Approximate optimal control for a class of nonlinear discrete-time systems with saturating actuators  

作　　者：Yanhong Luo Huaguang Zhang 

机构地区：[1]School of Information Science and Engineering, Northeastern University, Shenyang 110004, China

出　　处：《Progress in Natural Science:Materials International》2008年第8期1023-1029,共7页自然科学进展·国际材料（英文版）

基　　金：the National Natural Science Foundation of China (Grant Nos. 60534010, 60572070, 60774048, 60728307);the Program for Changjiang Scholars and Innovative Research Groups of China (60521003);the Research Fund for the Doctoral Program of China Higher Education (20070145015);the National High Technology Research and Development Program of China (2006AA04Z183).

摘　　要：In this paper, we solve the approximate optimal control problem for a class of nonlinear discrete-time systems with saturating actuators via greedy iterative Heuristic Dynamic Programming (GI-HDP) algorithm. In order to deal with the saturating problem of actuators, a novel nonquadratic functional is developed. Based on the nonquadratic functional, the GI-HDP algorithm is introduced to obtain the optimal saturated controller with a rigorous convergence analysis. For facilitating the implementation of the iterative algorithm, three neural networks are used to approximate the value function, compute the optimal control policy and model the unknown plant, respectively. An example is given to demonstrate the validity of the proposed optimal control scheme.In this paper, we solve the approximate optimal control problem for a class of nonlinear discrete-time systems with saturating actuators via greedy iterative Heuristic Dynamic Programming （GI-HDP） algorithm. In order to deal with the saturating problem of actuators, a novel nonquadratic functional is developed. Based on the nonquadratic functional, the GI-HDP algorithm is introduced to obtain the optimal saturated controller with a rigorous convergence analysis. For facilitating the implementation of the iterative algorithm, three neural networks are used to approximate the value function, compute the optimal control policy and model the unknown plant, respectively. An example is given to demonstrate the validity of the proposed optimal control scheme.

关 键 词：Saturating GI-HDP algorithm Nonquadratic functional Convergence analysis Neural networks 

分 类 号：TP389.1[自动化与计算机技术—计算机系统结构]