A New Algorithm for Zero-Sum Linear-Quadratic Stochastic Differential Games in Infinite Horizons

Abstract: We propose a new algorithm for Zero-Sum Linear-Quadratic Stochastic Differential Games by dissecting their inherent structures. Specifically, we construct dual-layer iterative matrix-increasing sequences, which reformulate the original problem into a set of mutually interconnected subproblems. By sequentially computing the stabilizing solutions to the associated algebraic Riccati equations within each subproblem, we derive the stabilizing solutions for the original problem and rigorously establish the convergence of the proposed algorithm. Numerical simulations further validate the effectiveness of the method. This work extends classical setting and provides the first complete unified numerical framework for solving a broader class of stochastic Game-Theoretic algebraic Riccati equations.