Computer Engineering and Applications ›› 2008, Vol. 44 ›› Issue (4): 219-223.

• 工程与应用 • Previous Articles     Next Articles

Partner selection approach based on multi-Agent negotiation for virtual enterprise

WU Yan,ZENG Jian-chao,BAI Shang-wang   

  1. School of Computer Science and Technology,Taiyuan University of Science and Technology,Taiyuan 030024
  • Received:2007-05-28 Revised:2007-08-02 Online:2008-02-01 Published:2008-02-01
  • Contact: WU Yan

基于多Agent协商的虚拟企业伙伴选择方法

武 妍,曾建潮,白尚旺   

  1. 太原科技大学 计算机科学与技术学院,太原 030024
  • 通讯作者: 武 妍

Abstract: Partner selection is the core problem in the process of virtual enterprise creation.Based on analysis of characteristics of virtual enterprise and negotiation in virtual enterprise,a negotiation model for virtual enterprise is presented,which is a multi-agent multi-issue negotiation model supporting multi-stage negotiation.As an important factor for proposals,the types of agents are defined in this model.The agent updates the information through Bayesian learning method under an incomplete information constraint,and speculates on the type of opponent agent through the sequence of opponent’s proposals to guide itself to select the strategies and tactics,so the proposals avoid blindness,and become more effective.The negotiation model will reduce the time and improve the efficiency of negotiation,so the appropriate partner can be selected in a short time.

Key words: multi-Agent, virtual enterprise, partner selection, negotiation

摘要: 伙伴选择是虚拟企业建立过程中的核心问题,分析了虚拟企业的特点、虚拟企业环境下协商问题的特点,提出了一个适合于虚拟企业环境的多Agent协商模型。该模型支持多Agent多议题的多轮谈判,并将Agent类型引入到协商中来,作为指导协商Agent提议的一个重要因素。在不完全信息的条件下,应用贝叶斯学习的方法,更新既有信息,并通过分析对方Agent的历史提议序列,推测其类型,来指导自身的提议策略和战术,使自己的提议更具有针对性,避免了盲目性,从而节约协商时间,提高了协商的效率,使得盟主企业能在尽短的时间里寻找到理想的合作伙伴。

关键词: 多Agent, 虚拟企业, 伙伴选择, 协商