多Agent系统中信任的动态性处理

Dynamic Trust Processing in Multi-Agent Systems

信任是多Agent系统中进行决策和交互的重要内容。收集必要的信息确定信任关系,动态地管理、维护信任关系,以及监控和重估已有的信任关系是多Agent系统中信任管理的关键问题。虽然研究者对上述关键问题提出了一系列解决方案,但依然存在一些问题有待进一步解决。本文针对信任的动态性处理这一问题,在分析现有的典型信任模型基础之上,提出一个具有动态性的Confidence-Reputation信任模型。模型中,我们不仅考虑了Agent的直接交互历史(Confidence)和信誉(Reputation),同时也考虑了信任的本体性。

Trust is one of the most important concepts guiding decision making and contracting in multi-agent sys- tems. Collecting necessary information to make a trust relationship decision, continuously managing and maintaining rust relationships overtime as well as monitoring and re-evaluating existing trust relationships are the key issues for trust management in multi-agent systems. While previous work provides some solutions to these key issues, there are still some problems remained unsolved especially the trust dynamic processing problem. In this paper we propose a coofidence-reputation model with trust network for trust dynamic processing so that agents can autonomously deal with deception and identify trustworthy parties in multi-agent systems. When evaluating the trustworthiness of the target agent, our approach not only considers the confidence information (based on direct prior interactions with the target agent) and the reputation information from trust network, but also considers the trust's ontological property.