that v = 0 . Analogously, if v = 0 , then u = 0 .

Case (a): u 0 and v 0 .

For each m N , let ( u m , u t m , v m , v t m ) be the solution of Problem (3.12). Taking (3.6), (3.7), and (3.11) into consideration, we have that

(3.17) ρ ( x ) u t t ( K ( x ) u x ) x + α ( u v ) + f 1 ( u , v ) = 0 , in ( 0 , L ) × ( 0 , T ) , ρ ( x ) v t t ( K ( x ) v x ) x α ( u v ) + f 2 ( u , v ) = 0 , in ( 0 , L ) × ( 0 , T ) , u ( 0 , t ) = u ( L , t ) = v ( 0 , t ) = v ( L , t ) = 0 , t ( 0 , T ) , u t = v t = 0 , in ω × ( 0 , T ) .

Taking the derivatives of (3.17) 1 and (3.17) 2 with respect to t and defining w = u t and z = v t , we deduce that

(3.18) ρ ( x ) w t t ( K ( x ) w x ) x + α ( w z ) + u f 1 ( u , v ) w + v f 1 ( u , v ) z = 0 , in ( 0 , L ) × ( 0 , T ) , ρ ( x ) z t t ( K ( x ) z x ) x α ( w z ) + v f 2 ( u , v ) z + u f 2 ( u , v ) w = 0 , in ( 0 , L ) × ( 0 , T ) , w ( 0 , t ) = w ( L , t ) = z ( 0 , t ) = z ( L , t ) = 0 , t ( 0 , T ) , w = z = 0 , in ω × ( 0 , T ) .

If we consider the following functions:

V 1 ( x , t ) = u f 1 ( u , v ) + α , V 2 ( x , t ) = v f 2 ( u , v ) + α , V 3 ( x , t ) = α v f 1 ( u , v ) , V 4 ( x , t ) = α u f 2 ( u , v ) ,

by using Assumptions (2.5) and (3.7), one can easily obtain V i L 2 ( 0 , T ; L 2 ( 0 , L ) ) , i = 1 , 2 , 3 , 4 . Thereby in light of Lemma 3.1, we conclude that

w = z 0 , a.e. in ( 0 , L ) × ( 0 , T ) ,

and then

u t = v t 0 , a.e. in ( 0 , L ) × ( 0 , T ) ,

which implies u t t = v t t 0 a.e. in ( 0 , L ) × ( 0 , T ) . Thus, (3.17) can be written as

(3.19) ( K ( x ) u x ) x + α ( u v ) + f 1 ( u , v ) = 0 , in ( 0 , L ) × ( 0 , T ) , ( K ( x ) v x ) x α ( u v ) + f 2 ( u , v ) = 0 , in ( 0 , L ) × ( 0 , T ) , u ( 0 , t ) = u ( L , t ) = v ( 0 , t ) = v ( L , t ) = 0 , t ( 0 , T ) .

Finally, multiplying (3.19) 1 by u plus (3.19) 2 by v , integrating the resultant equations over ( 0 , L ) and observing (3.19) 3 , we deduce that

(3.20) 0 L K ( x ) ( u x 2 + v x 2 ) d x + α u v 2 + 0 L ( f 1 ( u , v ) u + f 2 ( u , v ) v ) d x = 0 .

Again using Assumptions (2.7) and (2.8), we infer that

(3.21) π 2 K 1 L 2 2 β ( u 2 + v 2 ) K 1 ( u x 2 + v x 2 ) + α u v 2 2 β ( u 2 + v 2 ) 0 ,

so u = v = 0 , which is contradiction.

Case (b): ( u , v ) = ( 0 , 0 ) .

From (3.7), (3.8), and (3.9), now we have that

(3.22) ( u m , v m ) ( 0 , 0 ) weakly-star in L ( 0 , T ; ( H 0 1 ( 0 , L ) ) 2 ) , as m + ,

(3.23) ( u t m , v t m ) ( 0 , 0 ) weakly-star in L ( 0 , T ; ( L 2 ( 0 , L ) ) 2 ) , as m + ,

(3.24) ( u m , v m ) ( 0 , 0 ) strongly in L ( 0 , T ; ( L q ( 0 , L ) ) 2 ) , as m + , for all q [ 2 , + ) .

Step 2. Assert ( w m , z m ) ( 0 , 0 ) strongly in L ( 0 , T ; ( L q ( 0 , L ) ) 2 ) , as m + .

Now, we define

(3.25) λ m U m ( 0 ) , w m u m λ m and z m v m λ m .

Using (3.5) and the definition of λ m , we obtain

(3.26) lim m + 0 T 0 L β 1 ( x ) ( w t m ) 2 + ( g 1 ( λ m w t m ) ) 2 λ m 2 + β 2 ( x ) ( z t m ) 2 + ( g 2 ( λ m z t m ) ) 2 λ m 2 d x d t = 0 .

On the other hand, for each m N , U ˜ m ( w m , w t m , z m , z t m ) is a solution of

(3.27) ρ ( x ) w t t m ( K ( x ) w x m ) x + α ( w m z m ) + 1 λ m β 1 ( x ) g 1 ( λ m w t m ) + 1 λ m f 1 ( λ m w m , λ m z m ) = 0 , in ( 0 , L ) × ( 0 , T ) , ρ ( x ) z t t m ( K ( x ) z x m ) x α ( w m z m ) + 1 λ m β 2 ( x ) g 2 ( λ m z t m ) + 1 λ m f 2 ( λ m w m , λ m z m ) = 0 , in ( 0 , L ) × ( 0 , T ) , w m ( 0 , t ) = w m ( L , t ) = z m ( 0 , t ) = z m ( L , t ) = 0 , t ( 0 , T ) , w m ( x , 0 ) = u 0 m λ m , w t m ( x , 0 ) = u 1 m λ m , z m ( x , 0 ) = v 0 m λ m , z t m ( x , 0 ) = v 1 m λ m , x ( 0 , L ) .

The energy functional associated with (3.27) is given by

(3.28) U ˜ m ( t ) 1 2 U ˜ m 2 + 1 λ m 2 F ( λ m w m , λ m z m ) .

Taking in the same way as (2.16), we have

(3.29) d d t U ˜ m ( t ) = 0 L β 1 ( x ) g 1 ( λ m w t m ) λ m w t m + β 2 ( x ) g 2 ( λ m z t m ) λ m z t m d x 0 .

Moreover, it is not difficult to check that 0 U ˜ m ( t ) = 1 λ m 2 U m ( t ) for all t 0 . Then, in particular,

(3.30) U ˜ m ( 0 ) = 1 λ m 2 U m ( 0 ) = 1 , for all m N .

In order to achieve a contradiction, our main goal is to prove that

(3.31) lim m + U ˜ m ( 0 ) = 0 .

This will complete the proof (3.3) as desired. Indeed, since U ˜ m ( t ) U ˜ m ( 0 ) = 1 , we deduce there exists a subsequence of { U ˜ m } m N (without changing notation), such that

(3.32) ( w m , z m ) ( w , z ) weakly-star in L ( 0 , T ; ( H 0 1 ( 0 , L ) ) 2 ) , as m + ,

(3.33) ( w t m , z t m ) ( w t , z t ) weakly-star in L ( 0 , T ; ( L 2 ( 0 , L ) ) 2 ) , as m + .

Since H 0 1 ( 0 , L ) L q ( 0 , L ) ( q 2 ) compactly, from the Aubin-Lions theorem we have

(3.34) ( w m , z m ) ( w , z ) strongly in L ( 0 , T ; ( L q ( 0 , L ) ) 2 ) , as m + , for all q [ 2 , + ) .

Next let us prove that w = z 0 a.e. in ( 0 , L ) × ( 0 , T ) . Since λ m = U m ( 0 ) > 0 and U m ( 0 ) R , there exists a subsequence { λ m } m N such that λ m λ as m + . At this moment, we shall divide our proof into two cases: Case (A): λ > 0 , and Case (B): λ = 0 .

Case (A): λ > 0 .

Passing to the limit as m 0 in (3.27), together with (3.26) and (3.32)–(3.34), it follows that

(3.35) ρ ( x ) w t t ( K ( x ) w x ) x + α ( w z ) + 1 λ f 1 ( λ w , λ z ) = 0 , in ( 0 , L ) × ( 0 , T ) , ρ ( x ) z t t ( K ( x ) z x ) x α ( w z ) + 1 λ f 2 ( λ w , λ z ) = 0 , in ( 0 , L ) × ( 0 , T ) , w ( 0 , t ) = w ( L , t ) = z ( 0 , t ) = z ( L , t ) = 0 , t ( 0 , T ) , w t ( x , t ) = z t ( x , t ) = 0 , in ω × ( 0 , T ) .

Hence, taking derivatives over t in the equations of (3.35), and naming ψ = w t , ζ = z t , we conclude that

(3.36) ρ ( x ) ψ t t ( K ( x ) ψ x ) x + α ( ψ ζ ) + R 1 ( x , t ) ψ + R 3 ( x , t ) ζ = 0 , in ( 0 , L ) × ( 0 , T ) , ρ ( x ) ζ t t ( K ( x ) ζ x ) x α ( ψ ζ ) + R 2 ( x , t ) ζ + R 4 ( x , t ) ψ = 0 , in ( 0 , L ) × ( 0 , T ) , ψ ( 0 , t ) = ψ ( L , t ) = ζ ( 0 , t ) = ζ ( L , t ) = 0 , t ( 0 , T ) , ψ ( x , t ) = ζ ( x , t ) = 0 , in ω × ( 0 , T ) ,

where

R 1 ( x , t ) 1 λ w f 1 ( λ w , λ z ) , R 2 ( x , t ) 1 λ z f 2 ( λ w , λ z ) , R 3 ( x , t ) 1 λ z f 1 ( λ w , λ z ) , R 4 ( x , t ) 1 λ w f 2 ( λ w , λ z ) .

If we denote the following functions:

V 1 ( x , t ) = α + R 1 ( x , t ) , V 2 ( x , t ) = α + R 2 ( x , t ) , V 3 ( x , t ) = α R 3 ( x , t ) , V 4 ( x , t ) = α R 4 ( x , t ) ,

we can rewrite (3.36) as

(3.37) ρ ( x ) ψ t t ( K ( x ) ψ x ) x + V 1 ( x , t ) ψ = V 3 ( x , t ) ζ , in ( 0 , L ) × ( 0 , T ) , ρ ( x ) ζ t t ( K ( x ) ζ x ) x + V 2 ( x , t ) ζ = V 4 ( x , t ) ψ , in ( 0 , L ) × ( 0 , T ) , ψ ( 0 , t ) = ψ ( L , t ) = ζ ( 0 , t ) = ζ ( L , t ) = 0 , t ( 0 , T ) , ψ ( x , t ) = ζ ( x , t ) = 0 , in ω × ( 0 , T ) .

Moreover, by (2.5) and (3.32), it is easy to obtain V i L 2 ( 0 , T ; L 2 ( 0 , L ) ) . Thereby employing Lemma 3.1, it follows that ( ψ , ζ ) = ( 0 , 0 ) . Consequently, by similar arguments to those made previously, we obtain w = z = 0 . As a consequence w = z = 0 in all the convergence in (3.32)–(3.34).

Case (B): λ = 0 .

From (3.24), ( w m , z m ) ( H 0 1 ( 0 , L ) ) 2 and Assumption (A2), we can obtain that

(3.38) ( λ m w m , λ m z m ) = ( u m , v m ) ( 0 , 0 ) strongly in L ( 0 , T ; ( L q ( 0 , L ) ) 2 ) , as m + , q [ 2 , + ) ,

(3.39) 1 λ m f 1 ( λ m w m , λ m z m ) 0 strongly in L 2 ( 0 , T ; L 2 ( 0 , L ) ) , as m + ,

(3.40) 1 λ m f 2 ( λ m w m , λ m z m ) 0 strongly in L 2 ( 0 , T ; L 2 ( 0 , L ) ) , as m + .

Taking (3.26), (3.32), (3.33), (3.38), (3.39), and (3.40) into account, passing to the limit in (3.27) arrive at

(3.41) ρ ( x ) w t t ( K ( x ) w x ) x + α ( w z ) = 0 , in ( 0 , L ) × ( 0 , T ) , ρ ( x ) z t t ( K ( x ) z x ) x α ( w z ) = 0 , in ( 0 , L ) × ( 0 , T ) , w ( 0 , t ) = w ( L , t ) = z ( 0 , t ) = z ( L , t ) = 0 , t ( 0 , T ) , w t ( x , t ) = z t ( x , t ) = 0 , in ω × ( 0 , T ) .

For ψ = w t and ζ = z t , from (3.41), we infer that

(3.42) ρ ( x ) ψ t t ( K ( x ) ψ x ) x + α ( ψ ζ ) = 0 , in ( 0 , L ) × ( 0 , T ) , ρ ( x ) ζ t t ( K ( x ) ζ x ) x α ( ψ ζ ) = 0 , in ( 0 , L ) × ( 0 , T ) , ψ ( 0 , t ) = ψ ( L , t ) = ζ ( 0 , t ) = ζ ( L , t ) = 0 , t ( 0 , T ) , ψ ( x , t ) = ζ ( x , t ) = 0 , in ω × ( 0 , T ) .

Denoting V 1 ( x , t ) = V 2 ( x , t ) = V 3 ( x , t ) = V 4 ( x , t ) = α L 2 ( 0 , T ; L 2 ( 0 , L ) ) and again employing Lemma 3.1, it implies that ψ = ζ = 0 . Consequently based on (3.41), we can obtain w = z = 0 .

Then, in both cases λ > 0 and λ = 0 , we obtain that w = z = 0 in all the convergence in (3.32)–(3.34). Thus, taking (3.34) and Hölder’s inequality into account, it is easy to see that

(3.43) 1 λ m f 1 ( λ m w m , λ m z m ) 0 strongly in L 2 ( 0 , T ; L 2 ( 0 , L ) ) , as m + ,

(3.44) 1 λ m f 2 ( λ m w m , λ m z m ) 0 strongly in L 2 ( 0 , T ; L 2 ( 0 , L ) ) , as m + .

Step 3. Assert ( w t m , z t m ) ( 0 , 0 ) strongly in ( L 2 ( ( 0 , L ) × ( 0 , T ) ) ) 2 as m + .

Let us denote

P ρ ( x ) t 2 x ( K ( x ) x ) .

Taking into account (3.26), (3.34), (3.43), and (3.44), we deduce that as m +

(3.45) P t w m = t P w m = t α ( z m w m ) 1 λ m f 1 ( λ m w m , λ m z m ) 1 λ m β 1 ( x ) g 1 ( λ m w t m ) 0

and

(3.46) P t z m = t P z m = t α ( w m z m ) 1 λ m f 2 ( λ m w m , λ m z m ) 1 λ m β 2 ( x ) g 2 ( λ m z t m ) 0

strongly in H loc 1 ( ( 0 , L ) × ( 0 , T ) ) .

Let us also denote by μ w the microlocal defect measures associated with { w t m } m N in L loc 2 ( ( 0 , L ) × ( 0 , T ) ) . In view of (3.45), we deduce that

  1. From Theorem 6.2 (Appendix), the support of the measures μ w is contained in the characteristic set of the wave operator. That is,

    supp ( μ w ) ( t , x , τ , ξ ) : τ 2 K ( x ) ρ ( x ) ξ 2 = 0 .

Our wish is to propagate the convergence of w t m 0 from L 2 ( ω × ( 0 , T ) ) to the whole L 2 ( ( 0 , L ) × ( 0 , T ) ) . From (3.45), we can also obtain:
  1. μ w propagates along the bicharacteristic flow of this operator, which signifies, particularly, that if some point ω 0 = ( t 0 , x 0 , τ 0 , ξ 0 ) does not belong to the supp ( μ w ) , the whole bicharacteristic issued from ω 0 is out of supp ( μ w ) .

Indeed, from Theorem 6.4 and Proposition 6.1, it follows that supp ( μ w ) in ( 0 , T ) × ( 0 , L ) × S 1 is a union of curves like

(3.47) t I ( 0 , + ) m ± ( t ) = t , x ( t ) , ± 1 1 + x ˙ 2 , G ( x ) x ˙ 1 + x ˙ 2 ,

where t I x ( t ) ( 0 , L ) is a geodesic associated with the metric G = K ρ 1 .

From the convergence w t m 0 strongly in L 2 ( ω × ( 0 , T ) ) as m + , it follows that μ w is supported in the set ( 0 , T ) × ( 0 , L ) \ ω × S 1 . We affirm that supp μ w = . Indeed, let ( t 0 , x 0 , τ 0 , ξ 0 ) supp ( μ w ) and x be a geodesic of G = K ρ 1 defined near t 0 . Once the geodesics inside ( 0 , L ) \ ω enters necessarily in the region ω , thus, for all geodesics of the metric G , with 0 I , there exists t 0 such that x ( t ) meets ω , or, in other words, every geodesic enters in the region ω . Thus, we deduce from (3.47) that m ± ( t ) supp ( μ w ) and as a consequence, ( t 0 , x 0 , τ 0 , ξ 0 ) supp ( μ w ) . Therefore, supp ( μ w ) is empty.

From Remark 6.1, it follows that w t m 0 in L loc 2 ( ( 0 , L ) × ( 0 , T ) ) . Since, w t m 0 strongly in L 2 ( ω × ( 0 , T ) ) and as in the approach in [7] ((4.26)–(4.28) in [7]), we deduce that

(3.48) w t m 0 in L 2 ( ( 0 , L ) × ( 0 , T ) ) .

Proceeding analogously as above, we also deduce

(3.49) z t m 0 in L 2 ( ( 0 , L ) × ( 0 , T ) ) .

Step 4. Derive contradictory point lim m + U ˜ m ( 0 ) = 0 .

Next it is easy to show that U ˜ m ( 0 ) converges to zero. Indeed, multiplying (3.27) 1 by θ w m and adding (3.27) 2 with θ z m , where θ ( t ) is such that 0 θ ( t ) 1 and θ ( t ) = 1 in ( ε , T ε ) for ε ( 0 , T ) fixed and arbitrary, and integrating in ( 0 , L ) × ( 0 , T ) , we obtain

(3.50) 0 T θ ( t ) 0 L K ( x ) w x m 2 d x d t + 0 T θ ( t ) 0 L K ( x ) z x m 2 d x d t + 1 λ m 0 T θ ( t ) 0 L [ f 1 ( λ m w m , λ m z m ) w m + f 2 ( λ m w m , λ m z m ) z m ] d x d t = 0 T θ ( t ) 0 L ρ ( x ) w m w t m d x d t + 0 T θ ( t ) 0 L ρ ( x ) w t m 2 d x d t + 0 T θ ( t ) 0 L ρ ( x ) z m z t m d x d t + 0 T θ ( t ) 0 L ρ ( x ) z t m 2 d x d t α 0 T θ ( t ) 0 L ( w m z m ) 2 d x d t 1 λ m 0 T θ ( t ) 0 L β 1 ( x ) g 1 ( λ m w t m ) w m d x d t 1 λ m 0 T θ ( t ) 0 L β 2 ( x ) g 2 ( λ m z t m ) z m d x d t .

Therefore, taking the limit as m , from (3.26), (3.34), (3.38), (3.43), (3.44), (3.48), and (3.50), we obtain

(3.51) lim m + 0 T θ 0 L K ( x ) w x m 2 d x d t = lim m + 0 T θ 0 L K ( x ) z x m 2 d x d t = 0 ,

which implies that

(3.52) lim m + ε T ε 0 L K ( x ) w x m 2 d x d t = lim m + ε T ε 0 L K ( x ) z x m 2 d x d t = 0 .

Moreover, we have that

(3.53) lim m + 1 λ m ε T ε 0 L [ f 1 ( λ m w m , λ m z m ) w m + f 2 ( λ m w m , λ m z m ) z m ] d x d t = lim m + 1 λ m 2 ε T ε 0 L [ f 1 ( u m , v m ) u m + f 2 ( u m , v m ) v m ] d x d t = 0 ,

then, from (2.8) we conclude

(3.54) lim m + 1 λ m 2 ε T ε 0 L F ( u m , v m ) d x d t = 0 .

Combining the above convergence we obtain that

(3.55) lim m + ε T ε U ˜ m ( t ) d t = 0 .

Then, by nonincreasing property of the energy function, we obtain

(3.56) lim m + ( T 2 ε ) U ˜ m ( T ε ) = 0 .

Combining the energy identity

(3.57) U ˜ m ( T ε ) U ˜ m ( ε ) = 1 λ m ε T ε 0 L β 1 ( x ) g 1 ( λ m w t m ) w t m d x d t 1 λ m ε T ε 0 L β 2 ( x ) g 2 ( λ m z t m ) z t m d x d t ,

along with (3.26), we can obtain that

(3.58) lim m + U ˜ m ( ε ) = 0 .

Therefore, we have that for m +

(3.59) U ˜ m ( 0 ) = 1 λ m 0 ε 0 L [ β 1 ( x ) g 1 ( λ m w t m ) w t m + β 2 ( x ) g 2 ( λ m z t m ) z t m ] d x d t + U ˜ m ( ε ) 0 ,

which is a contradiction with (3.30) and it allows us to conclude that (3.3) holds.□

Remark 3.3

The combination of the Microlocal analysis method and GCC condition is mainly used to prove that w t m and z t m strongly converge to 0 in L 2 ( ( 0 , L ) × ( 0 , T ) ) as m + .

3.2 Stability result

In order to characterize decay rates for the energy, we need to introduce several special functions, which in turn will depend on the growth of g i , i = 1 , 2 near the origin and at infinity.

Proceed as in [27,34], since g 1 and g 2 are continuous monotone increasing functions vanishing at zero, there exist continuous, increasing, concave functions vanishing at the origin h i : R R , i = 1 , 2 , such that

(3.60) h i ( g i ( s ) s ) s 2 + g i ( s ) 2 , s 1 , i = 1 , 2 .

We define h ¯ : R R by

(3.61) h ¯ ( s ) h 1 ( s ) + h 2 ( s ) , s R .

In this way, h ¯ ( s ) has the same properties as its composing functions h 1 and h 2 . However, when dealing with nonlinear weak damping terms at infinity in the sublinear or superlinear cases, we need the solutions to have a better regularity [9,16]. Therefore, we make the following regularity assumptions:

Assumption 3.2

(Regularity assumptions)

  • If g 1 grows sublinearly at infinity, that is q 1 < 1 , we assume the existence of θ 1 > 2 such that

    u t L ( 0 , + ; L θ 1 ( 0 , L ) ) D 0

    for some D 0 > 0 .

  • If g 2 grows sublinearly at infinity, that is q 2 < 1 , we assume the existence of θ 2 > 2 such that

    v t L ( 0 , + ; L θ 2 ( 0 , L ) ) D 0

    for some D 0 > 0 .

  • If g 1 grows superlinearly at infinity, that is q 1 > 1 , we assume the existence of θ 1 > 2 q 1 such that

    u t L ( 0 , + ; L θ 1 ( 0 , L ) ) D 0

    for some D 0 > 0 .

  • If g 2 grows superlinearly at infinity, that is q 2 > 1 , we assume the existence of θ 2 > 2 q 2 such that

    v t L ( 0 , + ; L θ 2 ( 0 , L ) ) D 0

    for some D 0 > 0 .

Next as in [9,16,34], setting

(3.62) h = h ˜ + h ¯ 1 L T I ,

where h ˜ is defined in Table 1.

In this setting, the following technical lemma due to Lasiecka and Tataru [34] is given.

Lemma 3.3

[34, Lemma 3.3] Let p be a positive, increasing function such that p ( 0 ) = 0 . Since p is increasing, we can define an increasing function q, q ( x ) x ( I + p ) 1 ( x ) . Consider a sequence s n of positive numbers which satisfies

s m + 1 + p ( s m + 1 ) s m .

Then, s m S ( m ) , where S ( t ) is a solution of the differential equation

d d t S ( t ) + q ( S ( t ) ) = 0 , S ( 0 ) = s 0 .

Moreover, if p ( x ) > 0 for x > 0 , then lim t + S ( t ) = 0 .

We are now in a position to state and briefly prove the main result of our work. It reads as follows.

Theorem 3.1

Let Assumption 3.2 hold and T 0 > 0 be given by Assumption 3.1. Let h be the functions defined in (3.62). Suppose U ( t ) = ( u , u t , v , v t ) is the solution to the nonlinear problems (1.1) and (1.2), then

(3.63) U ( t ) S t T 0 1 , for a l l t > T 0 ,

with lim t + S ( t ) = 0 , where S ( t ) is a solution of the differential equation

(3.64) d S d t ( t ) + q ( S ( t ) ) = 0 , t > 0 , S ( 0 ) = s 0 ,

where

(3.65) q = I ( I + h 1 ( K ˜ I ) ) 1 , s 0 = U ( 0 ) ,

the constant K ˜ > 0 to be specified later. Since h is strictly increasing concave function with h ( 0 ) = 0 , then q is a monotone increasing function vanishing at zero.

  1. ( g 1 , g 2 ) is the combination (L, L), then

    K ˜ = C ˜ 1 ( β 1 , β 2 , T , L T ) .

  2. ( g 1 , g 2 ) is the combination (Sub, Sub), then

    K ˜ = C ( β 1 , β 2 , T , L T ) D 0 max θ 1 ( 1 q 1 ) θ 1 q 1 1 , θ 2 ( 1 q 2 ) θ 2 q 2 1 1 .

  3. ( g 1 , g 2 ) is the combination (Super, Super), then

    K ˜ = C ( β 1 , β 2 , T , L T ) D 0 max θ 1 ( q 1 1 ) θ 1 q 1 1 , θ 2 ( q 2 1 ) θ 2 q 2 1 1 .

  4. ( g 1 , g 2 ) is the combination (L, Sub), then

    K ˜ = C ( β 1 , β 2 , T , L T ) D 0 θ 2 ( 1 q 2 ) θ 2 q 2 1 1 .

  5. ( g 1 , g 2 ) is the combination (L, Super), then

    K ˜ = C ( β 1 , β 2 , T , L T ) D 0 θ 2 ( q 2 1 ) θ 2 q 2 1 1 .

  6. ( g 1 , g 2 ) is the combination (Sub, Super), then

    K ˜ = C ( β 1 , β 2 , T , L T ) D 0 max θ 1 ( 1 q 1 ) θ 1 q 1 1 , θ 2 ( q 2 1 ) θ 2 q 2 1 1 .

  7. ( g 1 , g 2 ) is the combination (Sub, L), then

    K ˜ = C ( β 1 , β 2 , T , L T ) D 0 θ 1 ( 1 q 1 ) θ 1 q 1 1 1 .

  8. ( g 1 , g 2 ) is the combination (Super, L), then

    K ˜ = C ( β 1 , β 2 , T , L T ) D 0 θ 1 ( q 1 1 ) θ 1 q 1 1 1 .

  9. ( g 1 , g 2 ) is the combination (Super, Sub), then

    K ˜ = C ( β 1 , β 2 , T , L T ) D 0 max θ 1 ( q 1 1 ) θ 1 q 1 1 , θ 2 ( 1 q 2 ) θ 2 q 2 1 1 .

Proof

Proceed as in Daoulatli et al.’s work [16] (Lemma 3.1 of the referred paper) adapted to our context. We define

Σ u { ( x , t ) ( 0 , L ) × ( 0 , T ) : u t ( x , t ) > 1 } and Γ u = ( 0 , L ) × ( 0 , T ) \ Σ u ,

Σ v { ( x , t ) ( 0 , L ) × ( 0 , T ) : v t ( x , t ) > 1 } and Γ v = ( 0 , L ) × ( 0 , T ) \ Σ v ,

D 0 T ( g 1 ( s ) ; u t ) 0 T 0 L β 1 ( x ) g 1 ( u t ) u t d x d t and D 0 T ( g 2 ( s ) ; v t ) 0 T 0 L β 2 ( x ) g 2 ( v t ) v t d x d t .

The proof will require the following inequalities:

Case (1): Damping near the origin. Let U = ( u , u t , v , v t ) . Using Inequality (3.60) for function h 1 and (the reverse) Jensens inequality, we obtain

(3.66) Γ u β 1 ( x ) ( u t 2 + g 1 ( u t ) 2 ) d x d t Γ u β 1 ( x ) h 1 ( g 1 ( u t ) u t ) d x d t ( 1 + β 1 ) Γ u h 1 β 1 ( x ) 1 + β 1 g 1 ( u t ) u t d x d t ( 1 + β 1 ) Γ u h 1 ( β 1 ( x ) g 1 ( u t ) u t ) d x d t L T ( 1 + β 1 ) h 1 1 L T D 0 T ( g 1 ( s ) ; u t ) .

Analogously, we can conclude that

(3.67) Γ v β 2 ( x ) ( v t 2 + g 2 ( v t ) 2 ) d x d t L T ( 1 + β 2 ) h 2 1 L T D 0 T ( g 2 ( s ) ; v t ) .

Since each h i is an increasing function, then combining (3.66) and (3.67), we have

(3.68) Γ u β 1 ( x ) ( u t 2 + g 1 ( u t ) 2 ) d x d t + Γ v β 2 ( x ) ( v t 2 + g 2 ( v t ) 2 ) d x d t L T i = 1 2 ( 1 + β i ) h 1 1 L T D 0 T ( g 1 ( s ) ; u t ) + h 2 1 L T D 0 T ( g 2 ( s ) ; v t ) L T i = 1 2 ( 1 + β i ) h ¯ 1 L T ( D 0 T ( g 1 ( s ) ; u t ) + D 0 T ( g 2 ( s ) ; v t ) ) = L T i = 1 2 ( 1 + β i ) k ( D 0 T ( g 1 ( s ) ; u t ) + D 0 T ( g 2 ( s ) ; v t ) ) ,

where k ( s ) h ¯ 1 L T s .

Case (2): ( g 1 , g 2 ) is the combination (L, L). From Assumption 2.1, we have

(3.69) Σ u β 1 ( x ) ( u t 2 + g 1 ( u t ) 2 ) d x d t ( m 1 1 + M 1 ) D 0 T ( g 1 ( s ) ; u t ) .

Analogously, we can conclude that

(3.70) Σ v β 2 ( x ) ( v t 2 + g 2 ( v t ) 2 ) d x d t ( m 2 1 + M 2 ) D 0 T ( g 2 ( s ) ; v t ) .

Thus, by using (3.69) and (3.70), we obtain

(3.71) Σ u β 1 ( x ) ( u t 2 + g 1 ( u t ) 2 ) d x d t + Σ v β 2 ( x ) ( v t 2 + g 2 ( v t ) 2 ) d x d t i = 1 2 ( m i 1 + M i ) ( D 0 T ( g 1 ( s ) ; u t ) + D 0 T ( g 2 ( s ) ; v t ) ) i = 1 2 ( m i 1 + M i ) h ˜ ( D 0 T ( g 1 ( s ) ; u t ) + D 0 T ( g 2 ( s ) ; v t ) ) .

Case (3a): ( g 1 , g 2 ) is the combination (Sub, Sub). Inequality (2.4) and q 1 < 1 allow us to conclude that g 1 ( s ) M 1 s q 1 M 1 s , for s > 1 . Thus, we have

(3.72) Σ u β 1 ( x ) g 1 ( u t ) 2 d x d t M 1 Σ u β 1 ( x ) g 1 ( u t ) u t d x d t M 1 2 Σ u β 1 ( x ) u t 2 d x d t .

On the other hand, let δ ( 0 , 1 ) . Using Hölder’s inequality gives

(3.73) Σ u β 1 ( x ) u t 2 d x d t = Σ u β 1 ( x ) u t 2 δ u t 2 2 δ d x d t Σ u β 1 ( x ) u t θ 1 d x d t 2 δ θ 1 Σ u β 1 ( x ) u t 2 ( 1 δ ) θ 1 θ 1 2 δ d x d t θ 1 2 δ θ 1 ,

and choosing δ ( 0 , 1 ) , such that

2 ( 1 δ ) θ 1 θ 1 2 δ = q 1 + 1 ,

that is,

δ = ( 1 q 1 ) θ 1 2 ( θ 1 q 1 1 ) .

Therefore, from (2.4) we have

(3.74) Σ u β 1 ( x ) u t 2 d x d t Σ u β 1 ( x ) u t θ 1 d x d t 1 q 1 θ 1 q 1 1 Σ u β 1 ( x ) u t q 1 + 1 d x d t θ 1 2 θ 1 q 1 1 m 1 2 θ 1 θ 1 q 1 1 β 1 1 q 1 θ 1 q 1 1 u t L ( 0 , + ; L θ 1 ( 0 , L ) ) θ 1 ( 1 q 1 ) θ 1 q 1 1 ( D 0 T ( g 1 ( s ) ; u t ) ) θ 1 2 θ 1 q 1 1 .

Combining (3.72) and (3.74), we conclude that

(3.75) Σ u β 1 ( x ) ( u t 2 + g 1 ( u t ) 2 ) d x d t ( M 1 2 + 1 ) m 1 2 θ 1 θ 1 q 1 1 β 1 1 q 1 θ 1 q 1 1 u t L ( 0 , + ; L θ 1 ( 0 , L ) ) θ 1 ( 1 q 1 ) θ 1 q 1 1 ( D 0 T ( g 1 ( s ) ; u t ) ) θ 1 2 θ 1 q 1 1 .

Similar to previous calculations, we also conclude that

(3.76) Σ v β 2 ( x ) ( v t 2 + g 2 ( v t ) 2 ) d x d t ( M 2 2 + 1 ) m 2 2 θ 2 θ 2 q 2 1 β 2 1 q 2 θ 2 q 2 1 v t L ( 0 , + ; L θ 2 ( 0 , L ) ) θ 2 ( 1 q 2 ) θ 2 q 2 1 ( D 0 T ( g 2 ( s ) ; v t ) ) θ 2 2 θ 2 q 2 1 .

Summing (3.75) and (3.76), we obtain

(3.77) Σ v β 1 ( x ) ( u t 2 + g 1 ( u t ) 2 ) d x d t + Σ v β 2 ( x ) ( v t 2 + g 2 ( v t ) 2 ) d x d t i = 1 2 ( M i 2 + 1 ) m i 2 θ i θ i q i 1 β 1 1 q 1 θ 1 q 1 1 u t L ( 0 , + ; L θ 1 ( 0 , L ) ) θ 1 ( 1 q 1 ) θ 1 q 1 1 + β 2 1 q 2 θ 2 q 2 1 v t L ( 0 , + ; L θ 2 ( 0 , L ) ) θ 2 ( 1 q 2 ) θ 2 q 2 1 × h ˜ ( D 0 T ( g 1 ( s ) ; u t ) + D 0 T ( g 2 ( s ) ; v t ) ) .

Case (3b): ( g 1 , g 2 ) is the combination (Super, Sper). Using again Inequality (2.7) and q 1 > 1 , it follows that g 1 ( s ) m 1 s q 1 m 1 s , for s > 1 . Thus, we obtain

(3.78) Σ u β 1 ( x ) u t 2 d x d t 1 m 1 2 Σ u β 1 ( x ) g 1 ( u t ) 2 d x d t .

Next let δ ( 0 , 1 ) . Using Hölder’s inequality, we observe that

(3.79) Σ u β 1 ( x ) g 1 ( u t ) 2 d x d t = Σ u β 1 ( x ) g 1 ( u t ) 2 δ g 1 ( u t ) 2 2 δ d x d t Σ u β 1 ( x ) g 1 ( u t ) θ 1 q 1 d x d t 2 δ q 1 θ 1 Σ u β 1 ( x ) g 1 ( u t ) 2 ( 1 δ ) θ 1 θ 1 2 δ q 1 d x d t θ 1 2 δ q 1 θ 1 ,

and choosing δ ( 0 , 1 ) , such that

2 ( 1 δ ) θ 1 θ 1 2 δ q 1 = q 1 + 1 q 1 ,

that is,

δ = ( q 1 1 ) θ 1 2 q 1 ( θ 1 q 1 1 ) .

From (2.4), we know g 1 ( s ) 1 q 1 M 1 1 q 1 s for s > 1 . So, Inequality (3.79) allows us to conclude that

(3.80) Σ u β 1 ( x ) g 1 ( u t ) 2 d x d t Σ u β 1 ( x ) g 1 ( u t ) θ 1 q 1 d x d t q 1 1 θ 1 q 1 1 Σ u β 1 ( x ) g 1 ( u t ) q 1 + 1 q 1 d x d t θ 1 2 q 1 θ 1 q 1 1 M 1 ( q 1 1 ) θ 1 ( θ 1 q 1 1 ) q 1 β 1 q 1 1 θ 1 q 1 1 Σ u u t θ 1 d x d t q 1 1 θ 1 q 1 1 Σ u β 1 ( x ) g 1 ( u t ) u t d x d t θ 1 2 q 1 θ 1 q 1 1 M 1 ( q 1 1 ) θ 1 ( θ 1 q 1 1 ) q 1 β 1 q 1 1 θ 1 q 1 1 u t L ( 0 , + ; L θ 1 ( 0 , L ) ) θ 1 ( q 1 1 ) θ 1 q 1 1 ( D 0 T ( g 1 ( s ) ; u t ) ) θ 1 2 q 1 θ 1 q 1 1 .

Now, using (3.78) and (3.80), we have

(3.81) Σ u β 1 ( x ) ( u t 2 + g 1 ( u t ) 2 ) d x d t ( 1 + m 1 2 ) Σ u β 1 ( x ) g 1 ( u t ) 2 d x d t M 1 ( q 1 1 ) θ 1 ( θ 1 q 1 1 ) q 1 ( 1 + m 1 2 ) β 1 q 1 1 θ 1 q 1 1 u t L ( 0 , + ; L θ 1 ( 0 , L ) ) θ 1 ( q 1 1 ) θ 1 q 1 1 ( D 0 T ( g 1 ( s ) ; u t ) ) θ 1 2 q 1 θ 1 q 1 1 .

Analogously, we can conclude that

(3.82) Σ u β 2 ( x ) ( v t 2 + g 2 ( v t ) 2 ) d x d t M 2 ( q 2 1 ) θ 2 ( θ 2 q 2 1 ) q 2 ( 1 + m 2 2 ) β 2 q 2 1 θ 2 q 2 1 v t L ( 0 , + ; L θ 2 ( 0 , L ) ) θ 2 ( q 2 1 ) θ 2 q 2 1 ( D 0 T ( g 2 ( s ) ; v t ) ) θ 2 2 q 2 θ 2 q 2 1 .

Then, combining (3.87) and (3.88), we obtain

(3.83) Σ u β 1 ( x ) ( u t 2 + g 1 ( u t ) 2 ) d x d t + Σ v β 2 ( x ) ( v t 2 + g 2 ( v t ) 2 ) d x d t i = 1 2 M i ( q i 1 ) θ i ( θ i q i 1 ) q i ( 1 + m i 2 ) β 1 q 1 1 θ 1 q 1 1 u t L ( 0 , + ; L θ 1 ( 0 , L ) ) θ 1 ( q 1 1 ) θ 1 q 1 1 + β 2 q 2 1 θ 2 q 2 1 u t L ( 0 , + ; L θ 2 ( 0 , L ) ) θ 2 ( q 2 1 ) θ 2 q 2 1 × h ˜ ( D 0 T ( g 1 ( s ) ; u t ) + D 0 T ( g 2 ( s ) ; v t ) ) .

Case (3c): ( g 1 , g 2 ) is the combination (L, Sub). Since q 1 = 1 and q 2 < 1 , proceeding as in Case (2) and Case (3a), we obtain

(3.84) Σ u β 1 ( x ) ( u t 2 + g 1 ( u t ) 2 ) d x d t + Σ v β 2 ( x ) ( v t 2 + g 2 ( v t ) 2 ) d x d t ( m 1 1 + M 1 ) D 0 T ( g 1 ( s ) ; u t ) + ( M 2 2 +1 ) m 2 2 θ 2 θ 2 q 2 1 β 2 1 q 2 θ 2 q 2 1 v t L ( 0 , + ; L θ 2 ( 0 , L ) ) θ 2 ( 1 q 2 ) θ 2 q 2 1 ( D 0 T ( g 2 ( s ) ; v t ) ) θ 2 2 θ 2 q 2 1 ( m 1 1 + M 1 ) + ( M 2 2 +1 ) m 2 2 θ 2 θ 2 q 2 1 β 2 1 q 2 θ 2 q 2 1 v t L ( 0 , + ; L θ 2 ( 0 , L ) ) θ 2 ( 1 q 2 ) θ 2 q 2 1 × h ˜ ( D 0 T ( g 1 ( s ) ; u t ) + D 0 T ( g 2 ( s ) ; v t ) ) .

Case (3d): ( g 1 , g 2 ) is the combination (L, Super). Since q 1 = 1 and q 2 > 1 , proceeding as in Case (2) and Case (3b), we have that

(3.85) Σ u β 1 ( x ) ( u t 2 + g 1 ( u t ) 2 ) d x d t + Σ v β 2 ( x ) ( v t 2 + g 2 ( v t ) 2 ) d x d t M 2 ( q 2 1 ) θ 2 ( θ 2 q 2 1 ) q 2 ( 1 + m 2 2 ) β 2 q 2 1 θ 2 q 2 1 v t L ( 0 , + ; L θ 2 ( 0 , L ) ) θ 2 ( q 2 1 ) θ 2 q 2 1 ( D 0 T ( g 2 ( s ) ; v t ) ) θ 2 2 q 2 θ 2 q 2 1 + ( m 1 1 + M 1 ) D 0 T ( g 1 ( s ) ; u t ) ( m 1 1 + M 1 ) + M 2 ( q 2 1 ) θ 2 ( θ 2 q 2 1 ) q 2 ( 1 + m 2 2 ) β 2 q 2 1 θ 2 q 2 1 v t L ( 0 , + ; L θ 2 ( 0 , L ) ) θ 2 ( q 2 1 ) θ 2 q 2 1 × h ˜ ( D 0 T ( g 1 ( s ) ; u t ) + D 0 T ( g 2 ( s ) ; v t ) ) .

Case (3e): ( g 1 , g 2 ) is the combination (Sub, Super). Since q 1 < 1 and q 2 > 1 , proceeding as in Case (3a) and Case (3b), we can obtain

(3.86) Σ u β 1 ( x ) ( u t 2 + g 1 ( u t ) 2 ) d x d t + Σ v β 2 ( x ) ( v t 2 + g 2 ( v t ) 2 ) d x d t ( M 1 2 +1 ) m 1 2 θ 1 θ 1 q 1 1 β 1 1 q 1 θ 1 q 1 1 u t L ( 0 , + ; L θ 1 ( 0 , L ) ) θ 1 ( 1 q 1 ) θ 1 q 1 1 ( D 0 T ( g 1 ( s ) ; u t ) ) θ 1 2 θ 1 q 1 1 + + M 2 ( q 2 1 ) θ 2 ( θ 2 q 2 1 ) q 2 ( 1 + m 2 2 ) β 2 q 2 1 θ 2 q 2 1 v t L ( 0 , + ; L θ 2 ( 0 , L ) ) θ 2 ( q 2 1 ) θ 2 q 2 1 ( D 0 T ( g 2 ( s ) ; v t ) ) θ 2 2 q 2 θ 2 q 2 1 ( M 1 2 +1 ) m 1 2 θ 1 θ 1 q 1 1 β 1 1 q 1 θ 1 q 1 1 u t L ( 0 , + ; L θ 1 ( 0 , L ) ) θ 1 ( 1 q 1 ) θ 1 q 1 1 + M 2 ( q 2 1 ) θ 2 ( θ 2 q 2 1 ) q 2 ( 1 + m 2 2 ) β 2 q 2 1 θ 2 q 2 1 v t L ( 0 , + ; L θ 2 ( 0 , L ) ) θ 2 ( q 2 1 ) θ 2 q 2 1 × h ˜ ( D 0 T ( g 1 ( s ) ; u t ) + D 0 T ( g 2 ( s ) ; v t ) ) .

Case (3f): ( g 1 , g 2 ) is the combination (Sub, L). Similarly as Case (3c), we conclude that

(3.87) Σ u β 1 ( x ) ( u t 2 + g 1 ( u t ) 2 ) d x d t + Σ v β 2 ( x ) ( v t 2 + g 2 ( v t ) 2 ) d x d t ( M 1 2 +1 ) m 1 2 θ 1 θ 1 q 1 1 β 1 1 q 1 θ 1 q 1 1 u t L ( 0 , + ; L θ 1 ( 0 , L ) ) θ 1 ( 1 q 1 ) θ 1 q 1 1 ( D 0 T ( g 1 ( s ) ; u t ) ) θ 1 2 θ 1 q 1 1 + ( m 2 1 + M 2 ) D 0 T ( g 2 ( s ) ; v t ) ( M 1 2 +1 ) m 1 2 θ 1 θ 1 q 1 1 β 1 1 q 1 θ 1 q 1 1 u t L ( 0 , + ; L θ 1 ( 0 , L ) ) θ 1 ( 1 q 1 ) θ 1 q 1 1 + ( m 2 1 + M 2 ) × h ˜ ( D 0 T ( g 1 ( s ) ; u t ) + D 0 T ( g 2 ( s ) ; v t ) ) .

Case (3g): ( g 1 , g 2 ) is the combination (Super, L). Similarly as Case (3d), we have that

(3.88) Σ u β 1 ( x ) ( u t 2 + g 1 ( u t ) 2 ) d x d t + Σ v β 2 ( x ) ( v t 2 + g 2 ( v t ) 2 ) d x d t M 1 ( q 1 1 ) θ 1 ( θ 1 q 1 1 ) q 1 ( 1 + m 1 2 ) β 1 q 1 1 θ 1 q 1 1 u t L ( 0 , + ; L θ 1 ( 0 , L ) ) θ 1 ( q 1 1 ) θ 1 q 1 1 ( D 0 T ( g 1 ( s ) ; u t ) ) θ 1 2 q 1 θ 1 q 1 1 + ( m 2 1 + M 2 ) D 0 T ( g 2 ( s ) ; v t ) M 1 ( q 1 1 ) θ 1 ( θ 1 q 1 1 ) q 1 ( 1 + m 1 2 ) β 1 q 1 1 θ 1 q 1 1 u t L ( 0 , + ; L θ 1 ( 0 , L ) ) θ 1 ( q 1 1 ) θ 1 q 1 1 + m 2 1 + M 2 × h ˜ ( D 0 T ( g 1 ( s ) ; u t ) + D 0 T ( g 2 ( s ) ; v t ) ) .

Case (3h): ( g 1 , g 2 ) is the combination (Super, Sub). As in Case (3e), we obtain

(3.89) Σ u β 1 ( x ) ( u t 2 + g 1 ( u t ) 2 ) d x d t + Σ v β 2 ( x ) ( v t 2 + g 2 ( v t ) 2 ) d x d t M 1 ( q 1 1 ) θ 1 ( θ 1 q 1 1 ) q 1 ( 1 + m 1 2 ) β 1 q 1 1 θ 1 q 1 1 u t L ( 0 , + ; L θ 1 ( 0 , L ) ) θ 1 ( q 1 1 ) θ 1 q 1 1 ( D 0 T ( g 1 ( s ) ; u t ) ) θ 1 2 q 1 θ 1 q 1 1 × ( M 2 2 +1 ) m 2 2 θ 2 θ 2 q 2 1 β 2 1 q 2 θ 2 q 2 1 v t L ( 0 , + ; L θ 2 ( 0 , L ) ) θ 2 ( 1 q 2 ) θ 2 q 2 1 ( D 0 T ( g 2 ( s ) ; v t ) ) θ 2 2 θ 2 q 2 1 M 1 ( q 1 1 ) θ 1 ( θ 1 q 1 1 ) q 1 ( 1 + m 1 2 ) β 1 q 1 1 θ 1 q 1 1 u t L ( 0 , + ; L θ 1 ( 0 , L ) ) θ 1 ( q 1 1 ) θ 1 q 1 1 + ( M 2 2 +1 ) m 2 2 θ 2 θ 2 q 2 1 β 2 1 q 2 θ 2 q 2 1 v t L ( 0 , + ; L θ 2 ( 0 , L ) ) θ 2 ( 1 q 2 ) θ 2 q 2 1 × h ˜ ( D 0 T ( g 1 ( s ) ; u t ) + D 0 T ( g 2 ( s ) ; v t ) ) .

Finally, combining the above estimates, the observability inequalities (3.3) and (3.62), we deduce that

(3.90) U ( T ) C ˜ ( k + h ˜ ) ( D 0 T ( g 1 ( s ) ; u t ) + D 0 T ( g 1 ( s ) ; v t ) ) = C ˜ h ( D 0 T ( g 1 ( s ) ; u t ) + D 0 T ( g 1 ( s ) ; v t ) ) ,

where the constant C ˜ is different for Cases (2) and (3a)–(3g), depending on T , L , β i ( i = 1 , 2 ) , D 0 , and k ( s ) = h ¯ 1 L T s , h ˜ is defined in Table 1. Using the energy equality (2.14) and defining K ˜ = C ˜ 1 , we have

(3.91) h 1 ( K ˜ U ( T ) ) U ( 0 ) U ( T ) .

To finish the proof, we replace T by ( m + 1 ) T (respectively, 0 by m T ) in (3.91), m N , in order to obtain

h 1 ( K ˜ U ( ( m + 1 ) T ) ) + U ( ( m + 1 ) T ) U ( m T ) , for m N .

Thus, using Lemma 3.3 with s m = U ( m T ) , we can conclude that

U ( m T ) S ( m ) , for m N .

Finally, observing that for every t > T we can find m N and τ [ 0 , T ] such that t = m T + τ , then

U ( t ) U ( m T ) S ( m ) S t τ T S t T 1 , for all t > T 0 ,

where we have used the fact that the solution S of ordinary differential equation (ODE) (3.64) is dissipative. The proof of decay for U is now complete.□

Remark 3.4

If g 1 and g 2 are bounded above and below by linear functions near the infinity, then K ˜ is independent of D 0 . K ˜ depended on D 0 for other cases. Moreover, Assumption 3.2 can be satisfied to a certain extent by starting with smooth initial data, i.e., (3.63) and conclusions of 1–9 hold for strong solution.

3.3 Some examples

While Theorem 3.1 provides an abstract estimate of uniform stability, the specific decay behavior is determined by the growth properties of the nonlinear damping function g i , i = 1 , 2 near the origin. In this section, we primarily present some concrete examples that satisfy the decay inequality (3.63). To achieve this, if g 1 and g 2 are bounded above and below by linear or superlinear functions near the origin, that is, for all s < 1 ,

(3.92) c i s κ i g i ( s ) c ¯ i s κ i ,

where κ i 1 , c i > 0 , and c ¯ i > 0 , i = 1 , 2 . Then, we can select

(3.93) h i ( s ) = c i 2 κ i + 1 ( 1 + c ¯ i 2 ) s 2 κ i + 1 ,

which satisfies Inequality (3.60). However, if the dampings are bounded by sublinear functions near the origin, namely, for all s < 1 ,

(3.94) c i s k i g i ( s ) c ¯ i s k i ,

where k i < 1 , c i > 0 and c ¯ i > 0 , i = 1 , 2 . Then, we can select

(3.95) h i ( s ) = c i 2 k i k i + 1 ( 1 + c ¯ i 2 ) s 2 k i k i + 1 ,

which also satisfies Inequality (3.60). Therefore, by (3.93) and (3.95), there exists constant C i such that

(3.96) h i ( s ) = C i s K i ,

where K i = 2 κ i + 1 or 2 k i k i + 1 .

Next, we rely on a corollary provided by Lasiecka and Toundykov [35] or Cavalcanti et al. [9] for calculations and estimations.

Corollary 3.1

Suppose that the assumptions of Theorem 3.1 are satisfied and that the function h in (3.62) can be expressed as h = h b + h s , where h b and h s are concave, monotone increasing, zero at the origin, and, in addition, h s = o ( h b ) . The latter means that lim x 0 + h s ( x ) h b ( x ) = 0 and h b has no upper linear bound on [ 0 , 1 ) . Then, given any positive χ < 1 , there exists t 0 = t 0 ( χ ) T such that the following energy estimate holds:

(3.97) U ( t ) S ¯ t T 0 1 , for a l l t 2 t 0 ,

where S ¯ is the solution of the following nonlinear ODE:

(3.98) d d t S ¯ ( t ) + h b 1 ( χ K ˜ S ¯ ) = 0 , S ¯ ( 0 ) = U ( 0 ) ,

with K ˜ and T 0 as in Theorem 3.1.

Proof

The proof is similar to Corollary 1 in [35] or Corollary 2 in [9].□

Finally, we will provide some specific examples of energy stability based on Corollary 3.1.

Example 3.1

If ( g 1 , g 2 ) is the combination (L, L) and g 1 and g 2 are all linearly bounded near the origin, we have

h ( s ) = s + c 1 1 ( 1 + c ¯ 1 2 ) L T s + c 2 1 ( 1 + c ¯ 2 2 ) L T s .

Therefore, according to (3.64) and (3.65), the function q is also linear, and it implies that U ( t ) decays exponentially.

Example 3.2

If at least one of g 1 and g 2 are not linearly bounded near the origin or at infinity, then the decay rate of U ( t ) is algebraic. Indeed, by using Table 1, we can rewrite that

h ˜ ( s ) = s Q 1 + s Q 2 ,

where Q i = 1 or θ i 2 θ i q i 1 or θ i 2 q 1 θ i q i 1 , i = 1 , 2 . Let Q = min { Q 1 , Q 2 , K 1 , K 2 } , if Q = 1 , then this situation follows Example 3.1. If Q < 1 , then according to Corollary 3.1, we select h b ( s ) s Q . Thus,

h b 1 ( s ) s 1 Q ,

which satisfies ODE

d d t S ¯ ( t ) + C ˆ S ¯ 1 Q = 0 , S ¯ ( 0 ) = U ( 0 ) ,

where C ˆ depends on χ , K ˜ . By direct calculation, we can obtain that

S ¯ ( t ) = C ˆ 1 Q 1 t + U ( 0 ) 1 1 Q Q 1 Q .

Then, it is known from (3.97) that the decay rate of U ( t ) is algebraic and the order of the polynomial decay rate is Q 1 Q .

4 Global attractors

In this section, our main task is to study the existence of global attractors for Problems (1.1) and (1.2). In order to obtain a nontrivial global attractor, we need to modify Assumptions (2.4), (2.7), and (2.8) as follows:

Assumption 4.1

  1. Suppose that q 1 = q 2 = 1 for (2.4), that is,

    (4.1) m i s 2 g i ( s ) s M i s 2 , for i = 1 , 2 , s 1 .

    Moreover, we assume that

    (4.2) g i ( s ) M i , s R .

  2. There exists constants 0 β < K π 2 2 L 2 and m F > 0 such that for all u , v R ,

    (4.3) F ( u , v ) β ( u 2 + v 2 ) m F ,

    (4.4) F ( u , v ) ( u , v ) F ( u , v ) β ( u 2 + v 2 ) m F .

Remark 4.1

  1. Assumption (4.1) can be replaced with a weaker assumption

    liminf s g i ( s ) > 0 , for i = 1 , 2 .

  2. A simple example of F satisfying all above assumption (G2) is

    F ( u , v ) = u + v 4 u + v 2 + 2 u v 2 ,

    and by a simple calculation, conditions (4.3) and (4.4) hold with m F = 1 4 and p = 3 .

  3. Assumption (G2) guarantees existence of nontrivial global attractors. Moreover, Theorem 2.1 also holds under the assumption (G2). Then, Problems (1.1) and (1.2) generate a dynamical system in the space with the evolution operator S α ( t ) : as

    S α ( t ) U 0 = U ( t ) , t 0 ,

    where U ( t ) = ( u ( t ) , u t ( t ) , v ( t ) , v t ( t ) ) is the weak solution to (1.1) and (1.2) with initial data U 0 = ( u 0 , u 1 , v 0 , v 1 ) .

4.1 Gradient system

In this subsection, we prove that the dynamical system ( , S α ( t ) ) is a gradient and the set of stationary solutions is bounded.

Lemma 4.1

Suppose that (A1), (A2) of Assumptions 2.1 and 4.1 hold. Then, the dynamical system ( , S α ( t ) ) is a gradient, i.e., there exists a strict Lyapunov function Φ defined. In addition,

(4.5) Φ ( U ) + if a n d o n l y i f U + .

Proof

Let U , and we denote S α ( t ) U = ( u ( t ) , u t ( t ) , v ( t ) , v t ( t ) ) as the corresponding solution of (1.1) and (1.2). Let us define the function Φ : R by

(4.6) Φ ( S α ( t ) U ) = 1 2 ( u ( t ) , u t ( t ) , v ( t ) , v t ( t ) ) 2 + 0 L F ( u ( t ) , v ( t ) ) d x .

From (2.3) and (2.16), we have that

(4.7) d d t Φ ( S α ( t ) U ) = 0 L ( β 1 ( x ) g 1 ( u t ) u t + β 2 ( x ) g 2 ( v t ) v t ) d x 0 , t 0 .

Thus, the function t Φ ( S α ( t ) U ) is nonincreasing.

Next, we suppose that Φ ( S α ( t ) U ) = Φ ( U ) for all t 0 . Then, (4.7) implies that

0 L ( β 1 ( x ) g 1 ( u t ) u t + β 2 ( x ) g 2 ( v t ) v t ) d x = 0 , t 0 .

Then, using (4.1), we can deduce for all t 0 that

0 L ( u t 2 + v t 2 ) d x = 0

and

0 L ( β 1 ( x ) g 1 ( u t ) 2 + β 2 ( x ) g 2 ( v t ) 2 ) d x = 0 ,

which shows that for all T > 0

u t = v t = 0 a.e. in ω × ( 0 , T )

and

β 1 ( x ) g 1 ( u t ) = β 2 ( x ) g 2 ( v t ) = 0 a.e. in ( 0 , L ) × ( 0 , T ) .

This means that U = ( u ( t ) , u t ( t ) , v ( t ) , v t ( t ) ) is a solution of

(4.8) ρ ( x ) u t t ( K ( x ) u x ) x + α ( u v ) + f 1 ( u , v ) = 0 , in ( 0 , L ) × ( 0 , T ) , ρ ( x ) v t t ( K ( x ) v x ) x α ( u v ) + f 2 ( u , v ) = 0 , in ( 0 , L ) × ( 0 , T ) , u t = v t = 0 , in ω × ( 0 , ) .

By density arguments, we can assume U = ( u ( t ) , u t ( t ) , v ( t ) , v t ( t ) ) is a strong solution. Then, taking the derivative of equation (4.8) with respect to the variable t in distributional sense and denoting w = u t and z = v t , we obtain the linear system

(4.9) ρ ( x ) w t t ( K ( x ) w x ) x + α ( w z ) + t [ f 1 ( u , v ) ] = 0 , in ( 0 , L ) × ( 0 , T ) , ρ ( x ) z t t ( K ( x ) z x ) x α ( w z ) + t [ f 2 ( u , v ) ] = 0 , in ( 0 , L ) × ( 0 , T ) , w = z = 0 , in ω × ( 0 , ) .

If we apply to (2.5) and set the following functions:

V 1 ( x , t ) = α + f 1 ( u , v ) u L 2 ( 0 , T ; L 2 ( 0 , L ) ) , V 2 ( x , t ) = α + f 2 ( u , v ) v L 2 ( 0 , T ; L 2 ( 0 , L ) ) , V 3 ( x , t ) = α f 1 ( u , v ) v L 2 ( 0 , T ; L 2 ( 0 , L ) ) , V 4 ( x , t ) = α f 2 ( u , v ) u L 2 ( 0 , T ; L 2 ( 0 , L ) ) ,

we can rewrite (4.9) as

(4.10) ρ ( x ) w t t ( K ( x ) w x ) x + V 1 ( x , t ) u = V 3 ( x , t ) v , in ( 0 , L ) × ( 0 , T ) , ρ ( x ) z t t ( K ( x ) z x ) x + V 2 ( x , t ) v = V 4 ( x , t ) u , in ( 0 , L ) × ( 0 , T ) , w = z = 0 , in ω × ( 0 , ) .

Using Lemma 3.1, we conclude that ( u t , v t ) = ( w , z ) = ( 0 , 0 ) . Therefore, U = ( u 0 , 0 , v 0 , 0 ) is a stationary solution of ( , S ( t ) ) . This proves that Φ is a strict Lyapunov function.

Now, it follows from (4.3) and Poincaré inequality that

(4.11) 0 L F ( u , v ) d x β ( u 2 + v 2 ) L m F β L 2 π 2 ( u x 2 + v x 2 ) L m F β L 2 π 2 U 2 L m F ,

and therefore, (4.11) together with 0 β < K 1 π 2 2 L 2 leads to

(4.12) Φ ( t ) 1 2 β L 2 π 2 K 1 U 2 L m F .

On the other hand, using the embedding H 0 1 ( 0 , L ) L ( 0 , L ) and (2.8), we deduce that

0 L F ( u , v ) d x C ( 1 + u x p + 1 + v x p + 1 ) .

So, we obtain

(4.13) Φ ( t ) C ( 1 + U p + 1 ) .

Combining (4.12) and (4.13), we have

(4.14) 1 2 β L 2 π 2 K 1 U 2 L m F Φ ( t ) C ( 1 + U p + 1 ) , t 0 ,

and it implies that (4.5) holds. The proof is complete.□

Lemma 4.2

Suppose that (A1), (A2) of Assumption 2.1 and Assumption 4.1 hold. Then, the set N of the stationary points of ( , S α ( t ) ) is bounded in uniformly in α .

Proof

Stationary points U = ( u , 0 , v , 0 ) satisfy the system

(4.15) ( K ( x ) u x ) x + α ( u v ) + f 1 ( u , v ) = 0 , in ( 0 , L ) × ( 0 , T ) , ( K ( x ) v x ) x α ( u v ) + f 2 ( u , v ) = 0 , in ( 0 , L ) × ( 0 , T ) .

Multiplying the first equation in equation (4.15) by u and second by v , respectively, taking the sum and integrating over ( 0 , L ) , we obtain

(4.16) 0 L [ K ( x ) ( u x 2 + v x 2 ) + α ( u v ) 2 ] d x = 0 L ( f 1 ( u , v ) u + f 2 ( u , v ) v ) d x .

By Poincare inequality, (4.3), and (4.4), we have

(4.17) 0 L ( f 1 ( u , v ) u + f 2 ( u , v ) v ) d x 2 β ( u 2 + v 2 ) + 2 L m F 2 β L 2 π 2 ( u x 2 + v x 2 ) + 2 L m F 2 β L 2 π 2 K 1 0 L [ K ( x ) ( u x 2 + v x 2 ) + α ( u v ) 2 ] d x + 2 L m F .

Then, using 0 β < K 1 π 2 2 L 2 , we conclude that

U 2 = 0 L [ K ( x ) ( u x 2 + v x 2 ) + α ( u v ) 2 ] d x 2 π 2 L K 1 m F π 2 K 1 2 β L 2 , U N .

The proof is complete.□

4.2 Asymptotic smoothness

In this subsection, we will show that the dynamical system ( , S α ( t ) ) is asymptotically smooth.

Lemma 4.3

Under the hypotheses Lemma 4.1 given a bounded subset of , suppose K ( x ) K , ρ ( x ) ρ are constants. Let S α ( t ) U i = ( u i , u t i , v i , v t i ) , i = 1 , 2 be two weak solutions of Problems (1.1) and (1.2) with initial data U 1 , U 2 . Then, for any ε > 0 , there exists a positive constant C ε , independent of α , such that for T > 0 sufficiently large, we have

(4.18) E U ( T ) ε + C ε , T + C , T , ε 0 T ( u ( t ) 2 p 2 + v ( t ) 2 p 2 ) d t ,

where u = u 1 u 2 , v = v 1 v 2 , U = ( u , u t , v , v t ) , and E U ( t ) 1 2 U 2 .

Proof

For u = u 1 u 2 , we use the following notation:

G i ( u ) = g i ( u 1 ) g i ( u 2 ) and F i ( u ) = f i ( u 1 ) f i ( u 2 ) .

Then, ( u , u t , v , v t ) solves the system

(4.19) ρ u t t ( K u x ) x + α ( u v ) + β 1 ( x ) G 1 ( u t ) + F 1 ( u , v ) = 0 , ρ v t t ( K v x ) x α ( u v ) + β 1 ( x ) G 2 ( v t ) + F 2 ( u , v ) = 0 ,

with the Dirichlet boundary condition and initial condition

( u ( 0 ) , u t ( 0 ) , v ( 0 ) , v t ( 0 ) ) = ( u 0 1 u 0 2 , u 1 1 u 1 2 , v 0 1 v 0 2 , v 1 1 v 1 2 ) .

Let us consider ε 0 , small enough, such that 0 < ε 0 ω 2 , where ω ( L 1 , L 2 ) . We define the auxiliary function, as in [8,11],

(4.20) h η ( x ) = ( η 1 ) x , x [ 0 , L 1 + ε 0 ) , η ( x L 1 ε 0 ) + L 1 L 2 + 2 ε 0 L ( L 1 + ε 0 ) , x [ L 1 + ε 0 , L 2 ε 0 ] , ( η 1 ) ( x L ) , x ( L 2 ε 0 , L ] ,

with η L ( L 2 L 1 2 ε 0 ) L [ 0 , 1 ) .

Now, multiplying the equations in (4.19) by h η u x , h η v x , respectively, and integrating by parts over [ 0 , L ] × [ 0 , T ] , we obtain

(4.21) 1 2 0 T 0 L [ ρ ( u t 2 + v t 2 ) + K ( u x 2 + v x 2 ) + α ( u v ) 2 ] h η d x d t = α 0 T 0 L ( u v ) 2 h η d x d t 0 L ρ ( u t u x + v t v x ) h η d x 0 T 0 T 0 L ( β 1 ( x ) G 1 ( u t ) u x + β 2 ( x ) G 2 ( v t ) v x ) h η d x d t 0 T 0 L ( F 1 ( u , v ) u x + F 2 ( u , v ) v x ) h η d x d t .

Then, observing that

(4.22) h η ( x ) = η 1 , x [ 0 , L 1 + ε 0 ) ( L 2 ε 0 , L ] , η , x [ L 1 + ε 0 , L 2 ε 0 ] ,

we deduce that

(4.23) ( 1 η ) 0 T E U ( t ) d t = 1 2 0 T L 1 + ε 0 L 2 ε 0 [ ρ ( u t 2 + v t 2 ) + K ( u x 2 + v x 2 ) + α ( u v ) 2 ] d x d t α 0 T 0 L ( u v ) 2 h η d x d t + 0 L ρ ( u t u x + v t v x ) h η d x 0 T + 0 T 0 L ( β 1 ( x ) G 1 ( u t ) u x + β 2 ( x ) G 2 ( v t ) v x ) h η d x d t + 0 T 0 L ( F 1 ( u , v ) u x + F 2 ( u , v ) v x ) h η d x d t .

Step 1. Now, let us estimate the left-hand side of (4.23). First, employing the embedding L 2 p ( 0 , L ) L 2 ( 0 , L ) , we have

(4.24) 0 T 0 L ( u v ) 2 h η d x d t 2 0 T 0 L ( u 2 + v 2 ) h η d x d t C 0 T ( u ( t ) 2 p 2 + v ( t ) 2 p 2 ) d t .

Using Hölder’s and Young’s inequalities yields that

(4.25) 0 L ρ ( u t u x + v t v x ) h η d x 0 T C ( E U ( T ) + E U ( 0 ) ) .

To estimate the damping terms, we use the assumption (G1), which implies that for any δ > 0 , there exists a constant C δ > 0 such that

(4.26) C δ G i ( s ) s + δ s 2 , for all s R .

Thus, by (4.2), (4.26), and Young’s inequality, we obtain for every ε > 0 ,

(4.27) 0 T 0 L β 1 ( x ) G 1 ( u t ) u x h η d x d t M 1 0 T 0 L β 1 ( x ) u t u x h η d x d t ε 2 0 T E U ( t ) d t + C ε 0 T 0 L β 1 ( x ) u t 2 d x d t ε 2 0 T E U ( t ) d t + C ε C δ 0 T 0 L β 1 ( x ) G 1 ( u t ) u t d x d + δ T L C ε β 1 .

Similarly, we obtain

(4.28) 0 T 0 L β 2 ( x ) G 2 ( v t ) v x h η d x d t ε 2 0 T E U ( t ) d t + C ε C δ 0 T 0 L β 2 ( x ) G 2 ( v t ) v t d x d t + δ T L C ε β 2 .

This allows us to conclude the following estimate:

(4.29) 0 T 0 L ( β 1 ( x ) G 1 ( u t ) u x + β 2 ( x ) G 2 ( v t ) v x ) h η d x d t C ε C δ 0 T 0 L ( β 1 ( x ) G 1 ( u t ) u t + β 2 ( x ) G 2 ( v t ) v t ) d x dt + ε 0 T E U ( t ) d t + δ T L C ε ( β 1 + β 2 ) .

Next we estimate the source terms. Using the assumption (2.5), Hölder’s inequality with exponents p 1 = 2 ( p + 1 ) p 1 , p 2 = 2 p , p 3 = 2 , and L 2 ( p + 1 ) ( 0 , L ) H 0 1 ( 0 , L ) , we obtain that

(4.30) 0 T 0 L F 1 ( u , v ) u x h η d x d t C 0 T 0 L ( 1 + u 1 p 1 + u 2 p 1 + v 1 p 1 + v 2 p 1 ) ( u + v ) u x d x d t C 0 T ( u ( t ) 2 p + v ( t ) 2 p ) u x ( t ) d t ε 2 0 T E U ( t ) d t + C ε , 0 T ( u ( t ) 2 p 2 + v ( t ) 2 p 2 ) d x d t ,

and analogously, we also have

(4.31) 0 T 0 L F 2 ( u , v ) v x h η d x d t ε 2 0 T E U ( t ) d t + C ε , 0 T ( u ( t ) 2 p 2 + v ( t ) 2 p 2 ) d t .

The above implies that

(4.32) 0 T 0 L ( F 1 ( u , v ) u x + F 2 ( u , v ) v x ) h η d x d t ε 0 T E U ( t ) d t + C ε , 0 T ( u ( t ) 2 p 2 + v ( t ) 2 p 2 ) d t .

Now, we combine (4.28), (4.29), (4.32) with (4.23). For ε > 0 small enough, we have

(4.33) 0 T E U ( t ) d t δ T L C ε ( β 1 + β 2 ) + C ε ( E U ( T ) + E U ( 0 ) ) + C ε , δ 0 T 0 L ( β 1 ( x ) G 1 ( u t ) u t + β 2 ( x ) G 2 ( v t ) v t ) d x dt + C ε 0 T L 1 + ε 0 L 2 ε 0 [ ρ ( u t 2 + v t 2 ) + K ( u x 2 + v x 2 ) + α ( u v ) 2 ] d x d t + C ε , 0 T ( u ( t ) 2 p 2 + v ( t ) 2 p 2 ) d t .

Step 2. Next we estimate the fourth integral on the right in (4.33). We consider a cutoff function γ C 0 ( 0 , L ) such that

(4.34) γ ( x ) = 0 , x [ 0 , L 1 ] [ L 1 , L ] , 1 , x [ L 1 + ε 0 , L 2 ε 0 ] .

Thus, multiplying the first and second equations on equation (4.19) by γ u and γ v , respectively, and performing integration by parts, we obtain that

(4.35) 2 0 T E U ( t ) γ d t = 0 L ρ ( u u t + v v t ) γ d x 0 T + 2 0 T 0 L ρ ( u t 2 + v t 2 ) γ d x d t 0 T 0 L K ( u u x + v v x ) γ d x d t 0 T 0 L ( β 1 ( x ) G 1 ( u t ) u + β 2 ( x ) G 2 ( v t ) v ) γ d x d t 0 T 0 L ( F 1 ( u , v ) u + F 2 ( u , v ) v ) γ d x d t .

We shall estimate the right-hand side of (4.35). Analogous to (4.25), we also have

(4.36) 0 L ρ ( u u t + v v t ) γ d x 0 T C ( E U ( T ) + E U ( 0 ) ) .

Again applying the estimate (4.26), we can conclude that

(4.37) 0 T 0 L ρ ( u t 2 + v t 2 ) γ d x dt ρ 0 T 0 L ( u t 2 + v t 2 ) d x dt ρ C δ 0 T 0 L ( G 1 ( u t ) u t + G 2 ( v t ) v t ) d x d t +2 δ T L .

Now, integrating by parts and using the embedding L 2 p ( 0 , L ) L 2 ( 0 , L ) , we have

(4.38) 0 T 0 L K ( u u x + v v x ) γ d x d t = K 2 0 T 0 L d d x ( u 2 + v 2 ) γ d x d t = K 2 0 T 0 L ( u 2 + v 2 ) γ d x d t C 0 T ( u ( t ) 2 p 2 + v ( t ) 2 p 2 ) d t .

Again using (4.2), (4.26), and Hölder’s and Young’s inequalities, we deduce that

(4.39) 0 T 0 L ( β 1 ( x ) G 1 ( u t ) u + β 2 ( x ) G 2 ( v t ) v ) γ d x d t 0 T 0 L ( M 1 β 1 ( x ) u t u + M 2 β 2 ( x ) v t v ) γ d x d t ε 0 T E U ( t ) γ d t + C , ε 0 T ( u ( t ) 2 p 2 + v ( t ) 2 p 2 ) d t .

Analogous to (4.30), we see that

(4.40) 0 T 0 L ( F 1 ( u , v ) u + F 2 ( u , v ) v ) γ d x d t C 0 T ( u ( t ) 2 p 2 + v ( t ) 2 p 2 ) d t .

Substituting the estimates (4.36)–(4.40) into (4.35) and choosing ε > 0 small enough, we obtain that

(4.41) 0 T E U ( t ) γ d t 4 δ T L C ε + C ε ( E U ( T ) + E U ( 0 ) ) + C ε , δ 0 T 0 L ( G 1 ( u t ) u t + G 2 ( v t ) v t ) d x d t + C ε , 0 T ( u ( t ) 2 p 2 + v ( t ) 2 p 2 ) d t .

Inserting (4.41) into (4.33), and choosing ε 0 small enough, we obtain

(4.42) 0 T E U ( t ) d t δ T L C ε ( β 1 + β 2 + 1 ) + C ε ( E U ( T ) + E U ( 0 ) ) + C ε , δ 0 T 0 L ( β 1 ( x ) G 1 ( u t ) u t + β 2 ( x ) G 2 ( v t ) v t ) d x d t + C ε , 0 T ( u ( t ) 2 p 2 + v ( t ) 2 p 2 ) d t .

Step 3. Next we estimate damping terms in (4.42). Multiplying the first and second equation of (4.19) by u t and v t , and integrating by parts over [ 0 , L ] × [ s , T ] such that

(4.43) s T 0 L ( β 1 ( x ) G 1 ( u t ) u t + β 2 ( x ) G 2 ( v t ) v t ) d x d t = E U ( s ) E U ( T ) s T 0 L ( F 1 ( u , v ) u t + F 2 ( u , v ) v t ) d x d t .

Analogous to (4.32), we can conclude that

(4.44) 0 T 0 L ( F 1 ( u , v ) u t + F 2 ( u , v ) v t ) d x d t ε 0 T E U ( t ) d t + C ε , 0 T ( u ( t ) 2 p 2 + v ( t ) 2 p 2 ) d t .

Consequently, combining (4.44) ( s = 0 ) in (4.43), we obtain

(4.45) 0 T 0 L ( β 1 ( x ) G 1 ( u t ) u t + β 2 ( x ) G 2 ( v t ) v t ) d x d t E U ( 0 ) E U ( T ) + ε 0 T E U ( t ) d t + C ε , 0 T ( u ( t ) 2 p 2 + v ( t ) 2 p 2 ) d t .

We return to (4.42) and using (4.45), with ε > 0 small enough, the following estimate is obtained:

(4.46) 0 T E U ( t ) d t δ T L C ε ( β 1 + β 2 + 1 ) + ( C ε C ε , δ ) E U ( T ) + ( C ε + C ε , δ ) E U ( 0 ) + C ε , δ , 0 T ( u ( t ) 2 p 2 + v ( t ) 2 p 2 ) d t .

Step 4. Integrating the energy equality (4.43) with respect to s yields

(4.47) T E U ( T ) = 0 T E U ( t ) d t 0 T s T 0 L ( β 1 ( x ) G 1 ( u t ) u t + β 2 ( x ) G 2 ( v t ) v t ) d x d t d s 0 T s T 0 L ( F 1 ( u , v ) u t + F 2 ( u , v ) v t ) d x d t d s .

Similarly (4.44), we find

(4.48) 0 T s T 0 L ( F 1 ( u , v ) u t + F 2 ( u , v ) v t ) d x d t d s 0 T s T 1 T E U ( t ) + C T , ( u ( t ) 2 p 2 + v ( t ) 2 p 2 ) d t d s 0 T E U ( t ) d t + C T , 0 T ( u ( t ) 2 p 2 + v ( t ) 2 p 2 ) d t .

By (4.26), and substituting (4.48) into (4.47) yields

(4.49) T E U ( T ) 2 0 T E U ( t ) d t + C T , 0 T ( u 2 p 2 + v 2 p 2 ) d t .

Step 5. Finally, by the estimates (4.46) and (4.49), we can find

(4.50) T E U ( T ) δ T L C ε ( β 1 + β 2 + 1 ) + ( C ε C ε , δ ) E U ( T ) + ( C ε + C ε , δ ) E U ( 0 ) + C ε , δ , T , 0 T ( u ( t ) 2 p 2 + v ( t ) 2 p 2 ) d t .

Therefore, we see that

(4.51) E U ( T ) δ L C ε ( β 1 + β 2 + 1 ) + ( C ε C ε , δ ) E U ( T ) + ( C ε + C ε , δ ) E U ( 0 ) T + C ε , δ , T , 0 T ( u ( t ) 2 p 2 + v ( t ) 2 p 2 ) d t .

Let δ > 0 sufficiently small and T sufficiently large, (4.18) holds. The proof is complete.□

Theorem 4.1

Under the assumptions of Lemma 4.3. Then, the system ( , S α ( t ) ) generated by Problems (1.1) and (1.2) is asymptotically smooth.

Proof

In order to prove Theorem 4.1, we will use the abstract result Theorem 6.5. Let be bounded positively invariant set, then by (4.18) and choosing T sufficiently large, such that

E U ( T ) ε + Ψ T , , ε ( U 1 , U 2 ) ,

where

Ψ T , , ε ( U 1 , U 2 ) = C , T , ε 0 T ( u 1 ( t ) u 2 ( t ) 2 p 2 + v 1 ( t ) v 2 ( t ) 2 p 2 ) d t 1 2 .

Let U n = ( u n , u t n , v n , v t n ) , by positive invariance of , we have

S α ( t ) U n 2 = ρ ( u t n 2 + v t n 2 ) + K ( u x n 2 + v x n 2 ) + α u n v n 2 C , t 0 .

Hence,

(4.52) { u n } , { v n } are bounded in L ( 0 , T ; H 0 1 ( 0 , L ) ) ,

(4.53) { u t n } , { v t n } are bounded in L ( 0 , T ; L 2 ( 0 , L ) ) .

Thus, by Alaoglu’s theorem, there exists a subsequence, reindexed again by n , such that

(4.54) ( u n , v n ) ( u , v ) weakly-star in L ( 0 , T ; ( H 0 1 ( 0 , L ) ) 2 ) ,

(4.55) ( u t n , v t n ) ( u t , v t ) weakly-star in L ( 0 , T ; ( L 2 ( 0 , L ) ) 2 ) .

By Simon’s compactness theorem [45], we conclude that

(4.56) ( u n , v n ) ( u , v ) strongly in C ( ( 0 , T ] ; ( H 0 1 ε 0 ( 0 , L ) ) 2 ) , ε 0 ( 0 , 1 ] .

Therefore,

lim m + lim n + 0 T ( u m u n 2 p 2 + v m v n 2 p 2 ) d t = 0 .

That is,

lim inf m lim inf n Ψ ε , , T ( U m , U n ) = 0 .

The asymptotic smoothness follows by Theorem 6.5. This completes the proof.□

Remark 4.2

All the constants in the proofs of Lemma 4.3 and Theorem 4.1 are independent of the variable α .

4.3 Existence of global attractors

In this section, we prove the existence of global attractors for the dynamical systems ( , S α ( t ) ) and study their properties.

Theorem 4.2

Under the assumptions of Lemma 4.3. Then, the system ( , S α ( t ) ) generated by equations (1.1) and (1.2) has a global attractor A α given by

A α = u ( N ) ,

where N is the set of stationary points of S α ( t ) and u ( N ) is the unstable manifold of N .

Proof

Using Lemma 4.1, Lemma 4.2, and Theorem 4.1, we know all the assumptions of [15, Corollary 7.5.7] are fulfilled and consequently the system ( , S α ( t ) ) has a global attractor A α = u ( N ) . Thus, the proof of Theorem 4.2 is complete.□

Remark 4.3

Suppose

(4.57) m i g i ( s ) M i , s R .

As in [26], if the proof process has been slightly modified, we can prove finiteness of fractal dimension and smoothness of global attractors by Quasi-stability ([14], Theorem 3.4.18). We can also establish that the singular limit of two wave equations connected in parallel is a single wave equation when α + . Moreover, upper-semicontinuity of global attractors is proved when α + .

5 Determining functionals

Determining functionals can be interpreted as some kinds of measurements of the state of the system. Based on the fourth section, we now present several results on the determining functionals of the dynamical system ( , S α ( t ) ) .

5.1 The concept and a related lemma of determining functionals

In order to state the main conclusion, we first provide the definitions of completeness defect and the determining functionals for the dynamical system ( , S α ( t ) ) , see [14,46] for more details.

Definition 5.1

Let = { l j : j = 1 , , N } be a set of continuous, linear functionals on ( H 0 1 ( 0 , L ) ) 2 . We say that is a determining set of functionals, if for any two trajectories S α ( t ) U i = ( u i , u t i , v i , v t i ) , i = 1 , 2 , we have

(5.1) lim t + l j ( u 1 ( t ) , v 1 ( t ) ) l j ( u 2 ( t ) , v 2 ( t ) ) 2 = 0 for j = 1 , , N ,

which implies that

(5.2) lim t + S α ( t ) U 1 S α ( t ) U 2 2 = 0 .

Definition 5.2

Let = { l j } j = 1 N be a finite set of linear functionals on ( H 0 1 ( 0 , L ) ) 2 . The completeness defect of on ( H 0 1 ( 0 , L ) ) 2 , with respect to ( L p ˆ ( 0 , L ) ) 2 ( p ˆ 2 p 2 ) is the value

(5.3) ε , p ˆ ε ( ( H 0 1 ( 0 , L ) ) 2 , ( L p ˆ ( 0 , L ) ) 2 ) = sup { ( u , v ) ( L p ˆ ( 0 , L ) ) 2 : l j ( u , v ) = 0 , ( u , v ) ( H 0 1 ( 0 , L ) ) 2 1 } .

Next, we prove the following key lemma.

Lemma 5.1

Let = { l j } j = 1 N be a finite set of linear functionals on ( H 0 1 ( 0 , L ) ) 2 . Then, there exists C > 0 such that for any ( u , v ) ( H 0 1 ( 0 , L ) ) 2 , we have

(5.4) ( u , v ) ( L p ˆ ( 0 , L ) ) 2 ε , p ˆ ( u , v ) ( H 0 1 ( 0 , L ) ) 2 + C max j = 1 , , N l j ( u , v ) .

Proof

Let { e j : j = 1 , , N } ( H 0 1 ( 0 , L ) ) 2 be an orthonormal system for , i.e., we have that l j ( e i ) = 0 if i j and l j ( e j ) = 1 . In this case, for any ( u , v ) ( H 0 1 ( 0 , L ) ) 2 , the element ( u ˆ , v ˆ ) = ( u , v ) j = 1 N l j ( u , v ) e j satisfies that

l j ( u ˆ , v ˆ ) = 0 for j = 1 , , N .

By the definition in (5.3), we have that ( u ˆ , v ˆ ) ( L p ˆ ( 0 , L ) ) 2 ε , p ˆ ( u ˆ , v ˆ ) ( H 0 1 ( 0 , L ) ) 2 . Therefore, we conclude that

(5.5) ( u , v ) ( L p ˆ ( 0 , L ) ) 2 ( u ˆ , v ˆ ) ( L p ˆ ( 0 , L ) ) 2 + ( u , v ) ( u ˆ , v ˆ ) ( L p ˆ ( 0 , L ) ) 2 ε , p ˆ ( u ˆ , v ˆ ) ( H 0 1 ( 0 , L ) ) 2 + ( u , v ) ( u ˆ , v ˆ ) ( L p ˆ ( 0 , L ) ) 2 ε , p ˆ [ ( u , v ) ( u ˆ , v ˆ ) ( H 0 1 ( 0 , L ) ) 2 + ( u , v ) ( H 0 1 ( 0 , L ) ) 2 ] + ( u , v ) ( u ˆ , v ˆ ) ( L p ˆ ( 0 , L ) ) 2 ε , p ˆ ( u , v ) ( H 0 1 ( 0 , L ) ) 2 + C max j = 1 , , N l j ( u , v )

for some C > 0 .□

5.2 Construction of determining functionals

According to Theorem 4.2, if there is a global attractor, then the dynamic system ( , S α ( t ) ) must have a bounded positive invariant absorb set ^ . We now present a important result on determining functionals as in [14, Theorem 3.4.20].

Theorem 5.1

Take the hypotheses from Theorem 4.2 and (4.57), and consider ( , S α ( t ) ) as above. Let = { l j } j = 1 N be a finite set of linear functionals on ( H 0 1 ( 0 , L ) ) 2 . Then, there exists a number ε ˆ with ε , 2 ε ˆ such that is a determining set of functionals for ( , S α ( t ) ) .

Proof

Let S α ( t ) U i = ( u i , u t i , v i , v t i ) , i = 1 , 2 be two trajectories for S α ( t ) U i ^ . We claim that if ε , 2 is sufficiently small, then

(5.6) lim t + l j ( u 1 ( t ) , v 1 ( t ) ) l j ( u 2 ( t ) , v 2 ( t ) ) 2 = 0 , j = 1 , , N ,

which implies that

(5.7) lim t + S α ( t ) U 1 S α ( t ) U 2 2 = 0 .

Indeed, suppose that the assumption in (5.6) holds and note that this is equivalent to

(5.8) A ( t ) sup s [ t , t + τ ] max j l j ( u 1 ( s ) , v 1 ( s ) ) l j ( u 2 ( s ) , v 2 ( s ) ) 2 = 0 , t .

Based on (4.50), Assumption (4.57), p ˆ 2 p , and the semigroup property, we can conclude that

(5.9) S α ( t + τ ) U 1 S α ( t + τ ) U 2 2 C ^ ( e ϖ t S α ( t ) U 1 S α ( t ) U 2 2 + sup t s t + τ ( u ( s ) , v ( s ) ) ( L p ˆ ( 0 , L ) ) 2 2 )

for some ϖ > 0 and C ^ > 0 , where u ( t ) = u 1 ( t ) u 2 ( t ) and v ( t ) = v 1 ( t ) v 2 ( t ) . By using Young’s inequality and (5.4), we know that for any δ > 0

(5.10) ( u , v ) ( L p ˆ ( 0 , L ) ) 2 2 ( 1 + δ ) ε , p ˆ 2 ( u , v ) ( H 0 1 ( 0 , L ) ) 2 2 + C , δ max j l j ( u , v ) 2 .

By using interpolation and the embedding H 0 1 ( 0 , L ) L ( 0 , L ) , we obtain

( u , v ) ( L p ˆ ( 0 , L ) ) 2 C ( u , v ) ( L p ˆ ( 0 , L ) ) 2 1 2 ( u , v ) ( H 0 1 ( 0 , L ) ) 2 1 2 .

Then, from [14, (3.3.9), p. 123] with V = ( H 0 1 ( 0 , L ) ) 2 , W = ( L p ˆ ( 0 , L ) ) 2 , and H = ( L 2 ( 0 , L ) ) 2 , where θ = 1 2 , and a θ = c ˆ , we infer that

(5.11) ε , p ˆ 2 c ˆ 2 ε , 2 c ˆ 4 [ ε ( ( L p ˆ ( 0 , L ) ) 2 , ( L 2 ( 0 , L ) ) 2 ) ] 2 .

It follows from (5.10) and (5.11) that we obtain

(5.12) ( u , v ) ( L p ˆ ( 0 , L ) ) 2 2 ( 1 + δ ) c ˆ 2 ε , 2 ( u , v ) ( H 0 1 ( 0 , L ) ) 2 2 + C , δ max j l j ( u , v ) 2 .

With the Lipschitz estimate on S α ( t ) in (2.15), we obtain from above,

(5.13) sup s [ t , t + τ ] ( u ( s ) , v ( s ) ) ( L p ˆ ( 0 , L ) ) 2 2 ( 1 + δ ) c ˆ 2 ε , 2 e C τ S α ( t ) U 1 S α ( t ) U 2 2 + C , δ A ( t ) .

Then, from (5.9) and (5.13), we conclude that

(5.14) S α ( t + τ ) U 1 S α ( t + τ ) U 2 2 C ^ [ e ϖ τ + ( 1 + δ ) c ˆ 2 ε , 2 e C τ ] S α ( t ) U 1 S α ( t ) U 2 2 + C ^ C , δ A ( t ) .

Choosing ϖ > 0 , τ > 0 sufficiently large, and ε , 2 sufficiently small, we can have that

C ^ [ e ϖ τ + ( 1 + δ ) c ˆ 2 ε , 2 e C τ ] 1 .

Thus, again using the semigroup property, we can iterate on intervals of size τ to obtain

S α ( t 0 + n τ ) U 1 S α ( t 0 + n τ ) U 2 2 { C ^ [ e ϖ τ + ( 1 + δ ) c ˆ 2 ε , 2 e C τ ] } n S α ( t 0 ) U 1 S α ( t 0 ) U 2 2 + C m = 0 n 1 { C ^ [ e ϖ τ + ( 1 + δ ) c ˆ 2 ε , 2 e C τ ] } n m 1 A ( t ) .

From here, taking n + , we obtain from (5.8) the desired conclusion in (5.7) and the proof of Theorem 5.1 is complete.□

Acknowledgments

The authors 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: This work was supported by the National Natural Science Foundation of China (Grant No. 12271227).

  2. Author contributions: Yunlong Gao: writing – original draft preparation, Chunyou Sun: writing - reviewing and editing, and Kaibin Zhang: writing – original draft preparation.

  3. Conflict of interest: The authors declare that there is no conflict of interests regarding the publication of this article.

  4. Data availability statement: Data sharing not applicable to this article as no data sets were generated or analyzed during the current study.

Appendix A Part I: Microlocal analysis

For the readers comprehension, we will recall some results which can be found in Burq and Gérard [4] and in Gérard [28] and were used in the proof of the stabilization.

Theorem A.1

[4, Theorem 5.14] Let { u n } n N be a bounded sequence in L loc 2 ( Ω ) such that it converges weakly to zero in L loc 2 ( Ω ) . Then, there exists a subsequence { u φ ( n ) } and a positive Radon measure μ on T 1 Ω Ω × S d 1 such that for all pseudo-differential operator A of order 0 on Ω which admits a principal symbol σ 0 ( A ) and for all χ C 0 ( Ω ) such that χ σ 0 ( A ) = σ 0 ( A ) , we have

(A1) ( A ( χ u φ ( n ) ) , χ u φ n ) L 2 n + Ω × S n 1 σ 0 ( A ) ( x , ξ ) d μ ( x , ξ ) .

Definition A.1

Under the circumstances of Theorem A.1, μ is called the microlocal defect measure (MDM) of the sequence { u φ ( n ) } n N .

Remark A.1

Theorem A.1 assures that for all bounded sequence { u n } n N of L loc 2 ( Ω ) which converges weakly to zero, there exists a subsequence admitting an MDM. We observe that from (A1) in the particular case when A = f C 0 ( O ) , it follows that

(A2) Ω f ( x ) u φ ( n ) ( x ) 2 d x n + O × S d 1 f ( x ) u φ ( n ) ( x ) 2 d μ ( x , ξ ) ,

so that u φ ( n ) converges to 0 strongly if and only if μ = 0 .

The second important result reads as follows.

Theorem A.2

[4, Theorem 5.18] Let P be a differential operator of order m on O and let u n be a bounded sequence of L loc 2 ( O ) which converges weakly to 0 and admits an MDM. μ . The following statements are equivalent:

  1. P u n n + 0 strongly i n H loc m ( O ) ( m > 0 ) .

  2. supp ( μ ) { ( x , ξ ) O × S n 1 : σ m ( P ) ( x , ξ ) = 0 } .

Theorem A.3

[4, Theorem 5.19] Let P be a differential operator of order m on Ω , verifying P * = P , and let u n be a bounded sequence in L loc 2 ( O ) which converges weakly to 0 and it admits an MDM. μ . Let us assume that P u n n + 0 strongly in H loc 1 m ( O ) . Then, for all function a C ( O × R n \ { 0 } ) homogeneous of degree 1 m in the second variable and with compact support in the first one,

(A3) Ω × S n 1 { a , p } ( x , ξ ) d μ ( x , ξ ) = 0 .

We complete this section by examining the case of the wave equation in an inhomogeneous medium:

P ( x , D ) u = ρ ( x ) t 2 u + i , j = 1 n x i ( K ( x ) x j u ) ,

whose principal symbol is given by

(A4) p ( t , x , τ , ξ ) = ρ ( x ) τ 2 + ξ K ( x ) ξ , where ξ = ( ξ 1 , , ξ d ) ,

where t R , x Ω R n , ( τ , ξ ) R d + 1 , ρ C ( Ω ) , 0 < α ρ ( x ) β < + , and K ( x ) = ( k i , j ( x ) ) 1 i , j d is a positive-definite matrix, verifying

a ξ 2 ξ K ( x ) ξ b ξ 2 ,

for 0 < a < b < + .

Proposition A.1

Unless a change in variables, the bicharacteristics of (A4) are curves of the form

t t , x ( t ) , τ , τ K ( x ( t ) ) ρ ( x ( t ) ) 1 x ˙ ( t ) ,

where t x ( t ) is a geodesic of the metric G = ( K ρ ) 1 on Ω , parameterized by the curvilinear abscissa.

The main result as follows:

Theorem A.4

[4, Theorem A.1] Let P be a self-adjoint differential operator of order m on O which admits a principal symbol p. Let { u n } n be a bounded sequence in L loc 2 ( O ) which converges weakly to zero, with a microlocal defect measure μ . Let us assume that P u n converges to 0 in H loc ( m 1 ) ( O ) . Then, the support of μ , supp ( μ ) , is a union of curves like s I x ( s ) , ξ ( s ) ξ ( s ) , where s I ( x ( s ) , ξ ( s ) ) is a bicharacteristic of p.

B Part II: Abstract results of dynamical systems

For the reader’s convenience, we present here some definitions and theorems related to dynamics of dissipative systems that can be found in classical works (e.g., [14,15]).

Definition B.2

Let { S ( t ) } t 0 be a semigroup on a metric space ( X , d ) . A subset A of X is called a global attractor for the semigroup, if A is compact and enjoys the following properties:

  1. A is invariant, that is, S ( t ) A = A , for all t 0 ;

  2. A attracts all bounded sets of X . That is, for any bounded subset B of X ,

    d ( S ( t ) B , A ) 0 , as t + ,

    where d ( B , A ) is the Hausdorff semi-distance.

Definition B.3

Let S ( t ) be an evolution operator on a complete metric space X . An evolution operator S ( t ) is said to be asymptotically smooth if the following condition is valid: for every bounded set D such that S ( t ) D D for t > 0 , there exists a compact set K in the closure D ¯ of D , such that

lim t + d X { S ( t ) D K } = 0 ,

where lim t + d X { A B } = sup x A dis t X ( x , B ) .

Theorem B.5

[14, Theorem 2.2.17] Let S ( t ) be an evolution operator on a complete metric space X. Assume that for any bounded forward invariant set B in X and for any ε > 0 , there exists T T ( ε ; B ) such that

d ( S ( T ) y 1 , S ( T ) y 2 ) ε + Ψ ε , B , T ( y 1 , y 2 ) , y i B ,

where Ψ ε , B , T ( y 1 , y 2 ) is a functional defined on B × B such that

lim inf m lim inf n Ψ ε , B , T ( y n , y m ) = 0 ,

for every sequence { y n } from B. Then, ( X , S ( t ) ) is an asymptotically smooth dynamical system.

Theorem B.6

[14, Theorem 2.3.5] Let ( X , S ( t ) ) be a dissipative asymptotically compact dynamical system on a complete metric space X. Then, S ( t ) possesses a unique compact global attractor A such that

A = ω ( B 0 ) = t > 0 τ t S ( τ ) B 0 ¯ ,

for every bounded absorbing set B 0 and

lim t + ( d X { S ( t ) B 0 A } + d X { A S ( t ) B 0 } ) = 0 ,

whereas above d X { A B } = sup x A dist ( x , B ) . Moreover, if there exists a connected absorbing bounded set, then A is connected.

Definition B.4

Let N be the sets of stationary points of ( H , S ( t ) ) , that is,

N { z H : S ( t ) z = z , t > 0 } .

Then, the unstable manifold emanating from N , represented by u ( N ) , is the set of all z H such that there is a full trajectory γ = { u ( t ) ; t R } satisfying

(A1) u ( 0 ) = z and lim t dis t H ( u ( t ) , N ) = 0 .

Definition B.5

A dynamical system ( H , S ( t ) ) is called gradient, if there exists a strict Lyapunov function on H , i.e., there exists a continuous function Φ such that t Φ ( S ( t ) z ) is nonincreasing for any z H , and if Φ ( S ( t ) z 0 ) = Φ ( z 0 ) for all t > 0 and some z 0 H , then z 0 is a stationary point of ( H , S ( t ) ) .

Theorem B.7

[15, Corollary 7.5.7] Let ( H , S ( t ) ) be an asymptotically compact gradient system on a Banach space H, with the corresponding Lyapunov functional denoted by Φ . Suppose that

(A2) Φ ( z ) + if a n d o n l y i f z H + ,

and that the set of stationary points N is bounded. Then, the system ( H , S ( t ) ) possesses a compact global attractor characterized by A = u ( N ) .

References

[1] C. Bardos, G. Lebeau, and J. Rauch, Sharp sufficient conditions for the observation, control, and stabilization of waves from the boundary, SIAM J. Control Optim. 30 (1992), no. 5, 1024–1065. 10.1137/0330055Suche in Google Scholar

[2] C. Buriol, L. G. Delatorre, V. H. Gonzalez Martinez, D. C. Soares, and E. H. G. Tavares, Asymptotic stability for a generalized nonlinear Klein-Gordon system, J. Differ. Equ. 280 (2021), 517–545. 10.1016/j.jde.2021.01.011Suche in Google Scholar

[3] N. Burq and P. Gérard, A necessary and sufficient condition for the exact controllability of the wave equation, C. R. Acad. Sci. Paris Seeeer. I Math. 325 (1997), no. 7, 749–752. 10.1016/S0764-4442(97)80053-5Suche in Google Scholar

[4] N. Burq and P. Gérard, Contróle Optimal des équations aux dérivés partielles, 2001, http://www.math.u-psud.fr/ burq/articles/coursX.pdf. Suche in Google Scholar

[5] M. M. Cavalcanti, V. N. Domingos Cavalcanti, R. Fukuoka, and J. A. Soriano, Asymptotic stability of the wave equation on compact manifolds and locally distributed damping: a sharp result, Arch. Ration. Mech. Anal. 197 (2010), 925–964. 10.1007/s00205-009-0284-zSuche in Google Scholar

[6] M. M. Cavalcanti, V. N. D. Domingos Cavalcanti, S. Mansouri, V. H. Gonzalez Martinez, Z. Hajjej, M. R. Astudillo Rojas, Asymptotic stability for a strongly coupled Klein-Gordon system in an inhomogeneous medium with locally distributed damping, J. Differ. Equ. 268 (2020), 447–489. 10.1016/j.jde.2019.08.011Suche in Google Scholar

[7] M. M. Cavalcanti, W. J. Corrêa, V. N. Domingos Cavalcanti, M. A. Jorge Silva, and J. P. Zanchetta, Uniform stability for a semilinear non-homogeneous Timoshenko system with localized nonlinear damping, Z. Angew. Math. Phys. 72 (2021), 191. 10.1007/s00033-021-01622-7Suche in Google Scholar

[8] M. M. Cavalcanti, V. N. Domingos Cavalcanti, F. A. F. Nascimento, I. Lasiecka, and J. H. Rodrigues, Uniform decay rates for the energy of Timoshenko system with the arbitrary speeds of propagation and localized nonlinear damping, Z. Angew. Math. Phys. 65 (2014), 1189–1206. 10.1007/s00033-013-0380-7Suche in Google Scholar

[9] M. M. Cavalcanti, V. N. Domingos Cavalcanti, V. H. Gonzalez Martinez, and T. Özsari, Decay rate estimates for the wave equation with subcritical semilinearities and locally distributed nonlinear dissipation, Appl. Math. Optim. 87 (2023), 2. 10.1007/s00245-022-09918-4Suche in Google Scholar

[10] M. M. Cavalcanti, V. N. Domingos Cavalcanti, V. H. Gonzalez Martinez, V. A. Peralta, and A. Vicente, Stability for semilinear hyperbolic coupled system with frictional and viscoelastic localized damping, J. Differ. Equ. 269 (2020), no. 10, 8212–8268. 10.1016/j.jde.2020.06.013Suche in Google Scholar

[11] W. Charles, J. A. Soriano, F. A. Falcao Nascimento, and J. H. Rodrigues, Decay rates for Bresse system with arbitrary nonlinear localized damping, J. Differ. Equ. 255 (2013), no. 8, 2267–2290. 10.1016/j.jde.2013.06.014Suche in Google Scholar

[12] Y. X. Chen and R. Z. Xu, Global well-posedness of solutions for fourth order dispersive wave equation with nonlinear weak damping, linear strong damping and logarithmic nonlinearity, Nonlinear Anal. 192 (2020), Paper no. 111664, 39 pp. 10.1016/j.na.2019.111664Suche in Google Scholar

[13] I. Chueshov, M. Eller, and I. Lasiecka, On the attractor for a semilinear wave equation with critical exponent and nonlinear boundary dissipation, Commun. Part. Differ. Equ. 27 (2002), 1901–1951. 10.1081/PDE-120016132Suche in Google Scholar

[14] I. Chueshov, Dynamics of Quasi-Stable Dissipative Systems, Springer International Publishing, Switzerland, 2015. 10.1007/978-3-319-22903-4Suche in Google Scholar

[15] I. Chueshov and I. Lasiecka, Von Karman Evolution Equations: Well-Posedness and Long Time Dynamics, Springer Monographs in Mathematics, Springer, New York, 2010. 10.1007/978-0-387-87712-9Suche in Google Scholar

[16] M. Daoulatli, I. Lasiecka, and D. Toundykov, Uniform energy decay for a wave equation with partially supported nonlinear boundary dissipation without growth restrictions, Discrete Contin. Dyn. Syst. Ser. S 2 (2009), no. 1, 67–94. 10.3934/dcdss.2009.2.67Suche in Google Scholar

[17] M. Daoulatli, Rate of decay of solutions of the wave equation with arbitrary localized nonlinear damping, Nonlinear Anal. 73 (2010), 987–1003. 10.1016/j.na.2010.04.026Suche in Google Scholar

[18] D. Dos Santos Ferreira, Sharp Lp carleman estimates and unique continuation, Journées équations aux Dérivés Partielles, vol. VI, 2003, p. 12. Suche in Google Scholar

[19] M. J. Dos Santos, B. Feng, D. S. Almeida, and M. L. Santos, Global and exponential attractors for a nonlinear porous elastic system with delay term, Discrete Contin. Dyn. Syst. Ser. B 22 (2020), no. 11, 1–24. Suche in Google Scholar

[20] J. C. O. Faria and A. Y. Souza Franco, Well-posedness and exponential stability for a Klein-Gordon system with locally distributed viscoelastic dampings in a past-history framework, J. Differ. Equ. 346 (2023), 108–144. 10.1016/j.jde.2022.11.022Suche in Google Scholar

[21] L. H. Fatori, M. A. Jorge Silva, and V. Narciso, Quasi-stability property and attractors for a semilinear Timoshenko system, Discrete Contin. Dyn. Syst. 36 (2016), no. 11, 6117–6132. 10.3934/dcds.2016067Suche in Google Scholar

[22] E. Zuazua, Exponential decay for the semilinear wave equation with locally distributed damping, Commun. Partial Differ. Equ. 15 (1990), 205–235. 10.1080/03605309908820684Suche in Google Scholar

[23] E. Feireisl and E. Zuazua, Global attractors for semilinear wave equations with locally distributed nonlinear damping and critical exponent, Commun. Partial Differ. Equ. 18 (1993), 1539–1555. 10.1080/03605309308820985Suche in Google Scholar

[24] J. S. Ferreira, Exponential decay for a nonlinear system of hyperbolic equations with locally distributed dampings, Nonlinear Anal. 18 (1992), no. 11, 1015–1032. 10.1016/0362-546X(92)90193-ISuche in Google Scholar

[25] M. M. Freitas, R. Q. Caljaro, and A. J. A. Ramos, Long-time dynamics of ternary mixtures with localized dissipation, J. Math. Phys. 63 (2022), 121508. 10.1063/5.0098498Suche in Google Scholar

[26] M. M. Freitas, R. Q. Caljaro, M. L. Santos, and A. J. A. Ramos, Singular limit dynamics and attractors for wave equations connected in parallel, Appl. Math. Optim. 85 (2022), 9. 10.1007/s00245-022-09849-0Suche in Google Scholar

[27] M. M. Freitas, M. L. Santos, and L. A. Langa, Porous elastic system with nonlinear damping and sources terms, J. Differ. Equ. 264 (2018), 2970–3051. 10.1016/j.jde.2017.11.006Suche in Google Scholar

[28] P. Gérard, Microlocal defect measures, Commun. Partial Differ. Equ. 16 (1991), 1761–1794. 10.1080/03605309108820822Suche in Google Scholar

[29] Y. Q. Guo and M. A. Rammaha, Global existence and decay of energy to systems of wave equations with damping and supercritical sources, Z. Angew. Math. Phys. 64 (2013), 621–658. 10.1007/s00033-012-0252-6Suche in Google Scholar

[30] Y. Q. Guo and M. A. Rammaha, Systems of nonlinear wave equations with damping and supercritical boundary and interior sources, Trans. Amer. Math. Soc. 366 (2014), 2265–2325. 10.1090/S0002-9947-2014-05772-3Suche in Google Scholar

[31] 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), no. 2, 214–244. 10.3934/cam.2023012Suche in Google Scholar

[32] L. F. Ho, Exact controllability of the one dimensional wave equation with locally distributed control, SIAM J. Control Optim. 28 (1990), no. 3, 733–748. 10.1137/0328043Suche in Google Scholar

[33] H. Koch and D. Tataru, Dispersive estimates for principally normal pseudodifferential operators, Commun. Pure Appl. Math. 58 (2005), no. 2, 217–284. 10.1002/cpa.20067Suche in Google Scholar

[34] I. Lasiecka and D. Tataru, Uniform boundary stabilization of semi-linear wave equation with nonlinear boundary dissipation, Differ. Integral Equ. 6 (1993), 507–533. 10.57262/die/1370378427Suche in Google Scholar

[35] I. Lasiecka and D. Toundykov, Energy decay rates for the semilinear wave equation with nonlinear localized damping and source terms, Nonlinear Anal. 64 (2006), no. 8, 1757–1797. 10.1016/j.na.2005.07.024Suche in Google Scholar

[36] Y. C. Li and C. M. Wang, On a partially synchronizable system for a coupled system of wave equations in one dimension, Commun. Anal. Mech. 15 (2023), no. 3, 470–493. 10.3934/cam.2023023Suche in Google Scholar

[37] W. Lian and R. Z. Xu, Global well-posedness of nonlinear wave equation with weak and strong damping terms and logarithmic source term, Adv. Nonlinear Anal. 9 (2020), no. 1, 613–632. 10.1515/anona-2020-0016Suche in Google Scholar

[38] T. F. Ma, R. N. Monteiro, and P. N. Seminario-Huertas, Attractors for locally damped Bresse systems and a unique continuation property, 2021, arXiv:2102.12025. Suche in Google Scholar

[39] M. Najafi and H. Wang, Exponential stability of wave equations coupled in parallel with viscous damping, In: Proceedings of 32nd IEEE Conference on Decision and Control, December 1993, pp. 15–17. Suche in Google Scholar

[40] M. Najafi, G. R. Sarhangi, and H. Wang, Stabilizability of coupled wave equations in parallel under various boundary conditions, IEEE Trans. Autom. Control 42 (1997), no. 9, 1308–1312. 10.1109/9.623099Suche in Google Scholar

[41] M. Najafi, Energy decay estimate for the boundary stabilization of the coupled wave equations, Math. Comput. Model. 48 (2008), 1796–1805. 10.1016/j.mcm.2008.06.002Suche in Google Scholar

[42] R. Rajaram and M. Najafi, Exact controllability of wave equations in Rn coupled in parallel, J. Math. Anal. Appl. 356 (2009), 7–12. 10.1016/j.jmaa.2009.02.034Suche in Google Scholar

[43] R. E. Showalter, Monotone Operators in Banach Spaces and Nonlinear Partial Differential Equations, vol. 49 of Mathematical Surveys and Monographs, American Mathematical Society, Providence, RI, 1997. Suche in Google Scholar

[44] D. H. Sattinger, On global solution of nonlinear hyperbolic equations, Arch. Ration. Mech. Anal. 30 (1968), 148–172. 10.1007/BF00250942Suche in Google Scholar

[45] J. Simon, Compact sets in the space Lp(0,T;B), Ann. Mat. Pura Appl. 146 (1987), no. 4, 65–96. 10.1007/BF01762360Suche in Google Scholar

[46] J. T. Webster, Attractors and determining functionals for a flutter model: finite dimensionality out of thin air, Pure Appl. Funct. Anal. 5 (2020), 85–119. Suche in Google Scholar

[47] X. C. Wang, Y. X. Chen, Y. B. Yang, J. H. Li, and R. Z. Xu, Kirchhoff-type system with linear weak damping and logarithmic nonlinearities, Nonlinear Anal. 188 (2019), 475–499. 10.1016/j.na.2019.06.019Suche in Google Scholar

[48] J. Q. Wu, S. J. Li, and S. G. Chai, Exact controllability of wave equations with variable coefficients coupled in parallel, Asian J. Control 12 (2010), no. 5, 650–655. 10.1002/asjc.179Suche in Google Scholar

[49] C. Yang, V. D. Rădulescu, 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), no. 1, 436–468. 10.1515/ans-2022-0024Suche in Google Scholar

[50] X. Y. Yang, S. B. Wang, and M. A. J. Silva, Lamé system with weak damping and nonlinear time-varying delay, Adv. Nonlinear Anal. 12 (2023), no. 1, 22. 10.1515/anona-2023-0115Suche in Google Scholar

Received: 2023-11-05
Revised: 2024-01-25
Accepted: 2024-04-04
Published Online: 2024-06-19

© 2024 the author(s), published by De Gruyter

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

Artikel in diesem Heft

  1. Regular Articles
  2. Existence of normalized peak solutions for a coupled nonlinear Schrödinger system
  3. Orbital stability of the trains of peaked solitary waves for the modified Camassa-Holm-Novikov equation
  4. Global boundedness in a two-dimensional chemotaxis system with nonlinear diffusion and singular sensitivity
  5. Existence and uniqueness of solution for a singular elliptic differential equation
  6. The bounded variation capacity and Sobolev-type inequalities on Dirichlet spaces
  7. Vanishing and blow-up solutions to a class of nonlinear complex differential equations near the singular point
  8. Existence, uniqueness, localization and minimization property of positive solutions for non-local problems involving discontinuous Kirchhoff functions
  9. Concentration phenomena for a fractional relativistic Schrödinger equation with critical growth
  10. Beyond the classical strong maximum principle: Sign-changing forcing term and flat solutions
  11. Monotonicity of solutions for parabolic equations involving nonlocal Monge-Ampère operator
  12. On a nonlinear Robin problem with an absorption term on the boundary and L1 data
  13. Multiple solutions for the quasilinear Choquard equation with Berestycki-Lions-type nonlinearities
  14. Gradient estimates for a class of higher-order elliptic equations of p-growth over a nonsmooth domain
  15. Global existence and decay estimates of the classical solution to the compressible Navier-Stokes-Smoluchowski equations in ℝ3
  16. Global existence and finite-time blowup for a mixed pseudo-parabolic r(x)-Laplacian equation
  17. Multiple positive solutions for a class of concave-convex Schrödinger-Poisson-Slater equations with critical exponent
  18. A minimization problem with free boundary for p-Laplacian weakly coupled system
  19. Multiplicity of semiclassical solutions for a class of nonlinear Hamiltonian elliptic system
  20. k-convex solutions for multiparameter Dirichlet systems with k-Hessian operator and Lane-Emden type nonlinearities
  21. Infinitely many solutions for Hamiltonian system with critical growth
  22. On optimal control in a nonlinear interface problem described by hemivariational inequalities
  23. Variational–hemivariational system for contaminant convection–reaction–diffusion model of recovered fracturing fluid
  24. Potential and monotone homeomorphisms in Banach spaces
  25. Critical fractional Schrödinger-Poisson systems with lower perturbations: the existence and concentration behavior of ground state solutions
  26. Supercritical Hénon-type equation with a forcing term
  27. Concentration of blow-up solutions for the Gross-Pitaveskii equation
  28. Mountain-pass-type solutions for Schrödinger equations in R2 with unbounded or vanishing potentials and critical exponential growth nonlinearities
  29. Regularity for critical fractional Choquard equation with singular potential and its applications
  30. Double phase anisotropic variational problems involving critical growth
  31. Normalized solutions of NLS equations with mixed nonlocal nonlinearities
  32. Nontrivial solutions for resonance quasilinear elliptic systems
  33. Standing waves for Choquard equation with noncritical rotation
  34. Low regularity conservation laws for Fokas-Lenells equation and Camassa-Holm equation
  35. Modified quasilinear equations with strongly singular and critical exponential nonlinearity
  36. Limit profiles and the existence of bound-states in exterior domains for fractional Choquard equations with critical exponent
  37. Nests of limit cycles in quadratic systems
  38. Regularity of minimizers for double phase functionals of borderline case with variable exponents
  39. Boundedness, existence and uniqueness results for coupled gradient dependent elliptic systems with nonlinear boundary condition
  40. Ground state solutions for magnetic Schrödinger equations with polynomial growth
  41. Nonoccurrence of Lavrentiev gap for a class of functionals with nonstandard growth
  42. Dynamics for wave equations connected in parallel with nonlinear localized damping
  43. Boussinesq's equation for water waves: Asymptotics in Sector I
  44. Optimal global second-order regularity and improved integrability for parabolic equations with variable growth
  45. Normalized solutions of Schrödinger equations involving Moser-Trudinger critical growth
  46. Normalized solutions for the double-phase problem with nonlocal reaction
  47. Normalized solutions for Sobolev critical fractional Schrödinger equation
  48. Well-posedness of Cauchy problem of fractional drift diffusion system in non-critical spaces with power-law nonlinearity
  49. Improved results on planar Klein-Gordon-Maxwell system with critical exponential growth
  50. Existence and multiplicity of solutions for a new p(x)-Kirchhoff equation
  51. Quasiconvex bulk and surface energies: C1,α regularity
  52. Time decay estimates of solutions to a two-phase flow model in the whole space
  53. Bounds for the sum of the first k-eigenvalues of Dirichlet problem with logarithmic order of Klein-Gordon operators
  54. The properties of a new fractional g-Laplacian Monge-Ampère operator and its applications
  55. A fractional profile decomposition and its application to Kirchhoff-type fractional problems with prescribed mass
  56. Cauchy problem for a non-Newtonian filtration equation with slowly decaying volumetric moisture content
  57. Multiplicity of normalized semi-classical states for a class of nonlinear Choquard equations
  58. The second Yamabe invariant with boundary
  59. Peaked solitary waves and shock waves of the Degasperis-Procesi-Kadomtsev-Petviashvili equation
  60. Normalized solutions for the Choquard equations with critical nonlinearities
  61. Boundedness and long-time behavior in a parabolic-elliptic system arising from biological transport networks
  62. Initial boundary value problem and exponential stability for the planar magnetohydrodynamics equations with temperature-dependent viscosity
  63. Nonexistence of mean curvature flow solitons with polynomial volume growth immersed in certain semi-Riemannian warped products
  64. Analysis of a vector-borne disease model with vector-bias mechanism in advective heterogeneous environment
  65. Normalized solutions for the Kirchhoff equation with combined nonlinearities in ℝ4
  66. Nonexistence of the compressible Euler equations with space-dependent damping in high dimensions
  67. Multiplicity of solutions for a nonhomogeneous quasilinear elliptic equation with concave-convex nonlinearities
  68. On semilinear inequalities involving the Dunkl Laplacian and an inverse-square potential outside a ball
  69. Besov regularity for the elliptic p-harmonic equations in the non-quadratic case
  70. Uniqueness and nondegeneracy of ground states for Δ u + u = ( I α u 2 ) u in R 3 when α is close to 2
  71. Existence of solutions for a class of quasilinear Schrödinger equations with Choquard-type nonlinearity
  72. Weighted Hardy-Adams inequality on unit ball of any even dimension
  73. Existence and regularity for a p-Laplacian problem in ℝN with singular, convective, and critical reaction
  74. Temporal periodic solutions of non-isentropic compressible Euler equations with geometric effects
  75. Nodal solutions for a zero-mass Chern-Simons-Schrödinger equation
  76. The modified quasi-boundary-value method for an ill-posed generalized elliptic problem
  77. Nonlocal heat equations with generalized fractional Laplacian
  78. Choquard equations with recurrent potentials
  79. Local and global well-posedness of the Maxwell-Bloch system of equations with inhomogeneous broadening
  80. The Riemann problem for two-layer shallow water equations with bottom topography
  81. Global stability and asymptotic profiles of a partially degenerate reaction diffusion Cholera model with asymptomatic individuals
  82. On Kirchhoff-Schrödinger-Poisson-type systems with singular and critical nonlinearity
  83. Energy-variational solutions for viscoelastic fluid models
  84. Stability on 3D Boussinesq system with mixed partial dissipation
  85. Existence and multiplicity results for non-autonomous second-order Hamiltonian systems
  86. Erratum
  87. Erratum to “Normalized solutions for the Kirchhoff equation with combined nonlinearities in ℝ4
Heruntergeladen am 8.9.2025 von https://www.degruyterbrill.com/document/doi/10.1515/anona-2024-0015/html
Button zum nach oben scrollen