透過您的圖書館登入
IP:3.128.120.95
  • 期刊

A Novel Policy Iteration Algorithm to Solve Differential Feedback Control for Continuous-Time System

摘要


In this paper, a novel synchronous policy iteration algorithm is proposed to solve the differential feedback control of the corresponding continuous-time infinite horizon linear systems. The continuous-time linear differential control problem depends on the numerical iterative solution of the related algebraic Riccati equation, which is a differential equation. Furthermore, based on reinforcement learning, the new policy iteration algorithm is executed by the structure of neural networks. And then, the evaluation neural network and the actor neural network will be constructed on the evaluator and controller respectively to complete policy evaluation and policy improvement. These two iterative steps are executed simultaneously until the overall performance of the system is stabilized. In each iterative, policy iterative algorithm is used to calculate the minimum infinite time-domain cost function associated with the updated policy. Finally, the relevant algebra Riccati equations corresponding to the system can be easily approached by the novel synchronous policy iteration algorithm. The computer simulations are given to demonstrate the feasibility and effectiveness of the novel algorithm.

參考文獻


D. L. Kleinman: On an iterative technique for riccati Equation computations, IEEE Trans.a.c, vol. 13 (1968), 114-115.
J. E. Potter: Matrix Quadratic Solutions, Siam Journal on Applied Mathematics, vol. 14 (1966), 496-501.
Ralph and Byers: Solving the Algebraic Riccati equation with the Matrix Sign Function, Linear Algebra and its Applications, vol. 85 (1987), 267-279.
Z. Yu: Linear-Quadratic Optimal Control and Nonzero-sum Differential Game of Forward-backward Stochastic System, Asian Journal of Control, vol. 14 (2012), 173-185.
C. H. Guo and P. Lancaster: Analysis and Modificaton of Newton's Method for Algebraic Riccati Equations, Mathematics of Computation, vol. 67 (1998), 1089-1105.

延伸閱讀