In recent years, the small initial boundary value problem of the Kirchhoff-type wave system attracts many scholars’ attention. However, the big initial boundary value problem is also a topic of theoretical significance. In this paper, we devote oneself to the well-posedness of the Kirchhoff-type wave system under the big initial boundary conditions. Combining the potential well method with an improved convex method, we establish a criterion for the well-posedness of the system with nonlinear source and dissipative and viscoelastic terms. Based on the criteria, the energy of the system is divided into different levels. For the subcritical case, we prove that there exist the global solutions when the initial value belongs to the stable set, while the finite time blow-up occurs when the initial value belongs to the unstable set. For the supercritical case, we show that the corresponding solution blows up in a finite time if the initial value satisfies some given conditions.
National Natural Science Foundation of China1107103311101102PhD Start-Up Fund of Liaoning Province of China2014113720141139Natural Science Foundation of Liaoning Province20170540004Liaoning BaiQianWan Talents Program20139210551. Introduction
This paper studies the initial boundary value problem for the following Kirchhoff-type wave system:(1)utt-MtΔu+∫0tg1t-sΔuds+gut=f1u,vin Ω×0,+∞,(2)vtt-MtΔv+∫0tg2t-sΔuds+gvt=f2u,vin Ω×0,+∞,(3)ux,0=u0x,utx,0=u1xin Ω,(4)vx,0=v0x,vtx,0=v1xin Ω,(5)ux,t=vx,t=0on Γ1×0,+∞,where Ω is a bounded domain in Rn (n=1,2,3) with a smooth boundary ∂Ω, g(ut)=utp-1ut, g(vt)=vtq-1vt, M(r) is a nonnegative C1 function like M(t)=m0+α∇u2+∇v2γ, with m0≥0,α≥0,m0+α>0,γ>0, and g1,g2: R+→R+,fi(·,·): R2→R,i=1,2, are given functions which will be specified later.
1.1. Historical Research
Kirchhoff-type wave system with nonlinear source and dissipative and viscoelastic terms have various applications in the field of physics and mechanics, which is the model to describe the motion of deformable solids. A single Kirchhoff-type wave equation is proposed:(6)utt-M∇u2Δu+∫0tgt-sΔuds+hut=fuin Ω×0,+∞,and (6) has its roots for the small amplitude vibrations of a string when g=0 and n=1, but the tension of the string can not be ignored (see, e.g., Carrier et al. [1]). While (6) is used to describe the dynamics of an elastic string with fading memory when g≠0, this equation shows that the dynamic equilibrium of the object depends on both the present state of deformation and the history of the deformation gradient. Pohozaev and Tesei [2] proved that the solution exists in time if the datum satisfy an analytic-type condition for the case g=0. This result of the case g≠0 was extended by Torrejon and Yong [3]; they obtained the existence of weakly asymptotic stable solution. Later, Munoz Rivera [4] showed the existence of global solutions for small initial value and the exponential decay of the total energy. Then Wu and Tsai [5] established the global existence and energy decay under the assumption g′(t)≤-rg(t),h(ut)=-Δut. Recently, this decay estimate was improved for a weaker condition on g′(t)≤0 in [6].
Problem (6) is simplified to the following format without viscoelastic term:(7)utt-M∇u2Δu+hut=fuin Ω×0,+∞,and some results of (7) concerning global well-posedness have been established in [7–11] for the case of M≡1. The above problem without source and dissipative terms is called Kirchhoff-type equation when M is not a constant function, which was first introduced by Kirchhoff [12]; it describes the nonlinear vibrations of an elastic string. up to now, there are numerous results related to global well-posedness, including global existence, decay result, and blow-up properties; we refer the reader to [13–17].
Most recently Xu and Yang considered the initial boundary value problem of the following equation in [18]; they gave a blow-up result under supercritical energy:(8)utt-Δu+∫0tgt-τΔuτdτ-Δut-Δutt+ut=up-1u.
Wave system such as (1) and (2) goes back to Reed [19] who proposed a similar system, but it does not contain M(t)=m0+α∇u2+∇v2γ and the viscoelastic terms ∫0tg1(t-s)Δuds,∫0tg2(t-s)Δuds. Subsequently, concerning blow-up and nonexistence, results in wave systems were discussed. Agre and Rammaha [20] studied the following concrete system:(9)utt-Δu+utm-1ut=f1u,v,vtt-Δv+vtr-1vt=f2u,v,in Ω⊂Rn×(0,T)(n=1,2,3) with initial and boundary conditions of Dirichlet type, where f1(u,v) and f2(u,v) satisfy (A1) and (A2). They obtained several results concerning global well-posedness of a weak solution and showed that any weak solution blows up in finite time at negative initial energy. Afterwards, Alives et al. made further efforts as regards (9) in [21]. They obtained the global existence, uniform decay rates, and blow-up of solutions in finite time by involving the Nehari manifold when the initial energy is nonnegative and less than the mountain pass level value. And this blow-up result was improved by Said-Houari [22] when the initial data are large enough. In [23], Rammaha and Sakuntasathien studied a more general case of (9) by degenerating damping terms. Several results on the existence of local and global solutions as well as uniqueness are obtained by considering the constraint on the parameters of the system. Furthermore, they proved that the weak solutions blow up in finite time whenever the initial energy is negative and the exponent of the source term is more dominant than the exponents of both damping terms. Moreover, many studies of the global well-posedness for wave systems with dissipative terms have been researched in [24–28].
Wave systems with viscoelastic terms and dissipative terms have not been fully studied. In [29] the following coupled nonlinear wave equations with dispersive terms, viscoelastic dissipative terms, and nonlinear weak damping terms are considered:(10)utt-Δu-eΔutt+∫0tgt-sΔusds+utm-1ut=f1u,v,vtt-Δv-fΔvtt+∫0tht-sΔvsds+utr-1vt=f2u,v.A global nonexistence theorem for certain solutions with positive initial energy is proved. Reference [30] further considered a fourth-order wave system similar to (10). In that case, the energy increases exponentially when time goes to infinity and the initial data are large enough.
Recently [31] considered a system of two coupled wave equations with dispersive and strong dissipative terms under Dirichlet boundary conditions:(11)utjutt-Δu-Δutt-Δut=f1u,v,vtjvtt-Δv-Δvtt-Δvt=f2u,v,where (12)f1u,v=-au+vr-2u+v-bur/2-2vr/2u,f2u,v=-au+vr-2u+v-bvr/2-2ur/2v,r>2if n=1,2;2<r≤2n-1n-2if n≥3.The global existence of weak solutions and uniform decay rates (exponential one) of the solution energy were established.
Many researches considered the initial boundary value problem with global existence and blow-up of solutions for the nonlinear wave equations as follows:(13)utt-Δu+∫0tg1t-sΔuds+gut=f1u,vin Ω×0,+∞,vtt-Δv+∫0tg2t-sΔuds+gvt=f2u,vin Ω×0,+∞,where Ω is a bounded domain in Rn (n=1,2,3) with a smooth boundary ∂Ω. When the viscoelastic terms gi(i=1,2) are absent in (13), [20] showed local and global existences of a weak solution that any weak solution blows up in finite time with negative initial energy as the same way used in [22]. Later, Said-Houari extended this blow-up result to positive initial energy. At the same time, Liu [32] studied the following Cauchy problem for the coupled system of nonlinear Klein-Gordon equations with damping terms:(14)utt-Δu+u+γ1ut=vrur-2in Rn×0,+∞,vtt-Δv+v+γ1vt=urvr-2in Rn×0,+∞,and the existence of standing wave with ground state was stated and a sharp criterion for global existence and blow-up of solutions when E(0)<d was established. The author introduced a family of potential wells and the invariant sets and vacuum isolating behavior of solutions for 0<E(0)<d and E(0)≤0, respectively. Furthermore, he proved the global existence and asymptotic behavior of solutions when the condition is 0<E(0)<d. Finally, a blow-up result with arbitrarily positive initial energy was obtained.
Conversely, in the presence of the memory term (gi≠0,i=1,2), some results related to the asymptotic behavior and blow-up of solutions of viscoelastic systems were discussed. For example, [33] studied problem (13) with g(ut)=-Δut and g(vt)=-Δvt and obtained the fact that the decay rate of the energy function is exponential under suitable conditions on the functions gi,fi,i=1,2, certain initial data in the stable set. On the contrary, when the initial data is in the unstable set, the solutions blow up in finite time under positive initial energy. For g(ut)=utm-1ut and g(vt)=vtr-1vt, [26] established local and global existence as well as finite time blow-up (the initial energy E(0)<0). The latter blow-up result has been improved by [29] by considering a class of positive initial data. On the other hand, Messaoudi and Tatar [34] considered the following similar wave equations without considering coupling coefficient of Δu and Δv:(15)utt-Δu+∫0tg1t-sΔuds+f1u,v=0in Ω×0,+∞,vtt-Δv+∫0tg2t-sΔuds+f2u,v=0in Ω×0,+∞,where the functions f1 and f2 satisfy the following assumptions: (16)fu,v≤duβ1+vβ2,hu,v≤duβ3+vβ4, and for some constant d>0 and 1≤βi≤n/n-2,i=1,2,3,4, the solution goes to zero with an exponential or polynomial rate which depending on the decay rate of the relaxation functions gi,i=1,2 was obtained. This result was improved in [35] to weaker conditions on the relaxation functions and more general coupling functions.
Additionally, Liu and Wang [36] considered the following nonlinear hyperbolic systems with damping and source terms(17)utt-a+b∇u2+b∇v2Δu+gut=fu,in Ω×0,T,vtt-a+b∇u2+b∇v2Δv+gvt=hv,in Ω×0,T.The author defined potential well and the outer manifold of the potential well associated with system (17) and got the global existence in the case of E(0)<d1/2 and discussed the global nonexistence of solutions for problem (1)–(5) in the case of E(0)≤0 and E(0)<Cα on page 88 of [36]. In [37], (1)–(5) was considered with M≡1 and without imposing the memory terms (g=h=0). The rate of decay of the exponential or polynomial energy of the damping terms was obtained.
Recently, Wu [38] considered the initial boundary value problem for system (1)–(5) with M(t)=m0+α∇u2+∇v2γ. The assumptions for problem (1)–(5) are as follows:
The nonlinear source terms f1(u,v), f2(u,v) satisfy(18)f1=∂F∂u=m+1au+vm-1u+v+bum-3/2vm+1/2u,f2=∂F∂v=m+1au+vm-1u+v+bvm-3/2um+1/2v,
where Fu,v=au+vm+1+2buvm+1/2 with a,b>0.
M(s) is a nonnegative C1 function for s≥0 satisfying (19)Ms=m0+α∇u2+∇v2γ,m0≥m1,α≥0,γ>0.
The nonlinearity of m,p,q satisfies(20)m>1,if n=1,2or 1<m≤3,if n=3,(21)p,q≥1,if n=1,2,or 1<p,q≤5,if n=3.
For s≥0 the relaxation functions g1 and g2 are of class C1 and satisfy (22)g1s≥0,m0-∫0∞g1sds=l>0,g2s≥0,m0-∫0∞g2sds=k>0,g1′s≤0,g2′s≤0.
The solutions are global in time when the functions g1, g2, and fi,i=1,2, satisfy suitable conditions and certain initial data is in the stable set. The author established the rate of decay of solutions by a difference inequality given by Nakao [39] and intended to study the blow-up phenomena of problem (1)–(5). The blow-up of solutions when the energy is negative or subcritical case was proved by adopting and modifying the methods used in [29]. In this way, the above results in [38] allowed a bigger region for the blow-up results and improved the results of Messaoudi and Said-Houari [29]. More specifically, the decay result in [38] extends the one in [16, 37] to problem (1)–(5), where M is not a constant function and the equations considered in [38] have more dissipations.
1.2. Unsolved Problems
It is well known that in the absence of the nonlinear source term the damping term ensures global existence. In addition, without the dissipative term, the nonlinear source term causes finite time blow-up of solution. Moreover, the viscoelastic materials possess a capacity of storage and dissipation of mechanical energy; therefore, it is interesting to investigate the well-posedness of solution for the viscoelastic equation with dissipative term and nonlinear source term.
We can see that problem (1)–(5) contains system (9) (M≡1,g1=g2=0), systems (10), (13), and (15) (M≡1), system (11) (j=0,M≡1,g1=g2=0), and system (17) (γ=1,M≡1,g1=g2=0) as special cases. In fact, the viscoelastic terms change the frame of the equations comparing without those viscoelastic conditions, which makes the structure of the equations and solutions more complex and which also makes energy decay faster. So the classical methods can not be applied to investigate properties of solutions. Therefore, much less is known for (1)–(5) with viscoelastic terms. We can note that in all of the above studies except [38] only the global well-posedness of solutions was proved in a relatively rough variational framework, but the global existence and finite time blow-up for problem (1)–(5) at the other initial energy levels have not been discussed yet. The global existence and blow-up results in [38] are under the assumption of negative initial energy E(0)<0 or E(0)<E1=m-1/2(m+1)1/η(m+1)2/m-1. In other words, to our knowledge there are no results on global well-posedness of solutions to the initial boundary value problem for a couple of nonlinear wave equations with coupling coefficient M(t)=m0+α(∇u2+∇v2)γ of Δu and Δv, viscoelastic terms ∫0tg1(t-s)Δuds,∫0tg2(t-s)Δuds, and nonlinear weak dissipative terms utk-1ut, vtr-1vt. It is natural to ask a question of how the solution behaves for problem (1)–(5), which is what we want to deal with in this paper. Moreover, regarding the initial energy level, the present paper is also a comprehensive study for low energy case and high energy situation. To our knowledge this is the first try to consider this problem. The most attractive one is that we attain a blow-up result with arbitrary positive initial energy for a wave system of Kirchhoff type.
By reviewing above known results and also [19–38], we will face the fact that the following unsolved problems arise naturally. Firstly, from [38] we know the global existence for the definitely positive energy, but we know less for the initial energy which may be negative. Secondly, for general initial energy, which means the initial energy is not necessarily definitely positive, what will happen for the solution when the initial energy E(0)<d or E(0)>d?
We restricts our attention to considering the global existence and blow-up at two different initial energy levels. Since the initial energy level plays a crucial role in dealing with the well-posedness of problem (1)–(5), the two cases are, respectively, tackled with different tools. For the subcritical case E(0)<d, there have been many tools tackling the hyperbolic problem without viscoelastic terms in [16, 17]. We may refer the tools to deal with (1)–(5) with viscoelastic terms. By the well-known works [40–44], we see that the supercritical case E(0)>d is not easy to deal with. Filippo and Marco [45] made the initial attempt to consider the global well-posedness of hyperbolic problem at high initial energy level utt-Δu-ωΔut+μut=up-2u. However, they focused on particular source term |u|p-1u, and our work is on the viscoelastic terms condition and the complex source term f1(u,v),f2(u,v). Based on the general comparison principle in [9], we try to resolve the above open problems with variational methods. In this paper, we consider the initial boundary value problem for system (10) with e=f=0 and the nonlinear source terms, coupling coefficient M(s), the nonlinearity of m,p,q, and the relaxation functions g1 and g2 satisfying the assumptions (A1)–(A4), respectively. In addition, we consider nonlinear damping terms of the form |ut|k-1ut and vtr-1vt as in the first equation and in the second equation of (10), respectively.
1.3. The Main Results and Organization of the Paper
In this paper, we mainly discuss the following problems.
Case E(0)<d: different from the method applied in [38], we introduce a family of potential wells to obtain the results: invariant sets, global existence, and finite time blow-up.
Case E(0)>d: we obtain the finite time blow-up of solutions for problem (1)–(5) whose initial data have arbitrarily high initial energy.
We can summarize our main conclusions in Table 1 and use the question mark “?” to indicate the open problem.
Main results.
Main results
Initial data
Global existence
Blow-up
E0<d
Theorem 10
Theorem 14
E0>d
?
Theorem 18
The organization of the paper is as follows.
In Section 2, we introduce some notations, assumptions, and preliminaries.
From Sections 3–5, we prove the main results.
2. Notations and Primary Lemmas
In this section, we shall give some lemmas and some notations which will be used throughout this work. We use the standard Lebesgue space Lp(Ω) and Sobolev space H01(Ω) with their usual norms and products as follows: (23)uLpΩ=up,uL2Ω=u,u,v=∫Ωuvdx,m1=maxl1,k1,l1=∫0∞g1sds,k1=∫0∞g2sds,β=∇u2+∇v2γ+1,g∘ϕt=∫0tgt-s∫Ωϕs-ϕt2dxds.
We will use the embedding H01(Ω)↪Lp(Ω) for 2≤p≤2n/n-2, if n≥3 or 2≤p, if n=1,2. In this case, the embedding constant is denoted by c∗; that is, (24)up≤c∗∇u.
From assumption (A1) one can easily verify that (25)uf1u,v+vf2u,v=m+1Fu,v. Moreover we have the following result. Note that the following conclusion (Lemma 1) was assumed throughout many papers (see [16–31, 33, 34]); however in our opinion we think this conclusion is a deduction of assumption (A1). Thus we present this conclusion as follows and similar proof can be found in [29].
Lemma 1.
There exist two positive constants c0 and c1 such that(26)c0um+1+vm+1≤Fu,v≤c1um+1+vm+1,∀u,v∈R2.
Proof.
We can see that taking c1=2ma+b then the right-hand side of inequality (26) is trivial. For the left-hand side, the result is also trivial if u=v=0. If, without loss of generality, v≠0, then either u≤v or u>v.
For u≤v, we have (27)Fu,v=vm+1a1+uvm+1+2buvm+1/2. Consider the continuous function (28)js=a1+sm+1+2bsm+1/2,s∈-1,1. So minj(s)≥0. If minj(s)=0 then, for some s0∈[-1,1], we have (29)js0=a1+s0m+1+2bs0m+1/2=0. This implies that 1+s0=s0=0, which is impossible. Thus minj(s)=2c0>0. Therefore (30)Fu,v≥2c0vm+1≥2c0um+1. Consequently, (31)2Fu,v≥2c0vm+1+um+1, and then (32)Fu,v≥c0vm+1+um+1. If u≥v, similarly we have (33)Fu,v≥c0um+1+vm+1. This leads to the desired result and completes the proof of Lemma 1.
As in [29], we still have the following results.
Lemma 2 (see [29]).
Suppose that (20) holds. Then there exists a positive constant η>0 such that, for any (u,v)∈H01(Ω)×H01(Ω), one has(34)u+vm+1+2uvm+1/2m+1/2≤ηl∇u2+k∇v2m+1/2.
We also need the following technical lemma in the course of the investigation.
Lemma 3 (see [29]).
For any g1∈C1 and ϕ∈H1(0,T), one has(35)-2∫0t∫Ωgt-sϕϕtdxds=ddtg∘ϕt-∫0tgsdsϕ2+gtϕ2-g′∘ϕt,where g∘ϕt=∫0tgt-s∫Ωϕs-ϕt2dxds.
Now, we are in a position to state the local existence result to problem (1)–(5), which can be established by combining arguments of [15, 17, 20, 26].
Theorem 4 (local existence).
Let (u0,v0)∈H01(Ω)×H01(Ω) and (u1,v1)∈L2(Ω)×L2(Ω) be given. Assume that (A2)–(A4) are satisfied. Then there exists a couple solution (u,v) of problem (1)–(5) such that (36)u,v∈C0,T,H2Ω×H01Ω,ut∈C0,T,H01Ω∩Lp+1Ω,vt∈C0,T,H01Ω∩Lq+1Ω, for some T>0.
Remark 5 (see [46]).
Condition (A1) is necessary to guarantee the hyperbolicity of the equation in (1) and (2) and condition (21) is needed to establish the local existence result.
Next for problem (1)–(5) we introduce potential energy functional:(37)Et≡Eu,v=12ut2+12vt2+12m0-l1∇u2+12m0-k1∇v2+αβ2γ+1+12g1∘∇ut+12g2∘∇vt-∫ΩFu,vdx.Potential energy functional:(38)Jt≡Ju,v=12m0-l1∇u2+12m0-k1∇v2+αβ2γ+1+12g1∘∇ut+12g2∘∇vt-∫ΩFu,vdx.Nehari functional:(39)It≡Iu,v=m0-l1∇u2+m0-k1∇v2+αβ+g1∘∇ut+g2∘∇vt-m+1∫ΩFu,vdx.For the definition of F(u,v) please see assumption (A1) in the beginning of this paper. Moreover we introduce the potential well (stable set)(40)W=u,v∈H01Ω×H01Ω∣Iu,v>0∪0,0,and the outer space of potential well (unstable set)(41)V=u,v∈H01Ω×H01Ω∣Iu,v<0.Moreover we define (42)d=infu,v∈H01Ω×H01Ω∖0,0supλ≥0Jλu,λv, or equivalently (43)d=infu,v∈NJu,v, where N=u,v∈H01Ω×H01Ω∖0,0∣Iu,v=0.
Lemma 6 (depth of potential well).
The depth of potential well d=((m-1)(m0-m1)/2(m+1))m0-m1/c1(m+1)C∗m+12/m-1, where c1 is defined in (26) and C∗ is the best imbedding constant from H01(Ω) into Lm+1(Ω).
Proof.
From the definition of d, we have (u,v)∈N; that is, I(u,v)=0. Then on the one hand from Lemma 1 we get (44)m0-l1∇u2+m0-k1∇v2+αβ+g1∘∇ut+g2∘∇vt=m+1∫ΩFu,vdx≤c1m+1um+1m+1+vm+1m+1≤c1m+1C∗m+1∇u2+∇v2m+1/2.Notice that, from assumption (A4) and the definitions β and m1, we have (45)m0-m1∇u2+∇v2≤c1m+1C∗m+1∇u2+∇v2m+1/2;that is(46)∇u2+∇v2≥m0-m1c1m+1C∗m+12/m-1.
On the other hand notice I(u,v)=0; moreover by virtue of (38), (39), and (78), we get (47)Ju,v=12-1m+1m0-l1∇u2+m0-k1∇v2+α2γ+1-αm+1β+12-1m+1g1∘∇ut+12-1m+1g2∘∇vt+1m+1Iu,v≥12-1m+1m0-l1∇u2+m0-k1∇v2≥12-1m+1m0-m1∇u2+∇v2≥12-1m+1m0-m1m0-m1c1m+1C∗m+12/m-1, and hence we have d=1/2-1/m+1(m0-m1)m0-m1/c1(m+1)C∗m+12/m-1.
Lemma 7 (nonincreasing energy).
Let (u,v) be a solution of problem (1)–(5); then E(t) is a nonincreasing function for t≥0; that is,(48)ddtEt=-utp+1p+1-vtq+1q+1+12g1′∘∇ut+12g2′∘∇vt-12g1t∇u2-12g2t∇v2≤0,∀t≥0.
Proof.
Multiplying (1) by ut and (2) by vt, integrating them over Ω, and then adding the results together and integrating by parts, it follows that (49)ddt12ut2+vt2+∇u2+∇v2+αγ+1∇u2+∇v2γ+1-∫ΩFu,vdx=-utp+1p+1-vtq+1q+1+∫0t∫Ωg1t-s∇us·∇utdxds+∫0t∫Ωg2t-s∇vs·∇vtdxds. Exploiting Lemma 3 on the third term and the fourth term on the right side of the above equality, we can obtain (48) for any regular solution.
3. Global Existence under the Case E0<d
Now we give the following definition of weak solution for problem (1)–(5).
Definition 8 (weak solution).
A function (u,v) is called a weak solution of problem (1)–(5) on Ω×[0,T], if it satisfies u,v∈L∞0,T,H01Ω×H01Ω with (ut,vt)∈L∞0,T,L2Ω×L2Ω∩L∞0,T,Lr+1Ω×Lr+1Ω and(50)ut,ω1-∫0tm0+α∇u2+∇v2γΔu,ω1dτ-∫0t∫0σg1σ-τ∇uτ,∇w1dτdσ+∫0tutp-1ut,ω1dτ=∫0tf1u,v,ω1dτ+u1,ω1,∀ω1∈H01Ω,vt,ω2-∫0tm0+α∇u2+∇v2γΔv,ω2dτ-∫0t∫0σg2σ-τ∇uτ,∇w2dτdσ+∫0tvtq-1vt,ω2dτ=∫0tf2u,v,ω2dτ+v1,ω2,∀ω2∈H01Ω,with (51)u0,x=u0x,ut0,x=u1x,v0,x=v0x,vt0,x=v1x.
Lemma 9 (invariant set W).
Let (u0,v0)∈H01(Ω)×H01(Ω), (u1,v1)∈L2(Ω)×L2(Ω), and (A1)–(A4) hold. Then all solutions of problem (1)–(5) with E(0)<d belong to W, provided (u0,v0)∈W.
Proof.
Let (u(t),v(t)) be any local weak solution of problem (1)–(5) with E(0)<d and (u0,v0)∈W and T be the existence time of (u(t),v(t)). Then it follows from Lemma 7 that E(u(t),v(t))≤E(0)<d. Thus if suffices to show that I(u(t),v(t))>0 for 0<t<T. Suppose that there exists t1∈(0,T) such that I(u(t1),v(t1))≤0. From the continuity of the solution in time, there exists t∗∈(0,T) such that I(u(t∗),v(t∗))=0. Then from the definition of d we have (52)d≤Jut∗,vt∗≤Eut∗,vt∗≤E0<d,which is a contradiction.
Then we give the global existence of solutions for problem (1)–(5) with low initial energy level E(0)<d.
Theorem 10 (global existence when E0<d).
Let (u0,v0)∈H01(Ω)×H01(Ω), (u1,v1)∈L2(Ω)×L2(Ω), and (A1)–(A4) hold. Assume that E(0)<d and (u0,v0)∈W. Then problem (1)–(5) admits a global weak solution u(t),v(t)∈L∞0,T;H01(Ω), ut(t),vt(t)∈L∞0,T;L2(Ω), and (u,v)∈W for 0≤t≤∞.
Proof.
Let {ωj} be a basis in H01(Ω) given by the eigenfunction of the operator -Δ and it constructs a complete orthogonal system such that ωj=1 for all j. Then {ωj} is orthogonal and complete in L2(Ω) and in H01(Ω). Let Vm be the space generated by {ω1,ω2,…,ωm}, m∈N. Construct the approximate solutions of problem (1)–(5):(53)umx,t=∑j=1mgjmtwjx,m=1,2,…,vmx,t=∑j=1mhjmtwjx,m=1,2,…, satisfying(54)umttt,ω+m0+α∇um2+∇vm2γ,∇ω-∫0tg1t-τ∇umτ,∇ωdτ+umtp-1umt,ω=f1umt,vmt,ω,∀ω∈Vm,(55)vmttt,ω+m0+α∇um2+∇vm2γ,∇ω-∫0tg2t-τ∇vmτ,∇ωdτ+vmtq-1vmt,ω=f2umt,vmt,ω,∀ω∈Vm,(56)um0=u0m=∑j=1mu0,ωjωj⟶u0in H01Ω,vm0=v0m=∑j=1mv0,ωjωj⟶v0in H01Ω,(57)umt0=u1m=∑j=1mu1,ωjωj⟶u1in L2Ω,vmt0=v1m=∑j=1mv1,ωjωj⟶v1in L2Ω.Multiplying (54) and (55) by gsm′t, hsm′t, respectively, and summing for s and adding these two equations, we can deduce(58)ddtEumt,vmt=12g1′∘∇umt-umtp+1p+1-12g1t∇um2+12g2′∘∇vmt-12g2t∇vm2-vmtq+1q+1. Integrating the above equation with respect to t, we have(59)Emt+∫0tumτp+1p+1+12g1τ∇um2-12g1′∘∇umτdτ+∫0tvmτq+1q+1+12g2τ∇vm2-12g2′∘∇vmτdτ=Em0,where(60)Emt≔12umt2+12vmt2+12m0-l1∇um2+12m0-k1∇vm2+αβ2γ+1+12g1∘∇umt+12g2∘∇vmt-∫ΩFum,vmdx=12umt2+vmt2+Jum,vm,From (u0,v0)∈H01(Ω)×H01(Ω), (56), and (57) we get that as m→∞(61)umt0⟶u1,vmt0⟶v1,∇um0⟶∇u0,∇vm0⟶∇v0.Therefore we have Em(0)→E(0) as m→0. Then for sufficiently large m we have(62)12umt2+vmt2+Jum,vm+∫0tumτp+1p+1+12g1τ∇um2-12g1′∘∇umτdτ+∫0tvmτq+1q+1+12g2τ∇vm2-12g2′∘∇vmτdτ<d.Note that(63)Ju,v=12-1m+1m0-l1∇u2+m0-k1∇v2+α2γ+1-αm+1β+12-1m+1g1∘∇ut+12-1m+1g2∘∇vt+1m+1Iu,v.Hence, from (62) and (63), we get(64)12umt2+vmt2+α2γ+1-αm+1β+1m+1Ium,vm+12-1m+1m0-l1∇um2+m0-k1∇vm2+12-1m+1g1∘∇umt+12-1m+1g2∘∇vmt<d.By (u0,v0)∈W and(65)12umt02+vmt02+Jum0,vm0=Em0,taking into account (56) and (57), we can get (um(0),vm(0))∈W for sufficiently large m. From (62) and an argument similar to the proof of Lemma 9 we can prove that (um(t),vm(t))∈W for 0≤t<∞ and sufficiently large m. Thus (64) gives(66)12umt2+vmt2+α2γ+1-αm+1β+12-1m+1m0-l1∇um2+m0-k1∇vm2+12-1m+1g1∘∇umt+12-1m+1g2∘∇vmt<d,for sufficiently large m and t∈[0,∞). Inequality (66) gives(67)um and vm are both bounded in L∞0,∞;H01Ω,(68)umt and vmt are both bounded in L∞0,∞;L2Ω.Furthermore, according to (68), the following results hold:(69)umtp-1umt is bounded in L∞0,∞;LrΩ,where r=p+1p,(70)vmtq-1vmt is bounded in L∞0,∞;LrΩ,where r=q+1q,(71)umm-1um and vmm-1vm are both bounded in L∞0,∞;LrΩ,where r=m+1m.Hence integrating (54) and (55) with respect to t, for every s∈H01(Ω) and 0≤t<∞, we have(72)umt,ws-∫0tm0+α∇um2+∇vm2γΔum,wsdτ-∫0t∫0σg1σ-τ∇umτ,∇wsdτdσ+∫0tumtp-1umt,wsdτ=∫0tf1um,vm,wsdτ+u1,ws,vmt,ws-∫0tm0+α∇um2+∇vm2γΔvm,wsdτ-∫0t∫0σg2σ-τ∇umτ,∇wsdτdσ+∫0tvmtq-1vmt,wsdτ=∫0tf2um,vm,wsdτ+v1,ws.Therefore, up to a subsequence, by (67)–(71), we may pass to the limit in (72) and obtain a weak solution (u,v) of problem (1)–(5) with the above regularity (67)–(71) and (50). On the other hand, from (56) and (57) we have u(x,0),v(x,0)=u0(x),v0(x) in H01(Ω)×H01(Ω) and (ut(x,0),vt(x,0))=(u1(x),v1(x)) in L2(Ω)×L2(Ω).
4. Finite Time Blow-Up When E0<d
Let us turn to discuss blow-up properties of solutions for system (1)–(5) when E(0)<d, g(ut)=ut,g(vt)=vt. We firstly give the following definition of finite time blow-up of weak solution for problem (1)–(5).
Definition 11 (finite time blow-up).
A solution (u,v) of problem (1)–(5) is called a blow-up solution if there exists a finite time T such that(73)limsupt→T-∫Ωu2+v2dx=∞.
By the same argument as Lemma 9, we can get the following lemma.
Lemma 12 (invariant set V).
Let (u0,v0)∈H01(Ω)×H01(Ω), (u1,v1)∈L2(Ω)×L2(Ω), and (A1)–(A4) hold. Then all solutions of problem (1)–(5) with E(0)<d belong to V, provided (u0,v0)∈V.
In order to prove Theorem 14 we state some relations of the depth of potential well d, norm ∇u2+∇v2, and function F(u,v) as follows.
Lemma 13.
Under the assumptions of Lemma 15, one has(74)d<m-1m0-m12m+1∇u2+∇v2.
Proof.
From Lemma 6 for the depth of potential well d, we have (75)d=m-1m0-m12m+1m0-m1c1m+1C∗m+12/m-1, where c1 is defined in (26) and C∗ is the best imbedding constant from H01(Ω) into Lm+1(Ω). By Lemma 15, we get (u,v)∈V; that is, I(u,v)<0. Moreover by Lemma 1 and Sobolev embedding inequality, I(u,v)<0 implies that (76)m0-l1∇u2+m0-k1∇v2+αβ+g1∘∇ut+g2∘∇vt<m+1∫ΩFu,vdx≤c1m+1um+1m+1+vm+1m+1≤c1m+1C∗m+1∇u2+∇v2m+1/2. Notice that, from assumption (A4) and the definitions β and m1, we have (77)m0-m1∇u2+∇v2<c1m+1C∗m+1∇u2+∇v2m+1/2; that is(78)∇u2+∇v2>m0-m1c1m+1C∗m+12/m-1.Hence we obtain (79)d<m-1m0-m12m+1∇u2+∇v2.
A finite time blow-up result of solutions for problem (1)–(5) is showed as follows.
Theorem 14 (finite time blow-up when E(0)<d).
Let (u0,v0)∈H01(Ω)×H01(Ω), (u1,v1)∈L2(Ω)×L2(Ω), and (A1)–(A4) hold; 1<p<m, 1<q<m hold. Assume that E(0)<ζd(ζ<1), (u0,v0)∈V, and m satisfy(80)m1≤m-11-ζm0m-11-ζ+1/m+1.Then the existence time of solution of problem (1)–(5) is finite.
Proof.
Let (u,v) be any solution of problem (1)–(5) with E(0)<d and (u0,v0)∈V. Next, we prove the solution of problem (1)–(5) blows up in finite time. Suppose by contradiction that the solution (u(t),v(t)) is global. Then, for any T0>0, we define a auxiliary function F(t) by(81)Ft=u2+v2+∫0tu2+v2dτ+T0-tu02+v02.Clearly F(t)>0 for all t∈[0,T0]. From the continuity of F(t) in t, it is easy to see that there exists ρ>0 (independent of the choice of T0) such that(82)Ft≥ρ,∀t∈0,T0.Then for t∈[0,T0] we have(83)F′t=2u,ut+2v,vt-u02+v02+u2+v2=2u,ut+2v,vt+2∫0tuτ,uττ+vτ,vττdτ,(84)F′′t=2ut2+vt2+2u,utt+2v,vtt+2u,ut+2v,vt=2ut2+vt2-2m0+α∇u2+∇v2γ∇u2+2∫0tg1t-τ∫Ω∇ut∇uτdxdτ+2f1u,v,u-2m0+α∇u2+∇v2γ∇v2+2∫0tg2t-τ∫Ω∇vt∇vτdxdτ+2f2u,v,v.Applying Young’s inequality to estimate the fourth term on the right side of (84), we have (85)2∫0tg1t-τ∫Ω∇ut∇uτdxdτ=2∫0tg1t-τ∇ut2dτ+2∫0tg1t-τ∫Ω∇ut∇uτ-∇utdxdτ≥2∫0tg1t-τ∇ut2dτ-2∫0tg1t-τ∇ut∇uτ-∇utdτ≥2∫0tg1t-τ∇ut2dτ-2η1g1∘∇ut-l12η1∇u2, for any η1>0. Similarly we have (86)2∫0tg2t-τ∫Ω∇vt∇vτdxdτ≥2∫0tg2t-τ∇vt2dτ-2η2g2∘∇vt-k12η2∇v2, for any η2>0. Then (84) arrives at(87)F′′t≥2ut2+vt2-2m0+α∇u2+∇v2γ∇u2+2∫0tg1t-τ∇ut2dτ-2η1g1∘∇ut-l12η1∇u2-2m0+α∇u2+∇v2γ∇v2+2∫0tg2t-τ∇vt2dτ-2η2g2∘∇vt-k12η2∇v2+2f1u,v,u+2f2u,v,v=2ut2+vt2-2m0+α∇u2+∇v2γ-l1∇u2-2η1g1∘∇ut-l12η1∇u2-2m0+α∇u2+∇v2γ-k1∇v2-2η2g2∘∇vt-k12η2∇v2+2f1u,v,u+2f2u,v,v=2ut2+vt2-2m0-l1∇u2-2m0-k1∇v2-2η1g1∘∇ut-l12η1∇u2-2αβ-2η2g2∘∇vt-k12η2∇v2+2m+1∫ΩFu,vdx.On the other hand from (83), we have(88)F′t2=4u,ut+v,vt2+4∫0tuτ,uττ+vτ,vττdτ2+8u,ut+v,vt∫0tuτ,uττ+vτ,vττdτ. Using the Schwarz inequality, (88) takes on the form (89)u,ut+v,vt2≤u2+v2ut2+vt2,∫0tuτ,uττdτ+∫0tvτ,vττdτ2≤∫0tu2+v2dτ∫0tuτ2+vτ2dτ,2u,ut+v,vt∫0tuτ,uττ+vτ,vττdτ≤uτ2+vτ2∫0tu2+v2dτ+u2+v2∫0tuτ2+vτ2dτ, and therefore (88) becomes(90)F′t2≤4u2+v2+∫0tu2+v2dτ·ut2+vt2+∫0tuτ2+vτ2dτ≤4Ftut2+vt2+∫0tuτ2+vτ2dτ.Then by (81), (87), and (90), we have (91)F′′tFt-p+34F′t2≥FtF′′t-p+3ut2+vt2+∫0tuτ2+vτ2dτ≥Ft2ut2+vt2-2m0-l1∇u2-2m0-k1∇v2-Ft2η1g1∘∇ut+l12η1∇u2+2αβ-Ft2η2g2∘∇vt+k12η2∇v2-2m+1∫ΩFu,vdx-Ftm+3ut2+vt2+∫0tuτ2+vτ2dτ. Now we define(92)ξt≔2ut2+vt2-2m0-l1∇u2-2m0-k1∇v2-2η1g1∘∇ut-l12η1∇u2-2αβ-2η2g2∘∇vt-k12η2∇v2+2m+1∫ΩFu,vdx-m+3ut2+vt2+∫0tuτ2+vτ2dτ.From the definition of E(t), (92) becomes(93)ξt≔m-1m0-l1∇u2+m-1m0-k1∇v2+m+1-2η1g1∘∇ut+m+1-2η2g2∘∇vt-k12η2∇v2-l12η1∇u2+m+1r+1-2αβ-m+3∫0tuτ2+vτ2dτ-2m+1Et.Then from Lemma 2 with p=q=1, (93) arrives at(94)ξt≔m-1m0-l1∇u2+m-1m0-k1∇v2+m+1-2η1g1∘∇ut-l12η1∇u2+m+1-2η2g2∘∇vt-k12η2∇v2+m+1r+1-2αβ-2m+1E0+m+1∫0tg1s∇us2+g2s∇vs2ds-m+1∫0tg1′∘∇us+g2′∘∇vsds.From assumption (A4) on g1 and g2 we can derive(95)ξt≥m-1m0-l1∇u2+m-1m0-k1∇v2+m+1-2η1g1∘∇ut-l12η1∇u2+m+1-2η2g2∘∇vt-k12η2∇v2+m+1r+1-2αβ-2m+1E0=m-1m0-m-1+12η1l1∇u2+m-1m0-m-1+12η2k1∇v2+m+1-2η1g1∘∇ut+m+1-2η2g2∘∇vt+m+1r+1-2αβ-2m+1E0=m-1m0-m-1+12η1l1∇u2+m-1m0-m-1+12η2k1∇v2+m+1-2η1g1∘∇ut+m+1-2η2g2∘∇vt-ζm-1m0-m1∇u2+∇v2+ζm-1m0-m1∇u2+∇v2-2m+1ζd+m+1r+1-2αβ+2m+1ζd-2m+1E0=ξ1+ξ2+ξ3,where (96)ξ1=m-1m0-m-1+12η1l1∇u2+m-1m0-m-1+12η2k1∇v2+m+1-2η1g1∘∇ut+m+1-2η2g2∘∇vt-ζm-1m0-m1∇u2+∇v2,ξ2=ζm-1m0-m1∇u2+∇v2-2m+1ζd,ξ3=m+1r+1-2αβ+2m+1ζd-2m+1E0. We next estimate the terms ξ1, ξ2, and ξ3 one by one as follows. For the term ξ1 from (97)m1=maxl1,k1, we have(98)ξ1=m-1m0-m-1+12η1l1∇u2+m-1m0-m-1+12η2k1∇v2+m+1-2η1g1∘∇ut+m+1-2η2g2∘∇vt-ζm-1m0-m1∇u2+∇v2≥m-1m0-m-1+12η1m1∇u2+m-1m0-m-1+12η2m1∇v2+m+1-2η1g1∘∇ut+m+1-2η2g2∘∇vt-ζm-1m0-m1∇u2+∇v2=m-11-ζm0-m-11-ζ+12η1m1∇u2+m-11-ζm0-m-11-ζ+12η2m1∇v2+m+1-2η1g1∘∇ut+m+1-2η2g2∘∇vt.Here by taking 2η1=m+1 and 2η2=m+1 and by (80) we have (99)ξ1>0. From Lemma 16 we can derive (100)ξ2>0. With the fact E0<ζd we have (101)ξ3>0. So from (95) we have ξ(t)>σ1>0. Therefore we can derive (102)F′′tFt-p+34F′t2≥ρσ1>0,t∈0,T0. Setting y(t)=Ft-p-1/4, this inequality becomes (103)y′′t≤-p-14σ1ρytp+7/p-1,t∈0,T0. This proves that y(t) reaches 0 in finite time, say t→T∗. Since T∗ is independent of the initial choice of T0, we may assume that T∗<T0. This tells us that (104)limt→T∗Ft=+∞, which completes the proof.
5. A Finite Time Blow-Up When E0>0
We first present the following lemmas in order to prove Theorem 18.
Lemma 15.
Let condition (A4) hold and the nonlinear viscoelastic terms g1 and g2 satisfy(105)∫0tws∫0ses-τ/2g1s-τwτdτds≥0,∀w∈C10,∞,∀t>0,(106)∫0tws∫0ses-τ/2g2s-τwτdτds≥0,∀w∈C10,∞,∀t>0.If H(t) is a twice continuously differentiable function and satisfies the inequality(107)H′′t+H′t>∫0tg1t-τ∫Ω∇ux,τ∇ux,tdxdτ+∫0tg2t-τ∫Ω∇vx,τ∇vx,tdxdτand the initial condition(108)H0>0,H′0>0,for every t∈[0,T0), where (u(t),v(t)) is the corresponding solutions of problem (1)–(5) with (u0,v0) and (u1,v1), then the function H(t) is strictly increasing on [0,T0).
Proof.
Consider the following auxiliary ordinary differential equation:(109)h′′t+h′t=∫0tg1t-τ∫Ω∇ux,τ∇ux,tdxdτ+∫0tg2t-τ∫Ω∇vx,τ∇vx,tdxdτ,with the initial condition(110)h0=H0,h′0=0,for every t∈[0,T0).
Clearly we can find the following function:(111)ht=h0+∫0te-ξ-e-te-ξ∫0ξg1ξ-τ∫Ω∇ux,τ∇ux,ξdxdτdξ+∫0te-ξ-e-te-ξ∫0ξg2ξ-τ∫Ω∇vx,τ∇vx,ξdxdτdξ,as a solution of the ODE (109) and (110) for every t∈[0,T0).
Now in order to show that (112)H′t>0,t≥0, we need to prove that(113)H′t>h′t≥0,t≥0.From (105) a direct computation on (111) yields(114)h′t=∫0teξ-t∫0ξg1ξ-τ∫Ω∇ux,τ∇ux,ξdxdτdξ+∫0teξ-t∫0ξg2ξ-τ∫Ω∇vx,τ∇vx,ξdxdτdξ=e-t∫Ωeξ∫0t∫0ξg1ξ-τ∇ux,τ∇ux,ξdτdξdx+e-t∫Ωeξ∫0t∫0ξg2ξ-τ∇vx,τ∇vx,ξdτdξdx=e-t∫Ω∫0t∫0ξeξ-τ/2g1ξ-τeτ/2∇ux,τeξ/2∇ux,ξdτdξdx+e-t∫Ω∫0t∫0ξeξ-τ/2g2ξ-τeτ/2∇vx,τeξ/2∇vx,ξdτdξdx=e-t∫Ω∫0teξ/2∇ux,ξ∫0ξeξ-τ/2g1ξ-τeτ/2∇ux,τdτdξdx+e-t∫Ω∫0teξ/2∇vx,ξ∫0ξeξ-τ/2g2ξ-τeτ/2∇vx,τdτdξdx≥0,for every t∈[0,T0), which says that (115)ht≥h0=H0. Moreover from (108) and (114) it implies that (116)H′0>0=h′0. Suppose by contradiction that the first inequality of (113) is invalid; then there exists t1∈[0,T0) such that (117)H′t1≤h′t1. From the continuity of the solution in time, there exists t0∈[0,T0) such that(118)H′t0=h′t0.On the other hand we have the following ordinary differential inequality (119)H′′t-h′′t+H′t-h′t>0,H0-h0=0,H′0-h′0>0, for every t∈[0,T0). This ordinary differential inequality can be solved as (120)H′t0-h′t0>e-t0H′0-h′0>0, which contradicts (118). Thus we prove the first inequality of (113), which together with (114) states (113). So we complete this proof.
Lemma 16.
Let (u0,v0)∈H01(Ω)×H01(Ω), (u1,v1)∈L2(Ω)×L2(Ω), p=q=1, and (u,v) be the solution of problem (1)–(5) with the initial data (u0,v0) and (u1,v1). Assume the initial data satisfy(121)u0,u1+v0,v1≥0;then the map (122)t⟼ut2+vt2 is strictly increasing as long as (u,v)∈V.
Proof.
Let(123)Ht=u2+v2;then we have(124)H′t=2u,ut+2v,vt,(125)H′′t=2ut2+vt2+2u,utt+2v,vtt=2ut2+vt2-2u,ut-2v,vt+2∫0tg1t-τ∫Ω∇ut∇uτdxdτ+2∫0tg2t-τ∫Ω∇vt∇vτdxdτ-2m0+α∇u2+∇v2γ∇u2-2m0+α∇u2+∇v2γ∇v2+2f1u,v,u+2f2u,v,v=2ut2+vt2-2u,ut-2v,vt-2Iu,v+2∫0tg1t-τ∫Ω∇ut∇uτdxdτ+2∫0tg2t-τ∫Ω∇vt∇vτdxdτ.Adding (124) and (125) we have(126)H′′t+H′t=2ut2+vt2-2Iu,v+2∫0tg1t-τ∫Ω∇ut∇uτdxdτ+2∫0tg2t-τ∫Ω∇vt∇vτdxdτ,which, from the fact that (u,v)∈V, implies that(127)H′′t+H′t≥2∫0tg1t-τ∫Ω∇ut∇uτdxdτ+2∫0tg2t-τ∫Ω∇vt∇vτdxdτ.Therefore, applying Lemma 15 with the fact that (128)H′0=2∫Ωu0u1dx+2∫Ωv0v1dx≥0, we can obtain that the map (129)t⟼ut2+vt2 is strictly increasing.
In the following, we show the invariance of the unstable set under the flow of the problem (1)–(5).
Lemma 17.
Let (u0,v0)∈H01(Ω)×H01(Ω), (u1,v1)∈L2(Ω)×L2(Ω), p=q=1, and (u,v) be the solution of problem (1)–(5) with the initial data (u0,v0) and (u1,v1). Assume that the nonlinear viscoelastic terms g1 and g2 satisfy(130)m1<m-1m0m-1+1/m+1,and the initial data satisfy (121) and(131)u02+v02>2m+1ACE0,where (132)A=m-1m0-m-1+1m+1m1,C=minC1,C2.C1 is the coefficient of Poincaré inequality ∇u2≥C1u2 and C2 is the coefficient of Poincaré inequality ∇v2≥C2v2. Then all solutions of problem (1)–(5) with E(0)>0 belong to V, provided I(u0,v0)<0.
Proof.
We prove (u(t),v(t))∈V. If it is false, let t0∈(0,T) be the first time such that I(u(t),v(t))=0; that is, I(u(t),v(t))<0, for t∈[0,t0) and I(u(t0),v(t0))=0. Now let H(t) be defined by (123) above. Hence from Lemma 16, we get that H(t) and H′(t) are strictly increasing on the interval (0,t0). And then by (131), we have (133)Ht>v02+u02>2m+1ACE0,∀t∈0,t0.
Moreover, from the continuity of u(t) in t, we obtain(134)Ht0>2m+1ACE0.On the other hand, by (37) and (39) we can obtain(135)E0≥Et0=12utt02+12vtt02+αβt02γ+1+12m0-l1∇ut02+12m0-k1∇vt02+12g1∘∇ut0+12g2∘∇vt0-∫ΩFut0,vt0dx=12utt02+12vtt02+α2γ+1-αm+1βt0+12-1m+1m0-l1∇ut02+m0-k1∇vt02+12-1m+1g1∘∇ut0+12-1m+1g2∘∇vt0+1m+1Iut0,vt0.Note that Iut0,vt0=0; hence we have(136)E0≥12utt02+12vtt02+α2γ+1-αm+1βt0+12-1m+1m0-l1∇ut02+m0-k1∇vt02+12-1m+1g1∘∇ut0+12-1m+1g2∘∇vt0≥12-1m+1m0-l1∇ut02+m0-k1∇vt02≥12-1m+1m0-m1∇ut02+∇vt02.Using the Poincaré inequality (137)∇u2≥C1u2,∇v2≥C2v2,we have(138)∇u2+∇v2≥C1u2+C2v2≥Cu2+v2.By (138), we deduce (136) to(139)E0≥12-1m+1m0-m1∇ut02+∇vt02≥12-1m+1m0-m1Cut02+vt02≥12-1m+1m0-m1Cut02+vt02-m12m+12Cut02+vt02=m-1m0-m-1m1-m1/m+12m+1Cut02+vt02=AC2m+1ut02+vt02,which means(140)Ht0=ut02+vt02≤2m+1ACE0.It is obvious that (140) contradicts (131).
Theorem 18 (finite time blow-up under the case of E(0)>0 and p=q=1).
Let (u0,v0)∈H01(Ω)×H01(Ω), (u1,v1)∈L2(Ω)×L2(Ω), and (A1)–(A4) hold. Assume that the nonlinear viscoelastic terms g1 and g2 satisfy (105), (106), and (130) and the initial data satisfy (121), (131), and (u0,v0)∈V. Then the solution of problem (1)–(5) with p=q=1 and E(0)>0 blows up in finite time.
Proof.
Recalling the auxiliary function F(t) defined as (81) and the proof of Theorem 14, we have(141)ξt≥m-1m0-l1∇u2+m-1m0-k1∇v2+m+1-2η1g1∘∇ut-l12η1∇u2+m+1-2η2g2∘∇vt-k12η2∇v2+m+1r+1-2αβ-2m+1E0=m-1m0-m-1+12η1l1∇u2+m-1m0-m-1+12η2k1∇v2+m+1-2η1g1∘∇ut+m+1-2η2g2∘∇vt+m+1r+1-2αβ-2m+1E0≥m-1m0-m-1+12η1m1∇u2+m-1m0-m-1+12η2m1∇v2+m+1-2η1g1∘∇ut+m+1-2η2g2∘∇vt+m+1r+1-2αβ-2m+1E0.Since (141) holds for any 0<η1,η2≤m+1/2, we can choose η1=η2=m+1/2; then (141) becomes(142)ξt≥m-1m0-m-1+1m+1m1∇u2+∇v2+m+1r+1-2αβ-2m+1E0≥m-1m0-m-1+1m+1m1∇u2+∇v2-2m+1E0.Then from Lemma 17 and Poincaré inequality, we conclude that(143)ξt≥m-1m0-m-1+1m+1m1∇u2+∇v2+m+1r+1-2αβ-2m+1E0≥m-1m0-m-1+1m+1m1∇u2+∇v2-2m+1E0≥m-1m0-m-1+1m+1m1Cu2+v2-2m+1E0,which means ξ(t)>σ>0. Similar to the proof of Theorem 14, by the concavity argument, we conclude the result.
Conflicts of Interest
The authors declare that they have no conflicts of interest.
Acknowledgments
This work was supported by the National Natural Science Foundation of China (11071033 and 11101102), the PhD Start-Up Fund of Liaoning Province of China (nos. 20141137 and 20141139), the Natural Science Foundation of Liaoning Province of China (no. 20170540004), and the Liaoning BaiQianWan Talents Program (no. 2013921055).
CarrierG. O.PegramB. L.CarrierO.Jr.Antagonistic effects of reserpine on d-tubocurarine action on motor function of rabbits1969621251282-s2.0-001451399910.1016/0014-2999(69)90208-8PohozaevS. I.TeseiA.Nonexistence of local solutions to semilinear partial differential inequalities200421448750210.1016/j.anihpc.2003.06.002MR2069634TorrejonR.YongJ. M.On a quasilinear wave equation with memory1991161617810.1016/0362-546X(91)90131-JMR1086826Munoz RiveraJ. E.Global solution on a quasilinear wave equation with memory199482289303MR1278337WuS.-T.TsaiL.-Y.On global existence and blow-up of solutions for an integro-differential equation with strong damping20061049791014MR222963710.11650/twjm/1500403889Zbl1147.353342-s2.0-33750044074WuS.-T.Exponential energy decay of solutions for an integro-differential equation with strong damping20103642609617MR257621110.1016/j.jmaa.2009.11.046Zbl1205.450122-s2.0-72149129470GeorgievV.TodorovaG.Existence of a solution of the wave equation with nonlinear damping and source terms19941092295308MR127330410.1006/jdeq.1994.1051Zbl0803.350922-s2.0-0000028946LevineH. A.Instability and nonexistence of global solutions to nonlinear wave equations of the form Dutt=Au+F(u)197419212110.1090/S0002-9947-1974-0344697-2MR0344697LevineH. A.Some additional remarks on the nonexistence of global solutions to nonlinear wave equations19745113814610.1137/0505015MR0399682Zbl0243.35069LevineH. A.ParkS. R.SerrinJ.Global existence and global nonexistence of solutions of the Cauchy problem for a nonlinearly damped wave equation19982281181205MR165990510.1006/jmaa.1998.6126Zbl0922.350942-s2.0-0001324542VitillaroE.Global nonexistence theorems for a class of evolution equations with dissipation19991492155182MR171914510.1007/s002050050171Zbl0934.351012-s2.0-0033416133KirchhoffG.1883Leipzig, GermanyTeubnerOnoK.Blowing up and global existence of solutions for some degenerate nonlinear wave equations with some dissipation199730744494457MR1603589OnoK.Global existence, decay, and blowup of solutions for some mildly degenerate nonlinear Kirchhoff strings19971372273301MR145659810.1006/jdeq.1997.3263Zbl0879.351102-s2.0-0000366546WuS. T.TsaiL.2004National Chengchi UniversityWuS. T.TsaiL. Y.Blow-up of solutions for some non-linear wave equations of Kirchhoff type with some dissipation200665224326410.1016/j.na.2004.11.023MR2228427WuS.-T.TsaiL.-Y.On a system of nonlinear wave equations of Kirchhoff type with a strong dissipation2007381120MR2321028Zbl1132.354192-s2.0-34250863257XuR.YangY.Finite time blow-up for the nonlinear fourth-order dispersive-dissipative wave equation at high energy level20122352-s2.0-8486013379210.1142/S0129167X1250060712500607Zbl1267.33016ReedM.1976Springer-Verlag, Berlin-New Yorkvi+128Lecture Notes in Mathematics, Vol. 507MR0605679Zbl0317.35002AgreK.RammahaM. A.Systems of nonlinear wave equations with damping and source terms2006191112351270MR2278006AlvesC. O.CavalcantiM. M.Domingos CavalcantiV. N.RammahaM. A.ToundykovD.On existence, uniform decay rates and blow up for solutions of systems of nonlinear wave equations with damping and source terms200923583608MR252576910.3934/dcdss.2009.2.583Zbl1183.350332-s2.0-72649100143Said-HouariB.Global nonexistence of positive initial-energy solutions of a system of nonlinear wave equations with damping and source terms2010231-27992MR25888032-s2.0-73449126567RammahaM. A.SakuntasathienS.Global existence and blow up of solutions to systems of nonlinear wave equations with degenerate damping and source terms20107252658268310.1016/j.na.2009.11.013MR2577827LiG.SunY.LiuW.Global existence, uniform decay and blow-up of solutions for a system of Petrovsky equations20117441523153810.1016/j.na.2010.10.025MR2746830RammahaM. A.SakuntasathienS.Critically and degenerately damped systems of nonlinear wave equations with source terms201089812011227MR268144010.1080/00036811.2010.483423Zbl1198.351472-s2.0-77954421500HanX.WangM.Global existence and blow-up of solutions for a system of nonlinear viscoelastic wave equations with damping and source200971115427545010.1016/j.na.2009.04.031MR2560212WuJ.LiS.ChaiS.Existence and nonexistence of a global solution for coupled nonlinear wave equations with damping and source201072113969397510.1016/j.na.2010.01.028MR2606756Said-HouariB.MessaoudiS. A.GuesmiaA.General decay of solutions of a nonlinear system of viscoelastic wave equations2011186659684MR286125910.1007/s00030-011-0112-7Zbl1246.350402-s2.0-81455135489MessaoudiS. A.Said-HouariB.Global nonexistence of positive initial-energy solutions of a system of nonlinear viscoelastic wave equations with damping and source terms2010365127728710.1016/j.jmaa.2009.10.050MR2585099Said-HouariB.Exponential growth of positive initial-energy solutions of a system of nonlinear viscoelastic wave equations with damping and source terms20116211151332-s2.0-7915147146010.1007/s00033-010-0082-3Zbl1339.35169LiuW.Global existence and uniform decay of solutions for a system of wave equations with dispersive and dissipative terms201053555574MR266052910.1007/s11464-010-0060-2Zbl1223.352192-s2.0-77954215425LiuW.Global existence, asymptotic behavior and blow-up of solutions for coupled Klein-Gordon equations with damping terms201073124425510.1016/j.na.2010.03.017MR2645848LiangF.GaoH.Exponential energy decay and blow-up of solutions for a system of nonlinear viscoelastic wave equations with strong damping20112011:22, 19MR2834324Zbl1272.35142MessaoudiS. A.TatarN.-e.Uniform stabilization of solutions of a nonlinear system of viscoelastic equations2008873247263MR240182010.1080/00036810701668394Zbl1139.35320de Lima SantosM.Decay rates for solutions of a system of wave equations with memory2002No.\ 38, 17MR1907714Zbl1010.35012LiuL.WangM.Global existence and blow-up of solutions for some hyperbolic systems with damping and source terms2006641699110.1016/j.na.2005.06.009MR2183830Said-HouariB.Global existence and decay of solutions of a nonlinear system of wave equations2012913475489MR287673810.1080/00036811.2010.549475Zbl1246.350392-s2.0-84856113276WuS.-T.On decay and blow-up of solutions for a system of nonlinear wave equations20123941360377MR292622710.1016/j.jmaa.2012.04.054Zbl1258.350342-s2.0-84861699059NakaoM.A difference inequality and its application to nonlinear evolution equations1978304747762MR51308210.2969/jmsj/03040747Zbl0388.350072-s2.0-84972531349WangY.A global nonexistence theorem for viscoelastic equations with arbitrary positive initial energy2009229139414002-s2.0-6734924040710.1016/j.aml.2009.01.052Zbl1173.35575MaJ.MuC.ZengR.A blow up result for viscoelastic equations with arbitrary positive initial energy20112011:6, 10MR2821485Zbl1270.35134ZengR.MuC.ZhouS.A blow-up result for Kirchhoff-type equations with high energy2011344479486MR2791488Zbl1217.350372-s2.0-7995175446010.1002/mma.1374XuR.YangY.LiuY.Global well-posedness for strongly damped viscoelastic wave equation2013921138157MR300792810.1080/00036811.2011.601456Zbl1282.352262-s2.0-84874319950XuR.YangY.ChenS.SuJ.ShenJ.HuangS.Nonlinear wave equations and reaction-diffusion equations with several nonlinear source terms of different signs at high energy level2013543153170MR3085810FilippoG.MarcoS.Global solutions and finite time blow up for damped similinear wave equations20067675702MessaoudiS. A.Blow-up of positive-initial-energy solutions of a nonlinear viscoelastic hyperbolic equation20063202902915MR222600310.1016/j.jmaa.2005.07.022Zbl1098.350312-s2.0-33646341732