【摘要】 针对资源共享系统中节点的恶意行为,提出一种基于证据理论的群组信任管理模型,在节点间服务与被服务的过程中建立信任关系,抑止恶意节点的破坏行为.首先根据改进的直方图法计算节点直接信任值,利用改进的冲突信任加权证据组合规则计算推荐信任值;然后融合本地信任值和其它节点推荐信任值完成对目标节点的最终信任评估;最后进一步通过服务节点选择算法选择可信度较高的节点作为服务提供节点,避免服务请求节点得到错误或破坏性的资源.通过仿真实验对此模型在对抗恶意攻击时的有效性进行了验证.更多还原

【Abstract】 A group trust model based on D-S theory was presented to overcome the Malicious behaviors in the distributed network system,in which trust among participants can be set up and the Malicious behaviors can be alleviated.first,An improved histogram and a Trust-Weighted proportional Conflict Combination methods were used to calculate the node′s local trust and other nodes′ recommended data respectively;then combine both of calculated data to get the final trust value of the object node;finally,a server selection algorithm was proposed to ensures that the node with higher credible be selected as service providers.Simulations show that the proposed model can effectively restrain the malicious behaviors in the system.更多还原