The finite speed of propagation for solutions to nonlinear stochastic wave equations driven by multiplicative noise (Q2636656)

The authors consider a stochastic wave equation \(d\dot X-\Delta X\,dt+g(X)\,dt=\sigma(X)\,dW\) with a homogeneous Dirichlet boundary condition in a bounded domain \(\mathcal O\) in \(\mathbb R^d\) with a smooth boundary, with a deterministic initial condition \(X(0)=x\), \(\dot X(0)=y\), where \(\sigma\) is a real Lipschitz continuous function with \(\sigma(0)=0\), \(g\) is a real continuous function satisfying \(|g(r)|\leq L(1+|r|^{p+1})\) and \(rg(r)\geq 0\) for every \(r\in\mathbb R\) and for some \(p\geq 0\) if \(d=1,2\) or \(0\leq p\leq\frac{2}{d-2}\) if \(d\geq 3\) and \(W\) is a Wiener process in \(L^2(\mathcal O)\) with a covariance operator subject to certain additional spectral conditions related to the spectral decomposition of the Dirichlet Laplace operator on \(L^2(\mathcal O)\). If \(K\) is a subset in \(\mathcal O\), then \(K_r\) denotes the set of points in \(\mathcal O\) whose Euclidean distance from \(K\) is at most \(r\). It is proved that if the supports of the initial conditions \(x\in H^1_0(\mathcal O)\) and \(y\in L^2(\mathcal O)\) are contained in a closed set \(K\) in \(\mathcal O\), then the support of any solution (unique or not) starting from this initial condition is contained in \(K_t\) for every \(t\geq 0\) almost surely. An analogous result is proved for a stochastic wave equation with an additive noise.