基于第三方仲裁的双边多议题自动谈判模型

Third-party mediation based bilateral multi-issue automated negotiation model

提出一种基于遗传算法且引入第三方仲裁Agent、L-K型同步报价协议、公平优先兼顾效率的评价判据C5、以及精英、偏好和随机选价方式的双边多议题自动谈判模型.实验结果表明:用该模型能在信息背对背条件下获得双赢的谈判结果;C5充分体现了公平优先兼顾效率的谈判原则,采用该判据容易找到最充分地体现了该原则的纳什解;L和K是影响谈判结果的关键因素,当L和K较小但取值适当时精英选价方式是最好的报价选择方式;增加两者都能改善谈判结果,彼此起到互补作用;L足够大时谈判结果不再受其他因素影响,纳什均衡报价将是唯一的成交报价.

A genetic algorithm based bilateral multi-issue automated negotiation model with a third-party mediator agent,a L-K type synchronous offer protocol,a fairness-first and efficiency-second evaluation criterion C5,as well as an elite,a preferential and a random offer choice method is proposed.The experimental results show that:The model can be used to obtain a win-win negotiation result under the condition of private information;C5 fully reflects the negotiation principle of fairness-first and efficiencysecond,and it is easy to find Nash solution that fully reflects the principle by using this criterion;L and K are the key factors that affect the negotiation result,when L and K are small but valued appropriately,the elite offer choice method is the best offer choice method;Increasing both L and K can improve the negotiation result,and they play a complementary role with each other;When L is large enough,the negotiation results will no longer be affected by other factors,and Nash equilibrium offer will be the unique closed offer.