Comparison of Rapid Action Value Estimation Variants for General Game Playing

Chiara F. Sironi; Mark H. M. Winands

doi:10.1109/CIG.2016.7860429

Comparison of Rapid Action Value Estimation Variants for General Game Playing

^*Corresponding author for this work

Networks and Strategic Optimization

Research output: Chapter in Book/Report/Conference proceeding › Conference article in proceeding › Academic › peer-review

Abstract

General Game Playing (GGP) aims at creating computer programs able to play any arbitrary game at an expert level given only its rules. The lack of game-specific knowledge and the necessity of learning a strategy online have made Monte-Carlo Tree Search (MCTS) a suitable method to tackle the challenges of GGP. An efficient search-control mechanism can substantially increase the performance of MCTS. The RAVE strategy and its more recent variant, GRAVE, have been proposed for this reason. In this paper we further investigate the use of GRAVE for GGP and compare its performance with the more established RAVE strategy and with a new variant, called HRAVE, that uses more global information. Experiments show that for some games GRAVE and HRAVE perform better than RAVE, with GRAVE being the most promising one overall.

Original language	English
Title of host publication	2016 IEEE Conference on Computational Intelligence and Games (CIG)
Publisher	IEEE
Pages	309-316
Number of pages	8
DOIs	https://doi.org/10.1109/CIG.2016.7860429
Publication status	Published - Sept 2016
Event	2016 IEEE Conference on Computational Intelligence and Games (CIG) - Petros M. Nomikos Conference Centre, Santorini, Greece Duration: 20 Sept 2016 → 23 Sept 2016

Publication series

Series	IEEE Conference on Computational Intelligence and Games
ISSN	2325-4270

Conference

Conference	2016 IEEE Conference on Computational Intelligence and Games (CIG)
Country/Territory	Greece
City	Santorini
Period	20/09/16 → 23/09/16

Keywords

CARLO TREE-SEARCH
STRATEGIES
OPERATORS

Access to Document

10.1109/CIG.2016.7860429

http://ieeexplore.ieee.org/document/7860429/

Cite this

@inproceedings{87cd275f3b02408d856367419bd2acde,

title = "Comparison of Rapid Action Value Estimation Variants for General Game Playing",

abstract = "General Game Playing (GGP) aims at creating computer programs able to play any arbitrary game at an expert level given only its rules. The lack of game-specific knowledge and the necessity of learning a strategy online have made Monte-Carlo Tree Search (MCTS) a suitable method to tackle the challenges of GGP. An efficient search-control mechanism can substantially increase the performance of MCTS. The RAVE strategy and its more recent variant, GRAVE, have been proposed for this reason. In this paper we further investigate the use of GRAVE for GGP and compare its performance with the more established RAVE strategy and with a new variant, called HRAVE, that uses more global information. Experiments show that for some games GRAVE and HRAVE perform better than RAVE, with GRAVE being the most promising one overall.",

keywords = "CARLO TREE-SEARCH, STRATEGIES, OPERATORS",

author = "Sironi, {Chiara F.} and Winands, {Mark H. M.}",

year = "2016",

month = sep,

doi = "10.1109/CIG.2016.7860429",

language = "English",

series = "IEEE Conference on Computational Intelligence and Games",

publisher = "IEEE",

pages = "309--316",

booktitle = "2016 IEEE Conference on Computational Intelligence and Games (CIG)",

address = "United States",

note = "2016 IEEE Conference on Computational Intelligence and Games (CIG) ; Conference date: 20-09-2016 Through 23-09-2016",

}

Sironi, CF & Winands, MHM 2016, Comparison of Rapid Action Value Estimation Variants for General Game Playing. in 2016 IEEE Conference on Computational Intelligence and Games (CIG). IEEE, IEEE Conference on Computational Intelligence and Games, pp. 309-316, 2016 IEEE Conference on Computational Intelligence and Games (CIG), Santorini, Greece, 20/09/16. https://doi.org/10.1109/CIG.2016.7860429

Comparison of Rapid Action Value Estimation Variants for General Game Playing. / Sironi, Chiara F.; Winands, Mark H. M.
2016 IEEE Conference on Computational Intelligence and Games (CIG). IEEE, 2016. p. 309-316 (IEEE Conference on Computational Intelligence and Games).

Research output: Chapter in Book/Report/Conference proceeding › Conference article in proceeding › Academic › peer-review

TY - GEN

T1 - Comparison of Rapid Action Value Estimation Variants for General Game Playing

AU - Sironi, Chiara F.

AU - Winands, Mark H. M.

PY - 2016/9

Y1 - 2016/9

N2 - General Game Playing (GGP) aims at creating computer programs able to play any arbitrary game at an expert level given only its rules. The lack of game-specific knowledge and the necessity of learning a strategy online have made Monte-Carlo Tree Search (MCTS) a suitable method to tackle the challenges of GGP. An efficient search-control mechanism can substantially increase the performance of MCTS. The RAVE strategy and its more recent variant, GRAVE, have been proposed for this reason. In this paper we further investigate the use of GRAVE for GGP and compare its performance with the more established RAVE strategy and with a new variant, called HRAVE, that uses more global information. Experiments show that for some games GRAVE and HRAVE perform better than RAVE, with GRAVE being the most promising one overall.

AB - General Game Playing (GGP) aims at creating computer programs able to play any arbitrary game at an expert level given only its rules. The lack of game-specific knowledge and the necessity of learning a strategy online have made Monte-Carlo Tree Search (MCTS) a suitable method to tackle the challenges of GGP. An efficient search-control mechanism can substantially increase the performance of MCTS. The RAVE strategy and its more recent variant, GRAVE, have been proposed for this reason. In this paper we further investigate the use of GRAVE for GGP and compare its performance with the more established RAVE strategy and with a new variant, called HRAVE, that uses more global information. Experiments show that for some games GRAVE and HRAVE perform better than RAVE, with GRAVE being the most promising one overall.

KW - CARLO TREE-SEARCH

KW - STRATEGIES

KW - OPERATORS

U2 - 10.1109/CIG.2016.7860429

DO - 10.1109/CIG.2016.7860429

M3 - Conference article in proceeding

T3 - IEEE Conference on Computational Intelligence and Games

SP - 309

EP - 316

BT - 2016 IEEE Conference on Computational Intelligence and Games (CIG)

PB - IEEE

T2 - 2016 IEEE Conference on Computational Intelligence and Games (CIG)

Y2 - 20 September 2016 through 23 September 2016

ER -