Towards Human-Competitive Game Playing for Complex Board Games with Genetic Programming

Abstract : Recent works have shown that Genetic Programming (GP) can be quite successful at evolving human-competitive strategies for games ranging from classic board games, such as chess, to action video games. However to our knowledge GP was never applied to modern complex board games, so-called eurogames, such as Settlers of Catan, i.e. board games that typically involve four characteristics: they are non zero-sum games, multiplayer, with hidden information and random elements. In this work we study how GP can evolve artificial players from low level attributes of a eurogame named " 7 Wonders " , that features all the characteristics of this category. We show that GP can evolve competitive artificial intelligence (AI) players against human-designed AI or against Monte Carlo Tree Search, a standard in automatic game playing.
Type de document :
Communication dans un congrès
International Conference on Artificial Evolution (Evolution Artificielle), EA 2015, Oct 2015, Lyon, France. Springer, International Conference on Artificial Evolution (Evolution Artificielle), EA 2015, 9954, pp.123 - 135, 2016, Lecture Notes in Computer Science (LNCS). 〈10.1007/978-3-319-31471-6_10〉
Liste complète des métadonnées

Littérature citée [17 références]  Voir  Masquer  Télécharger

https://hal.inria.fr/hal-01488191
Contributeur : Cyrille Fonlupt <>
Soumis le : lundi 13 mars 2017 - 14:15:05
Dernière modification le : vendredi 24 novembre 2017 - 01:02:36

Fichier

1445328959.pdf
Fichiers produits par l'(les) auteur(s)

Identifiants

Collections

Citation

Denis Robilliard, Cyril Fonlupt. Towards Human-Competitive Game Playing for Complex Board Games with Genetic Programming. International Conference on Artificial Evolution (Evolution Artificielle), EA 2015, Oct 2015, Lyon, France. Springer, International Conference on Artificial Evolution (Evolution Artificielle), EA 2015, 9954, pp.123 - 135, 2016, Lecture Notes in Computer Science (LNCS). 〈10.1007/978-3-319-31471-6_10〉. 〈hal-01488191〉

Partager

Métriques

Consultations de la notice

26

Téléchargements de fichiers

11