Automating Texas Hold'em poker with PLICAS

Abstract

Influenced by the possibilities of the Internet poker has become a popular online game. Spurred by this development, automated poker got into the focus of research in game theory (GT), artificial intelligence (AI) and multi-agent systems (MAS). This paper describes the development and evaluation of PLICAS, a poker bot designed for the 'Texas Hold'em Fixed Limit Heads-up' variant. The poker bot integrates approaches, such as opponent modeling, abstraction techniques, and case-based reasoning. PLICAS also introduces simulation-based methods for the exploitation of the opponent's play. Participation in the 2010 AAAI Computer Poker Competition (ACPC) shows that PLICAS has a lot of potential but suffers from a vulnerable opponent modeling strategy. © 2012 Springer-Verlag.