Abstract
In this paper, we study the global existence and blow up for the Cauchy problem for some hyperbolic system
Under certain conditions we prove the global existence of solutions by adapting the method of modified potential well in a functional setting of generalized Sobolev spaces, and we prove that the solution decays exponentially by introducing an appropriate Lyapunov function. By the concave method, we discuss the blow-up behavior of weak solution with certain conditions and give some estimates for the lifespan of solutions.
Similar content being viewed by others
Avoid common mistakes on your manuscript.
1 Introduction
In the paper, we study the following Cauchy problem
\(N\ge 1, T>0\), \(u=(u_1,u_2)\) is unknown, \(\delta \), \(\lambda \) is constant and \(\delta >0\), \(\lambda >0\), \(\beta \ge 2\), \(f_k\) is a known continuously differentiable function, \(\phi (x)>0\) is a known function. We study the behavior of solutions to (1.1) in the space \({\mathcal {X}}_0\times {\mathcal {X}}_0\), where \({\mathcal {X}}_0=D^{1,2}(R^N)\times L^2_g(R^N)\). Models of this type are of interest in applications in various areas of mathematical physics (see [9, 10, 23, 31]), as well as in geophysics and ocean acoustics, where, for example, the coefficient \(\phi (x)\) represents the speed of sound at the point \(x\in R^N\) (see [10]), and makes impossible the treatment of the system in the classical Sobolev space setting.
In the paper, we assume that the functions \(\phi (x)\) and \(g: R^N\rightarrow R\) satisfy the following condition:
-
(G) \(\quad \phi (x)>0\), \((\phi (x))^{-1}=: g(x)\in {\mathcal {C}}^{0,\gamma }(R^N)\), \(\gamma \in (0,1)\) and \(g\in L^{N/2}(R^N)\cap L^\infty (R^N)\).
Karahalios and Stavrakakis [9] consider the following initial value problem:
The questions of the Cauchy problem for nonlinear wave equations have been treated by many authors, see [5, 6, 9, 11, 20, 22, 25]. In general, global existence happens, when the damping terms dominate over the source terms, while blow-up appears in the opposite situation and under the assumption that the initial data is sufficiently large. In the papers [19] the problem is considered in \(R^N\) and the method of modified potential well is used to construct the global solutions. In the works [18, 33] decay properties of solutions of wave equations are discussed. Recently, Levine, Park, Pucci, Serrin and G.Todorova in [12,13,14, 21, 28, 29] studied global existence and nonexistence of solutions for unbounded domain cases and nonlinear damping. In [12, 29] nonexistence occurs for all negative initial energies. In [21] nonexistence results for abstract evolution equations have been obtained.
Aliev and Yusifova [2] studied the Cauchy problem for a system of semilinear hyperbolic equations.
The system describes the model of interaction of various fields with single masses [4]. The absence of global solutions with positive arbitrary initial energy for systems of semilinear hyperbolic equations was investigated in [1].
Reed [22] proposed an interesting problem for the following system of equations:
This system defines the motion of charged masses in an electro-magnetic field which was introduced by Segal [26]. Later, Jörgens [7], Makhankov [17] studied such systems to find the existence of weak solutions in a bounded domain. Further generalizations are also given in [15] by using Galerkin methods.
The presentation of this paper is as follows: In Sect. 2 we discuss some useful properties of the homogeneous Sobolev space and imbedding relations with some weighted spaces. In Sect. 3 we discuss the global solutions for (1.1). In Sect. 4 we obtain blow-up results for the solutions of the problem (1.1).
Notation: We will denote by \(B_R\) the open ball of \(R^N\) with center 0 and radius R. Sometimes for simplicity we use the symbols \(L^p\), \(1\le p\le \infty \) and \(D^{1,2}\), for the spaces \(L^p(R^N)\) and \(D^{1,2}(R^N)\), respectively; \(\Vert \cdot \Vert _p\) for the norm \(\Vert \cdot \Vert _{L^p(R^N)}\).
2 Preliminary
In the section, u is a scalar function. The space for the initial conditions and the solutions of the problem is the product space \({\mathcal {X}}_0=D^{1,2}(R^N)\times L^2_g(R^N)\). The space \(D^{1,2}(R^N)\) is defined as the closure of \(C^\infty _0 (R^N)\) functions with respect to the energy norm \(\Vert u\Vert _{D^{1,2}}^2=: \int _{R^N}|\nabla u|^2\,dx\). It is well known that
and that \(D^{1,2}\) is embedded continuously in \(L^{\frac{2N}{N-2}}\), there exists \(k>0\) such that
We shall use the following generalized Poincaré inequality
for all \(u\in C^\infty _0\) and \(g\in L^{N/2}\), where \(\alpha =: k^{-2}\Vert g\Vert ^{-1}_{N/2}\) (see [3]. It has been shown that \(D^{1,2}(R^N)\) is a separable Hilbert space. The space \(L^2_g(R^N)\) is defined to be the closure of \(u\in C^\infty _0(R^N)\) functions with respect to the inner product
Clearly, \(L^2_g(R^N)\) is a separable Hilbert space.
Lemma 1
(see [8]) Let \(g\in L^{N/2}(R^N\cap L^\infty (R^N))\), then the embedding \(D^{1,2}\subset L_g^2\) is compact.
Hence we are able to construct the evolution triple, which is necessary for our problem, namely
where all the embeddings are compact and dense.
Lemma 2
Let \(g\in L^{\frac{2N}{2N-pN+2p} }(R^N)\). Then we get the following continuous embedding \(D^{1,2}(R^N)\subset L^p_g (R^N)\), for all \(1\le p \le \frac{2N}{N-2}\) .
Proof
where \(a=\frac{2N}{2N -pN + 2p}\), \(b = \frac{2N}{(N-2)p}\). \(\square \)
Lemma 3
Let g satisfy condition (G) and \(1\le q< p < p^*= \displaystyle \frac{2N}{N-2}\), then there exists \(C_0 > 0\) such that the inequality
holds for all \(\theta \in (0,1)\) which satisfy \(\frac{1}{p}=\frac{1-\theta }{q}+ \frac{\theta }{p^*}\).
Proof
Using the interpolation inequality
(see [24]) and (2.1). Here \(C_0 = k^\theta \). \(\square \)
Lemma 4
Assume that \(g\in L^1(R^N)\cap L^\infty (R^N)\), then the following continuous embedding \(L^p_g(R^N)\subset L^q_g(R^N)\) is valid for any \(1\le q\le p<\infty \).
Proof
Using Hölder inequality, we get
where \(a = \frac{p}{p-q}\), \(b = \frac{p}{q}\). Hence for \(\sigma =\frac{p-q}{p}\), \(\tau =\frac{q}{p}\), we obtain the embedding inequality \(\Vert u\Vert _{L^q_g}\le C_*\Vert u\Vert _{L^p_g}\) , where \(C_* = \Vert g\Vert _1^{\frac{p-q}{pq}}\). \(\square \)
Lemma 5
Assume that \(1<a, b, c < \infty \), \(s\in [0, c^{-1})\) and \(\frac{1}{a}+ \frac{1}{b}+ \frac{1}{c}= 1\). Then for every \(u\in L^a_g\) , \(v\in L^b_g\), \(w\in L^c_g\) and every \(K > 0\) we have the inequality
Proof
The proof is a direct application of Levine et al. [12].
In order to deal with (1.1), We consider the equation
Since for every u, v in \(C_0^\infty (R^N)\)
and \(L^2_g(R^N)\) is defined as the closure of \(C_0^\infty (R^N)\) with respect to the inner product (2.3), we may consider equation (2.6) as an operator equation
Relation (2.7) implies that the operator \(A_0 = -\phi \Delta \) with domain of definition \(D(A_0) = C_0^\infty (R^N)\) is symmetric. Let us note that the operator \(A_0\) is not symmetric in the standard Lebesgue space \(L^2(R^N)\), because of the appearance of \(\phi (x)\). From (2.2) and (2.7), we have
From (2.7) and (2.9), we conclude that \(A_0\) is a symmetric, strongly monotone operator on \(L^2_g (R^N)\). Hence, the Friedrichs extension theorem (see [32]) is applicable. The energy scalar product given by (2.7) is
and the energy space is the completion of \(D(A_0)\) with respect to \((u, v)_E\). It is obvious that the energy space \(X_E\) is the homogeneous Sobolev space \(D^{1,2}(R^N)\). The energy extension \(A_E = -\phi \Delta \) of \(A_0\), namely
is defined to be the duality mapping of \(D^{1,2}(R^N)\). For every \( \eta \in D^{-1,2}(R^N)\) the Eq. (2.6) has a unique solution. Define D(A) to be the set of all solutions of the Eq. (2.6), for arbitrary \(\eta \in L^2_g (R^N)\). The Friedrichs extension A of \(A_0\) is the restriction of the energy extension \(A_E\) to the set D(A). The operator A is self-adjoint and therefore graph-closed. Its domain, D(A), is a Hilbert space with respect to the graph scalar product
The norm induced by the scalar product \((u, v)_{D(A)}\) is
which is equivalent to the norm
A consequence of the compactness of the embeddings in (2.4) is that for the eigenvalue problem .
there exists a complete system of eigensolutions \(\{w_n,\mu _n\}\) with the following properties
It can be shown, as in [3], that every solution of (2.11) is such that
uniformly with respect to x. Finally, we give the definition of weak solutions for the problem (1.1). \(\square \)
Definition 1
A weak solution of (1.1) is a function \(u(x,t)=(u_1(x,t),u_2(x,t))\) such that
-
(i)
\(u_k\!\in \! L^2[0,T;D^{1,2}(R^N)]\), \(u_{kt}\!\in \! L^2[0,T; L^2_g(R^N)]\), \(u_{ktt}\!\in \! L^2[0, T;D^{-1,2}(R^N)]\),
-
(ii)
for all \(v\in C^\infty _0([0,T]\times R^N)\), u satisfies the generalized formula
$$\begin{aligned}&\int ^T_0(u_{ktt}(\tau ), v(\tau ))_{L^2_g}d\tau +\delta \int ^T_0(u_{kt}(\tau ), v(\tau ))_{L^2_g}d\tau \nonumber \\&\quad +\,\int ^T_0\int _{R^N}\nabla u_k(\tau )\nabla v(\tau )dxd\tau -\lambda \int ^T_0(f(u(\tau )), v(\tau ))_{L^2_g}d\tau =0,\qquad \quad \end{aligned}$$(2.14)where \(f(s)=|s|^{\beta -1}s\), and
-
(iii)
u satisfies the initial conditions
$$\begin{aligned} u(x,0)=\varphi (x)\in D^{1,2}(R^N)\times D^{1,2}(R^N),\quad u_t(x,0)=\psi (x)\in L^2_g(R^N)\times L^2_g(R^N). \end{aligned}$$
3 Global existence
Now we make the following assumptions.
-
(A1)
$$\begin{aligned} \varphi (x)= & {} (\varphi _1(x),\varphi _2(x))\in D^{1,2}(R^N)\times D^{1,2}(R^N),\\ \psi (x)= & {} (\psi _1(x),\psi _2(x))\in L^2_g(R^N)\times L^2_g(R^N) \end{aligned}$$
-
(A2)
\(f_k:R^2\rightarrow R\) is continuously differentiable such that for each \(u=(u_1,u_2)\in {\mathcal {X}}_0\times {\mathcal {X}}_0\), \(u_kf_k(u)\in L^1(\Omega ),k= 1,2\), \(F(u)\in L^1(R^N)\), where \(F(u) =\displaystyle \int _0^{u_1}f_1(s,u_2)ds + \int _0^{u_2}f_2(0,s)ds\).
-
(A3)
\(f_k:{\mathcal {X}}_0\times {\mathcal {X}}_0\rightarrow L^2(R^N), k=1,2\), satisfies a local Lipschitz condition, i.e., for any \(\delta >0\), there exists a positive constant \(C(\delta )\) such that
\(\Vert f_k(u)-f_k(v)\Vert \le C(\delta )\Vert u-v\Vert _{{\mathcal {X}}_0\times {\mathcal {X}}_0}\) for \(u,v\in {\mathcal {X}}_0\times {\mathcal {X}}_0\) with
\(\Vert u\Vert _{{\mathcal {X}}_0\times {\mathcal {X}}_0}\le \delta \), \(\Vert v\Vert _{H_0^1\times H_0^1}\le \delta \).
-
(A4)
\(\displaystyle \frac{\partial f_1}{\partial u_2}=\frac{\partial f_2}{\partial u_1}\).
-
(A5)
\(u_1f_1+u_2f_2\ge F(u)\ge 0,\forall u_1,u_2\in R\).
-
(A6)
There exists a positive constant \(n_0\ge {1/p}\) such that
\(\displaystyle \max \{u_1f_1+u_2f_2,0\}\le \frac{1}{n_0}F(u),\forall u_1,u_2 \in R\).
We may consider the \(C^1\) functionals \(I_k,J_k:H_0^1(R^N)\rightarrow R\) defined by
The mountain pass value of \(J_k\) (also known as potential well depth) is defined as
It is readily seen (see [30]) that the mountain pass level \(d_k\) may also be characterized as
Finally, we consider the energy functional E(t) The energy of the problem (1.1) is defined as
F(u) will be defined in the following (A3).
From equation (1.1), we can have
Under certain assumptions on the initial data, solutions exist globally in the energy space \({\mathcal {X}}_0\). In addition to the principal condition (G) in the introduction, we shall use the following additional hypotheses for the function g and the nonlinearity exponent \(\beta \).
-
(G1)
\(g\in L^1(R^N)\) and \(1<\beta \le \frac{N}{N-2}\) , for all \(N\ge 3\).
-
(G2)
\(N\ge 3\) and \(\frac{N+2}{N}\le \beta \le \frac{N}{N-2}\).
-
(G3)
\(N=3,4\) and \(\frac{N+4}{N}\le \beta \le \frac{N}{N-2}\).
Let us note that since \(g\in L^{N/2}(R^N)\cap L^\infty (R^N)\) by hypothesis (G), then any g satisfying hypothesis (G1) belongs to all spaces \(L^p(R^N)\), for \(p\in [1,+\infty )\).
Proposition 1
Let the assumptions (A1)-(A5) be fulfilled. Then problem (1.1) admits a unique weak solution \((u_1,u_2)\) defined on \([0, T_{max})\), and at least one of the following statements is valid:
-
(1)
\(T_{max}=\infty \);
-
(2)
\(T_{max}<\infty \), and
$$\begin{aligned} \lim \limits _{t\rightarrow T_{max}}\sum _{k=1}^2\Vert u_{kt}(t)\Vert ^2_2+\Vert \nabla u_k\Vert _2^2=\infty \\ E(t)+ \sum _{k=1}^2\int _0^t\Vert u_{kt}(\tau )\Vert _*^2d\tau =E(0) \end{aligned}$$
Lemma 6
Let g, \(\beta \), N satisfy conditions (G1) or (G2). Suppose that the constants \(\delta >0\), \(\lambda <\infty \) and the initial conditions
are given. Then for sufficiently small \(T>0\) the problem (1.1) admits a unique (weak) solution such that
Proof
(a) Local Existence of the Restricted Problem on \(B_R\)
First we prove an existence result for the problem
where \(\varphi _k\in D^{1,2}(B_R)\) and \(\psi _k\in L^2_g(B_R)\). Let \(Z_k=\{z_k, z_{kt}\}\in C[0,T; \chi _0(B_R)]\) be given. In order to obtain solutions for (3.7) we first consider the following non-homogeneous problem
where \(Z=(Z_1,Z_2)\) , \(\varphi _k\in D^{1,2}(B_R)\) and \(\psi _k\in L^2_g (B_R)\). Existence of a unique (weak) solution for the problem (3.8) can be obtained by using Faedo–Galerkin approximations (see [16]).
For \(Z_k\in C[0,T; \chi _0(B_R)]\) we define the mapping
by \(U = {\mathbb {T}} (Z)\), where \(U =(u_1, u_2)\) is the unique solution of equation (3.8). It is clear that the map \({\mathbb {T}}\) is well defined. Next, we show that \({\mathbb {T}}\) maps the ball \(B_M\) to itself, where
and the space \(\chi _{0,T}\) is defined by
For \(Z_k\in B_M\), we multiply Eq. (3.8) by \(gu_t\) and integrate with respect to time and space on the set \((0, t)\times B_R\), for some \(t\in (0,T]\), to obtain
The positivity of the quantity \(\delta \int _0^t\Vert u_{kt}(.,s)\Vert ^2_{L^2_g}ds\) implies that
and
We use the assumption on Z and relation (3.10) to obtain
Choosing T sufficiently small and M sufficiently large, depending on the norm of the initial data, we have
The next step is to show that \({\mathbb {T}}\) is a contraction. Let \(Z, Z^*\in \chi _{0,T}\) such that \(U={\mathbb {T}}(Z)\), \(U^*={\mathbb {T}} (Z^*)\) and consider the difference \(W =: U-U^*\), which satisfies the equation
Following the procedure above, for the right hand side of equation (3.11), we get the estimates
From relations (3.11)–(3.13), we have
which is equivalent to the inequality
For \(T < C^{-1}(\lambda ) M^{1-\beta }\) the map \({\mathbb {T}}\) is a contraction. Then the result of existence for (3.7) is a direct consequence of the contraction mapping theorem.
(b) Extension of Solutions to \(R^N\). For \(R>R_0\), \(R\in N\), with \({\varphi _k, \psi _k}\in C_0^\infty (B_R)\times C_0^\infty (B_R)\) such that \(supp(\varphi _k)\subset B_{R_0}\) and \(supp(\psi _k)\subset B_{R_0}\) , we consider the approximating problem
The existence result in (a) holds for (3.14). We get that \(u_k^R\) is bounded in \(C[0, T; D^{1,2}(B_R)]\) and \(u^R_{kt}\) is bounded in \(C[0, T; L^2_g (B_R)]\), independently of R. Since, for any Banach space X, the following continuous embedding \(C[0, T; X] \subset L^p[0, T; X]\) is valid, for all \(1\le p <\infty \), we have \(u_k^R\), \(u^R_{kt}\) remain bounded in \(L^2[0, T; D^{1,2}(B_R)]\) and in \(L^2[0, T; L^2_g (B_R)]\), respectively. We extend \(u_k^R\), as
So that \({\widetilde{u}}_k^R\), \( {\widetilde{u}}_{kt}^R\) remain bounded in the above spaces with \(B_R\) replaced by \(R^N\). Using the assumptions on \(\beta \), we may easily check that \(f(u^R)\) is bounded in \(L^2[0, T; L^2_g (R^N)]\). From the relations (2.10) and (3.14) we obtain (as in [16], Remark 8.2, p. 265), that \(u^R{ktt}\) is bounded in \(L^2[0, T; D^{-1,2}(B_R)]\). Lemma 2.1 applied to [27] implies that \({\widetilde{u}}_k^R\) is relatively compact in \(C[0, T; L^2_g (R^N)]\). Therefore we get
Hence we may extract a subsequence of \({\widetilde{u}}_k^R\), denoted by \({\widetilde{u}}_k^{R_m}\), such that
Following the arguments in [8] we may see that \({\widetilde{u}}\) defines a unique weak solution of (1.1) with initial data satisfying (3.5). \(\square \)
Definition 2
Let \(T_{max}=\sup \{T>0:u=u(x,t)\) exists on \([0,T)\}\). If \(T_{max}<\infty \), we say that the solution to (1.1) blows up and that \(T_{max}\) is the blow up time. If \(T_{max}=\infty \), we say that the solution is global.
To obtain global existence, we adapt the method of modified potential well, as developed by Payne and Sattinger [20] and generalized to all of \(R^N\) by Nakao and Ono [19]. To this end we consider the potential well
where IntB denotes the interior of set B. It is easily seen that 0 is in W. Indeed, from Lemma 3, the Poincaré inequality (2.2) and hypothesis (G1) we have
Therefore, for any \(\lambda \in R^+\) we obtain
It is obvious that, if \( \Vert u_k\Vert _{D^{1,2}}\) is sufficiently small, then \(I_k(u_k)\ge 0 \) and 0 is in W. By the definition of W we have that
Multiply equation (1.1) by \(gu_{kt}\) and integrate over \(R^N\) to obtain
The energy of the problem is defined as
Let us note that \(E(t)\ge 0\) if \(u_k\in {\bar{W}}\) and \(u_k\notin {\bar{W}}\) if \(E(t)<0\). Lemma 2 and Proposition 1 imply that the functional E(t) is well defined. From (3.17), it is easy to obtain that \(E'(t)= -\sum _{k=1}^2\delta \Vert u_{kt}(t)\Vert _{L_g^2}^2\le 0\). Therefore E(t) is a nonincreasing function of t, i.e.,
The global existence result is given in the following theorem.
Theorem 1
Let condition (G3) be satisfied and \(u_0\in W\). Assume that the initial data satisfy (3.5) and they are sufficiently small in the sense
where \(p_1=\frac{2(\beta +1)-N(\beta -1)}{2}\), \(p_2=\frac{N\beta -N-4}{4}\) , then the (weak) solution of (1) is such that
And there exist two positive constants \({\widehat{C}}\) and \(\xi \), independent of t such that:
Proof
We shall show that the local solution given by Proposition 1, is in the modified potential well W, as long as it exists. We argue by contradiction. Assume that there exists some time \(T^* > 0\) , such that \(u_k(t)\in W\), where \(0\le t<T\) and \(u_k(T^*)\in \partial W\). Then \(I_k(u_k(T^*)) = 0\) and \(u(T^*)\ne 0\). We multiply equation (1.1) by \(gu_k\) and integrate over \(R^N\), to get the equation
We integrate over [0, t], for some \(t\in [0, T)\), to get the inequality
by applying Young inequality for \(\epsilon = \delta /2\). Since \(u_k(t)\) is in W, we have from (3.16) and (3.17)
Then from (3.23) and (3.24) we get the estimate
Using Lemma 3 and relation (3.25) we obtain the inequality
where \(\theta =\frac{N(\beta -1)}{2(\beta +1)}\) according to Lemma 3, \(p_1=(\beta +1)(1-\theta )\), \(p_2=\frac{(\beta +1)\theta }{2}-1\) and \(p_1\), \(p_2\) are positive by hypothesis (G3). Setting \(\delta _1 = C_0\mu _0^{p_2}E(0)^{p_1}\) then inequality (3.26) implies, for \(t = T^*\), that
under the assumption that \(\lambda < \displaystyle \frac{1}{\delta _1}\) [which is equivalent to the relation (3.19)] and the contradiction is achieved.
\(\forall t\ge 0\), \(u(t)\in W\), so we have
The proof of the other inequality relies on the construction of a Lyapunov functional by performing a suitable modification of the energy. To this end, for \(\varepsilon >0\), to be chosen later, we define
It is straightforward to see that L(t) and E(t) are equivalent in the sense that there exist two positive constants \(\beta _1\) and \(\beta _2\) depending on \(\varepsilon \) such that for \(t\ge 0\),
By taking the time derivative of the function L defined above , using problem (1.1), and performing several integration by parts, we get:
By using Young inequality and Sobolev inequality, we obtain, for any \(\gamma >0\),
here \(\alpha \) is the Sobolev constant.
Consequently, inserting (3.31) into (3.30), we have
By the condition \(E(0)^{p_2}c_0\lambda \mu _0^{p_1}<1\), let us choose \(\gamma \) small enough such that
From (3.32) we may find \(\eta >0\), which depends only on \(\gamma \) such that
Consequently, using the definition of the energy (3.3), for any positive constant M, we obtain
Choose \(M\alpha <\eta \), and \(\varepsilon \) small enough such that
inequality (3.35) becomes
On the other hand, by virtue of (3.29), setting \(\xi =\displaystyle \frac{M\varepsilon }{\beta _2}\), the last inequality becomes
Integrating the previous differential inequality (3.28) between 0 and t gives the following estimate for the function L
Consequently, by using (3.29) once again, we conclude
This completes the proof. \(\square \)
4 Blow-up
Theorem 2
Let the assumptions (A1–A4) and (A6) hold. If \(I_k(\varphi _k)<0, (k=1,2)\), \(E(0)<\min \{d_1, d_2\}\). Then the solution \((u_1, u_2)\) blows up in finite time, i.e., there exists T such that
and an upper bound for T is estimated
where \(\gamma = 2\left( \sum _{k=1}^2d_k-E(0)\right) \), \(\theta =\frac{\beta -1}{4}\)
Lemma 7
Suppose that \(F(s)\ge 0\) with \(s=(s_1,s_2)\in R^1\times R^1\). Let \((u_1,u_2)\) defined on \([0,T_{max})\) be a weak solution of problem (1.1). For each \(t\in [0,T_{max})\).
-
(i)
If for all \(k\in \{1,2\}\), \(I_k(u_k)\ge 0\) , then
$$\begin{aligned} \sum J_k(u_k)\ge \sum \frac{\beta -1}{2(\beta +1)}\Vert u_k\Vert ^2_{D^{1,2}} \end{aligned}$$(4.1) -
(ii)
If for all \(k\in \{1,2\}\), \(I_k(u_k)<0\) , then
$$\begin{aligned} \frac{\beta -1}{2(\beta +1)}\Vert u_k\Vert ^2_{D^{1,2}}> d_k \end{aligned}$$(4.2)
Proof
-
(i)
If \(I_k(u_k)\ge 0\) for each \(t\in [0,T_{max})\), we obtain
$$\begin{aligned} \begin{array}{llll} E(t)&{}\displaystyle \ge \sum J_k(u_k)=\frac{1}{2}\sum \Vert u_k\Vert ^2_{D^{1,2}}-\sum \frac{\lambda }{\beta +1}\int g|u_k|^{\beta +1}\\ &{}\displaystyle \ge \frac{1}{2}\sum \Vert u_k\Vert ^2_{D^{1,2}}-\sum \frac{1}{\beta +1}\Vert u_k\Vert ^2_{D^{1,2}} =\frac{\beta -1}{2(\beta +1)}\sum \Vert u_k\Vert ^2_{D^{1,2}} \end{array} \end{aligned}$$ -
(ii)
If \(I_k(u_k)<0\) for each \(t\in [0,T_{max})\), there exists \(t_0\in [0,T_{max})\) such that
$$\begin{aligned} \frac{\beta -1}{2(\beta +1)}\Vert u_1(t_0)\Vert ^2_{D^{1,2}}\le d_1\quad or \quad \frac{\beta -1}{2(\beta +1)}\Vert u_2(t_0)\Vert ^2_{D^{1,2}}\le d_2 \end{aligned}$$we then obtain
$$\begin{aligned} \lambda \Vert u_1(t_0)\Vert _{L_g^{\beta +1}}^{\beta +1}\le \Vert u_1(t_0)\Vert ^2_{D^{1,2}}\quad or\quad \lambda \Vert u_2(t_0)\Vert _{L_g^{\beta +1}}^{\beta +1}\le \Vert u_2(t_0)\Vert ^2_{D^{1,2}} \end{aligned}$$A contradiction with \(I_k(u_k)<0\).\(\square \)
Lemma 8
Suppose that \(F(s)\ge 0\) with \(s=(s_1,s_2)\in R^1\times R^1\). Let \((u_1,u_2)\) defined on \([0,T_{max})\) be a weak solution of problem (1.1). For any \(0\le t< T_{max}\).
-
(i)
If for all \(k\in \{1,2\}\), \(I_k(\varphi _k)\ge 0\) , \(E(0)<\min \{d_1,d_2\}\), then
$$\begin{aligned} \sum J_k(u_k)\ge \sum \frac{\beta -1}{2(\beta +1)}\Vert u_k\Vert ^2_{D^{1,2}} \end{aligned}$$(4.3) -
(ii)
If for all \(k\in \{1,2\}\), \(I_k(\varphi _k)<0\), \(E(0)<\min \{d_1,d_2\}\), then
$$\begin{aligned} \frac{\beta -1}{2(\beta +1)}\Vert u_k\Vert ^2_{D^{1,2}}>d_k \end{aligned}$$(4.4)
Proof
(i) Since \(I_k(\varphi _k)\ge 0\), by Lemma 7, we get that
Define
If \(T^*<T\) , then we have
Since \(E(0)\!<\!\min \{d_1,d_2\}\), Hence \(I_k(u_k(T^*))\!>\!0\). we get that \(\sum \!\frac{\beta -1}{2(\beta +1)}\Vert u_k(T^*)\Vert ^2_{D^{1,2}}\!<d_k\). By the continuity of \(\Vert u_k(t)\Vert ^2_{D^{1,2}}\) , there exists an interval \((T^*,{\hat{T}})\subset (T^*,T_{max})\) such that \(I_k(u_k(t))>0\), \(\forall t\in (T^*,{\hat{T}})\). Again using Lemma 7, it yields
This contradicts with (4.5). The contradiction implies that \(T^*\) meets with T and (4.3) holds.
(i) Let \(I_k(\varphi _k)<0\), then by Lemma 7, we get that \(\displaystyle \frac{\beta -1}{2(\beta +1)}\Vert u_k(t)\Vert ^2_{D^{1,2}}>d_k,\ k=1,2\). Define
If \(T_*<T\), then we have
or
Therefore, \(I_k(u_k(T_*))>0,\ k=1,2\). Then by use of Lemma 7, we obtain
Combining (3.3), (4.6) and noting that \(E(0)<\min \{d_1,d_2\}\), we get that \(\frac{\beta -1}{2(\beta +1)}\Vert u_k(T_*)\Vert ^2_{D^{1,2}}\le d_k\). A contradiction. Thus (4.4) is true.
We give the estimates in the following. For any \(T>0\) we may consider \(L:[0,T]\rightarrow R^+\) defined by
where \(\gamma \) and \(s_0>0\) are constants to be determined. \(\square \)
Lemma 9
Let the assumptions of Theorem 2 be satisfied. then
Proof
By (4.9), we get that
Using the energy equality and A6 and Lemma 8 , we have
By the assumptions \(\sum _{k=1}^2d_k-E(0)>0\). Now taking \(\gamma = 2\left( \sum _{k=1}^2d_k-E(0)\right) \)
we get that
Assume by contradiction that the solution u is global. Define
It is easy to check \(0<P\le L(t)\), \(Q=\displaystyle \frac{1}{2}L'(t)\), \(\displaystyle 0<U\le \frac{L''(t)}{2+p}\). For any real pair \( (\lambda ,\eta )\) and for all \(t\in [0,T]\), we have
Therefore, \(PU-Q^2\ge 0\). We infer from the above inequality that
(4.10) implies that \([L^{-\theta }(t)]''\le 0,\ t\in [0,T]\) where \(\theta =\frac{\beta -1}{4}\). Now taking \(s_0>-\frac{1}{\gamma }\sum (\varphi _k,\psi _k)_{L_g^2}\), we get that \(L'(0)>0\) and \((L^{-\theta })'(0)<0\) . Choosing \(T\ge -\frac{(L^{-\theta })(0)}{(L^{-\theta })'(0)}\), then by the concavity Lemma, there exists \(T_1\) satisfying
From (4.11), we see that \(\lim _{t\rightarrow T^-_1}L(t)\!=\!\infty \), which implies that \(\displaystyle \lim _{t\rightarrow T^-_1}\sum _{k=1}^2\Vert u_k(t)\Vert _{L_g^2}^2\!=\infty \). This leads to a contraction with \(T_{max}=\infty \). Now we give the estimate of T. If (4.11) holds, it suffices
where if it is necessary we may take \(s_0\) sufficiently large such that
Therefore, we only need to take
and so we have completed the proof of Theorem 2. \(\square \)
Availability of data and materials
Data sharing is not applicable to this article as no new data were created or analyzed in this study.
References
Aliev, A.B., Kazimov, A.A.: Nonexistence of global solutions of the Cauchy Problem for systems of Klein–Gordon equations with positive initial energy. Differ. Equ. 51(12), 1563–1568 (2015)
Aliev, A.B., Yusifova, G.I.: Nonexistence of global solutions of Cauchy problems for systems of semilinear hyperbolic equations with positive initial energy. Electron. J. Differ. Equ. 2017(211), 1–10 (2017)
Brown, K.J., Stavrakakis, N.: Global bifurcation results for a semilinear elliptic equation on all of \(R^N\). Duke Math. J. 85(1), 77–94 (1996)
Cocco, S., Barbi, M., Peyrard, M.: Vector nonlinear Klein-Gordon lattices: general derivation of small amplitude envelope soliton solutions. Phys. Lett. A 253(3–4), 161–167 (1999)
Georgiev, V., Todorova, G.: Existence of a solution of the wave equation with nonlinear damping and source terms. J. Differ. Equ. 109(2), 295–308 (1994)
Grillakis, M.G.: Regularity and asymptotic behavior of the wave equation with a critical nonlinearity. Ann. Math. 132(3), 485–509 (1990)
Jörgens, K.: Nonlinear Wave Equations. University of Colorado (1970)
Karachalios, N.I., Stavrakakis, N.M.: Existence of a global attractor for semilinear dissipative wave equations on \(R^N\). J. Differ. Equ. 157(1), 183–205 (1999)
Karachalios, N.I., Stavrakakis, N.M.: Global existence and blow-up results for some nonlinear wave equations on \(R^ N\). Adv. Differ. Equ. 6(2), 155–174 (2001)
Klibanov, M.V.: Global convexity in a three-dimensional inverse acoustic problem. SIAM J. Math. Anal. 28(6), 1371–1388 (1997)
Levine, H.A.: Instability and nonexistence of global solutions to nonlinear wave equations of the form \({\cal{P}}u_{tt}=-{\cal{A}}u+{\cal{F}}(u)\). Trans. Am. Math. Soc. 192, 1–21 (1974)
Levine, H.A., Park, S.R., Serrin, J.: Global existence and global nonexistence of solutions of the Cauchy problem for a nonlinearly damped wave equation. J. Math. Anal. Appl. 228(1), 181–205 (1998)
Levine, H.A., Serrin, J.: Global nonexistence theorems for quasilinear evolution equations with dissipation. Arch. Ration. Mech. Anal. 137(4), 341–361 (1997)
Levine, H., Todorova, G.: Blow up of solutions of the Cauchy problem for a wave equation with nonlinear damping and source terms and positive initial energy. Proc. Am. Math. Soc. 129(3), 793–805 (2001)
Li, M.R., Tsai, L.Y.: Existence and nonexistence of global solutions of some system of semilinear wave equations. Nonlinear Anal.: Theory, Methods Appl. 54(8), 1397–1415 (2003)
Lions, J.L., Magenes, E.: Nonhomogeneous Boundary Value Problems, vol. 2. Springer, Berlin (1975)
Makhankov, V.G.: Dynamics of classical solitons in non-integrable systems. Phys. Rep. 35(1), 1–128 (1978)
Nakao, M.: Decay of solutions of the wave equation with a local degenerate dissipation. Isr. J. Math. 95(1), 25–42 (1996)
Nakao, M., Ono, K.: Existence of global solutions to the Cauchy problem for the semilinear dissipative wave equations. Math. Z. 214(1), 325–342 (1993)
Payne, L.E., Sattinger, D.H.: Saddle points and instability of nonlinear hyperbolic equations. Isr. J. Math. 22(3), 273–303 (1975)
Pucci, P., Serrin, J.: Global nonexistence for abstract evolution equations with positive initial energy. J. Differ. Equ. 150(1), 203–214 (1998)
Reed, M.: Abstract Non Linear Wave Equations. Springer (2006)
Reed, M., Simon, B.: Methods of Modern Mathematical Physics III Scattering Theory. Academic Press (1979)
Rudin, W.: Real and Complex Analysis, 2nd edn. McGraw-Hill, New York (1974)
Sattinger, D.H.: On global solution of nonlinear hyperbolic equations. Arch. Ration. Mech. Anal. 30(2), 148–172 (1968)
Segal, I.: Nonlinear partial differential equations in quantum field theory. Proc. Symp. Appl. Math. AMS 17(965), 210–226 (1965)
Simon, J.: Compact sets in the space \(L^p(0, T;B)\). Ann. Mat. Pura Appl. 146(1), 65–96 (1986)
Todorova, G.: Cauchy problem for a nonlinear wave equation with nonlinear damping and source terms. Compt. Rendus l’Acad. Sci.-Ser. I-Math. 326(2), 191–196 (1998)
Todorova, G.: Stable and unstable sets for the Cauchy problem for a nonlinear wave equation with nonlinear damping and source terms. Compt. Rendus l’Acad. Sci. 328(2), 117–122 (1999)
Willem, M.: Minimax Theorems. Progress Nonlinear Differential Equations Appl, vol. 24. Birkhauser, Boston (1996)
Zauderer, E.: Partial Differential Equations of Applied Mathematics. Wiley (2011)
Zeidler, E.: Nonlinear Functional Analyssis and its Applications, Monotone Operators, vol. II. Springer, Berlin (1990)
Zuazua, E.: Exponential decay for the semilinear wave equation with localized damping in unbounded domains. J. Math. Pures Appl. 70(4), 513–529 (1991)
Acknowledgements
The authors are grateful to the referees for their very careful reading and useful comments which do improve the presentation of this article.
Author information
Authors and Affiliations
Corresponding author
Ethics declarations
Conflict of interest
The authors declare that they have no conflict of interest.
Additional information
Publisher's Note
Springer Nature remains neutral with regard to jurisdictional claims in published maps and institutional affiliations.
This work was supported by NSFC 11401043, Natural Science Foundation of Shaanxi Province of China 2018JQ1084, 2019GY202, The Fundamental Research Funds for the Central Universities 300102121101.
Rights and permissions
About this article
Cite this article
Wei, X. Global existence and blow up of solutions for the Cauchy problem of some nonlinear wave equations. Anal.Math.Phys. 12, 18 (2022). https://doi.org/10.1007/s13324-021-00625-x
Received:
Revised:
Accepted:
Published:
DOI: https://doi.org/10.1007/s13324-021-00625-x