摘要： 基于模糊相似关系，提出一种面向电子商务实际应用的多属性自动协商系统方法。该方法在交易过程中依据单调让步协议，按照一定规则作出适当让步，在保证自身效用前提下，利用启发函数搜索与对方上次提议最接近的提议，从而最大化某一社会福利函数。将该方法与博弈论中经典的Zeuthen策略做了对比实验，仿真方法验证了该方法的有效性，说明了参数对算法性能的影响。

关键词: 交易agent, 电子商务, 模糊相似, 自动协商系统

Abstract: This paper puts forward an electronic commercial oriented automatic negotiation system based on the fuzzy similarity function. In each period of negotiation, agent concedes according to certain rules. On condition of assuring the required utility, agent tries search for an offer which is the most similar to the offer given by the opponent in the last period. The effectiveness of the method is tested by comparing it with Zeuthen strategy, which is the typical and theoretical result in principled negotiation. The enffluence of some parameters of the algorithm is given at the end of the paper.

Key words: trading agent, electronic commerce, fuzzy similarity, automatic negotiation system

中图分类号: 

• TP18