基于改进D-S证据理论的信任评估模型

针对已有的信任证据模型不能快速有效地处理分布式网络中存在的恶意攻击，且缺乏关于三元信任关系组的信任归一方法，提出了一种基于改进D-S证据理论的信任模型，在此基础上，提出了基于持续序列的基本可信度函数和基于评估函数的信任评估方法，使得模型能更快地抑制恶意节点，并且评估结果更贴近现实值。通过分析与仿真，验证了本模型具有抑制聚集信任攻击的有效性和健壮性，同时信任评估方法更具合理性和准确性。

Trust evaluation model based on improved D-S evidence theory

The existing trust evidence models cannot deal with malicious attacks in the distributed network quickly and effectively, and are lack of a trust normalization method that can measure trust relationship described by a triple set, so a trust model based on the improved D-S evidence theory was proposed. On this basis, basic trust value function based on continuous sequences and trust evaluation method based on evaluation function were also proposed to inhibit malicious nodes in a higher speed, and to enable the prediction results to be closer to the fact. Analysis and simulation show that this model has better effectiveness and robustness to deal with the aggregating trust attack, and trust evaluation method has better reasonableness and accuracy.