Multi-Agent Coevolutionary Learning Method Based on Individual Rule Set

The process of cooperative or competitive co evolutionary learning exist among multiple species in the ecological and artificial system, which is one of the primary features for emergent phenomena and behaviors. In this paper, we propose a multi-agent co evolutionary learning method based on the individual rule set, define the co evolutionary learning framework. We design a typical "predator-prey" competition system as an example, define the rule set and implement the system by multi-agent modeling language, Net Logo. With simulation experiments of the birds-bugs prey problem, we analyze emergent behaviors in the process of co evolutionary learning.