The Research On Reliability Problems In Network Systems

Due to the rapid development of information technology industry and the fast rise of internet networks, the complexity of networks is increasing too, fault possibility of net-work components are rising steeply, the network failures will bring about a huge economic losses or even disaster. So both the research community and the government-user agencies attaches great importance to the study of network reliability.With the focus on resolving major problems connecting network reliability analysis, we endeavored to made two researches in our thesis:(1) based on cellular automaton, the network fuzzy reliability has been analyzed. (2) based on the research (1), the effect of network components on network reliability has been calculated.The primarily work and achievements of the paper is as follows:(1) The fuzzy reliability analysis of network systems. To incorporate the important effect of the different types of membership function for the network components'fail-ure rates, a method for solving the network fuzzy reliability has been put forward in this thesis. Meanwhile, a cellular automaton based fuzzy reliability calculation method has been proposed to overcome the complexity of modern network. What's more, our paper proceed a lot of numerical experiments, it reveals the the computational efficiency and calculation accuracy of the proposed methods.(2) Based on the topological decomposition model proposed in (1), three key param-eters which affect the quality of network component importance has been introduced in our paper:Birnbaum measurement, Critical importance, Risk increment factor. Firstly, a network component importance evaluation model has been proposed in our paper to cal-culate the weight of network components. Secondly, we proposed a method to analyze the component's risk increment factor and stability of reliability. The experiment shows that the proposed methods can exactly indicates the weakness and the fault component with the most probability in network systems.The theory and the practice all proved that, complex network connecting probabil-ity portrayed the complex network reliability. Fuzzy quantities were used to represent the components'failure rates and network reliability, which would be more accurate to reflect the true nature of network component and network reliability. The theoretical founda-tion of complex network protection has been strengthened through component importance analysis.