Bounded solutions of second order semilinear evolution equations and applications to the telegraph equation (Q1282445)

The authors consider the equation \[ \ddot u(t)+c\dot u(t)+ Au(t) +g(t,u)=0 \] for functions \(u:\mathbb{R} \supset J\to H\), where \(J\) is an interval and \(H\) is a Hilbert space. Here, \(c\) is a positive constant, \(A:H\supset D(A)\to H\) is a linear selfadjoint semi-definite unbounded operator such that \(\overline {D(A)}=H\) and for each \(\lambda<0\) the resolvent \((A-\lambda I)^{-1}\) exists and is a compact operator. The function \(g:\mathbb{R}\times H\to H\) is continuous, Lipschitzian with respect to the second variable and bounded i.e. \(\sup_{(t,u) \in \mathbb{R}\times H}\| g(t,u)\| <\infty\), where \(\|\cdot\|\) is the norm induced by the inner product \((\cdot,\cdot)\) in \(H\). A function \(u(t)\) is, by definition, a solution to the above equation if \(u\in C(\mathbb{R},V_1)\cap C^1 (\mathbb{R},H)\) and for each \(w\in V_1\) one has \[ {d^2\over dt^2}\bigl(u(t), w\bigr)+ c{d\over dt}\bigl(u(t), w\bigr)+ \bigl(u(t), w\bigr)_1+ \biggl(g\bigl( t,u(t) \bigr), w\biggr)= 0 \] in the sense of distributions. Here, \(V_1:=\{u\in H: \sum^\infty_{n=1} \lambda_n(u, \varphi_n)^2< \infty\}\), \((\lambda_i)^\infty_{i =1}\) and \((\varphi_i)\) denote eigenvalues and corresponding eigenvectors of \(A\), respectively, \((u,v)_1: =\sum^\infty_{n=1} \lambda_n(u,\varphi_n)\) \((v,\varphi_n)\) for \(u,v\in V_1\). The main result states that if the equation is dissipative (the definition is too long and technical for reproducing it here) then there exists a solution \(u\) defined and bounded on the whole line \(\mathbb{R}\). For the linear equation \(\ddot u+c\dot u+Au=f(t)\) both cases \(\lambda_1>0\) or \(\lambda_1 =0\) are investigated giving additional information on the solution. Finally, as an illustration of earlier abstract considerations the nonlinear telegraph equation \(u_{tt}+cu_t-u_{xx} +g(u)= f(t,x)\) \((t\in\mathbb{R},\;0<x<\pi)\) with Neumann boundary conditions \(u_x(t,0)= u_x(t,\pi)=0\) \((t\in\mathbb{R})\) is studied in detail.