Optimality Conditions for Parabolic Stochastic Optimal Control Problems with Boundary Controls
Piero Visconti
DOI: https://doi.org/10.1007/s00245-024-10204-8
2024-12-11
Applied Mathematics & Optimization
Abstract:In this paper, we study optimality conditions for a class of control problems driven by a cylindrical Wiener process, resulting in a stochastic maximum principle in differential form. The control acts on both the drift and volatility, potentially as unbounded operators, allowing for SPDEs with boundary control and/or noise. Through the factorization method, we establish a regularity property for the state equation, which, by duality, extends to the backward costate equation, understood in the transposition sense. Finally, we show that the cost functional is Gâteaux differentiable, with its derivative represented by the costate. The optimality condition is derived using results from set-valued analysis.
mathematics, applied