電気学会論文誌C(電子・情報・システム部門誌)
Online ISSN : 1348-8155
Print ISSN : 0385-4221
<ソフトコンピューティング>
RFCNによる連続値空間上での自律エージェントの行動制御
白川 真一長尾 智晴
著者情報
ジャーナル フリー

2007 年 127 巻 5 号 p. 762-769

詳細
抄録

Researches on action control of autonomous agents and multiple agents have attracted increasing attentions in recent years. The general method using action control of agents are neural network, genetic programming and reinforcement learning. In this study, we use neural network for action control of autonomous agents. Our method determines the structure and parameter of neural network in evolution. We proposed Flexibly Connected Neural Network (FCN) previously as a method of constructing arbitrary neural networks with optimized structures and parameters to solve unknown problems. FCN was applied to action control of an autonomous agent and showed experimentally that it is effective for perceptual aliasing problems. All of the experiments of FCN, however, are in only grid space. In this paper, we propose a new method based on FCN which can decide correct action in real and continuous valued space. The proposed method which called Real valued FCN (RFCN) optimizes input-output functions of each units, parameters of the input-output functions and speed of each units. In order to examine the effectiveness, we applied the proposed method to action control of an autonomous agent to solve continuous valued maze problems.

著者関連情報
© 電気学会 2007
前の記事 次の記事
feedback
Top