Research On Cooperation Of Muti-Agent Base On Game Theory

With the continuous development of the multi-agent technology, application of agents also gradually penetrated into the various fields. The difficulty of the agent to perform the task is increasing and the complexity of the application environment is also increasing. In some application fields, a single agent has been unable to meet the people’s needs, so the research on multi-agent systems becomes an important research direction of multi-agent technology.Multi-agent system is an important branch in the field of distributed artificial intelligence and it is also one of the hot issues worldwide in the area of artificial intelligence. And multi-agent cooperation is one of the core issues about the multi-agent system research.In the environment of the multi-agent, the choice of the agent strategy is affected by the other agent in the system. Game theory provides the mathematical tools for the mutual influence of decision-making behavior. In this paper, multi-agent cooperation was researched by using game theory. Specific contents are as follows:1. The research status of multi-agent cooperation at home and abroad was reviewed. The main research contents and the current hot topics of multi-agent cooperation were introduced. The existing results were analyzed and deficiencies were indicated.2. Several basic concepts about agent and Multi-agent system,game theory and classification, the definition of Nash equilibrium and the related theorem were introduced.3. Research on the formation of multi-agent cooperation team. For the issue of the collaborative team being broken easily at complex environment, the team formation algorithm was proposed based on game theory and credit value, and the punishment mechanism is used to punish the agents of trying to exit from the team, so that the stability of the collaborative team is ensured effectively in task-executing process.4. Research on the issues of the cooperative execution of multi-agent. For the collaborative problem of multi-agent at complex environment, the multi-agent cooperation algorithm was proposed based on game theory, and virtual operation method is used to predict the selection strategy of other members in the team, so that the agents can choose better behavior strategy to quickly reach cooperation equilibrium.5. Finally, the main work done in the paper was summarized, and the future research work was expounded.