Abstract: This paper proposes an automatic negotiation algorithm based on co-evolutionary algorithm, and simulates the strategy learning mechanism in the finite horizon alternating offer negotiation protocol. Experimental results indicate that the automatic negotiation based on co-evolution can produce the approximate strategy profile as that of the sub-game perfect equilibrium, and has a good prospect of application.
automatic negotiation; co-evolution; genetic algorithm; alternating offer

Key words: automatic negotiation, co-evolution, genetic algorithm, alternating offer

摘要： 提出了基于协同进化遗传算法的自动谈判算法，模拟了有限期轮流出价谈判协议中的策略学习机制。实验结果表明，基于协同进化的自动谈判能够生成近似于子博弈完美均衡的策略组合，具有良好的应用前景。
自动谈判；协同进化；遗传算法；轮流出价

关键词: 自动谈判, 协同进化, 遗传算法, 轮流出价

CLC Number: 

• TP391