The Study Of Repair Strategy On Complex Networks

With the discovery of small-world effect and scale-free nature, the fault-tolerant of complex network has become a hot research. Related research shows that complex networks have a strong robustness to response random attacks, and have vulnerabilitiness to response purpose attack. The network nodes suffering selective focus attack or virus infection may result in paralysis of the whole network. Thus,when complex networks are targeted deliberately, the repair strategy and research of complex networks become important.This paper first analyzed the existing repair strategy of complex networks,and summarized the existing repair strategy into four categories from topology, routing protocols, complex networks dynamics and complex networks power-law characteristics. The paper proposed matthew effect-based repair strategy according to the power-law characteristics of complex networks. Using linear detection algorithm, we collected test samples from the real network, and verified power-law properties of complex networks. Then the repair strategy gived simulation repair mainly from two aspects :selective attack on a single node and multi-node cluster attacks. Experimental results show that the repair rate of Matthew effect repaire strategy has reached above 95% in sample Internet network and the BA network. To the repaire degree of complex networks, the paper put forward the concept of stability. Experimental results show that, after a number of attackand repaire, the network can achieve a steady topology state. Based on routing protocols of IP network , we proposed MR-IP-R repair strategy. MR-IP-R repair strategy repaire the attacked route by setting the network nodes and link weights. Experimental results show that MR-IP-R repair strategy is localized, easy to algorithms implement, less impact on the network,and with the load allocation algorithm, after the repaire of the route, MR-IP-R repair strategy optimizes the load distribution. Experimental results show that this algorithm can avoid excessive network load, ,which can lead to network congestion thus affecting the network communication .