Potential theory for hyperbolic SPDEs. (Q1879816)

The authors study the following problem of potential theory for multiparameter stochastic processes on \(\mathbb{R}^d\): given \(E\subset \mathbb R^d\), does this process visit (or hit) \(E\) with positive probability? The objective is to relate this hitting probability to the Newtonian capacity of the set and to characterize the sets that, with probability one, are not visited. This is done for non-Gaussian processes that are solutions to the system of nonlinear hyperbolic stochastic partial differential equations \[ \frac{\partial^2 X_t^i}{\partial t_1 \partial t_2}=\sum_{j=1}^{d} \sigma_j^i(X_t) \frac{\partial^2 W_t^i}{\partial t_1 \partial t_2} + b^i(X_t),\quad t=(t_1,t_2)\in \mathbb{R}^2_{+}, \] \[ X_t^i=x_0 \quad \text{ if } t_1 t_2 =0,\; 1\leq i\leq d, \] where \(W\) is a two-parameter \(d\)-dimensional Wiener process and \(\sigma_j^i\), \(b^i\) are smooth functions on \(\mathbb{R}^d\). The work is motivated by \textit{D. Khoshnevisan} and \textit{Z. Shi} [Ann. Probab. 27, 1135--1159 (1999; Zbl 0962.60066)], where bounds of the hitting probabilities for the Brownian sheet are obtained. Gaussian-type lower bounds for the density of the random variable \(X_t\), for \(t\) away from the axes, are needed; for this, the authors apply a result of \textit{A. Kohatsu-Higa} [Probab. Theory Relat. Fields 126, 421--457 (2003; Zbl 1022.60056)].