Intelligence learning model of multi-agent system

In this paper, we do research based on the multi-agent system (MAS) with learning mechanism. Making use of agent perception ability, we enable the agents to collect environment information and store it into the shared knowledge base. According to the content of knowledge base, MAS with learning ability will optimize the independence service process; decrease the choosing process time of agent behavior. At last, we set the upper limit of knowledge base capacity to resolve the problem of knowledge base retrieval which will cause time increasing during the system running. It will proof that, the MAS with learning mechanism has a better efficiency.