On Iterative Adaptive Dynamic Programming

Linear system is very seldom in actual control works, so there is more engineering significant of researching on the actual nonlinear system. Time delay phenomenon is the objective phenomenon exists in nature. Neural network based on adaptive dynamic programming principle is selected to implement al...

Full description

Saved in:
Bibliographic Details
Published inApplied Mechanics and Materials Vol. 380-384; pp. 712 - 715
Main Authors Cui, Chang Qing, Yang, Chun Yan, Wang, Yi Qiang, Yang, Bao Sheng
Format Journal Article
LanguageEnglish
Published Zurich Trans Tech Publications Ltd 30.08.2013
Subjects
Time Delay
Iterative Algorithm
Optimal Control
Adaptive Dynamic Programming
Online AccessGet full text

Cover

More Information
Summary:Linear system is very seldom in actual control works, so there is more engineering significant of researching on the actual nonlinear system. Time delay phenomenon is the objective phenomenon exists in nature. Neural network based on adaptive dynamic programming principle is selected to implement algorithm. The algorithm contains model network training, H network training for time delay function, critic network training. Before running this iterative algorithm, training the model network first, the model uses a three-layer BP network to realize. Time delay function network H(K) is to approximate the functional relationship between the current control input and the delayed input. The critic network is used to approximate system performance function. The simulation results show that the proposed iterative adaptive dynamic programming can solve for the optimal control of delay nonlinear systems.
Bibliography:Selected, peer reviewed papers from the 2013 International Conference on Vehicle & Mechanical Engineering and Information Technology (VMEIT 2013), August 17-18, 2013, Zhengzhou, Henan, China
ISBN:3037858206
9783037858202
ISSN:1660-9336
1662-7482
1662-7482
DOI:10.4028/www.scientific.net/AMM.380-384.712