一种基于灰色系统理论的分布式信任模型

在开放多Agent系统中,引入信任模型有利于选择出保证收益的交互伙伴,发挥灰色系统理论在解决贫信息、不确定性问题中的优势,构建了GTrust信任模型,设计了对证人推荐行为进行评定的方法,用灰色定权聚类的方法判定对证人的信任,克服证人提供的不准确信息带来的负面影响,指导评价者Agent合理使用来自证人的推荐.仿真实验表明,GTrust忽略了对不准确信息来源的甄别,而是只以实际交互结果为依据的策略,针对复杂的环境是一种切合实际的选择,无论在静态环境中,还是动态环境中,均可以取得好的效果.

A distributed trust model based on grey system theory

The trust model can be introduced into open Multi-Agent System（MAS） to help an agent to choose interaction partners.A distributed trust model GTrust is proposed to resolve the inaccurate information problem,in which the grey system theory is adopted.In GTrust,in order to guide the agent to utilize the recommendation from witnesses reasonably and avoid the negative impact from all kinds of inaccurate factors,a mechanism is designed to rate the witnesses＇ recommendation action,and the grey fixed weight clustering approach is used to determine the trust to witnesses.GTrust makes decision only based on the actual interaction results and ignores the source of inaccuracy,which is a practical choice in the complex environment.Simulation results show that GTrust can achieve better utility than other trust models either in a static environment,or a dynamic environment.