Publication Details

Creating Action Heuristics for General Game Playing Agents

TRUTMAN Michal and SCHIFFEL Stephan. Creating Action Heuristics for General Game Playing Agents. In: Computer Games, CGW 2015. Communications in Computer and Information Science, vol. 614. Berlín: Springer Verlag, 2016, pp. 149-164. ISSN 1865-0929. Available from: https://link.springer.com/chapter/10.1007/978-3-319-39402-2_11

Czech title

Vytváření akční heuristiky pro obecné herní agenty

Type

conference paper

Language

english

Authors

Trutman Michal, Ing. (FIT BUT)
Schiffel Stephan (HR)

URL

https://link.springer.com/chapter/10.1007/978-3-319-39402-2_11

Keywords

Goal Condition, Heuristic Function, Game State, Game Tree, General Game

Abstract

Monte-Carlo Tree Search (MCTS) is the most popular search algorithm used in General Game Playing (GGP) nowadays mainly because of its ability to perform well in the absence of domain knowledge. Several approaches have been proposed to add heuristics to MCTS in order to guide the simulations. In GGP those approaches typically learn heuristics at runtime from the results of the simulations. Because of peculiarities of GGP, it is preferable that these heuristics evaluate actions rather than game positions. We propose an approach that generates heuristics that estimate the usefulness of actions by analyzing the game rules as opposed to the simulation results. We present results of experiments that show the potential of our approach.

Published

2016

Pages

149-164

Journal

Communications in Computer and Information Science, vol. 614, no. 1, ISSN 1865-0929

Proceedings

Computer Games, CGW 2015

Series

Communications in Computer and Information Science

Conference

GIGA: International Workshop on General Intelligence in Game-Playing Agents, Buenos Aires, Argentina, AR

Publisher

Springer Verlag

Place

Berlín, DE

DOI

10.1007/978-3-319-39402-2_11

UT WoS

000385788000011

EID Scopus

2-s2.0-84978664358

BibTeX

@INPROCEEDINGS{FITPUB12236,
   author = "Michal Trutman and Stephan Schiffel",
   title = "Creating Action Heuristics for General Game Playing Agents",
   pages = "149--164",
   booktitle = "Computer Games, CGW 2015",
   series = "Communications in Computer and Information Science",
   journal = "Communications in Computer and Information Science",
   volume = 614,
   number = 1,
   year = 2016,
   location = "Berl\'{i}n, DE",
   publisher = "Springer Verlag",
   ISSN = "1865-0929",
   doi = "10.1007/978-3-319-39402-2\_11",
   language = "english",
   url = "https://www.fit.vut.cz/research/publication/12236"
}