Pontryagin Maximum Principle for McKean-Vlasov Stochastic Reaction-Diffusion Equations

TL;DR

This paper develops a Pontryagin maximum principle for McKean-Vlasov stochastic reaction-diffusion equations, providing explicit gradient representations and existence results for optimal controls using advanced calculus and variational methods.

Contribution

It introduces a novel approach to the Lions derivative in Banach spaces, derives explicit gradient formulas, and proves existence and uniqueness results for controlled McKean-Vlasov SPDEs.

Findings

Derived explicit gradient representations of the cost functional.

Proved existence and uniqueness of solutions for linear McKean-Vlasov backward SPDE.

Established existence of optimal controls using a martingale and compactness approach.

Abstract

We consider the stochastic control of a semi-linear stochastic partial differential equations (SPDE) of McKean-Vlasov type. Based on a recent novel approach to the Lions derivative for Banach space valued functions, we prove the Gateaux differentiability of the control to state map and, using adjoint calculus, we derive explicit representations of the gradient of the cost functional and a Pontryagin maximum principle. On the way, we also prove a novel existence and uniqueness result for linear McKean-Vlasov backward SPDE. Furthermore, for deterministic controls, we prove the existence of optimal controls using a martingale approach and a novel compactness method. This result is complemented in the appendix with a rigorous proof of folklore results on the compactness method in the variational approach to SPDE. Our setting uses the variational approach to SPDE with monotone coefficients, allowing for a polynomial perturbation and allowing the drift and diffusion coefficients to depend on the state, the distribution of the state and the control.