Use of Genetic Algorithms to Solve Game Problems with Coalition Structure

Abstract

The possibility of using evolutionary computing to solve games theory problems with coalition structure is considered. The algorithm for solving such problems is proposed as well as a way to find a "fair" division among the members of the coalition, based on the distribution of the players reward in proportion to their useful deposition to the overall coalitions reward. The algorithm is demonstrated by the example of finding the optimal structure and the distribution of reward between players.