Date
2010
Dewey
Intelligence artificielle
Sujet
DAG; Transpositions; UCT; Monte-Carlo Tree Search
Conference name
TAAI 2010
Conference date
11-2010
Conference city
Hsinchu City
Conference country
Taïwan
Book title
International Conference on Technologies and Applications of Artificial Intelligence
Publisher
IEEE
Year
2010
ISBN
978-1-4244-8668-7
Author
Méhat, Jean
Cazenave, Tristan
Saffidine, Abdallah
Type
Communication / Conférence
Item number of pages
467-473
Abstract (EN)
In this paper we present a framework for testing various algorithms that deal with transpositions in Monte-Carlo Tree Search (MCTS). When using transpositions in MCTS, a Directed Acyclic Graph (DAG) is progressively developed instead of a tree. There are multiple ways to handle the exploration exploitation dilemma when dealing with transpositions. We propose parameterized ways to compute the mean of the child, the playouts of the parent and the playouts of the child. We test the resulting algorithms on LeftRight an abstract single player game and on Hex. For both games, original configurations of our algorithms improve on state of the art algorithms.