Multiagent-based Virtual Enterprise, Auto-negotiation

Virtual enterprise(VE) is a new management and operation mode in 21st century. Virtual enterprise has the character of dynamic structure in organization,distributed ranges in physics,restructurable and rapid response in grasping chance in the market.With the development of artificial intelligence and the technology of agent,people make use of the agent which has the ability of self-reasoning,self-determination and multi-agent systems which are make up of the agents to simulate,optimize,implement and control the operation of the VE supply chain,it is becoming one of the most important approach in research and implement VE supply chain.How to apply the technology of agents to realize a efficient,intelligent transaction model has become an important research direction in VE supply chain,so as to realize the online automatic negotiation among the members in VE.Thus this paper will do some research about VE automatic negotiation based on multi-agent.By reading a lot of literature,the author introduced the research actuality of the related theory,then summarized the theory of VE,agent and negotiation which were involved in the research subject,then proposesed a universal VE automatic negotiation framework and a new negotiation algorithm,and also do some good research about it.The author also designed a calculable example to verify the algorithm,based on the negotiation results the author proposed a method on partner selection in virtual supply chain based on index equilibrium and power interval numbers.At last the author developed a prototype system to simulate the prosess of automatic negotiation between the core enterprise and the leaguer enterprises in VE.In this negotiation model,the author introduced into the parameter of the number ratio of core enterprise and leaguer enterprise,it can indicate that the agent can apperceive the environment and can act synchronously, the author also introduced into the parameter of next expectation price and reserved price respectively,and use them to confirm the riskparameter or modify the likely price interval,it can make the model reflect the price attitude of the two negotiation parties,and can also simulate the iteration of offering price very well in reality.