The pathwise numerical approximation of stationary solutions of semilinear stochastic evolution equations (Q861505)

The authors consider a stochastic evolution equation with additive noise of the reaction-diffusion type under a one-sided dissipative Lipschitz condition on its drift. They prove the existence and uniqueness of a stochastic stationary solution for the SPDE, then for its Galerkin SDE and finally for their implicit Euler schemes. Then, they prove the pathwise convergence of these stationary solutions, first those of the Galerkin systems to that of the random PDE as the dimension increases and then of the stationary solutions of the numerical schemes to that of the corresponding Galerkin system as the time step tends to zero. In order to obtain these pathwise estimates they convert the equations in random PDEs, ODEs and difference equations by subtraction of appropiate Ornstein-Uhlenbeck stationary solutions.