Efficient Computing of Correlated Equilibria in Multi-Player Games

intelligent agents, multi-agent systems, decision making, game theory, correlated equilibrium

This paper deals with modelling and simulation of intelligent agents doing strategic decision making which we describe by terms of mathematical game theory. We have developed a package of computer models representing Central-European electricity markets having integrated intelligent behavior of electricity producers, traders and consumers. This work required a very detailed research in algorithmics and optimization of our game theoretic models because of their large time complexity. We adopted Aumann's correlated equilibrium as a basic concept of game equilibria. In this paper, we describe a very efficient general computerized algorithm of solving the multi-player game situations through the correlated equilibrium. Core of the presented method is based on iterative elimination of dominated strategies.