Title :
Evolution of the weight vectors in Mahjong non-player characters
Author_Institution :
Sch. of Sci. & Eng., Kindai Univ., Higashi-Osaka, Japan
Abstract :
The Mahjong game is classified as a imperfect information game. The Poker is also a famous game of such the imperfect information games. In comparison with the Poker, the Mahjong has complicated rules. In this paper, a mono-functional agent framework for the Mahjong game is proposed. As examples of the agent framework, three agents are introduced: the cautious agents, the Seven-Pairs agents, and the concealed-hand agents. The experimental results on the evolution of such agents show the evolutionary computation is able to improve the performance of the agents.
Keywords :
computer games; evolutionary computation; multi-agent systems; cautious agents; concealed-hand agents; evolutionary computation; imperfect information game; mahjong game; mahjong nonplayer characters; monofunctional agent framework; poker; seven-pairs agents; weight vectors evolution; Tiles;
Conference_Titel :
Nature and Biologically Inspired Computing (NaBIC), 2013 World Congress on
Conference_Location :
Fargo, ND
Print_ISBN :
978-1-4799-1414-2
DOI :
10.1109/NaBIC.2013.6617853