Infrastructure Network Reliability Analysis Research

Nowadays the scale of the infrastructure network is successively increasing, and people have a continually higher demand on the network reliability. Network reliability influences the layout and the planning of the infrastructure network construction as a key index. Hence how to calculate the network reliability rapidly and accurately has always been a focus of the trusted computing area.According to the character of the infrastructure network, this paper studied the network reliability based on both edge-failure model infrastructure network and the node-failure model infrastructure network. For the edge-failure model infrastructure network, we presented a redundant subnetwork deleting theory and a corresponding redundant subnetwork reducing technique. This simplification technique firstly found out all the articulate nodes with the help of Tarjan’s algorithm. Secondly, according to the distribution of Keynodes, the technique could finish these two operations one by one following a series of certain rules that deleting all edges belong to on the certain articulate node and then adding back the edges which could connect the keynodes. This network simplification technique could handle with the infrastructure networks to identify and delete the redundant subnetworks on a very small time cost. It helped a lot to ease the memory overflow and the combinatorial explosion problems caused by directly calculating the network reliabilities on these infrastructure networks with large scale and complex structure. This paper selected several infrastructure networks as experimental objects, and the experimental results showed that this pretreatment technique could significantly improve the efficiencies of the reliability analysis algorithms working on infrastructure networks.For the node-failure model infrastructure network, this paper noticed a network feature that a failure node usually leaded to many edges falling down. At the same time, we referenced Kuo and Hardy’s edge-failure model network reliability equivalence BDD construction algorithm, and presented a concept of nodes block, also we set some rules about how to choose the nodes block and how to choose the node within the nodes block. Thus, we actually presented a dynamic node ordering method for constructing the equivalent network reliability BDD for the node-failure and two-terminal infrastructure network. In addition, during the process of equivalent network reliability BDD constructing, we used the corresponding relation between the BDD node and the network structure to present an isomorphism BDD identification theory. Base on the dynamic node ordering method for equivalence BDD construction algorithm, it could be very quickly to judge the connectivity of the source and target, also it could reduce the computer memory pressure with the help of isomorphism BDD identification theory, thus improved the performance and efficiency of reliability analysis for the node-failure and two-terminal infrastructure network.