著者
牧野 泰裕 西野 順二 小高 知宏 小倉 久和 MAKINO Yasuhiro NISHINO Junji ODAKA Tomohiro OGURA Hisakazu
出版者
福井大学工学部
雑誌
福井大学工学部研究報告 (ISSN:04298373)
巻号頁・発行日
vol.45, no.1, pp.71-79, 1997-03

In this paper, we investigate of the repeated janken gameto estimate characteristic of game agents. In the repeated janken game, players repeat single"janken"reiteratively. The reppeated janken game has strategy. for examples, random strategy,fixed strategy, repeated strategy,history-based strategy that uses history of opponent hands. We analyze superiority of strategy by computer simulations. As a result, we show superiority of history-based strategy.