Learning a game strategy using pattern-weights and self-play

Abstract

This paper demonstrates the use of pattern-weights in order to develop a strategy for an automated player of a non-cooperative version of the game of Diplomacy. Diplomacy is a multi-player, zero-sum and simultaneous move game with imperfect information. Pattern-weights represent stored knowledge of various aspects of a game that are learned through experience. An automated computer player is developed without any initial strategy and is able to learn important strategic aspects of the game through self-play by storing pattern-weights and using temporal difference learning. © Springer-Verlag Berlin Heidelberg 2003.