多目的遺伝的アルゴリズムを用いた人工市場研究のための取引エージェントの構成(<論文特集>「エコノフィジックス」)  [in Japanese] Construction of Trading Agents for Artificial Market Study by Means of a Multi-Objective Genetic Algorithm(<Special Future>Econophysics)  [in Japanese]

Artificial markets (AM), agent-based simulation of markets, attract attention as novel methods for study of market economy. For successful AM study, to share market models having appropriate complexity among researches and to use trading agents having adequate rationality and diversity in simulation are important issues. The U-Mart is an AM system which is designed to share a market model among researchers. The present paper studies construction of trading agents having both rationality and diversity for the U-Mart. Multi-objective nature of preference to return and risk is considered as a factor of variety of agents, and agents are constructed using a multi-objective genetic algorithm.