Skip to Content
MathematicsMathematics
  • Article
  • Open Access

2 July 2026

26 Pages

Explosive for a Damped p-Laplacian-Type Stochastic Wave Equation with Logarithmic Nonlinearity

,
and
Department of Mathematics, Xi’an University of Science and Technology, Xi’an 710054, China
*
Author to whom correspondence should be addressed.

Abstract

In this paper, we consider a damped p-Laplacian-type wave equation with logarithmic nonlinearity driven by multiplicative noises. We first establish the local existence and uniqueness of a mild solution to the equation using the truncation technique and semigroup method and show that the local solution is global under certain conditions. Secondly, we show the blow-up properties of solutions using an appropriate energy inequality. Moreover, we also derive estimates of the upper bound of the blow-up time.

1. Introduction

In this paper, we consider the following stochastic p-Laplacian wave equation with logarithmic nonlinearity:
u t t − Δ p u + μ u t = | u | q − 2 u ln | u | + σ ( u , u t , ∇ u , x , t ) ∂ t W ( t , x ) x ∈ D , t > 0 , u ( x , t ) = 0 , x ∈ ∂ D , t > 0 , u ( x , 0 ) = u 0 ( x ) , u t ( x , 0 ) = v 0 ( x ) , x ∈ D ,
where D ∈ R n ( n ≥ 1 ) is a bounded domain with smooth boundary ∂ D , Δ p u = div ( | ∇ u | p − 2 ∇ u ) with q ≥ p > 2 . μ is a given positive constant that measures the strength of the damping. σ is a given function satisfying certain conditions to be specified later. W ( t , x ) : t ≥ 0 is a H-valued Q-Wiener process on the probability space with the variance Q satisfying operator T r Q < ∞ . Problem (1) arises naturally in the study of nonlinear wave propagation in random media, where the stochastic perturbation accounts for external random forces or fluctuating environmental parameters. The presence of the p-Laplacian operator Δ p captures the nonlinear diffusion effects commonly encountered in non-Newtonian fluid dynamics, glaciology, and porous medium flows. The logarithmic nonlinearity, on the other hand, appears in various physical contexts such as quantum field theory, nonlinear optics, and plasma physics, where it describes saturation effects or confined systems with exponential-type interactions [1]. Therefore, the stochastic p-Laplacian wave equation with logarithmic nonlinearity provides a unified mathematical framework for modeling complex wave phenomena under uncertainty.
In the deterministic case, many researchers have studied the following problems. Wu et al. [2] focused on the following initial boundary value problem for a damped p-Laplacian-type wave equation with logarithmic nonlinearity:
u t t − Δ p u − Δ u t + u t = | u | q − 2 u ln | u | , x ∈ D , t > 0 , u ( x , t ) = 0 , x ∈ ∂ D , t > 0 , u ( x , 0 ) = u 0 ( x ) , u t ( x , 0 ) = u 1 ( x ) , x ∈ D ,
they first discussed the local existence of weak solutions using the contraction mapping principle. Within the potential well framework, they demonstrate global existence, energy decay, and a sufficient condition for finite-time blow-up of solutions with subcritical initial energy, utilizing the Nehari manifold and concavity arguments. They then extend the results for the subcritical case to the critical case through scaling techniques. Additionally, for supercritical initial energy, they develop new methods to establish another finite-time blow-up criterion. When q = p > 2 in (1), Yang and Han [3] studied the global existence and finite-time blow-up results for solutions with subcritical and supercritical initial energy, respectively, and the lifespan of the weak solutions is estimated from above as well. Moreover, they give a lower bound for the finite blow-up time. In the q = p = 2 case of problem (1), Lian and Xu [4] proved global existence, energy decay, and infinite-time blow-up of weak solutions for subcritical and critical energy. Furthermore, they demonstrated that in the absence of strong damping, a criterion for infinite-time blow up with arbitrarily high initial energy is derived. There has also been considerable attention paid to the initial-boundary value problem for a mixed pseudo-parabolic p-Laplacian-type equation with logarithmic nonlinearity. Interested readers are referred to [5,6,7,8,9,10,11] and the literature cited therein.
Under the consideration of random environmental effects, several researchers have investigated the following stochastic wave equation. Liang [12] studied a class of damped stochastic nonlinear beam equations driven by multiplicative noise
u t t + γ Δ 2 u − m ∥ ∇ u ∥ 2 Δ u + u t = f ( u ) + σ u , u t , ∇ u , x , t ∂ W ( x , t ) ∂ t , x ∈ D , t ∈ ( 0 , T ) , u ( x , t ) = ∂ u ∂ ν = 0 , x ∈ ∂ D , t ∈ ( 0 , T ) , u ( x , 0 ) = u 0 ( x ) , u t ( x , 0 ) = v 0 ( x ) , x ∈ D .
By establishing a suitable energy inequality, the author proved that the local solutions to these stochastic equations blow up with positive probability or explosive in L 2 sense under some conditions. Moreover, the author give a lower bound for the finite blow-up time. Cheng et al. [13] investigated a nonlinear stochastic viscoelastic wave equation with nonlinear damping and source terms. They established the existence of a global solution and blow-up solutions and proved that finite-time blow-up with non-negative probability is explosive or it is explosive in an energy sense for p > q . Benramdane et al. [14] considered an initial boundary value problem of stochastic viscoelastic wave equation with nonlinear damping and logarithmic nonlinear source terms, and they proved a blow-up result for the solution with decreasing kernel. Bo et al. [15] focused on the following stochastic partial differential equation:
u t t ( t , x ) − Δ u ( t , x ) + A u t ( t , x ) = μ | u ( t , x ) | ρ u ( t , x ) + g u ( t , x ) , u t ( t , x ) , D u ( t , x ) ∂ ∂ t W ( t , x ) , u ( 0 , x ) = u 0 ( x ) , u t ( 0 , x ) = v 0 ( x ) , ( t , x ) ∈ ( 0 , ∞ ) × R d .
Using an appropriate energy inequality, they give sufficient conditions such that the local solutions of the stochastic equations are blow up with positive probability or explosive in L 2 sense. In [16], Chow examined a class of non-dissipative stochastic wave equations with polynomial nonlinearities and proved the global existence of solutions using energy inequality. Chow [17] studied the problem of explosive solutions for a class of nonlinear stochastic wave equation in a domain D ⊂ R d for d ≥ 3 . Brzézniak et al. [18] studied the global existence and stability of solutions for stochastic beam equations with white noise.
To the best of our knowledge, the stochastic p-Laplacian wave Equation (1) with logarithmic nonlinearity has not been previously investigated in the literature. Unlike the models in (3) and (4), Equation (1) generalizes the classical second-order linear Laplace operator Δ u or the fourth-order biharmonic operator Δ 2 u to the nonlinear p-Laplacian operator Δ p u . In this paper, we consider global existence and blow-up phenomena under appropriate conditions on the noise term σ ( u , u t , ∇ u , x , t ) and the nonlinear source terms.
This paper is organized as follows. In Section 2, we introduce some preliminaries related to assumptions and definitions that will be used throughout the paper. In Section 3, we provide detailed proofs of the local existence and uniqueness of a mild solution. In Section 4, we establish the blow up of the solution to our problem.

2. Preliminaries

In this section, we introduce some notations and lemmas that will be used throughout the paper. In what follows, we denote the norm in L r ( D ) as · r r ≥ 1 and the H = L 2 ( D ) inner product as ( · , · ) . When p > 1 , p ≠ 2 , we use W 0 1 , p ( D ) to denote the Sobolev space such that both u and | ▿ u | belong to L p ( D ) for any u ∈ W 0 1 , p ( D ) . We will equip W 0 1 , p ( D ) with the norm Z = ∥ u ∥ W 0 1 , p = ∥ ∇ u ∥ p , which is equivalent to the full one due to Poincaré’s inequality.
Let Ω , P , F be a complete probability space for which a { F t , t ≥ 0 } of sub- σ -fields of F is given. A point of Ω will be denoted by ω and E ( · ) stands for expectation with respect to probability measure P. Suppose that { W ( t , x ) : t ≥ 0 } is a H-valued Q-Wiener process on the probability space with the variance operator Q satisfying T r Q < ∞ . Correspondingly, we can assume that Q has the following form Q e i = ξ i e i , i = 1 , 2 , … , where ξ i are eigenvalues of Q satisfying ∑ i = 1 ∞ ξ i < ∞ and { e i } forms a complete orthonormal system of eigenfunctions with c 0 : = sup i ≥ 1 ∥ e i ∥ ∞ < ∞ (Where · ∞ denotes the super-norm). Moreover, we have E W ( x , t ) = 0 and a covariance function Q ( φ , ψ ) such that it satisfies
E [ ( W ( x ) , φ ) ∧ ( W ( y ) , ψ ) ] = ( x ∧ y ) ( Q φ , ψ )
for any φ , ψ ∈ L 2 ( D ) . In this case, W ( t , x ) = ∑ i = 1 ∞ ξ i B i ( t ) e i , where { B i ( t ) } is a sequence of standard one-dimensional Brownian motions. In this paper, we assume that
∑ j = 1 ∞ λ j ξ j < ∞ ,
where λ j are eigenvalues of − Δ e j = λ j e j , e j = 0 on ∂ D .
Meanwhile, assume that either p, q are satisfied
2 < p ≤ q ≤ n p 2 ( n − p ) + 1 .
Let σ : R × R × R n × R n × R + → R denotes a continuous function. Suppose that for any s , r , s ′ , r ′ ∈ R , t , t ′ ∈ R + and x , v , x ′ , v ′ ∈ R n ,
| σ ( s , r , v , x , t ) | 2 ≤ c 1 ( 1 + | s | 2 ( q − 1 ) + | r | 2 + | v | 2 )
and
| σ ( s , r , v , x , t ) − σ ( s ′ , r ′ , v ′ , x ′ , t ′ ) | 2 ≤ c 2 ( ( 1 + | s | 2 ( q − 2 ) + | s ′ | 2 ( q − 2 ) ) | s − s ′ | 2 + | r − r ′ | 2 + | v − v ′ | 2 ) ) ,
where c 1 and c 2 are positive constants.
Now, we introduce the following functionals
J ( u ) = 1 p ∥ ∇ u ∥ p p − 1 q ∫ D | u | q ln | u | d x + 1 q 2 ∥ u ∥ q q ,
I ( u ) = ∥ ∇ u ∥ p p − ∫ D | u | q ln | u | d x .
We define the energy function ε ( t ) associated to our system
ε ( t ) = 1 2 ∥ u t ∥ 2 2 + 1 p ∥ ∇ u ∥ p p − 1 q ∫ D | u | q ln | u | d x + 1 q 2 ∥ u ∥ q q .
Through a direct computation
J ( u ) = 1 p I ( u ) + ( ( 1 p − 1 q ) ) ∥ ∇ u ∥ p p + 1 q 2 ∥ u ∥ q q ,
and
ε ( t ) = 1 2 ∥ u t ∥ 2 2 + J ( u ) .
In the stochastic setting, the energy functional J ( u ) and the Nehari functional I ( u ) depend on the random terms through the solution u ( x , t ) , which is an F t -adapted stochastic process. Consequently, the classical deterministic Nehari manifold, defined pointwise by I ( u ) = 0 , is no longer directly applicable, as the equality depends on the realization of the noise. To overcome this difficulty, we adopt the expectation of I ( u ) to define the stochastic Nehari manifold in (12). This approach allows us to characterize the sign of the fibering map λ ↦ E J ( λ u ) in the mean sense, which is sufficient for applying the potential well method in the stochastic framework. We emphasize that the manifold N defined in (12) is time-independent, consistent with its deterministic counterpart. Define the Nehari manifold
N = u ∈ W 0 1 , p ( D ) ∣ E I ( u ) = 0 ,
and the depth of the well
d = inf u ∈ N E J ( u ) .
In addition, we set
N − = u ∈ W 0 1 , p ( D ) ∣ E I ( u ) < 0 ,
N + = u ∈ W 0 1 , p ( D ) ∣ E I ( u ) > 0 ,
where N − is an unstable set, and N + is a stable set.
Lemma 1 
([19]). Assume u ∈ W 0 1 , p ( D ) , then
(1) 
lim λ → 0 + J ( λ u ) = 0 , lim λ → + ∞ J ( λ u ) = − ∞ ;
(2) 
There exists a unique λ * = λ * ( u ) such that d d λ J ( λ u ) | λ = λ * = 0 ;
(3) 
J ( λ u ) is strictly increasing on 0 < λ < λ * , strictly decreasing on λ * < λ < + ∞ , and achieves the maximum value at λ = λ * ;
(4) 
I ( λ u ) > 0 for 0 < λ < λ * , I ( λ u ) < 0 for λ * < λ < + ∞ , and I ( λ * u ) = 0 .
Lemma 2 
([20,21,22]). Suppose that a positive, twice-differentiable function ψ ( t ) satisfies the inequality
ψ ″ ( t ) ψ ( t ) − ( 1 + θ ) ( ψ ′ ( t ) ) 2 ≥ 0 ,
where 0 > 0 . If ψ ( 0 ) > 0 and ψ ′ ( 0 ) > 0 , then ψ ( t ) → ∞ as t → t * ≤ t * = ψ ( 0 ) θ ψ ′ ( 0 ) .
Lemma 3 
([2,7]). Assume that s < n p n − p , i.e., s < ∞ for n ≤ p and r < s < n p n − p for n > p and r ≥ 1 . Then, for any function u ∈ W 0 1 , p ( D ) , we have the inequality
u s ≤ C G ∇ u p α u r 1 − α ,
where α ∈ ( 0 , 1 ) is determined by α = 1 r − 1 s 1 n − 1 p + 1 r − 1 and C G is a positive constant depends on n , p , s and r.

3. Local Existence and Uniqueness

In this section, we address local existence and uniqueness of the solution for problem (1). Let X = Z × H be a space with norm ∥ U ∥ X = ∥ ∇ u ∥ p 2 + ∥ v ∥ 2 2 1 2 for any U = u , v T ∈ X . It is well known that Equation (1) is equivalent to the following Ito system
d u = v d t , d v = ( Δ p u − μ v + | u | q − 2 u ln | u | ) d t + σ ( u , v , ∇ u , x , t ) d W ( x , t ) , u ( x , t ) = 0 , u ( x , 0 ) = u 0 ( x ) , v ( x , 0 ) = v 0 ( x ) .
Definition 1. 
The process ( u , v ) is said to be a solution on [ 0 , T ] to (1) if u and v are progressively measurable,
u ∈ L 2 ( Ω ; C ( [ 0 , T ] ; W 0 1 , p ( D ) ) ) , v ∈ L 2 ( Ω ; C ( [ 0 , T ] ; L 2 ( D ) ) ) ,
and the following is satisfied for almost all ω ∈ Ω : for any ϕ ∈ W 0 1 , p ( D )
( u ( t ) , ϕ ) = ( u ( 0 ) , ϕ ) + ∫ 0 t ( v ( s ) , ϕ ) d s , ( v ( t ) , ϕ ) = ( v ( 0 ) , ϕ ) + ∫ 0 t ( Δ p u ( s ) , ϕ ) d s − μ ∫ 0 t ( v ( s ) , ϕ ) d s + ∫ 0 t ( f ( u ( s ) ) , ϕ ) d s + ∫ 0 t ( σ ( s , u ( s ) , v ( s ) ) d W ( s ) , ϕ ) , u ( 0 ) = u 0 , v ( 0 ) = v 0 , u ( t ) | ∂ D = 0 .
Lemma 4 
([23]). For any given ε > 0 , there exists A > 0 , such that the real function
j ( s ) = | s | q − 2 log | s | , q > 2
satisfies
| j ( s ) | ≤ A + | s | q − 2 + ε .
Condition 1. 
The functions f and σ satisfy the Lipschitz condition and the growth condition:
| f ( t , z 1 ) − f ( t , z 2 ) | 2 ≤ L f | z 1 − z 2 | 2 ,
| σ ( t , z 1 , w 1 ) − σ ( t , z 2 , w 2 ) | 2 ≤ L σ ( | z 1 − z 2 | 2 + | w 1 − w 2 | 2 ) ,
| f ( t , z 1 ) | 2 ≤ d f + L f | z 1 | 2 ,
| σ ( t , z 1 , w 1 ) | 2 ≤ d σ + L σ ( | z 1 | 2 + | w 1 | 2 ) ,
for z 1 , z 2 ∈ W 0 1 , p , w 1 , w 2 ∈ L 2 ( D ) , where d f , L f , d σ and L σ are positive constants.
Assume that σ satisfies
E ∫ 0 t | ∇ σ ( s , ζ ( s ) , η ( s ) ) | 2 d s < ∞ .
Let ζ ( t ) and η ( t ) be progressively measurable processes. Under Condition 1 we consider the following stochastic wave equation with an additive noise:
d u ( t ) = v ( t ) d t d v ( t ) = Δ p u ( t ) d t − μ v ( t ) d t + f ( ζ ( t ) ) d t + σ ( t , ζ ( t ) , η ( t ) ) d W ( t ) , 0 ≤ t ≤ T u ( 0 ) = u 0 , v ( 0 ) = v 0 .
Lemma 5. 
The processes u 0 and v 0 are F 0 -measurable with u 0 ∈ L 2 ( Ω ; W 0 1 , p ( D ) ∩ W 0 2 , p ( D ) ) , v 0 ∈ L 2 ( Ω ; H 0 1 ( D ) ) . Assume that ( 5 ) , Condition 1 and ( 23 ) are satisfied. Then, there exists a unique solution ( u ( t ) , v ( t ) ) to (24) on [ 0 , T ] with the initial value ( u 0 , v 0 ) such that
( u , v ) ∈ L 2 ( Ω ; C [ 0 , T ] ; W 0 1 , p ( D ) ∩ W 0 2 , p ( D ) ) × L 2 ( Ω ; C [ 0 , T ] ; H 0 1 ( D ) ) .
Proof. 
Let { e j } j = 1 ∞ be an orthogonal basis in W 0 1 , p ( Ω ) . We define finite dimensional subspace V m = span { e 1 , e 2 , … , e m } . For each m ≥ 1 , define
u m ( t ) = ∑ j = 1 m a m , j ( t ) e j , v m ( t ) = ∑ j = 1 m a m , j ′ ( t ) e j ,
satisfy the follow finite-dimensional stochastic system
d u m ( t ) = v m ( t ) d t , d v m ( t ) = Δ p u m ( t ) d t − μ v m ( t ) d t + f ( ζ ( t ) ) d t + σ ( t ) d W ( t ) , u m ( 0 ) = u 0 , v m ( 0 ) = v 0 .
Then, a m , j form a solution of the following system of stochastic differential equations
a m , j ″ ( t ) = − ∑ i = 1 m a m , i ( s ) Δ p e i , e j − μ a m , j ′ ( s ) + f ( ζ ( t ) ) , e j + σ ( t , ζ ( t ) , η ( t ) ) W ( t ) , e j , 0 < j ≤ m ,
with initial conditions a m , j ( 0 ) = ( u 0 , e j ) and a m , j ′ ( 0 ) = ( v 0 , e j ) .
By the Ito formula, we have
∥ v m ( t ) ∥ 2 2 + 2 p ∥ ∇ u ∥ p p + 2 μ ∫ 0 t ∥ v m ( s ) ∥ 2 2 d s = ∥ v m ( 0 ) ∥ 2 2 + 2 p ∥ ∇ u m ( 0 ) ∥ p p + 2 ∫ 0 t f ( ζ ( s ) ) , v m ( s ) d s + 2 ∫ 0 t σ ( s , ζ ( s ) , η ( s ) ) d W ( s ) , v m ( s ) + c 0 2 Tr ( Q ) ∫ 0 t ∥ σ ( s , ζ ( s ) , η ( s ) ) ∥ 2 2 d s .
From (25), we have
∇ v m ( t ) = ∇ Δ p u m ( t ) d t − μ ∇ v m ( t ) d t + ∇ f m ( t ) d t + ∇ σ m ( t ) d W ( t ) .
Thus, we get
∥ ∇ v m ( t ) ∥ 2 2 + 2 p ∥ Δ u m ( t ) ∥ p p + 2 μ ∫ 0 t ∥ ∇ v m ( s ) ∥ 2 2 d s = ∥ ∇ v m ( 0 ) ∥ 2 2 + 2 p ∥ Δ u m ( 0 ) ∥ p p + 2 ∫ 0 t ∇ f ( ζ ( s ) ) , ∇ v m ( s ) d s + 2 ∫ 0 t ∇ σ ( s , ζ ( s ) , η ( s ) ) d W ( s ) , ∇ v m ( s ) + c 0 2 Tr ( Q ) ∫ 0 t ∥ ∇ σ ( s , ζ ( s ) , η ( s ) ) ∥ 2 2 d s .
Since
∫ D ( f ( ζ ( s ) ) , v m ( s ) ) d x ≤ μ 2 ∥ f ( ζ ) ∥ 2 ∥ v m ∥ 2 .
By the Burkholder–Davis–Gundy inequality, we get
E sup t ∈ [ 0 , T ] | ∫ 0 t ( σ d W s , v m ( s ) ) | ≤ C E sup t ∈ [ 0 , T ] ∥ v m ∥ 2 ∑ i = 1 ∞ ∫ 0 T ( σ ( x , t ) Tr ( Q ) e i , σ ( x , t ) e i ) d t 1 2 ≤ μ 2 E sup t ∈ [ 0 , T ] ∥ v m ∥ 2 2 + μ C c 0 2 2 Tr ( Q ) E ∫ 0 T ∥ σ ( t ) ∥ 2 2 d t ,
and
2 E sup t ∈ [ 0 , T ] | ∫ 0 t ∇ σ ( s , ζ ( s ) , η ( s ) ) d W ( s ) , ∇ v m ( s ) | ≤ 2 E sup t ∈ [ 0 , T ] | ∫ 0 t ( ∇ σ ( s , ζ ( s ) , η ( s ) ) ∑ k = 1 ∞ ξ k 1 2 d β k ( s ) e k , ∇ v m ( s ) ) | + 2 E sup t ∈ [ 0 , T ] | ∫ 0 t ( σ ( s , ζ ( s ) , η ( s ) ) ∑ k = 1 ∞ λ k 1 2 ξ k 1 2 d β k ( s ) e k , ∇ v m ( s ) ) | ≤ 1 4 E sup t ∈ [ 0 , T ] ∥ ∇ v m ( t ) ∥ 2 2 + k c 0 2 Tr ( Q ) E ∫ 0 t ∥ ∇ σ ( s , ζ ( s ) , η ( s ) ) ∥ 2 2 d s + k c 0 2 ∑ k = 1 ∞ λ k ξ k E ∫ 0 t ∥ σ ( s , ζ ( s ) , η ( s ) ) ∥ 2 2 d s .
Combining (30)–(32), we have
E sup t ∈ [ 0 , T ] ∥ ∇ v m ∥ 2 2 + sup t ∈ [ 0 , T ] ∥ u m ∥ W 0 1 , p ( D ) ∩ W 0 2 , p ( D ) + μ ∫ 0 T ∥ ∇ v m ∥ 2 d s ≤ C T .
Thus, the proof of the lemma by the same method as in Lemmas 6 and 7 [24] is carried out. So, we omit the proof. □
Next, we consider the stochastic equation
d u ( t ) = v ( t ) d t d v ( t ) = Δ p u ( t ) d t − μ v ( t ) d t + f ( u ( t ) ) d t + σ ( t , u ( t ) , v ( t ) ) d W ( t ) , 0 ≤ t ≤ T u ( 0 ) = u 0 , v ( 0 ) = v 0 .
Lemma 6. 
The processes u 0 and v 0 are F 0 -measurable with u 0 ∈ L 2 ( Ω ; W 0 1 , p ( D ) ∩ W 0 2 , p ( D ) ) , v 0 ∈ L 2 ( Ω ; H 0 1 ( D ) ) . Let ( 5 ) , Condition 1 and ( 23 ) hold. Then, there exists a unique solution ( u ( t ) , v ( t ) ) to (34) satisfying
u ∈ L 2 ( Ω ; C [ 0 , T ] ; W 0 1 , p ( D ) ∩ W 0 2 , p ( D ) ) , v ∈ L 2 ( Ω ; C [ 0 , T ] ; H 0 1 ( D ) ) .
Proof. 
Let u 0 ( t ) = u 0 and v 0 ( t ) = v 0 . If ( u n ( t ) , v n ( t ) ) is progressively measurable, n = 1 , 2 , 3 , ⋯ , then by using Lemma 5, we have the unique solution ( u n + 1 ( t ) , v n + 1 ( t ) ) to following stochastic equation:
d u n + 1 ( t ) = v n + 1 ( t ) d t , d v n + 1 ( t ) = Δ p u n + 1 ( t ) d t − μ v n + 1 ( t ) d t + f ( u n ( t ) ) d t + σ ( t , u n ( t ) , v n ( t ) ) d W ( t ) , u n + 1 ( 0 ) = u 0 , v n + 1 ( 0 ) = v 0 .
Then, ( u n + 1 ( t ) , v n + 1 ( t ) ) is also progressively measurable. Thus, by (35) we can define the sequence
{ ( u n ( t ) , v n ( t ) ) } ( n = 1 , 2 , 3 , ⋯ ) , t ∈ [ 0 , T ] .
Set
δ u n ( t ) : = u n + 1 ( t ) − u n ( t ) , δ v n ( t ) : = v n + 1 ( t ) − v n ( t ) .
Subtracting the Equation (35), we obtain
d ( δ u n ( t ) ) = δ v n ( t ) d t , d ( δ v n ( t ) ) = Δ p u n + 1 ( t ) − Δ p u n ( t ) d t − μ δ v n ( t ) d t + f ( u n ( t ) ) − f ( u n − 1 ( t ) ) d t + σ ( t , u n ( t ) , v n ( t ) ) − σ ( t , u n − 1 ( t ) , v n − 1 ( t ) ) d W ( t ) , δ u n ( 0 ) = 0 , δ v n ( 0 ) = 0 .
By Ito’s formula, we have
∥ δ v n ( t ) ∥ 2 2 − 2 ∫ 0 t Δ p u n + 1 ( s ) − Δ p u n ( s ) , δ v n ( s ) d s + 2 μ ∫ 0 t ∥ δ v n ( s ) ∥ 2 2 d s = 2 ∫ 0 t f ( u n ( s ) ) − f ( u n − 1 ( s ) ) , δ v n ( s ) d s + ∫ 0 t ∥ σ ( s , u n ( s ) , v n ( s ) ) − σ ( s , u n − 1 ( s ) , v n − 1 ( s ) ) ∥ 2 2 d s + 2 ∫ 0 t δ v n ( s ) , [ σ ( s , u n ( s ) , v n ( s ) ) − σ ( s , u n − 1 ( s ) , v n − 1 ( s ) ) ] d W ( s ) .
We recall the following elementary vector inequalities that are used frequently: for all a , b ∈ R n and 2 < p < ∞ , we have
2 2 − p | a − b | p ≤ ( | a | p − 2 a − | b | p − 2 b ) .
Thus, we have
∥ δ v n ( t ) ∥ 2 2 + 2 p ∥ ∇ δ u n ( t ) ∥ p p + 2 μ ∫ 0 t ∥ δ v n ( s ) ∥ 2 2 d s = 2 ∫ 0 t f ( u n ( s ) ) − f ( u n − 1 ( s ) ) , δ v n ( s ) d s + ∫ 0 t ∥ σ ( s , u n ( s ) , v n ( s ) ) − σ ( s , u n − 1 ( s ) , v n − 1 ( s ) ) ∥ L 2 0 2 d s + 2 ∫ 0 t δ v n ( s ) , [ σ ( s , u n ( s ) , v n ( s ) ) − σ ( s , u n − 1 ( s ) , v n − 1 ( s ) ) ] d W ( s ) .
It follows from (38) that
E sup 0 ≤ s ≤ t ∥ δ v n ( s ) ∥ 2 2 + 2 p ∥ ∇ δ u n ( s ) ∥ p p + 2 μ E ∫ 0 t ∥ δ v n ( s ) ∥ 2 2 d s ≤ 2 E sup 0 ≤ τ ≤ t ∫ 0 τ f ( u n ) − f ( u n − 1 ) , δ v n d s + E sup 0 ≤ t ≤ T c 0 2 Tr ( Q ) ∫ 0 t ∥ σ ( u n , v n ) − σ ( u n − 1 , v n − 1 ) ∥ 2 2 d s + 2 E sup 0 ≤ τ ≤ t ∫ 0 τ δ v n , [ σ ( u n , v n ) − σ ( u n − 1 , v n − 1 ) ] d W ( s ) = : I 1 + I 2 + I 3 .
For I 1 , by 5 and Young’s inequality,
I 1 ≤ 2 L f 1 2 E ∫ 0 t ∥ ∇ δ u n − 1 ( s ) ∥ p ∥ δ v n ( s ) ∥ 2 d s ≤ 4 L f T ∫ 0 t E ∥ ∇ δ u n − 1 ( s ) ∥ p 2 d s + 1 4 E sup 0 ≤ s ≤ t ∥ δ v n ( s ) ∥ 2 2 .
For I 2 , by Condition 1,
I 2 ≤ Tr ( Q ) L σ E ∫ 0 t ∥ ∇ δ u n − 1 ( s ) ∥ p 2 + ∥ δ v n − 1 ( s ) ∥ 2 2 d s .
For I 3 , by the Burkholder–Davis–Gundy inequality, there exists a constant κ > 0 such that
I 3 ≤ 1 4 E sup 0 ≤ s ≤ t ∥ δ v n ( s ) ∥ 2 2 + κ Tr ( Q ) L σ E ∫ 0 t ∥ ∇ δ u n − 1 ( s ) ∥ p 2 + ∥ δ v n − 1 ( s ) ∥ 2 2 d s .
Substituting (40)–(42) into (39), we obtain
E sup 0 ≤ s ≤ t ∥ δ v n ( s ) ∥ 2 2 + 4 p ∥ ∇ δ u n ( s ) ∥ p p + + 4 μ E ∫ 0 t ∥ δ v n ( s ) ∥ 2 2 d s ≤ ( 8 L f T + 2 Tr ( Q ) ( 1 + κ ) L σ ) ∫ 0 t E ∥ δ v n − 1 ( s ) ∥ 2 2 + ∥ ∇ δ u n − 1 ( s ) ∥ p p d s ,
i.e.,
E sup 0 ≤ s ≤ t p ∥ δ v n ( s ) ∥ 2 2 + 4 ∥ ∇ δ u n ( s ) ∥ p p + + 4 μ p E ∫ 0 t ∥ δ v n ( s ) ∥ 2 2 d s ≤ p ( 8 L f T + 2 Tr ( Q ) ( 1 + κ ) L σ ) ∫ 0 t E ∥ δ v n − 1 ( s ) ∥ 2 2 + ∥ ∇ δ u n − 1 ( s ) ∥ p p d s .
Set
A n ( t ) : = E sup 0 ≤ s ≤ t ∥ δ v n ( s ) ∥ 2 2 + ∥ ∇ δ u n ( s ) ∥ p p + 4 μ p E ∫ 0 t ∥ δ v n ( s ) ∥ 2 2 d s .
Then, we have
A n ( t ) ≤ p ( 8 L f T + 2 Tr ( Q ) ( 1 + κ ) L σ ) ∫ 0 t A n − 1 ( s ) d s , n = 1 , 2 , 3 , ⋯ ,
hence
A n ( t ) ≤ p ( ( 8 L f T + 2 Tr ( Q ) ( 1 + κ ) L σ ) T ) n n ! A 0 ( T ) , 0 ≤ t ≤ T ,
where
A 0 ( T ) : = E sup 0 ≤ s ≤ T ∥ v 1 ( s ) − v 0 ( s ) ∥ 2 2 + ∥ ∇ u 1 ( s ) − ∇ u 0 ( s ) ∥ p p + 4 μ p ∫ 0 t ∥ δ v n ( s ) ∥ 2 2 d s
is a positive constant. Therefore, { u n ( s ) } and { v n ( s ) } are Cauchy sequences and hence there exist the processes u ( t ) and v ( t ) such that
u n ( t ) → u ( t ) , strongly in L 2 ( Ω ; C ( [ 0 , T ] ; W 0 1 , p ( D ) ) ) ,
v n ( t ) → v ( t ) , strongly in L 2 ( Ω ; C ( [ 0 , T ] ; L 2 ( D ) ) ) ,
as n → ∞ . Consequently, ( u ( t ) , v ( t ) ) is the solution to (34).
Finally, we prove the uniqueness of the solution. Let ( u 1 , v 1 ) and ( u 2 , v 2 ) be two solutions of (34). Set δ u = u 1 − u 2 , δ v = v 1 − v 2 . Repeating the same estimates as (45), we obtain
E sup 0 ≤ s ≤ t ∥ δ v ( s ) ∥ 2 2 + ∥ ∇ δ u ( s ) ∥ p p + 4 μ p E ∫ 0 t ∥ δ v n ( s ) ∥ 2 2 d s ≤ p ( 8 L f T + 2 Tr ( Q ) ( 1 + κ ) L σ ) ∫ 0 t E sup 0 ≤ r ≤ s ∥ δ v ( s ) ∥ 2 2 + ∥ ∇ δ u ( s ) ∥ p p d s .
By Gronwall’s inequality, the left of above inequality is zero, and uniqueness is proven. Consequently, the proof of the theorem is complete. □
Next, we use the truncation method to deal with the local existence and uniqueness of solution for problem (1). For each N ≥ 1 , define a C 1 function k N by
k N ( x ) = 1 , if x ≤ N , ∈ ( 0 , 1 ) , if N < x < N + 1 , 0 , if x ≥ N + 1 ,
and further assume that ∥ k N ′ ∥ ∞ ≤ 2 . Let f ( u ) = | u | q − 2 u ln | u | . Define
f N ( u ) = k N ( ∥ ∇ u ∥ p ) f ( u ) ,
and
σ N ( t , u , v ) : = k N ( ∥ ∇ u ∥ p ) σ ( t , u , v ) .
Theorem 1. 
Assume that u 0 and v 0 are F 0 -measurable with u 0 ∈ L 2 ( Ω ; W 0 1 , p ( D ) ) and v 0 ∈ L 2 ( Ω ; L 2 ( D ) ) . Let (5)–(7) and ( 23 ) hold. Then, there exists a unique solution ( u ( t ) , v ( t ) ) to (1) satisfying
u ∈ L 2 ( Ω ; C ( [ 0 , T ] ; W 0 1 , p ( D ) ) ) , v ∈ L 2 ( Ω ; C ( [ 0 , T ] ; L 2 ( D ) ) ) .
Proof. 
Let us choose sequences u 0 m and v 0 m such that
u 0 m ∈ W 0 1 , p ( D ) ∩ W 0 2 , p ( D ) ) v 0 m ∈ H 0 1 ( D ) ,
and as m → ∞ , u 0 m → u 0 strongly in W 0 1 , p ( D ) , v 0 m → v 0 strongly in L 2 ( D ) . Fixed N, for each m ≥ 1 , we consider the existence of solutions to
d u ( t ) = v ( t ) d t d v ( t ) = Δ p u ( t ) d t − μ v ( t ) d t + f N ( u ( t ) ) d t + σ N ( t , u ( t ) , v ( t ) ) d W ( t ) , 0 ≤ t ≤ T u ( 0 ) = u 0 m , v ( 0 ) = v 0 m .
Without the loss of generality, assume that ∥ ∇ u ∥ p ≤ ∥ ∇ u ′ ∥ p for any U ′ = ( u ′ , v ′ ) T ∈ X . Then, we have
∥ f N ( u ) − f N ( u ′ ) ∥ L 2 ≤ ∥ ( k N ( ∥ ∇ u ∥ p ) − k N ( ∥ ∇ u ′ ∥ p ) ) f ( u ) ∥ L 2 + k N ( ∥ ∇ u ′ ∥ p ) ∥ f ( u ) − f ( u ′ ) ∥ L 2 .
We aim to estimate the logarithmic term in the above equation, i.e., ∥ f ( u ) − f ( u ′ ) ∥ 2 2 . From Lemma 4 and the mean value theorem, we have, for 0 < θ 1 < 1
| u | q − 2 u ln | u | − | u ′ | q − 2 u ′ ln | u ′ | = 1 + ( q − 1 ) ln | θ 2 u + ( 1 − θ 2 ) u ′ | · | θ 2 u + ( 1 − θ 2 ) u ′ | q − 2 | u − u ′ | ≤ θ 2 u + ( 1 − θ 2 ) u ′ | q − 2 | u − u ′ | + ( q − 1 ) A u − u ′ + ( q − 1 ) θ 2 u + ( 1 − θ 2 ) u ′ | q − 2 + ε | u − u ′ | ≤ | u | + | u ′ | q − 2 | u − u ′ | + q − 1 A | u − u ′ | + q − 1 | u | + | u ′ | q − 2 + ε | u − u ′ | .
By (5), we have the Sobolev embedding W 0 1 , p ↪ L 2 ( q − 1 ) . Then, by using Hölders inequality, we deduce
∫ D [ ( | u | + | u ′ | ) q − 2 | u − u ′ | ] 2 d x ≤ C ∫ D | u | + | u ′ | 2 ( q − 1 ) d x ( q − 2 ) / ( q − 1 ) × ∫ D | u − u ′ | 2 ( q − 1 ) d x 1 / ( q − 1 ) ≤ C ∥ u ∥ L 2 ( q − 1 ) 2 ( q − 1 ) + ∥ u ′ ∥ L 2 ( q − 1 ) 2 ( q − 1 ) ( q − 2 ) / ( q − 1 ) ∥ u − u ∥ L 2 ( q − 1 ) 2 ≤ C ∥ ∇ u ∥ p 2 ( q − 1 ) + ∥ ∇ u ′ ∥ p 2 ( q − 1 ) ( q − 2 ) / ( q − 1 ) ∥ u − u ′ ∥ W 0 1 , p ( D ) 2 ≤ 2 ( q − 2 ) / ( q − 1 ) C ∇ u ′ p 2 ( q − 2 ) u − u ′ W 0 1 , p ( D ) 2 .
By a similar argument, we have
∫ D | u | + u ′ q − 2 + ε u − u ′ 2 d x ≤ C ∫ D | u | + u ′ 2 ( q − 2 + ε ) ( q − 1 ) / ( q − 2 ) d x ( q − 2 ) / ( q − 1 ) × ∫ D u − u ′ 2 ( q − 1 ) d x 1 / ( q − 1 ) ≤ C ∫ D | u | + u ′ 2 ( q − 2 + ε ) ( q − 1 ) / ( q − 2 ) d x ( q − 2 ) / ( q − 1 ) u − u ′ L 2 ( q − 1 ) 2 .
Then, from Lemma 3, we can choose sufficiently small ε > 0 such that q ¯ = 2 ( q − 1 ) + 2 ε ( q − 1 ) ( q − 2 ) < n p n − p , which implies W 0 1 , p ↪ L q ¯ ( D ) , i.e.,
∥ w ∥ q ¯ ≤ S ε ∥ ∇ w ∥ p , ∀ w ∈ W 0 1 , p ( D ) ,
where S ε is the optimal embedding constant. So, combining (52), we can obtain
∫ D [ ( | u | + | u ′ | ) q − 2 + ε | u − u ′ | ] 2 d x ≤ C ∥ u ∥ L q ¯ ( D ) q ¯ + ∥ u ′ ∥ L q ¯ ( D ) q ¯ ( q − 2 ) / ( q − 1 ) ∥ u − u ′ ∥ L 2 ( q − 1 ) 2 ≤ C S ε q ¯ ∥ ∇ u ∥ p q ¯ + S ε q ¯ ∥ ∇ u ′ ∥ p q ¯ ( q − 2 ) / ( q − 1 ) ∥ u − u ′ ∥ L 2 ( q − 1 ) 2 ≤ 2 ( q − 2 ) / ( q − 1 ) C ( ε ) ∥ ∇ u ′ ∥ p q − 1 + ε ∥ u − u ′ ∥ W 0 1 , p ( D ) 2 .
By combining (50)–(53), we can deduce that
| u | q − 2 u ln | u | − | u ′ | q − 2 u ′ ln | u ′ | L 2 2 ≤ 2 ( q − 2 ) / ( q − 1 ) ( C ∥ ∇ u ′ ∥ p 2 ( q − 2 ) + C ( ε ) ∥ ∇ u ′ ∥ p q − 1 + ε + 1 ) ∥ u − u ′ ∥ W 0 1 , p ( D ) 2 .
Therefore, by combining Lemma 4 and (54), we have
f N ( u ) − f N ( u ′ ) L 2 2 ≤ 2 ∥ k N ( ∥ ∇ u ∥ p ) − k N ( ∥ ∇ u ′ ∥ p ) f ( u ) ∥ L 2 2 + 2 k N 2 ( ∥ ∇ u ′ ∥ p ) ∥ f ( u ) − f ( u ′ ) ∥ L 2 2 ≤ 4 ∥ k N ′ ∥ ∞ 2 ∥ ∇ u ∥ p − ∥ ∇ u ′ ∥ p 2 ( A 2 + ∥ u ∥ 2 ( q − 1 + ε ) 2 ( q − 1 + ε ) ) η ∥ ∇ u ′ ∥ p ≤ N + 1 + 2 ( p − 2 ) / ( p − 1 ) k N 2 ( ∥ ∇ u ′ ∥ p ) ( C ∥ ∇ u ′ ∥ p 2 ( q − 2 ) + C ( ε ) ∥ ∇ u ′ ∥ p q − 1 + ε + 1 ) ∥ u − u ′ ∥ W 0 1 , p ( D ) 2 ≤ 4 ∥ u − u ′ ∥ W 0 1 , p ( D ) 2 ( A 2 + C ′ ∥ u ∥ W 0 1 , p ( D ) 2 ( q − 1 + ε ) ) η { ∥ ∇ u ′ ∥ p ≤ N + 1 } + 2 ( q − 2 ) / ( q − 1 ) k N 2 ( ∥ ∇ u ′ ∥ p ) ( C ∥ ∇ u ′ ∥ p 2 ( q − 2 ) + C ( ε ) ∥ ∇ u ′ ∥ p q − 1 + ε + 1 ) ∥ u − u ′ ∥ W 0 1 , p ( D ) 2 ≤ C 1 ( N , p ) ∥ u − u ′ ∥ W 0 1 , p ( D ) 2 ,
where η B denotes the indicator of some set B and C ′ is the best constant of the Sobolev embedding W 0 1 , p ( D ) ↪ L ( q − 1 + ε ) . Here we choose 0 < ε ≤ n p n − p + 1 − q , q − 1 < n p n − p . For any U = u , v T , U ′ = u ′ , v ′ T ∈ X , combining (55), we have
∥ f N ( u ) − f N ( u ′ ) + μ ( v ′ − v ) ∥ L 2 2 ≤ 2 ( ∥ f N ( u ) − f N ( u ′ ) ∥ L 2 2 + ∥ μ ( v ′ − v ) ∥ L 2 2 ) ≤ C 2 ( N , p ) ∥ U − U ′ ∥ X 2 .
Similarly as in (55), (56) and Lemma 4, one has
∥ f N ( u ) − μ v ∥ L 2 2 ≤ 2 k N 2 ( ∥ ∇ u ∥ p ) ( A + ∥ u ∥ 2 ( q − 1 + ε ) 2 ( q − 1 + ε ) ) + 2 μ 2 ∥ v ∥ L 2 2 ≤ 2 C 1 k N 2 ( ∥ ∇ u ∥ p ) ( A + ∥ u ∥ W 0 1 , p ( D ) 2 ( q − 1 + ε ) ) + 2 μ 2 ∥ v ∥ L 2 2 ≤ C 3 ( N , p ) ( A + U X 2 ) .
On the other hand, from (6) and and the Sobolev embedding, we deduce
T r σ N ( U ) Q σ N * ( U ) = ∑ i = 1 ∞ σ N ( U ) Q e i , σ N ( U ) e i = ∑ i = 1 ∞ ξ i σ N ( U ) e i , σ N ( U ) e i ≤ c 1 ∑ i = 1 ∞ ξ i k N 2 ( ∥ ∇ u ∥ p ) ∫ D e i 2 ( x ) ( 1 + | u ( x ) | 2 ( q − 1 ) + | ∇ u | 2 + v 2 ) d x ≤ c 1 ∑ i = 1 ∞ ξ i k N 2 ( ∥ ∇ u ∥ p ) ∥ e i ∥ L 2 2 + sup i ≥ 1 ∥ e i ∥ ∞ 2 ( ∥ u ∥ L 2 ( q − 1 ) 2 ( q − 1 ) + ∥ ∇ u ∥ L 2 2 + ∥ v ∥ L 2 2 ) ≤ c 1 ∑ i = 1 ∞ ξ i k N 2 ( ∥ ∇ u ∥ p ) 1 + c 0 ( C 1 ∥ ∇ u ∥ p 2 ( q − 1 ) + C 2 ∥ ∇ u ∥ p 2 + ∥ v ∥ 2 2 ) ≤ C 4 ( N , p ) ( 1 + ∥ U ∥ X 2 ) .
Similarl to the above, one can conclude that
Tr σ N ( U ) − σ N ( U ′ ) Q σ N ( U ) − σ N ( U ′ ) * = ∑ i = 1 ∞ ξ i ∥ ( σ N ( U ) − σ N ( U ′ ) ) e i ∥ 2 2 ≤ 2 ∑ i = 1 ∞ ξ i ∥ ( k N ( ∥ ∇ u ∥ p ) − k N ( ∥ ∇ u ′ ∥ p ) ) σ N ( U ) e i ∥ 2 2 + 2 k N 2 ( ∥ ∇ u ′ ∥ p ) sup i ≥ 1 ∥ e i ∥ ∞ 2 Tr Q ∥ σ ( u , v , ∇ u , x , t ) − σ ( u ′ , v ′ , ∇ u ′ , x ′ , t ′ ) ∥ 2 2 ≤ 8 ∑ i = 1 ∞ ξ i ∥ ∇ u − ∇ u ′ ∥ p 2 ∥ σ N ( U ) e i ∥ 2 2 η { ∥ ∇ u ′ ∥ p ≤ N + 1 } + 2 c 2 c 0 2 k N 2 ( ∥ ∇ u ′ ∥ p ) Tr Q ( ∥ u − u ′ ∥ 2 2 + ∥ | u | q − 2 | u − u ′ | ∥ 2 2 + ∥ | u ′ | q − 2 | u − u ′ | ∥ 2 2 + ∥ ∇ u − ∇ u ′ ∥ 2 2 + ∥ v − v ′ ∥ 2 2 ) ≤ C 5 ( N , p ) ∥ U − U ′ ∥ X 2 .
Combining the estimates (56)–(59), then applying Lemma 6 we get that the truncated SDE (48) has a unique mild solution
u m ∈ L 2 ( Ω ; C [ 0 , T ] ; W 0 1 , p ( D ) ∩ W 0 2 , p ( D ) ) , v m ∈ L 2 ( Ω ; C [ 0 , T ] ; H 0 1 ( D ) )
for each fixed N > 0 . Similarly as Lemma 6, we have { u m } and { v m } are Cauchy sequences and hence there exist the processes u N ( t ) and v N ( t ) such that
u m ( t ) → u N ( t ) , strongly in L 2 ( Ω ; C ( [ 0 , T ] ; W 0 1 , p ( D ) ) ) ,
v m ( t ) → v N ( t ) , strongly in L 2 ( Ω ; C ( [ 0 , T ] ; L 2 ( D ) ) ) ,
as m → ∞ , Consequently ( u N ( t ) , v N ( t ) ) is the solution of the following equation
d u ( t ) = v ( t ) d t d v ( t ) = Δ p u ( t ) d t − μ v ( t ) d t + f N ( u ( t ) ) d t + σ N ( t , u ( t ) , v ( t ) ) d W ( t ) , 0 ≤ t ≤ T u ( 0 ) = u 0 , v ( 0 ) = v 0 .
For the uniqueness of (60), the proof is similar to Lemma 6, so we omit it here. Next define an F t t ≥ 0 -stopping time τ N as follows
τ N = inf t ≥ 0 ; ∥ ∇ u ∥ p ≥ N ,
then U ( t ) = U N ( t ) on the event { t < τ N } . If let ζ = lim N → ∞ τ N , then by the path continuity of t → U ( t ) , U ( t ) is the unique local solution with lifespan ζ . For uniqueness, suppose that there exists another solution U 1 ( t ) for t < τ , where τ is a stopping time. It follows that U ( t ) = U 1 ( t ) for t < τ ∞ and τ = τ ∞ . □
To obtain a global solution, it is necessary to impose further condition on the nonlinearity σ that ensure a certain energy bound can be established to prevent unlimited growth. Hence, we will replace the condition (6) with
| σ ( u , v , ∇ u , x , t ) | 2 ≤ ( p − 2 ) μ p c 0 2 T r Q | v | 2 ,
where c 0 is defined in Section 2. We define
e ( t ) = ∥ v ∥ 2 2 + ( 2 p − 2 q ) ∥ ∇ u ∥ p p + 2 q 2 ∥ u ∥ q q .
Lemma 7. 
Assume (61) hold, ( u 0 , v 0 ) ∈ W 0 1 , p ( D ) × L 2 ( D ) . Let U ( t ) = ( u ( t ) , v ( t ) ) be a mild solution of system (17). Then,
E ε ( t ) ≤ E ε ( 0 ) − μ E ∫ 0 t v 2 2 d τ + 1 2 c 0 2 Tr Q E ∫ 0 t ∫ D σ 2 ( u , v , ∇ u , x , t ) d x d t .
Proof. 
Applying Itô formula to ∥ v ∥ 2 2 , we obtain
∥ v ∥ 2 2 = ∥ v 0 ∥ 2 2 + 2 ∫ 0 t v , d v + ∫ 0 t d v , d v = ∥ v 0 ∥ 2 2 + 2 ∫ 0 t v , Δ p u d τ − 2 ∫ 0 t v , μ v d τ + 2 ∫ 0 t v , | u | q − 2 u ln | u | d τ + 2 ∫ 0 t v , σ d W τ + ∑ i = 1 ∞ ∫ 0 t ( σ Q e i , σ e i ) d τ .
Since
2 ∫ 0 t v , Δ p u d τ = 2 p ∥ ∇ u 0 ∥ p p − 2 p ∥ ∇ u ( t ) ∥ p p ,
2 ∫ 0 t v , | u | q − 2 u ln | u | d τ = 2 q ∫ D | u ( t ) | q ln | u ( t ) | d x − 2 q ∫ D | u 0 | p ln | u 0 | d x 2 ∫ 0 t v , | u | q − 2 u log | u | d τ = + 2 q 2 ∥ u 0 ∥ q q − 2 q 2 ∥ u ( t ) ∥ q q .
Therefore, combining (63)–(65), we have
v 2 2 = 2 ε ( 0 ) − 2 p ∥ ∇ u t ∥ p p − 2 μ ∫ 0 t ∥ v ∥ 2 2 d τ + 2 q ∫ D | u ( t ) | q ln | u ( t ) | d x − 2 q 2 ∥ u ( t ) ∥ q q + 2 ∫ 0 t v , σ W ( τ ) + ∑ i = 1 ∞ ξ i ∫ 0 t ∫ D σ 2 u , v , ∇ u , x , τ e i 2 x d x d τ .
Note that
T r Q = ∑ i = 1 ∞ ξ i , c 0 : = sup i ≥ 1 e i ∞ < ∞ .
By substituting (67) into (66) and taking the expectation of (66), we arrive at (62). □
Remark 1. 
If σ ( u , v , ∇ u , x , t ) satisfies (61), then it follows from (62) that the energy function is nonincreasing with respect to time variable t, that is
d d t E ε ( t ) ≤ − μ ( q + 2 ) 2 q E v 2 2 ≤ 0 .
Now, define
W = { u ∈ W 0 1 , p ( D ) ∣ E I ( u ) > 0 , E J ( u ) < d } .
We demonstrate that the set W is invariant under the semi-flow of problem (17).
Lemma 8. 
Let u ( x , t ) is the mild solution of problem (17), u 0 ( x ) ∈ W 0 1 , p ( D ) , v 0 ( x ) ∈ L 2 ( D ) and E ε ( 0 ) < d . If u 0 x ∈ W and the condition (61) holds, then we have u ( t ) ∈ W for all t ∈ 0 , T .
Proof. 
It follows from (10) and (68) that
1 2 E v 2 2 + E J u ≤ 1 2 E v 0 2 2 + E J u 0 < d , for all t ∈ [ 0 , T ) .
Arguing by contradiction, we assume that there exists a number t 0 ∈ 0 , T such that u t ∈ W for 0 , t 0 and u t 0 ∉ W . Then, by the continuity of E J ( u ( t ) ) and E I ( u ( t ) ) in the variable t, we have either (i) E J ( u ( t 0 ) ) = d or (ii) E I ( u ( t 0 ) ) = 0 and E u t 0 ≠ 0 . It follows from (70) that (i) is impossible. If (ii) holds, then by (14), we get d ≤ E J u t 0 , which contradicts with (70). Therefore, we have u ( t ) ∈ W for all t ∈ 0 , T . Lemma 8 has been proven. □
Theorem 2. 
Assume that conditions for Theorem 1 hold. Let u 0 ( x ) ∈ W 0 1 , p ( D ) , v 0 ( x ) ∈ L 2 ( D ) . If u 0 x ∈ W and (61) holds, then for any T > 0 , there exists a unique solution U ( t ) of (1) on the interval 0 , T satisfying
E sup 0 ≤ t ≤ T e t < ∞ .
Proof. 
For any T > 0 , we will prove that U N t = U t ∧ τ N → U t a.s. as N → ∞ for any t ≤ T , thereby extending the local solution to a global one. To this end, it suffices to show that τ N → ∞ as N → ∞ with a probability of 1.
Recall that for t ∈ 0 , τ N ∧ T , U N t = U t ∧ τ N is the local solution of (1). According to (9) and (66), the following energy equation holds:
e ( u ( t ∧ τ N ) ) = e ( u 0 ) + 2 q I ( u 0 ) − 2 q I ( u ) − 2 μ ∫ 0 t ∧ τ N ∥ v ( s ) ∥ 2 2 d s + 2 ∫ 0 t ∧ τ N ( v ( s ) , σ d W s ) + ∫ 0 t ∧ τ N T r ( σ ( u , v ) Q σ * ( u , v ) ) d s .
From (61) and (72), we arrive at
e ( u ( t ∧ τ N ) ) ≤ e ( u 0 ) + 2 q I ( u 0 ) − 2 q I ( u ) + 2 ∫ 0 t ∧ τ N ( v ( s ) , σ d W s ) + c 0 2 T r Q ∫ 0 t ∧ τ N σ ( s ) 2 2 d s ≤ e ( u 0 ) + 2 q I ( u 0 ) − 2 q I ( u ) + 2 ∫ 0 t ∧ τ N ( v ( s ) , σ d W s ) + K ∫ 0 t e ( u ( s ∧ t N ) ) d s .
for some constant K. By taking the expectation of (73) and applying Lemma 8, we get
E e ( u ( t ∧ τ N ) ) ≤ E e ( u 0 ) + 2 q E I ( u 0 ) + K E ∫ 0 t e ( u ( s ∧ t N ) ) d s .
Using the Gronwall inequality to (74), it follows that
E e ( u ( T ∧ τ N ) ) ≤ ( E e ( u 0 ) + 2 q E I ( u 0 ) ) e K T ≤ C T ,
for some constant C T independent of N. On the other hand, we have
P ( τ N ≤ T ) = P sup 0 ≤ s ≤ T ∥ ∇ u ( s ∧ τ N ) ∥ p ≤ E sup 0 ≤ s ≤ T ∥ u ( s ∧ τ N ) ∥ p p N p ≤ C E sup 0 ≤ s ≤ T e ( u ( T ∧ τ N ) ) N p .
In view of (76), we can obtain
P ( ζ ≤ T ) ≤ P ( τ N ≤ T ) ≤ C T N p ,
so that, by the Borel–Cantelli lemma implies that P ( ζ ≤ T ) = 0 or lim N → ∞ τ N = ∞ a.s. Hence, u = lim N → ∞ u N ( t ) is the global solution as claimed on 0 , ζ ∧ T = 0 , T .
In order to verify the energy bound (71), we take the limit ( N → ∞ ) in (72) to obtain the energy equation
e ( u ( t ) ) = e ( u 0 ) + 2 q I ( u 0 ) − 2 q I ( u ) − 2 μ ∫ 0 t ∥ v ( s ) ∥ 2 2 d s + 2 ∫ 0 t ( v ( s ) , σ d W s ) d s + ∫ 0 t T r ( σ ( u , v ) Q σ * ( u , v ) ) d s .
Similar to (74), from (77) and Lemma 8, we have
E sup 0 ≤ t ≤ T e ( u ( t ) ) ≤ E e ( u 0 ) + 2 q E I ( u 0 ) + 2 E sup 0 ≤ t ≤ T ∫ 0 t ( v , σ ) d W s + E ∫ 0 T T r ( σ ( u , v ) Q σ * ( u , v ) ) d s .
By the Burkholder–Davis–Gundy inequality, we get
E sup 0 ≤ t ≤ T ∫ 0 t ( v , σ d W s ) ≤ C 1 E sup 0 ≤ t ≤ T v 2 2 ∫ 0 T T r σ ( u , v ) Q σ * ( u , v ) d s 1 2 ≤ 1 4 E sup 0 ≤ t ≤ T v 2 2 + C 2 E ∫ 0 T T r σ ( u , v ) Q σ * ( u , v ) d s ,
for some constant C 1 , C 2 > 0 . In view of (61), we have
E ∫ 0 T T r σ ( u , v ) Q σ * ( u , v ) d s = E ∑ i = 1 ∞ ∫ 0 T ( σ Q e i , σ e i ) d t ≤ c 0 2 T r Q ∫ 0 T E σ 2 2 d t ≤ C 3 ∫ 0 T E sup 0 ≤ t ≤ T e u ( t ) d t ,
where C 3 > 0 is positive constant. Substituting (79) and (80) into (78), we obtain
E sup 0 ≤ t ≤ T e u ( t ) ≤ C 4 + C 5 ∫ 0 T sup 0 ≤ s ≤ T e ( u ) d s ,
where C 4 , C 5 are positive constants depending on c 0 , T etc. By applying the Gronwall inequality to (81), we obtain the following
E sup 0 ≤ t ≤ T e u ( t ) ≤ C 4 e C 5 T ,
which implies the energy bound (71). □

4. Blow-Up Phenomenon

In this section, we focus on analyzing the blow-up behavior of solutions to problem (1) under the condition of low initial energy and high energy, and give the upper bounds for the blow-up time. We first demonstrate that the set V remains invariant under the semi-flow associated with problem (1). This property is crucial for exploring the blow-up phenomena in this context. Here
V = { u ∈ W 0 1 , p ( D ) ∣ E I ( u ) < 0 , E J ( u ) < d } ,
where d is the depth of the potential well defined in (14).
Definition 2 
(Finite-time Blow-up). The solution u of problem (1) blows up in finite time in the sense of mean W 0 1 , p -norm if
P lim t → T | | ∇ u ( t , · ) | | p = ∞ = 1
for some random time T with P ( T < ∞ ) > 0 .
Lemma 9. 
Let u 0 ( x ) ∈ V and v 0 ( x ) ∈ L 2 ( D ) . If E ε ( 0 ) < d and the condition (61) hold, then u ( t ) ∈ V and d < ( 1 p − 1 q ) E ∇ u p p + 1 q 2 E u q q for all t ∈ [ 0 , T m a x ) .
Proof. 
Now, we claim that E I ( u ( t ) ) < 0 for all t ∈ [ 0 , T m a x ) . In fact, if it is false, then there would exist a t 0 ∈ ( 0 , T m a x ) such that E I ( u ( t ) ) < 0 for ∀ t ∈ [ 0 , t 0 ) and E I ( u ( t 0 ) ) = 0 . From (8), (10) and (68), we arrive at
E J ( u ( t ) ) ≤ E ε ( t ) ≤ E ε ( 0 ) < d , t ∈ [ 0 , T max ) .
By definition of (14), we have d ≤ E J ( u ( t 0 ) ) , which contradicts (83). Thus, u ( t ) ∈ V for all t ∈ [ 0 , T m a x ) .
Next, we prove d < ( 1 p − 1 q ) E ∇ u p p + 1 q 2 E u q q . From u ( t ) ∈ V , we get E I ( u ) < 0 , then according to Lemma 1 we can obtain: there exists a λ * ∈ ( 0 , 1 ) such that E I ( λ * u ) = 0 . Then, combining (11) and (14), we arrive at
d ≤ E J ( λ * u ) = 1 q E I ( λ * u ) + ( λ * ) p ( 1 p − 1 q ) E ∇ u p p + 1 q 2 ( λ * ) q E ∥ u ∥ q q < ( 1 p − 1 q ) E ∇ u p p + 1 q 2 E u q q .
Lemma 9 has been proven. □
Lemma 10. 
Let ( u , v ) be a mild solution of system (17) with the initial data ( u 0 , v 0 ) ∈ W 0 1 , p ( D ) × L 2 ( D ) . Then,
E ( u ( t ) , v ( t ) ) = E ( u 0 ( x ) , v 0 ( x ) ) − E ∫ 0 t ∥ ∇ u ∥ p p d τ − 1 2 μ E ∥ u ( t ) ∥ 2 2 + 1 2 μ E ∥ u 0 ∥ 2 2 + E ∫ 0 t ∫ D | u | q ln | u | d x d τ + E ∫ 0 t ∥ v ( τ ) ∥ 2 2 d τ .
Proof. 
If ( u , v ) is a global mild solution to system (17), then for each n ≥ 1 , { ( u ( t ) , ξ n ) ; t ≥ 0 } is a continuous { F t , t ≥ 0 } -adapted finite variation process and { ( v ( t ) , ξ n ) ; t ≥ 0 } is a continuous { F t , t ≥ 0 } -adapted semi-martingale, where { ξ n } n ≥ 1 is an orthonormal basis of L 2 ( D ) . Then, by the Itô formula, we have
( u ( t ) , ξ n ) ( v ( t ) , ξ n ) = ( u 0 , ξ n ) ( v 0 , ξ n ) + ∫ 0 t ( u ( s ) , ξ n ) d ( v ( s ) , ξ n ) + ∫ 0 t ( v ( s ) , ξ n ) d ( u ( s ) , ξ n ) ,
which implies that
( u ( t ) , v ( t ) ) = ( u 0 , v 0 ) + ∫ 0 t ( u ( τ ) , d v ( τ ) ) + ∫ 0 t ( v ( τ ) , d u ( τ ) ) = ( u 0 , v 0 ) − ∫ 0 t ∥ ∇ u ∥ p p d τ − 1 2 μ ∥ u ( t ) ∥ 2 2 + 1 2 μ ∥ u 0 ∥ 2 2 + ∫ 0 t ∫ D | u | q log | u | d x d τ + ∫ 0 t ∥ v ∥ 2 2 d τ + ∫ 0 t ( u , σ ( u , v , ∇ u , x , τ ) d W ( τ ) ) .
Then, (84) follows from taking the expectation for (85). □

4.1. Low Initial Energy E ε ( 0 ) < d

Theorem 3. 
Let (61) hold, u 0 ( x ) ∈ V and v 0 ( x ) ∈ L 2 ( D ) . If E ε ( 0 ) < d , then the solution u ( t ) of problem (1) blows up in finite time and blow-up time T m a x satisfies
T max ≤ 4 [ ( a 2 + ( q − 2 ) 2 ( d − E ε ( 0 ) ) E ∥ u 0 ∥ 2 2 ) 1 / 2 + a ] ( q − 2 ) 2 ( d − E ε ( 0 ) ) ,
where a = 2 μ E ∥ u 0 ∥ 2 2 − ( q − 2 ) E ( u 0 , v 0 ) .
Proof. 
This theorem is proven using Levine’s concavity argument. For any T > 0 , b > 0 , and η > 0 , we define
L ( t ) : = E u 2 2 + μ E ∫ 0 t u 2 2 d τ + ( T − t ) μ E u 0 2 2 + b ( t + η ) 2 .
Then, we have
L ′ ( t ) = 2 E ( u , v ) + 2 μ E ∫ 0 t ( u , v ) d τ + 2 b ( t + η ) ,
and
L ″ ( t ) = − 2 E ∥ ∇ u ∥ p p + 2 E ∥ v ∥ 2 2 + 2 E ∫ D | u | q ln | u | d x + 2 b .
By combining (10), (68) and (88), we can obtain
L ″ ( t ) = ( q + 2 ) E ∥ v ∥ 2 2 − 2 q E ε ( t ) + 2 ( q p − 1 ) E ∇ u p p + 2 q E ∥ u ∥ q q + 2 b ≥ ( q + 2 ) E ∥ v ∥ 2 2 + 2 q ( d − E ε ( 0 ) ) + ( q + 2 ) μ E ∫ 0 t ∥ v ∥ 2 2 d τ + 2 b
According to E ε ( 0 ) < d , we choose b = d − E ε ( 0 ) > 0 , which, together with (89) and q > 2 , implies
L ″ t > ( q + 2 ) [ E ∥ v ∥ 2 2 + μ E ∫ 0 t v 2 2 d τ + b ] .
Using (87), a direct calculation yields
( L ′ ( t ) ) 2 = 4 [ ( E ( u , v ) ) 2 + 2 μ E ( u , v ) E ∫ 0 t ( u , v ) d τ + 2 b ( t + η ) E ( u , v ) + ( μ E ∫ 0 t ( u , v ) d τ ) 2 + 2 b μ ( t + η ) E ∫ 0 t ( u , v ) d τ + b 2 ( t + η ) 2 ]
By the Cauchy–Schwarz inequality, it follows that
E ( u , v ) ≤ E ∥ u ∥ 2 2 1 2 E ∥ v ∥ 2 2 1 2 ,
E ∫ 0 t ( u , v ) d τ ≤ E ∫ 0 t ∥ u ∥ 2 2 d τ 1 2 E ∫ 0 t ∥ v ∥ 2 2 d τ 1 2 .
So, by (92) and (93) combined with Cauchy’s inequality, we have
( L ′ ( t ) ) 2 ≤ 4 E ∥ u ∥ 2 2 + μ E ∫ 0 t ∥ u ∥ 2 2 d τ + b ( t + η ) 2 E ∥ v ∥ 2 2 + μ E ∫ 0 t ∥ v ∥ 2 2 d τ + b .
In view of (86), (90) and (94), we derive
L ( t ) L ″ ( t ) − q + 2 4 ( L ′ ( t ) ) 2 ≥ ( q + 2 ) ( T − t ) E ∥ u 0 ∥ 2 2 E ∥ v ∥ 2 2 + μ E ∫ 0 t ∥ v ∥ 2 2 d τ + b ≥ 0 , t ∈ [ 0 , T ] .
we can take
η > max 0 , 2 μ E ∥ u 0 ∥ 2 2 − ( q − 2 ) E ( u 0 , v 0 ) b ( q − 2 ) .
Then, we can use (86) and (87) to get
L ( 0 ) = E ∥ u 0 ∥ 2 2 + T μ E ∥ u 0 ∥ 2 2 + b η 2 > 0 ,
L ′ ( 0 ) = 2 E ( u 0 , v 0 ) + 2 b η > 0 .
According to Lemma 2, there exists a t * > 0 satisfying
t * ≤ 4 L ( 0 ) ( q − 2 ) L ′ ( 0 ) = 2 E ∥ u 0 ∥ 2 2 + T μ E ∥ u 0 ∥ 2 2 + b η 2 ( q − 2 ) E ( u 0 , v 0 ) + b η
such that L ( t ) → ∞ as t → t * − .
This contradicts with the assumption that L ( t ) is well defined on [ 0 , T ] for any T > 0 . Therefore, we have proven the finite-time blow up result of the solution. To estimate T m a x from above, for any T ∈ ( 0 , T m a x ) , we define L ¯ ( t ) similarly to the above L ( t ) as follow
L ¯ ( t ) = μ E ∫ 0 t ∥ u ( τ ) ∥ 2 2 d τ + E ∥ u ( τ ) ∥ 2 2 + ( T m a x − t ) μ E ∥ u 0 ∥ 2 2 + b ( t + η ) 2 , t ∈ [ 0 , T ] .
According to the foregoing arguments and a straightforward calculation yield
T max ≤ T ( η ) = 2 ( E ∥ u 0 ∥ 2 2 + b η 2 ) ( q − 2 ) [ E ( u 0 , v 0 ) + b η ] − 2 μ E ∥ u 0 ∥ 2 2 .
Then, we minimize T ( η ) for η > max { 0 , 2 μ E ∥ u 0 ∥ 2 2 − ( q − 2 ) E ( u 0 , v 0 ) b ( q − 2 ) } to get
T max ≤ T min ( η ) = T ( η 0 ) = 4 [ ( a 2 + ( q − 2 ) 2 b E ∥ u 0 ∥ 2 2 ) 1 2 + a ] ( q − 2 ) 2 a ,
where a = 2 μ E ∥ u 0 ∥ 2 2 − ( q − 2 ) E ( u 0 , v 0 ) and η 0 = [ a 2 + ( q − 2 ) 2 b E ∥ u 0 ∥ 2 2 ] 1 / 2 + a ( q − 2 ) b . Theorem 3 has been proven. □
Corollary 1. 
Let (61) hold, p > 2 , u 0 ∈ V and v 0 ∈ L 2 ( D ) . If E ε ( 0 ) < 0 , then the solution u ( x , t ) of problem (1) blows up at finite time.
Proof. 
It is easily seen from (8)–(10) that E ε ( 0 ) < 0 implies E I ( u 0 ) < 0 and E J ( u 0 ) < 0 < d , which means all the conditions in Theorem 3 are satisfied. Using similar arguments to those in the proof of Theorem 3, the conclusion holds obviously. The proof of Corollary 1 has been completed. □
Lemma 11. 
If E ε ( 0 ) = d and ( u 0 , v 0 ) > 0 , then there exists a t 0 ∈ ( 0 , T ) such that ∫ 0 t 0 v 2 2 d τ > 0 .
Proof. 
Arguing by contradiction, we suppose that ∫ 0 t v 2 2 d τ ≡ 0 for ∀ t ∈ 0 , T , which means v 2 2 ≡ 0 , i.e., u ( t ) = u 0 . Combining (12) and the condition E ε ( 0 ) = d , we get E J ( u 0 ) = E ( 0 ) = d , which contradicts with u 0 ∈ V . So, we have ∫ 0 t 0 v 2 2 d τ > 0 for t 0 ∈ ( 0 , T ) . □
Lemma 12. 
Let u ( x , t ) be a mild solution of problem (1.1), u 0 ( x ) ∈ V and v 0 ( x ) ∈ L 2 ( D ) . If E ε ( 0 ) = d , ( u 0 , v 0 ) > 0 and condition (61) holds. Then, we have u ( t ) ∈ V for ∀ t ∈ [ 0 , T m a x ) .
Proof. 
Arguing by contradiction, we claim that E I ( u 0 ) < 0 for ∀ t ∈ [ 0 , T ) . In fact, if it is not true, then there exists a t 0 ∈ ( 0 , T ) such that E I ( u ( t ) ) < 0 for t ∈ [ 0 , t 0 ) and E I ( u ( t 0 ) ) = 0 . By (12) and (68) we have E J ( u ( t 0 ) ) + 1 2 E v ( t 0 ) 2 2 + μ ( q + 2 ) 2 q E ∫ 0 t 0 v 2 2 d τ ≤ E ε ( 0 ) = d . From (14), we get J ( u ( t 0 ) ) ≥ d , that means E ∫ 0 t v 2 2 d τ and E v ( t ) 2 2 = 0 , ∀ t ∈ [ 0 , t 0 ) . Thus, u ( t 0 ) = u 0 , that is, I ( u ( t 0 ) ) = I ( u 0 ) < 0 , which contradicts with I ( u ( t 0 ) ) = 0 . So, we have u ∈ V for ∀ t ∈ [ 0 , T ) . The proof of Lemma 12 is complete. □
Theorem 4. 
Let u ( x , t ) be a mild solution of problem (1.1), u 0 ( x ) ∈ V and v 0 ( x ) ∈ L 2 ( D ) . If E ε ( 0 ) = d , ( u 0 , v 0 ) > 0 and condition (61) holds. Then, the solution u ( t ) of problem (1) blows up in finite time.
Proof. 
By the condition E ε ( 0 ) = d , u 0 ( x ) ∈ V and Lemma 12, we know that there is a t 0 ∈ ( 0 , T ) such that ∫ 0 t 0 v 2 2 d τ > 0 and u ( t 0 ) ∈ V . So, combining (68) and E ε ( 0 ) = d , we obtain
E ε ( t 0 ) ≤ E ε ( 0 ) − μ ( q + 2 ) 2 q E ∫ 0 t 0 v 2 2 d τ < d .
Let t 0 be the initial time, by the similar arguments as those in the proof of Theorem 3, the conclusion holds obviously. The proof of Theorem 4 is complete. □

4.2. High Initial Energy

In this section, we consider the blow-up properties of solutions to problem (1) in the high energy. We first establish two lemmas demonstrating that the unstable set N − remains invariant under the semi-flow of problem by (1).
Lemma 13. 
Let u 0 ∈ W 0 1 , p ( D ) , v 0 ∈ L 2 ( D ) and u = u ( t ) be a solution to problem (1) such that u ( t ) ∈ N − on [ 0 , T m a x ) . Then,
{ t ↦ μ E ∥ u ( t ) ∥ 2 2 + 2 E ( u , v ) }
is strictly increasing on ( 0 , T m a x ) .
Proof. 
We define
K ( t ) = μ E u ( t ) 2 2 + 2 E ( u , v ) , t ∈ 0 , T max .
From (84), we have
K ′ ( t ) = 2 E v 2 2 − 2 E ∥ ∇ u ∥ p p + 2 ∫ D | u | q ln | u | d x = 2 ( E ∥ v ∥ 2 2 − I ( u ) ) > 0 ,
which implies that K ( t ) is strictly increasing on ( 0 , T m a x ) . Here we have used the assumption that E I ( u ( t ) ) < 0 for t ∈ [ 0 , T m a x ) to deduce the last inequality. The proof is complete. □
Lemma 14. 
Let (61) hold, u 0 ( x ) ∈ W 0 1 , p ( D ) , v 0 ( x ) ∈ L 2 ( D ) . Assume that u 0 ∈ N − and the initial data satisfy
μ E ∥ u 0 ∥ 2 2 + 2 E ( u 0 , v 0 ) > E ε ( 0 ) C ¯ > 0 ,
where C ¯ = min 1 2 , C * 1 + μ and C * = 1 p − 1 q μ 1 ℓ 1 | D | 2 p − 1 + ℓ 2 q 2 | D | 2 q − 1 . Then, the solution u ( t ) to problem (1) belongs to N − for ∀ t ∈ [ 0 , T m a x ) .
Proof. 
We claim that N − for all t ∈ [ 0 , T m a x ) . In fact, if it does not hold, then there would exist a t 1 ∈ ( 0 , T m a x ) such that E I ( u ( t ) ) < 0 , t ∈ [ 0 , t 1 ) and E I ( u ( t 1 ) ) = 0 . Then, it follows from Lemma 13 that K ( t ) = E u ( t ) 2 2 + 2 E ( u , v ) is strictly increasing on [ 0 , t 1 ) , we have
μ E ∥ u ( t ) ∥ 2 2 + 2 E ( u , v ) > μ E ∥ u 0 ∥ 2 2 + 2 E ( u 0 , v 0 ) > E ε ( 0 ) C ¯ > 0 , t ∈ ( 0 , t 1 ) .
By the continuity and monotonicity of K ( t ) , we obtain
μ E ∥ u ( t 1 ) ∥ 2 2 + 2 E ( u ( t 1 ) , v ( t 1 ) > E ε ( 0 ) C ¯ > 0 .
From (68), we get E ε ( 0 ) ≥ E ε ( t 1 ) . Combining (11) and (12), μ 1 = inf ∇ u p p u p p and Hölder’s inequality, we have
E ε ( 0 ) ≥ E ε ( t 1 ) = 1 2 E u t t 1 2 2 + 1 p − 1 q E ∇ u t 1 p p + 1 q 2 E u t 1 q q ≥ 1 2 E u t t 1 2 2 + 1 p − 1 q μ 1 E u t 1 p p + 1 q 2 E u t 1 q q ≥ 1 2 E u t t 1 2 2 + 1 p − 1 q μ 1 ℓ 1 E u t 1 2 p + ℓ 2 q 2 E u t 1 2 q ≥ 1 2 E u t t 1 2 2 + 1 p − 1 q μ 1 ℓ 1 | D | 2 p − 1 E u ( t 1 ) 2 2 + ℓ 2 q 2 | D | 2 q − 1 E u ( t 1 ) 2 2 ≥ 1 2 E u t t 1 2 2 + 1 p − 1 q μ 1 ℓ 1 | D | 2 p − 1 + ℓ 2 q 2 | D | 2 q − 1 E u ( t 1 ) 2 2 ,
where ℓ 1 , ℓ 2 is a constant in the embedding L p ( D ) , L q ( D ) ↪ L 2 ( D ) , q ≥ p > 2 . Namely,
E ε ( 0 ) ≥ E ε ( t 1 ) ≥ 1 2 E u t t 1 2 2 + C * E u t 1 2 2 ,
where C * = 1 p − 1 q μ 1 ℓ 1 | D | 2 p − 1 + ℓ 2 q 2 | D | 2 q − 1 . According to (102), we have
E ε ( 0 ) ≥ E ε ( t 1 ) ≥ 1 2 E u t t 1 2 2 + C * E u t 1 2 2 ≥ 1 2 E u t t 1 2 2 + C * 1 + μ E u ( t 1 ) 2 2 + C * μ 1 + μ E u ( t 1 ) 2 2 ≥ C ¯ 2 E ( u t ( t 1 ) , u ( t 1 ) ) + μ E u ( t 1 ) 2 2 ,
where C ¯ = min 1 2 , C * 1 + μ . Thus, this inequality contradicts with (100). The proof of Lemma 14 is complete. □
Theorem 5. 
Let 2 < p ≤ q < n p 2 ( n − p ) + 1 , u 0 ∈ W 0 1 , p ( D ) , v 0 ∈ L 2 ( D ) and (61) hold. Assume that u 0 ∈ N − and the initial data satisfy
μ E u 0 2 2 + 2 E ( u 0 , v 0 ) > E ε ( 0 ) C ̲ ,
where C ̲ = 1 2 min q − 2 2 q , C a q ( 1 + μ ) , C a = 2 q C * , and C * defined as in Lemma 14. Then, the solution u ( t ) to problem (1) blow up in finite time. Moreover, T m a x can be estimated by
T m a x ≤ 8 ( h 2 + k 0 ( q − 2 ) 2 E u 0 2 2 ) 1 / 2 + h k 0 ( q − 2 ) 2 ,
where h , k 0 are constants that will be determined in the proof.
Proof. 
First, from the definition of C ¯ and C ̲ , it follows that C ̲ ≤ C ¯ . This, combining with (104), implies that
μ E ∥ u 0 ∥ 2 2 + 2 E ( u 0 , v 0 ) > E ε ( 0 ) C ̲ ≥ E ε ( 0 ) C ¯ > 0 .
For any T > 0 , k > 0 and ξ > 0 , we define
F ( t ) = E ∥ u ( t ) ∥ 2 2 + μ E ∫ 0 t ∥ u ( τ ) ∥ 2 2 d τ + ( T − t ) μ E ∥ u 0 ∥ 2 + k ( t + ξ ) 2 , t ∈ [ 0 , T ] ,
Following an argument similar to that in the proof of Theorem 3, we obtain
F ″ ( t ) = − 2 E ∥ ∇ u ∥ p p + 2 E ∥ v ∥ 2 2 + 2 E ∫ D | u | q ln | u | d x + 2 k = 2 E ( ∥ v ∥ 2 2 − I ( u ) ) + 2 k
and
( F ′ ( t ) ) 2 ≤ 4 F ( t ) E v 2 2 + μ E ∫ 0 t v 2 2 + k .
Then from (105) and (106), we derive
F ( t ) F ″ ( t ) − q + 6 8 ( F ′ ( t ) ) 2 ≥ F ( t ) F ″ ( t ) − q + 6 2 E ∥ v ∥ 2 2 + μ E ∫ 0 t ∥ v ∥ 2 2 d τ + k = F ( t ) H ( t ) ,
where H t = − q + 2 2 E v 2 2 − 2 E I ( u ) − q + 6 2 μ E ∫ 0 t v 2 2 d τ − q + 2 2 k . By combining (11), (12) and (68) and Cauchy–Schwarz inequality, we arrive at
H ( t ) = q − 2 2 E v 2 2 + 2 q E u q q − 2 q E ε ( t ) − q + 6 2 μ E ∫ 0 t v 2 2 d τ − q + 2 2 k . ≥ q − 2 2 E ∥ v ∥ 2 2 + 2 q E ∥ u ∥ q q + 2 ( q p − 1 ) E ∇ u p p − 2 q E ε ( 0 ) − q + 2 2 k + q − 2 2 μ E ∫ 0 t ∥ v ∥ 2 2 d τ ≥ q − 2 2 E v 2 2 + C a 1 + μ E u 2 2 + C a μ 1 + μ E u 2 2 − 2 q E ε ( 0 ) − q + 2 2 k ≥ 2 q 1 2 min q − 2 2 q , C a q ( 1 + μ ) 2 E ( u , v ) + μ E ∥ u ∥ 2 2 − 2 q E ε ( 0 ) − q + 2 2 k = 2 q C ̲ 2 E ( u , v ) + μ E ∥ u ∥ 2 2 − E ε ( 0 ) C ̲ − q + 2 2 k ,
where C a = 2 q C * , C ̲ = 1 2 min q − 2 2 q , C a q ( 1 + μ ) . Choosing
k ∈ 0 , 4 q C ̲ q + 2 2 E u 0 , v 0 + μ E u 0 2 2 − E ( 0 ) C ̲ ,
then from Lemmas 13 and 14, we have
H t ≥ 2 q C ̲ 2 E u 0 , v 0 + E u 0 2 2 − E ( 0 ) C ̲ − q + 2 2 k ≥ 0 .
Thus, in view of (107) and (109) and the non-negativity of F ( t ) we obtain
F ( t ) F ″ ( t ) − q + 6 8 ( F ′ ( t ) ) 2 ≥ 0 .
Now, we choose
ξ > max 0 , 4 μ E ∥ u 0 ∥ 2 2 − ( q − 2 ) E ( u 0 , v 0 ) k ( q − 2 ) ,
which is independent of T, then
F ( 0 ) = E ∥ u 0 ∥ 2 2 + T μ E ∥ u 0 ∥ 2 2 + k ξ 2 > 0 ,
F ′ ( 0 ) = 2 E ( u 0 , v 0 ) + 2 k ξ > 0 .
Hence, from (112), (113) and by choosing θ = q − 2 8 in Lemma 2, we derive
8 F ( 0 ) ( q − 2 ) F ′ ( 0 ) = 4 E u 0 2 2 + T μ E u 0 2 2 + k ξ 2 ( q − 2 ) E u 0 , v 0 + k ξ ,
for enough large T. So, there exists a t * > 0 satisfying
t * ≤ F ( 0 ) θ F ′ ( 0 ) = 8 F ( 0 ) ( q − 2 ) F ′ ( 0 )
such that F ( t ) → ∞ as t → t * − . This contradicts with the assumption that F ( t ) is well defined on [ 0 , T ] for any T > 0 . Therefore, the solution u(t) blows up in finite time. Finally, we focus on the upper bound for the blow up time T m a x . For any T ∈ ( 0 , T m a x ) , we define F ¯ ( t ) similarly to the above F ( t ) by
F ¯ ( t ) = μ E ∫ 0 t ∥ u ( τ ) ∥ 2 2 d τ + E ∥ u ( τ ) ∥ 2 2 + ( T max − t ) μ E ∥ u 0 ∥ 2 2 + k ( t + ξ ) 2 , t ∈ [ 0 , T ] .
Similarly to the foregoing arguments, one can show that
T ≤ 4 [ E ∥ u 0 ∥ 2 2 + T max μ E ∥ u 0 ∥ 2 2 + k ξ 2 ] ( q − 2 ) [ E ( u 0 , v 0 ) + k ξ ] ,
where ξ still is defined in (111). It then follows the arbitrariness T < T m a x that
T m a x ≤ 4 [ E ∥ u 0 ∥ 2 2 + T max μ E ∥ u 0 ∥ 2 2 + k ξ 2 ] ( q − 2 ) [ E ( u 0 , v 0 ) + k ξ ] ,
which guarantees
T max ≤ T ξ , k = 4 E u 0 2 2 + k ξ 2 q − 2 E u 0 , v 0 + k ξ − 4 μ E u 0 2 2 .
Fix a k ∈ 0 , 4 q C ̲ q + 2 2 E ( u 0 , v 0 ) + μ E ∥ u 0 ∥ 2 2 − E ε ( 0 ) C ̲ . For ξ > max 0 , 4 μ E ∥ u 0 ∥ 2 2 − ( q − 2 ) E ( u 0 , v 0 ) k ( q − 2 ) , minimizing T ( ξ , k ) , we have
T min ξ , k = T ξ 0 , k = 8 h 2 + k ( q − 2 ) 2 E u 0 2 2 1 2 + h k q − 2 2 ,
where h = 4 μ E u 0 2 2 − q − 2 E u 0 , v 0 and ξ 0 = [ h 2 + k ( q − 2 ) 2 E ∥ u 0 ∥ 2 2 ] 1 / 2 + h k ( q − 2 ) .
Minimizing T ( ξ 0 , k ) for k ∈ 0 , 4 q C ̲ q + 2 2 E ( u 0 , v 0 ) + μ E ∥ u 0 ∥ 2 2 − E ε ( 0 ) C ̲ , we obtain
T min ξ 0 , k = T ξ 0 , k 0 = 8 h 2 + k 0 q − 2 2 E ∥ u 0 ∥ 2 2 1 2 + h k 0 q − 2 2 ,
where k 0 = 4 q C ̲ q + 2 2 E u 0 , v 0 + μ E ∥ u 0 ∥ 2 2 − E ε ( 0 ) C ̲ .
In conclusion,
T max ≤ 8 h 2 + k 0 ( q − 2 ) 2 E u 0 2 2 1 2 + h k 0 q − 2 2 .
□

Author Contributions

Formal analysis, F.L.; Writing—original draft, F.L. and Y.Z. (Yidi Zhang); Writing—review & editing, Y.Z. (Yizhuo Zhao). The article is a joint work of three authors who contributed equally to the final version of the paper. All authors have read and agreed to the published version of the manuscript.

Funding

Supported in part by China NSF Grant No. 42271309, and the Natural Science Basic Research Plan in Shaanxi Province of China No. 2025JC-YBMS-083.

Institutional Review Board Statement

Not applicable.

Data Availability Statement

No data sets were generated or analyzed during the current study.

Acknowledgments

The authors are indebted to the editor for giving some important suggestions which improved the presentations of this paper.

Conflicts of Interest

The authors declare that they have no competing interests.

References

  1. Boumaza, N.; Gheraibia, B.; Liu, G. Global well-posedness of solutions for the p-Laplacian hyperbolic type equation with weak and strong damping terms and logarithmic nonlinearity. Taiwan. J. Math. 2022, 26, 1235–1255. [Google Scholar]
  2. Wu, X.L.; Yang, X.X.; Cheng, L.B. Global well-posedness of wave equation with weak and strong damping terms p-Laplacian and logarithmic nonlinearity source term. J. Nonlinear Anal. Real. World Appl. 2023, 74, 103958. [Google Scholar]
  3. Yang, H.; Han, Y.Z. Blow-up for a damped p-Laplacian type wave equation with logarithmic nonlinearity. J. Differ. Equ. 2022, 306, 569–589. [Google Scholar] [CrossRef] [Scilit]
  4. Lian, W.; Xu, R.Z. Global well-posedness of nonlinear wave equation with weak and strong damping terms and logarithmic source term. Adv. Nonlinear Anal. 2020, 9, 613–632. [Google Scholar]
  5. Cao, Y.; Liu, C. Initial boundary value problem for a mixed pseudo-parabolic p-Laplacian type equation with logarithmic nonlinearity. Electron. J. Differ. Equ. 2018, 116, 1–19. [Google Scholar]
  6. Chen, H.; Tian, S. Initial boundary value problem for a class of semilinear pseudo-parabolic equations with logarithmic nonlinearity. J. Differ. Equ. 2015, 258, 4424–4442. [Google Scholar] [CrossRef] [Scilit]
  7. Le, C.N.; Le, X.T. Global solution and blow-up for a class of p-Laplacian evolution equations with logarithmic nonlinearity. Acta Appl. Math. 2017, 151, 149–169. [Google Scholar] [CrossRef] [Scilit]
  8. Le, C.N.; Le, X.T. Global solution and blow-up for a class of pseudo p-Laplacian evolution equations with logarithmic nonlinearity. Comput. Math. Appl. 2017, 73, 2076–2091. [Google Scholar]
  9. Sattinger, D.H. On global solution of nonlinear hyperbolic equations. Arch. Ration. Mech. Anal. 1968, 30, 148–172. [Google Scholar] [CrossRef] [Scilit]
  10. Liu, Y.C. On potential wells and vacuum isolating of solutions for semilinear wave equations. J. Differ. Equ. 2003, 192, 155–169. [Google Scholar] [CrossRef] [Scilit]
  11. Liu, Y.C.; Zhao, J.S. On potential wells and applications to semilinear hyperbolic equations and parabolic equations. Nonlinear Anal. 2006, 64, 2665–2687. [Google Scholar] [CrossRef] [Scilit]
  12. Liang, F. Explosive solutions of stochastic nonlinear beam equations with damping. J. Math. Anal. Appl. 2014, 419, 849–869. [Google Scholar] [CrossRef] [Scilit]
  13. Cheng, S.L.; Guo, Y.T.; Tang, Y.B. Stochastic viscoelastic wave equations with nonlinear damping and source terms. J. Appl. Math. 2014, 2014, 450289. [Google Scholar] [CrossRef] [Scilit]
  14. Benramdane, A.; Mezouar, N.; Alqawba, M.S.; Boulaaras, S.M.; Cherif, B.B. Blow-Up for a Stochastic Viscoelastic Lamé Equation with Logarithmic Nonlinearity. J. Funct. Spaces 2021, 2021, 9943969. [Google Scholar] [CrossRef] [Scilit]
  15. Bo, L.J.; Tang, D.; Wang, Y.J. Explosive solutions of stochastic wave equations with damping on. J. Differ. Equ. 2008, 244, 170–187. [Google Scholar] [CrossRef] [Scilit]
  16. Chow, P.L. Stochastic wave equations with polynomial nonlinearity. Ann. Appl. Probab. 2002, 12, 361–381. [Google Scholar] [CrossRef] [Scilit]
  17. Chow, P.L. Nonlinear stochastic wave equations: Blow-up of second moments in. Ann. Appl. Probab. 2009, 19, 2039–2046. [Google Scholar] [CrossRef] [Scilit]
  18. Brzézniak, Z.; Maslowski, B.; Seidler, J. Stochastic nonlinear beam equations, Probab. Theory Relat. Fields 2005, 132, 119–149. [Google Scholar]
  19. He, Y.; Gao, H.; Wang, H. Blow-up and decay for a class of pseudo-parabolic p-Laplacian equation with logarithmic nonlinearity. Comput. Math. Appl. 2018, 75, 459–469. [Google Scholar]
  20. Han, Y.; Gao, W.; Sun, Z.; Li, H. Upper and lower bounds of blow-up time to a parabolic type Kirchhoff equation with arbitrary initial energy. Comput. Math. Appl. 2018, 76, 2477–2483. [Google Scholar] [CrossRef] [Scilit]
  21. Levine, H.A. Some nonexistence and instability theorems for solutions of formally parabolic equations of the form Put = −Au + F(u). Arch. Ration. Mech. Anal. 1973, 51, 371–386. [Google Scholar]
  22. Levine, H.A. Remarks on the growth and nonexistence of solutions to nonlinear wave equations. In A Seminar on PDEs-1973; Rutgers University: New Brunswick, NJ, USA, 1973; pp. 59–70. [Google Scholar]
  23. Liu, G.W. The existence, general decay and blow-up for a plate equation with nonlinear damping and a logarithmic source term. Electron. Res. Arch. 2020, 28, 263–289. [Google Scholar] [CrossRef] [Scilit]
  24. Gao, H.; Liang, F.; Guo, B. Stochastic wave equations with nonlinear damping and source terms. Infin. Dimens. Anal. Quantum Probab. Relat. Top. 2013, 16, 1350013-1–1350013-29. [Google Scholar] [CrossRef] [Scilit]
Disclaimer/Publisher’s Note: The statements, opinions and data contained in all publications are solely those of the individual author(s) and contributor(s) and not of MDPI and/or the editor(s). MDPI and/or the editor(s) disclaim responsibility for any injury to people or property resulting from any ideas, methods, instructions or products referred to in the content.

Article Metrics

Citations

Article Access Statistics

Multiple requests from the same IP address are counted as one view.