A data-driven ADP with RBF network and LSM learning algorithm

ADP is an effective optimal method. However, the optimality depends on its network structure and training algorithm. This paper adopts RBF neural network to realize its critic and action networks after a detailed analysis on ADP. The LSM method is introduced as training algorithm, and a novel basis...

Full description

Saved in:
Bibliographic Details
Published in2017 6th Data Driven Control and Learning Systems (DDCLS) pp. 346 - 349
Main Authors Zhijian Huang, Yudong Li, Wentao Chen, Qin Zhang, Qili Wu, Qinmin Tan, Zhiyuan Yang
Format Conference Proceeding
LanguageEnglish
Published IEEE 01.05.2017
Subjects
Approximate dynamic programming
Approximation algorithms
Artificial neural networks
Convergence
Dynamic programming
LSM
Nonlinear
Optimal control
Radial basis function networks
RBF
Training
Utility function
Online AccessGet full text

Cover

More Information
Summary:ADP is an effective optimal method. However, the optimality depends on its network structure and training algorithm. This paper adopts RBF neural network to realize its critic and action networks after a detailed analysis on ADP. The LSM method is introduced as training algorithm, and a novel basis function is defined, which achieves global optimization and online control. The validity is verified by finding the optimal point through local minimums.
DOI:10.1109/DDCLS.2017.8068095