A study on strategy acquisition on imperfect information game by UCT search

Abstract

In recent years, using Monte Carlo Tree Search and UCT for game decision making is becoming a popular technique. However, the mainstream target of these researches is a perfect information game. There are perfect information game and imperfect information game in the classification of the game, research on incomplete information games is also important. In this study, the UCT search is applied to make a decision for the Hanafuda game as an incomplete information game that is a Japanese traditional card game. Hanafuda game is also one of typical games with imperfect information. We attempt to acquire a strategy of this game by using UCT search. In order to evaluate the performance of UCT, we carry out a simulated competition experiment. Obtained experimental results illustrate the effectiveness of UCT search.