Lyapunov exponent of a stochastic wave equation (Q4029029)

[For the entire collection see Zbl 0754.00020.] The author treats the stochastic wave equation \[ u_{tt}+2\beta u_ t+[\alpha+\varepsilon N(t)]u=c^ 2\Delta u \] in a smooth, bounded domain \(D\subset\mathbb{R}^ d\) or in full Euclidean space, with appropriate boundary conditions. Here \(N(t)\) is a noise process, either white noise or a centered function of an ergodic Markov process. Define the energy of the solution by \[ E(t)={1\over 2}\int_ M\bigl(u^ 2_ t+c^ 2|\nabla u|^ 2+\alpha u^ 2\bigr)dx \] where \(M=D\) or \(=\mathbb{R}^ d\). The Lyapunov exponent of \(E(t)\) and its behavior for \(\varepsilon\downarrow 0\) is considered. The approach is to obtain the Fourier components \(u_ n(t)\) through eigenfunction expansion, resulting in a system of ordinary differential equations with corresponding expressions for the energy. Now the results of \textit{E. Pardoux} and \textit{V. Wihstutz} [SIAM J. Appl. Math. 48, No. 2, 442-457 (1988; Zbl 0641.60065)] imply the existence of the proper Lyapunov exponents and their behavior for \(\varepsilon\downarrow 0\).