Research On Cooperation Of Multi-Agents Medical Diagnosis System

The medical diagnosis system is a complicated application system. Introduction of Agent and Multi-Agent System theory into the medical diagnosis system is one of the foreland of Distributed Artificial Intelligent. It is an effective way to solve complex problem system which disassembles the problem into simple one. The purpose of this paper is to build the cooperation model of multi-agents which is used in medical diagnosis system and to realize the cooperation in repast.net environment using the key technology of autonomy, sociality and Pro-activeness of Agent and MAS.The research of cooperation of multi-agents has some significative productions. Sycara constructed a model based on the labor and capital negotiation, Kraus created a common negotiation framework considering sociality and cooperation in negotiation. Another model given by Jennings took the changing of the mental state and joint-commitment into account. But several defects should be solved in real problems. The paper mainly analyses the production in cooperation mechanism based on studying multi-agent theory, and compare with several methods such as Functionally Accurate Cooperative, Coalition Formation, Joint Intentions and Contract Net Protocol. To overcome several defects of Contract Net Protocol, a concept of relation web model is introduced. The extension of the model organizes the agent society with Complete Balanced Ternary Tree, improves the negotiation of the Contract Net Protocol, introduce the cooperation factor and forgetting factor to modify trust degree. Finally, the article does a test in Repast.NET to prove the strategy of negotiation based on new model can reduce cost of communication, and can compute effectively. The aim is to afford an efficient cooperation model for medical diagnosis system.