Abstract:

In daily life, complex networks mainly undergo two types of damages: random failure and selective attack. The structural features of various network models make the capability of complex networks that resist attacks very different. Considering the invulnerability parameters and structural properties of complex networks, we have a comprehensive evaluation on of a small-world network by measuring node average degree, max degree, average betweenness centrality and max betweenness centrality under various attacks. With the removing of nodes, we estimate the invulnerability of the network by analyzing the relationship of node degree and betweenness centrality. The experimental results show that small-world network is more vulnerable under RD, while has strong robustness under other attack strategies. The experimental results show that the relationship between node degree and betweenness centrality of a small-world network is linear in some range, while it is irregular in most regions. The strong robustness of WS small-world networks has an important influence on building real networks. According to the structural features of small-world networks, we should not put the function of networks on a few Hub nodes. To improve more efficient strategies, we will make a further research.

Key words: complex network, small-world network, invulnerability, betweenness-degree correlation, attack strategy