摘要： 针对现有模型在节点行为变化时对诚实推荐节点的误判问题，提出一种基于可信推荐节点集合的P2P信誉模型。在该模型中，节点依据其推荐性能被划分为2个集合：可信推荐节点集合及考察节点集合。计算节点信誉值时仅使用前者的推荐信息，对后者中的节点进行实时考察以纠正被误判的诚实节点。网络节点依据其推荐性能的变化在2个集合内动态转化。仿真实验结果验证了该模型的有效性。

关键词: 对等网络, 信誉模型, 可信, 推荐

Abstract: A credible recommendation node set-based reputation model is proposed in this paper which solves the problem in the existent models that the miscarriage of justice for the honest recommenders when peers change their behavior. It divides the recommenders into two sets: the credible recommender set and the set to be taken a further look. The recommendations from the former set constitute the computation of peers’ reputation value. The behavior of the latter set is observed to correct the honest recommender that has been done wrong. Peers transform dynamically between these two sets according to their recommendation performance. Experimental results show the model’s validity.

Key words: P2P network, reputation model, credible, recommendation

中图分类号: 

• N945