Genetic learning using adaptive action value tables

Authors:
Masaya Yoshikawa;Takeshi Kihira;Hidekazu Terai
Affiliations:
Department of Information Engineering, Meijo University, Nagoya, Aichi, Japan;Department of VLSI System Design, Ritsumeikan University, Kusatsu, Shiga, Japan;Department of VLSI System Design, Ritsumeikan University, Kusatsu, Shiga, Japan
Venue:
EC'08 Proceedings of the 9th WSEAS International Conference on Evolutionary Computing
Year:
2008

Citing 6
Cited 0

Integrated architecture for learning, planning, and reacting based on approximating dynamic programming

Proceedings of the seventh international conference (1990) on Machine learning
Adaptation in natural and artificial systems

Adaptation in natural and artificial systems
Transfer of Learning by Composing Solutions of Elemental Sequential Tasks

Machine Learning
Using transitional proximity for faster reinforcement learning

ML92 Proceedings of the ninth international workshop on Machine learning
Genetic Algorithms in Search, Optimization and Machine Learning

Genetic Algorithms in Search, Optimization and Machine Learning
Combination of online clustering and Q-value based GA for reinforcement fuzzy system design

IEEE Transactions on Fuzzy Systems

Quantified Score

Hi-index	0.00

Visualization

Abstract

Reinforcement learning is applied to various fields such as robotics and mechatronics control. The reinforcement learning is an efficient method to control in unknown environment. This paper discusses a new reinforcement learning algorithm using adaptive action value tables. The proposed learning algorithm is based on Genetic Algorithm and enables sharing knowledge among agents. Regarding sharing knowledge, it introduces hierarchical evolutionary mechanism and the knowledge is inherited in one generation. As a result, the proposed algorithm achieves not only effective learning but also robustness. Experiments using pong simulator prove the effectiveness of the proposed algorithm.