Evolutionary learning in agent-based modeling

This paper develops a general model for evolutionary learning in agent-based modeling. The central concepts of the general model lie in internal model principle and mutual learning of agent's internal models in an evolutionary way. This paper particularly presents network-type dynamic hypergame as a model to describe an evolutionary learning process in multi-agent situation and a simulation method by genetic algorithm to perform a network-type dynamic hypergame. The experimental results given in this paper show some requisite conditions to progress the learning process effectively.