Researches On The Complex Networks Attack Efficiency Of PageRank And HITS Algorithms

Complex networks exist everywhere in our life, therefore, it is a very important issue in the scientific field of network science for protecting them from the attacks outside or collapse. In order to protect complex network from malicious attacks, we need to have a thorough understanding of the vulnerabilities of complex network and the attack strategies of attacker. In the field of Complex Networks Attack, the traditional strategies have some shortcomings. One problem is the computation speed. The algorithms may be too slow for large-scale networks. The other problem is that the effect of network attack may be not good enough. Therefore, looking for a better attack strategy that can solve those problems simultaneously, became a new topic in the research of complex network.For each kind of selective attack strategy, there must exist a rank ordering mapping into the selection strategy of nodes/edge. The sorting algorithm is the core of the selective attack strategy. Page Rank and HITS algorithm can be considered as a node sorting algorithm in the search engine. Therefore, those two algorithms can be used to construct a selective attack strategy. Those two kinds of algorithms have a suitable computation speed for large networks. We introduced the Page Rank algorithm and HITS algorithm for network attack strategy in this thesis. In addition, we also make a deeply discus the applicability and effectiveness of these two kinds of attack strategy in the complex network attack area.To compare the performance of Page Rank and HITS, this thesis use selective node degree attack strategy and selective betweenness node attack strategy as the references in the simulation. Moreover, we choose six real networks as the experimental networks. In the simulation experiments we use selective nodes attack model under considering the cost against the attacks. We also use quantitative evaluation model to analyze the attack effect of network against the attack.The experiment results show that the betweenness attack strategy is not stable; HITS strategy presents a poor performance; the degree attack strategy is quite stable, however, it is not the best for the effect; Page Rank attack strategy showed a great performance in six experiments.Thus it comes to the conclusion that: Page Rank attack strategy is suitable for the research of network attack.