信任路径过滤性搜索与合成算法分析-filtering search and synthesis algorithm analysis of trust path.docxVIP

摘 要随着计算机与网络技术的高速发展，原有的集中式网络系统已不能满足社会需 求，现在的网络新模式，如 P2P 系统、网格计算、普适计算等在资源共享、信息交 换、协同合作等方面显示了极大的优势。与此同时，大规模分布式网络网络动态性、 开放性、自治性等一些特点使得网络系统面临安全的考验。对信任以及信任模型的 研究便是解决安全问题的有效方法。本文首先介绍了信任的定义并描述了信任性质、度量方法等，介绍了大规模分 布式网络的信任模型分类，初步介绍了信任模型。其次，针对建模中信任路径搜索 过程，充分考虑并分析影响信任链搜索的多种因素，包括信任传递衰减，路经成环 等。再次，针对建模中信任路径合成过程，分析信任多条路径在合成信任信息时的 主导思想，同时分析了在合成过程中可能会遇到的三种情况。最后以多 Agent 交互 系统为背景进行验证。现有模型在信任路径搜索方面考虑因素不全面，在信任合成计算方面，将各种 因素的影响同一而论，未区分重要程度，同时忽略了交互次数的重要性。基于以上 几点本文提出了一种路径过滤性搜索算法，并引入了基于交互次数的诚实可信度且 融合多因素，提高了搜索效率，进而提出基于诚实可信度的权重因子，提高了信任 合成的准确性。关键词：信息；信任模型；路径搜索；诚实可信度；信任合成IAbstractWith the rapid development of computer and network technology, the existing centralized network system can not meet the needs of society. And now the new network model, such as P2P systems, grid computing, pervasive computing, have shown a great advantage in resource sharing, information exchange, collaboration and some other aspects. At the same time, the large-scale distributed networks have their own characteristics: dynamic, open, autonomy and so on. These systems with a number of features are facing with the test of making network security. Study of trust and trust model is an effective way to solve the security problems.The paper introduces the definition of trust and describes the nature of trust, the measurement of trust, describes the classification of large-scale distributed network trust model, the some trust model. Secondly, in the path searching of trust model, consider and analyze many factors fully, including the transmission loss of trust, the closed pass. Again, in the path synthesis of trust model, analysis the dominant ideology in the synthesis of multiple paths and analyzes the three cases in the synthesis process that it may encounter. Finally, it verified in the background of the multi Agent interaction system.Trust models that already exist have some deficiencies in searching the trust path: factors affecting the trust value are not considered fully in the searching, and in computing trusted sy