Minority Game and the Wealth Distribution in the Artificial Market

Abstract

We focus on the wealth-distribution problem in an artificial market in the context of multi-agent modeling in which trading activity is represented by the minority game. This model is widely investigated in the community of computational intelligence and econo-physics, due to its rich contents including the emergence of self-organization among agents, natural incorporation of limited rationality of agents, and good affinity to evolutional computation. However, the major defect of this kind of models is the difficulty in quantitatively explaining the constitution of the society in comparison to the real society. In particular, the wealth distribution among agents does not follow the scale-invariant distribution known as Pareto’s law. We consider two new elements to be added to the standard minority game in order to make the resulting wealth distribution of the society to fit the real-world value.