一种基于完全信息扩展博弈的自动信任协商策略

A Dynamic Game Based Automated Trust Negotiation Model

提出了一个基于完全信息扩展博弈的自动信任协商策略,该策略依据子博弈精炼纳什均衡来制定.首先,通过分析比较,将自动信任协商过程转化为完全信息扩展博弈过程,重点讨论了信任博弈树的构造算法和支付函数的计算方法;然后讨论了协商双方效用函数的计算,以此为基础提出了一种新的自动信任协商策略,按照该协商策略进行协商可以使协商双方利益最大化.

In this paper we model the ATN process as an Extensive Game with Perfect Information （EGPI） and conclude that the equilibrium of the game is the most desirable result of the ATN process. A EGPI based ATN is modeled as a quadruple. After the model is presented, the algorithm of building a Trust Game Tree and the method to calculate the payoff function is heavily discussed. Then the subgame perfect equilibrium set is found using the technique of Backward Induction. The ATN model presented in this paper can effectively prevent the privacy of negotiators from unnecessary exposure and save negotiation time by lesser exchange of credentials.