节点文献
基于加权紧密度的信任模型研究
Research on the Trust Model Based on Weighted Closeness
【作者】 鲁玉臻;
【导师】 田俊峰;
【作者基本信息】 河北大学 , 计算机应用技术, 2010, 硕士
【摘要】 分布式网络在信息交换、资源共享、协同工作和并行处理等方面显示出的优势,使其迅速成为网络发展的方向。分布式网络具有开放性、动态性、自治性等特性,相对于传统的C/S网络应用模式,分布式网络更注重节点之间的协作。在分布式网络平台的支持下,像普适计算、网格计算、P2P计算、Ad Hoc网络、P2P电子商务、电子政务等大规模分布式系统被广泛应用。分布式网络的广泛应用极大提高了网络的应用范围和活力,同时也对网络管理,特别是对安全管理方面提出了新的要求。分布式网络的性能始终无法达到理论上的最佳状态,其中一个主要原因是节点之间缺乏信任。传统的安全技术已经不能很好的解决分布式系统中出现的安全威胁。信任机制是解决这些问题的有效手段,为保证分布式环境的安全提供了新的思路和方法,已经成为当前的一个研究热点。信任模型是信任机制的基础,由于信任本身的复杂性,构建高效、高精度的信任模型仍然是一个具有挑战性也非常有意义的课题。目前,很多学者采用不同的理论和方法对信任相关理论和技术进行了卓有成效的研究,提出了很多的信任模型。分析现有的信任模型,虽然在信任量化方面做了大量的工作,但是对信任关系的研究还远远不够。本文对信任模型的理论及其关键技术进行了研究,针对目前信任度量中信任的合并问题研究不足的现状,提出一种基于加权紧密度的信任模型,旨在为分布式网络安全和信任机制提供支撑。针对当前信任网络构建合理性问题,给出了一种基于信任源的网络构建方法,并提出用邻接表的方式存储信任相关信息,这种网络构建和存储方法为信任的搜索奠定基础。然后给出了信任搜索和选择规则,以降低信任搜索和选择的时间复杂度。利用加权紧密度实现信任路径的合并,充分考虑了网络中的信任路径和节点间的直接信任和推荐信任,使信任值更加真实的反映实体的客观可信性。同时给出了抑制节点恶意推荐的策略,讨论了信任的时间衰减。最后通过仿真实验证明了模型的有效性。
【Abstract】 Due to the advantages at the information exchange, resources sharing, collaborative work and parallel processing, the distributed network has rapidly become the direction of network developing. The distributed network is characterized by its openness, dynamic property and autonomy. Compared with the traditional C/S network application mode, the distributed network pays more attention to the collaboration between nodes. In support of distributed network platform, such as pervasive computing, grid computing, P2P computing, Ad Hoc networks, P2P e-commerce, e-government and other large-scale distributed systems are widely used. The wide application of the distributed network greatly improves the application range and vitality. And at the same time, it puts forward new requirements on network management and security management in particular. Distributed network performance has never been able to achieve the theoretically best state. One of the main reasons is the lack of trust between the nodes.Traditional security technology has not been a good solution to the security threats arose from a distributed system. Trust mechanism is an effective way to address these issues. It provides new thinking and methods to ensure the security of distributed environments and has become a current research hotspot. Trust model is the basis of trust mechanism. Due to the complexity of the trust itself, building efficient, high-precision trust model is still a challenging and very significant issue.At present, by using different theories and methods, many scholars has carried out fruitful research on the theory and technology of trust and a lot of trust model has been proposed. By analyzing the existing trust model, we can see that although a lot work has been done on the trust quantification, the research on trust is far from enough. The research on the theoretical and key technology of the trust model is carried out in this thesis. Aiming at the current situation that the research on the trust combination in trust measure is not enough, a trust model based on weighted closeness is proposed, in order to support the distributed network security and trust mechanisms. For the current rationality problem of trust network construction, a trust source network construction method based on trust source is given. The adjacency list is used to store the trust information, this kind of network building and storing method lays the foundation for the trust search. The trust search and selection rules are given to reduce the time complexity. The weighted closeness is used in the combination of the trust paths. Considering the trust paths and the direct trust and recommendation trust among the nodes in the network, the trust value reflects the entity’s objective credibility more truly. The Strategy for inhibiting the malicious recommendation is given and the time decay of trust is discussed. The simulation results show validity of the model.