Agent-based modeling of corporate behaviors with evolutionary computation

Agent-based simulation has become one of the promising gears for computational social sciences including business management strategy studies. This paper describes an agent-based simulation model for analyzing corporate behaviors of competing firms. The simulator employs evolutionary computation to evolve the characteristics of the firms. This paper also addresses a novel technology of agent-based modeling with genetic algorithms. Our method is characterized by (1) decision making agents or competing companies with strategic parameters to be optimized; (2) a multiobjective optimisation framework to evolve the artificial simulated society; (3) grounding the simulation conditions with marketing survey data in the real world, and (4) validating the strategic parameters of the agents after simulation via statistical analysis of the individual genes. The proposed method enables us to investigate the strategic measures or balanced scoreboards of competing companies from the agents' actions in the simulator.