A Study Of The Modeling Of Agent-Based Electronic Commerce And Multi-Attribute Negotiation

Although agent technology is believed the enabling technology to implement next generation intelligent electronic commerce systems, there is a long way to go to get it applied in reality. The major problems in this area include to eastablish appropriate models for agent-based electronic commerce, to provide end-users flexible tools to support the configurations of agent negotiation strategies and so on.Aiming at the construction of an agent-based electronic commerce system that can be applied in reality, this thesis studies the problem of modeling of agent-based electronic commerce and multi-attribute negotiation. The major contributions are as follows. First, this thesis proposes an agent-based E-commerce model that is conformed to the FIPA specifications and is open and interoperable. The model is mainly oriented to B2C E-commerce, or retail E-commerce in other word j. In the model, Buyer agents and Seller agents can provide intelligent proxy services to consumers and sellers participating in the transaction, and accomplish the tasks of searching . comparison . negotiation and payment of the goods or services, which alleviates and even solves the contradiction of the explosive increase of business information on the Internet and the increasing complexity of network environment consequently. In addition, a reasonable agent transaction flow and the comparison-shopping are designed, in which consumers can win the best goods or services. Second, this thesis studies the multi-attribute negotiation in agent transaction. We analyze the utility of Seller agents, Buyer agents and the goods in the negotiation using multi-attribute utility theory, discuss the negotiation strategy in standard English reverse auction, and solve the problem such as the optimal bidding strategy of seller agents, determination of the winner agent and the public utility function. In particuhir, we design the Buyer agent's dynamic bidding strategy, which can maximize Buyer agent' s revenue by the dynamic adjustment of public utility function in each round of the auction. At the end of this paper, we deduce the method of determining the optimal attribute values of Seller agent's goods or services.