Playing PuyoPuyo: Two search algorithms for constructing chain and tactical heuristics

Author

Ikeda, Kokolo ; Tomizawa, Daisuke ; Viennot, Simon ; Tanaka, Yuu

Author_Institution

Japan Adv. Inst. of Sci. & Technol., Ishikawa, Japan

fYear

2012

fDate

11-14 Sept. 2012

Firstpage

71

Lastpage

78

Abstract

Tetris is one of the most famous tile-matching video games, and has been used as a test bed for artificial intelligence techniques such as machine learning. Many games have been derived from such early tile-matching games, in this paper we discuss how to develop AI players of "PuyoPuyo". PuyoPuyo is a popular two-player game, and where the main point is to construct a "chain" longer than the opponent. We introduce two tree search algorithms and some tactical heuristics for improving the performance. We were able to reach an average chain length of 11, notably higher than that of the commercial Als.

Keywords

computer games; learning (artificial intelligence); search problems; trees (mathematics); AI players; PuyoPuyo; Tetris; artificial intelligence techniques; chain length; machine learning; performance improvement; tactical heuristics; test bed; tile-matching video games; tree search algorithms; two-player game; Computational intelligence; Conferences; Games;

fLanguage

English

Publisher

ieee

Conference_Titel

Computational Intelligence and Games (CIG), 2012 IEEE Conference on

Conference_Location

Granada

Print_ISBN

978-1-4673-1193-9

Electronic_ISBN

978-1-4673-1192-2

Type

conf

DOI

10.1109/CIG.2012.6374140

Filename

6374140