Linear quadratic optimal control problems of infinite-dimensional mean-field type with jumps

Abstract

In this paper, we formulate and investigate a framework for the theory of the linear quadratic optimal control problem (LQ problem) for infinite-dimensional mean-field stochastic evolution systems with jumps. We ensure the well-posedness of the investigated problems by establishing the existence, uniqueness, and a priori estimates for mild solutions to general infinite-dimensional mean-field forward stochastic evolution equations (MF-SEE) and mean-field backward stochastic evolution equations (MF-BSEE) with jumps. Leveraging the Yosida approximation theory, we establish a dual theory between MF-SEE and MF-BSEE with jumps, overcoming the challenge posed by the inapplicability of Itô's formula in the context of mild solutions. Our main results regarding the existence and uniqueness of the optimal control, along with corresponding dual characterizations and state feedback representations, are obtained through convex analysis techniques, our established dual theory, and decoupling methods.