|May 26, 2015
*세미나 장소: 25동 114호
We propose and analyze a multilevel Monte Carlo finite element method (MLMC-FEM) for stochastic optimal control problem. Stochastic optimal control problem constrained by stochastic partial differential equations (PDEs) consists in finding a control function which minimizes a cost functional subject to a constraint in the form of PDEs. The proposed method leads to decrease to log-linear work and memory in the number of unknowns of a single level calculation. We analyze the MLMC-FEM of the solution based on the nested mesh for Finite Element spaces. We demonstrate the accuracy and usefulness of our method through numerical experiments.