Id: arxiv/2611.15823 Url: https://arxiv.org/abs/2611.15823 Tags: math.OC Score: 0 Title: Adaptive linear quadratic optimal control for discrete-time stochastic systems Domain: math.OC Source: arxiv Authors: Nian Liu, Cheng Zhao Reviews: 0 Subject: math.OC Category: research Comments: 0 Disputed: false Division: math Mentions: 0 Discussed: 0 Published: 2026-11-11 Page: https://delphi.science/r/arxiv/2611.15823 In this paper, we consider the adaptive linear quadratic optimal control problem for a class of discrete-time stochastic systems, where the system matrix pair $(A,B)$ is unknown and the state weighting matrix $Q$ in the cost function is not necessarily positive semidefinite. Our proposed adaptive optimal control only assumes that the corresponding algebraic Riccati equation has a stabilizing solution, which weakens the standard controllable and observable assumptions used in the literature. To address this problem, we modify a weighted least squares algorithm by incorporating a random…