基于社会网络与声誉信任机制的移动多Agent系统信任模型

摘要

基于信任机制的移动多Agent系统中,代理Agent一般通过直接信誉值和推荐信誉值来判断对于另一个Agent的信任程度.由于系统相对巨大,直接信誉值通常难以获得,判断的正确性很大程度上依赖于推荐信誉值的准确性和可靠性.通过对整个多Agent系统进行社会网络的挖掘,用以得到与代理Agent存在潜在社会关系的一组Agent.对这组Agent提供的推荐信息充分信任,并优先使用这些Agent提供的信息进行推荐信誉值的计算.最后通过双方直接交易的多寡判断综合信任值中直接信誉值与推荐信誉值的权重.通过实验验证了该模型的有效性.%In trust mechanism-based mobile multi-Agent system, proxy Agent usually judges the trust degree of another Agent by the values of direct reputation and recommended reputation. Because of the relative large of the system, direct reputation is often difficult to obtain, so the correctness of the judgment depends on to a great extent the accuracy and reliability of the recommended reputation. In the paper, by mining the whole social network for entire multi-Agent system, a group of Agent which have potential community relationship with the proxy Agent is gotten. The recommendation information provided by this group is thoroughly trusted, and the information offered by these Agents is employed in prior to calculate the recommended reputation value. At last, the weight of the values of direct reputation and recommended reputation in composite trust value are calculated through the amount of direct transaction between both sides. The validity of the model has been verified by the experiment.