Thumbnail - No thumbnail

Small and large MCTS playouts applied to Chinese Dark Chess stochastic game

Jouandeau, Nicolas; Cazenave, Tristan (2014), Small and large MCTS playouts applied to Chinese Dark Chess stochastic game, in Cazenave, Tristan; Winands, Mark H. M.; Björnsson, Yngvi, ECAI Computer Games Workshop, Springer, p. 78-89. 10.1007/978-3-319-14923-3_6

Type
Communication / Conférence
External document link
https://hal.archives-ouvertes.fr/hal-02317151
Date
2014
Conference title
3rd Workshop on Computer Games, CGW 2014, Held in Conjunction with the 21st European Conference on Artificial Intelligence, ECAI 2014
Conference date
2014-08
Conference city
Prague
Conference country
Czech Republic
Book title
ECAI Computer Games Workshop
Book author
Cazenave, Tristan; Winands, Mark H. M.; Björnsson, Yngvi
Publisher
Springer
ISBN
978-3-319-14923-3
Number of pages
163
Pages
78-89
Publication identifier
Metadata
Show full item record
Author(s)
Jouandeau, Nicolas cc

Cazenave, Tristan
Laboratoire d'analyse et modélisation de systèmes pour l'aide à la décision [LAMSADE]
Abstract (EN)
Monte-Carlo Tree Search is a powerful paradigm for deterministic perfect-information games. We present various changes applied to this algorithm to deal with the stochastic game Chinese Dark Chess. We experimented with group nodes and chance nodes using various configurations: with different playout policies, with different playout lengths, with true or estimated wins. Results show that extending playout length over the real draw condition is beneficial to group nodes and to chance nodes. It also shows that using an evaluation function can reduce the number of draw games with group nodes and can be increased with chance nodes.
Subjects / Keywords
Root Node; Group Node; Stochastic Game; Good Node; Classic Node