A Research On Agent Automated Negotiation Based On Market-driven

With the development of Intemet,Electronic Commerce has been an important business mode. In order to resolve the conflicts and disputations about different attributes in commerce activities, negotiation is used in most cases to reach an agreement.Traditional negotiation is usually conducted face to face by the buyers and sellers,which will surely waste lots of manpower,materials and time. Sometimes it may be influenced by the environment and human factors which may lead to some unfairness and randomicity,and even failure.Automated negotiation can resolve such problems.In order to assist people in solving these problems,Intelligent Agent(IA)technologies are used in Electronic Commerce.Recently,it is greatly booming in Artificial Intelligence(AI).By and by, Agent technologies attract the enormous interests of many research fields.As an intelligent computing entity,Agent has many characteristics such as independence,autonomy,spontaneity, adaptability and so on.In the Agent-based Electronic Commerce systems,agents act as the representatives of merchants or customers to accomplish business tasks,such as helping users look through the commodities information,selecting and filtering the products,negotiating with the opponents and so on.In this paper,our research focuses on the Agent-based automatic negotiation technology and its application in the Electronic Commerce.We propose the negotiation model(NM)based on Market-driven,negotiation protocol(SSNP,Six-Stage-Negotiation Protocol)and algorithm.In our NM,the content of negotiation is based on multi-item and multi-attribute utility.The SSNP includes filtering and selecting products,matching between offers and count-offers,negotiation strategies and so on.The negotiation strategies are planed based on user's subjective loves and objective environment of changing marketplace.So our negotiation represents sufficiently the self-determination and dynamics.Finally,we design and implement a prototype system(MAANBM,Multi-Agent Autonomic Negotiation Based on Market-driven)in which trading agents help merchants(or customers)publish products(or selecting commodities),and then negotiate automatically between merchants and customers.We simulate the progress of the bargaining for mobile phone between buyers and sellers. Experimental results show that the proposed approach for multi-agent based automatic negotiation has generally well-outcomes.