System transformation and model-free value iteration algorithms for continuous-time linear quadratic stochastic optimal control problems DOI
Guangchen Wang, Heng Zhang

International Journal of Systems Science, Год журнала: 2024, Номер unknown, С. 1 - 10

Опубликована: Авг. 18, 2024

In this paper, we investigate a continuous-time linear quadratic stochastic optimal control (LQSOC) problem in an infinite horizon, where diffusion and drift terms of the corresponding system depend on both state variables. light theory, LQSOC is reduced to solving generalised algebraic Riccati equation (GARE). With help existing model-based value iteration (VI) algorithm, propose two data-driven VI algorithms solve GARE. The first one relies transforming into deterministic then by data system. Consequently, algorithm does not need information coefficients has lower complexity. second directly uses generated system, thus it circumvents requirement all coefficients. We also provide convergence proofs these validate through simulation examples.

Язык: Английский

Partially observed linear quadratic stochastic optimal control problem in infinite horizon: A data-driven approach DOI
Xun Li, Guangchen Wang, Jie Xiong

и другие.

Systems & Control Letters, Год журнала: 2025, Номер 198, С. 106050 - 106050

Опубликована: Фев. 24, 2025

Язык: Английский

Процитировано

0
Inverse reinforcement learning by expert imitation for the stochastic linear-quadratic optimal control problem DOI
Zhongshi Sun, Guangyan Jia

Neurocomputing, Год журнала: 2025, Номер unknown, С. 129758 - 129758

Опубликована: Фев. 1, 2025

Язык: Английский

Процитировано

0
Online Q-learning for stochastic linear systems with state and control dependent noise DOI
Hongxu Zhu, Wei Wang, Xiaoliang Wang

и другие.

Applied Soft Computing, Год журнала: 2024, Номер 167, С. 112417 - 112417

Опубликована: Ноя. 6, 2024

Язык: Английский

Процитировано

1
System transformation and model-free value iteration algorithms for continuous-time linear quadratic stochastic optimal control problems DOI
Guangchen Wang, Heng Zhang

International Journal of Systems Science, Год журнала: 2024, Номер unknown, С. 1 - 10

Опубликована: Авг. 18, 2024

In this paper, we investigate a continuous-time linear quadratic stochastic optimal control (LQSOC) problem in an infinite horizon, where diffusion and drift terms of the corresponding system depend on both state variables. light theory, LQSOC is reduced to solving generalised algebraic Riccati equation (GARE). With help existing model-based value iteration (VI) algorithm, propose two data-driven VI algorithms solve GARE. The first one relies transforming into deterministic then by data system. Consequently, algorithm does not need information coefficients has lower complexity. second directly uses generated system, thus it circumvents requirement all coefficients. We also provide convergence proofs these validate through simulation examples.

Язык: Английский

Процитировано

0