\(H_2/H_\infty\) Control for Continuous-Time Mean-Field Stochastic Systems with Affine Terms

TL;DR

This paper develops a comprehensive framework for designing optimal control strategies for continuous-time mean-field stochastic systems with affine terms, using the MF-SBRL and MF-FBSDE to ensure H2/H-infinity performance over finite horizons.

Contribution

It introduces a novel approach linking open-loop and closed-loop control solutions via coupled Riccati equations and stochastic differential equations for mean-field systems.

Findings

Control problem solvable via coupled Riccati equations

State-feedback gains derived from Riccati solutions

Framework ensures H-infinity norm bounds in stochastic systems

Abstract

This paper discusses the \( H_2/H_{\infty} \) control problem for continuous-time mean-field linear stochastic systems with affine terms over a finite horizon. We employ the Mean-Field Stochastic Bounded Real Lemma (MF-SBRL), which provides the necessary and sufficient conditions to ensure that the \( H_{\infty} \) norm of system perturbations remains below a certain level. By utilizing the Mean-Field Forward-Backward Stochastic Differential Equations (MF-FBSDE), we establish the equivalence conditions for open-loop \( H_2/H_{\infty} \) control strategies. Furthermore, the paper demonstrates that the control problem is solvable under closed-loop conditions if solutions exist for four coupled Difference Riccati Equations (CDREs), two sets of backward stochastic differential equations (BSDEs) and ordinary equations (ODEs). The state-feedback gains for the control strategy can be derived from these solutions, thereby linking the feasibility of open-loop and closed-loop solutions.