基于演化博弈的讨价还价策略研究  

Evolutionary game analysis on bargaining strategies

作　　者：詹文杰[1] 邹轶[1] 

机构地区：[1]华中科技大学管理学院,武汉430074

出　　处：《系统工程理论与实践》2014年第5期1181-1187,共7页Systems Engineering-Theory & Practice

基　　金：国家自然科学基金(70871045);教育部留学回国人员科研启动基金

摘　　要：通过建立"双种群"复制动态模型,研究了有限理性假设下"多对多"讨价还价的策略演化问题,证明了只有严格纳什均衡才能成为"多对多"讨价还价的演化稳定策略.并利用计算机仿真发现:当买卖双方种群的初始策略为随机分布时,讨价还价的演化稳定策略以最大概率收敛到对称纳什均衡,产生买卖双方最大初始期望收益乘积的纳什均衡可以比较准确地预测"多对多"讨价还价的演化稳定策略,且演化过程不必是单调的.研究内容有助于理解"多对多"讨价还价的达成协议的一般规律,为设计多边谈判支持系统提供参考和借鉴.By constructing a model of replicator dynamics with two disjoint populations, this paper studies the evolution of strategies in a many-to-many bargaining based on the assumption that traders are boundedly rational. It is proved that only strict Nash equilibriums can be the evolutionarily stable strategy （ESS） in the bargaining. Simulation results show that the ESS converges to the symmetric Nash equilibrium with the maximal probability if fractions of buyers＇ and sellers＇ strategies are randomly initialized. Moreover, the Nash equilibrium, which generates the maximal product of initial expected return of two sides, accurately predicts the ESS in most cases. Finally, the evolutionary process of ESS is not necessary to be monotonous. The result helps to understand the general pattern in which a many-to-many bargaining comes to terms, and serves as reference for designing multilateral negotiation support system.

关 键 词：讨价还价 演化博弈 纳什均衡 演化稳定策略 

分 类 号：F063.1[经济管理—政治经济学]