基于GA和Agent组合自动谈判方案的研究  

On the Automatic Negotiation Scheme Combining GA with Agent

作　　者：王贵和 姜兴宇[1] 范利敏[2] 王宛山[1] 

机构地区：[1]东北大学机械工程与自动化学院,辽宁沈阳110004 [2]辽东学院,辽宁丹东118000

出　　处：《东北大学学报（自然科学版）》2007年第11期1636-1639,共4页Journal of Northeastern University(Natural Science)

基　　金：辽宁省重大科技攻关项目(2003220025)

摘　　要：分析了Agent的性能特点及其工作原理,研究了谈判收益问题;然后对谈判过程进行描述,建立了基于Agent自动谈判数学模型,并对模型中的参数以及自动谈判目标进行分析;提出了用遗传算法对模型进行求解,并对算法的编码方法、参数的设定等进行研究,分析了求解过程,得到了模型的最优解.结果表明,该方法可以对模型不断优化,最终使参与谈判的Agent双方实现双赢,从而证明了该自动谈判方案的正确性.To solve the problem of automatic negotiation, the characteristics and operating principle of an agent and the proceeds through negotiation were analyzed. A model based on agent was developed with the process of negotiation described including relevant parameters and goal. A strategy based on genetic algorithm was presented and its coding method and parameter setting and the solution process were studied with an optimum solution given to the model. The results showed that this strategy can optimize continuously the model and enable both agents who negotiated with each other to realize a win-win situation, thus verifying the validity of the scheme of automatic negotiation.

关 键 词：代理商 自动谈判 遗传算法 谈判收益 最优解 

分 类 号：TP18[自动化与计算机技术—控制理论与控制工程]