MP-Draughts: a multiagent reinforcement learning system based on MLP and Kohonen-SOM neural networks

  • Authors:

  • Valquiria Aparecida Rosa Duarte;Rita Maria Silva Julia;Ayres Roberto Araujo Barcelos;Alana Bueno Otsuka

  • Affiliations:

  • Graduate student at the Computer Science Department, Federal University of Uberlandia, Uberlandia, Brazil;Professor at the Computer Science Department, Federal University of Uberlandia, Uberlandia, Brazil;Graduate student at the Computer Science Department, Federal University of Uberlandia, Uberlandia, Brazil;Undergraduate student at the Computer Science Department, Federal University of Uberlandia, Uberlandia, Brazil

  • Venue:
  • SMC'09 Proceedings of the 2009 IEEE international conference on Systems, Man and Cybernetics
  • Year:
  • 2009

Quantified Score

Hi-index 0.00

Visualization

Abstract

This paper presents MP-Draughts (MultiPhase-Draughts): a multiagent environment for Draughts, where one agent - named IIGA- is built and trained such as to be specialized for the initial and the intermediate phases of the games and the remaining ones for the final phases of them. Each agent of MP-Draughts is a neural network which learns almost without human supervision (distinctly from the world champion agent Chinook). MP-Draughts issues from a continuous activity of research whose previous product was the efficient agent VisionDraughts.Despite its good general performance, VisionDraughts frequently does not succeed in final phases of a game, even being in advantageous situation compared to its opponent (for instance, getting into endgame loops). In order to try to reduce this misbehavior of the agent during endgames, MP-Draughts counts on 25 agents specialized for endgame phases, each one trained such as to be able to deal with a determined cluster of endgame board-states. These 25 clusters are mined by a Kohonen Network from a Data Base containing a large quantity of endgame board-states. After trained, MP-Draughts operates in the following way: first, VisionDraughts is used as IIGA; next, the endgame agent that represents the cluster which better fits the current endgame board-state will replace it up to the end of the game. This paper shows that such a strategy significantly improves the general performance of the player agents.