General Doubly Stochastic Maximum Principle and Its Applications to Optimal Control of SPDEs (1009.6061v4)

Abstract: In this paper, we prove the necessary and sufficient maximum principles (NSMPs in short) for the optimal control of systems described by a quasilinear stochastic heat equation within convex control domains, which all the coefficients contain control variables. For that, the optimal control problem of fully coupled forward-backward doubly stochastic system is studied. We apply our NSMPs to treat a kind of forward-backward doubly stochastic linear quadratic optimal control problems and an example of optimal control of stochastic partial differential equations (SPDEs in short) as well.