泛在网中多终端协同信任机制

发布时间：2018-06-05 12:29

本文选题：泛在网 + 模糊推理　；参考：《南京理工大学》2014年硕士论文

【摘要】：当前,国内外的许多城市都在推进“泛在城市”、“智慧城市”研究与建设,希望以此为契机带动整个信息产业,乃至促进经济和社会的发展。泛在网中,最为核心的问题是如何确保异构终端间进行可信的通信。只有让异构终端实现了可信交互,才能实现多种异构终端间的通信,才能真正实现“泛在”。现在的大部分信任模型中,推荐服务仅提供了一种单一的信任值作为推荐。但是由推荐者提供的这种单一信任值代表了它对未知实体的主观观点,不能很好的反映出在某种特定环境下的信任等级。为解决这个问题,本文提出了一种有效的基于可信度的间接信任计算方法。该间接信任计算方法不仅能计算出每个诚实推荐的可信度,还可以对不诚实的推荐进行过滤。其次,本文提出了一种基于信誉和模糊理论的综合信任模型CFMBRF,该模型采用“基于可信度的间接信任计算方法”作为间接推荐的方法,并在信誉模型的基础上,引入模糊推理子系统,使其能处理“交互重要性”、“不确定区域的决策”和“交互结果的设定”等主观概念,可以人性化的执行不确定性的决策。最后进行了大量的仿真实验。一方面,对“基于可信度的间接信任计算方法”的有效性进行验证,结果证明该间接信任计算方法不仅能够过滤不诚实的推荐,还能计算出每个诚实推荐的可信度。另一方面,对没有使用任何模糊子系统的网络和使用了模糊子系统的网络进行模拟和对比分析,并测试了CFMBRF模型的系统开销和稳定性。实验结果表明使用了模糊子系统之后,终端间交互的正确性大幅提高,信誉向量的查询次数明显减少,证明了CFMBRF模型的有效性。
[Abstract]:At present, many cities at home and abroad are promoting the research and construction of "ubiquitous cities" and "wise cities", hoping to take this opportunity to drive the whole information industry and even promote the development of economy and society. In ubiquitous networks, the core problem is how to ensure credible communication between heterogeneous terminals. Only when the heterogeneous terminal realizes the trusted interaction, can the communication between the heterogeneous terminals be realized, and the "ubiquitous" can be realized. In most trust models today, recommendation services provide only a single trust value as a recommendation. However, the single trust value provided by the referees represents its subjective view of the unknown entity, and does not reflect the trust level in a certain environment. In order to solve this problem, this paper proposes an effective indirect trust calculation method based on credibility. The indirect trust calculation method can not only calculate the credibility of each honest recommendation, but also filter dishonest recommendation. Secondly, a comprehensive trust model based on reputation and fuzzy theory is proposed in this paper. The model adopts "indirect trust calculation method based on credibility" as the indirect recommendation method, and based on the credibility model. The fuzzy inference subsystem is introduced to deal with the subjective concepts such as "interaction importance", "decision of uncertain region" and "setting of interactive result", which can humanize the execution of uncertain decision. Finally, a large number of simulation experiments are carried out. On the one hand, the validity of "indirect trust computing method based on credibility" is verified. The results show that the indirect trust calculation method can not only filter dishonest recommendations, but also calculate the credibility of each honest recommendation. On the other hand, the network without any fuzzy subsystem and the network with fuzzy subsystem are simulated and compared, and the system overhead and stability of CFMBRF model are tested. The experimental results show that the correctness of the interaction between terminals is greatly improved and the query times of the reputation vector are obviously reduced after the fuzzy subsystem is used, which proves the validity of the CFMBRF model.
【学位授予单位】：南京理工大学
【学位级别】：硕士
【学位授予年份】：2014
【分类号】：TP393.08

【参考文献】