Research On Multi-agent Negotiation Based On Web Services

With the rapid development of computer and Internet technology, the way people buy goods has fundamentally changed. It changes from traditional face-to-face communication to people visiting related websites, according to their demands to order. But the web browsing– order model still can not satisfy customers, for example, they can't find satisfactory goods or the price they spend is too high, so multi-agent negotiation techniques emerge.In order to resolve conflicts between buyer and seller, in most cases we use negotiation to reach a consensus. Negotiation is a process which is a number of entities who discuss some issues and reach a consensus. With the autonomous, automatic and self-aware features, intelligent agents are very suitable for achieving automatic negotiation technique to solve the various conflicts in trading. Therefore, the process of negotiation between people in trading becomes the process of exchanging information on certain issues between intelligent agents.Multiagent technology and web services technology are two very important technologies for e-commerce development. To combine web services technology which supports cross-platform, loosely and interoperable and multiagent technology that has a high degree of intelligence, coordination and initiative in order to bring greater convenience to customers. Web service based multi-agent systems provide a highly adaptive, scalable and interactive ennvironment to users.This thesis analyzes multi-agent negotiation and web services and proposes a multi-issue negotiation model-WSBANM. The subjects of negotiation are not only the price in this thesis, but can negotiate over multiple issues simultaneously. The WSBANM can make it easier for negotiation parties to find suitable products, rather than using just the price to select goods; this thesis uses fuzzy theory to formalize the various attributes of negotiation proposals, which is more suitable for general consumers because consumers generally use "good", "general" and other words to evaluate products. The thesis proposes a negotiation strategy which can be dynamically adjusted; in negotiation process, the negotiation parties can estimate opponent's behaviors or trends by calculating opponent's concession strategy; the thesis presentes an assessment strategy to find out one negotiation partner from all partners at a limited time to complete the transaction.; This thesis examines and proposes a web service based multi-agent negotiation model and realizes a negotiation algorithm; This thesis provides a simulate trading process by typical examples and then it implemented a prototype system.The web service based mulitagent negotiation model proposed in this thesis can be applied to other negotiation environments when slightly modified, which is practically valuable and makes a necessary theory and practice preparation for further realizing complex e-commerce system.