Discretization of a distributed optimal control problem with a stochastic parabolic equation driven by multiplicative noise (Q831252)

The authors consider a system that is governed by a parabolic stochastic partial differential equation with multiplicative noise and distributed control action. Homogeneous Dirichlet boundary conditions are prescribed. The objective function is given by the expected value of the sum of a squared \(L_2\)-norm tracking term and a quadratic control cost. The authors contribute to the numerical analysis of problems of this type. The numerical treatment requires the numerical analysis of the forward stochastic parabolic equation and the numerical analysis of the adjoint equation, which is a backwards stochastic parabolic equation. As a discretization, the authors suggest to use the continuous piecewise linear finite element method in space and the backwards Euler method in time. The main result states that the convergence rate is quadratic in space and of the order of the square root of the time step.