Applying assimilation and accommodation for cooperative learning of RoboCup agent

Adapting learning is the essential ability to improve the convergence rate and learning quality in the multi-agent system. This paper integrates three adapting learning methods to make agent learns efficiently. Reinforcement learning is used to compute strategies for multi-agent soccer teams. The accommodation technology attaches the new knowledge from external information. As a conflict between the external knowledge and agent´s knowledge, we utilize the assimilation technology to adjust the agent´s knowledge. Finally, our method compares with UvA team and be verified on the RoboCup simulator.