基于合同网协议的多智能体虚拟企业协商研究

协商是虚拟企业成员之间协作的实现方式。本文通过对多智能体的协商框架、协商模型和协商过程等进行深入的分析,在合同网协议的基础上构建多智能体协商体系框架,并结合贝叶斯决策方法来建立自学习协商模型。根据协商中的更新信息实时更新智能体信念,实现多智能体协商过程的自学习功能。从而有利于提高协商效率,改善网络通信,保证虚拟企业的敏捷性和协商双方的利益。

Negotiation Research in Virtual Enterprises Based on Multi-agents

With the demand of modern global market competence,negotiation has been used as a mechanism to solve the task distribution,coordination and conflicts in virtual enterprises. Contract net protocol tends to bring communication congestion when negotiation is used. In this paper,the Bayesian decision making is imported to extend the contract net protocol,and a multi-agent negotiation system architecture is constructed. Agents with incomplete information about their opponents can update their beliefs when considerating what other agents are thinking about during their interactions. The initial experimental results show that the learning model reduces the information exchange frequency between agents and increases their utility. This ensures that the virtual enterprise is agile and fulfills win-win during negotiation.