Lorenz Equilibrium: Concept and Evolutionary Detection

摘要

The most popular solution concepts in game theory, Nash and Pareto equilibrium, have some limitations when applied in real life. Nash equilibrium rarely assures maximal payoff and the Pareto equilibrium is a set of solutions that is often too hard to process. Our aim is to find an equilibrium concept that provides a small set of efficient solutions and is equitable for all players. The Lorenz dominance relation is considered and a crowding based differential evolution method is proposed to detect the Lorenz-optimal solutions. The Lorenz equilibrium for games is proposed. Experimental results show the effectiveness of the proposed approach.