IMPACT OF STRUCTURAL CENTRALITY BASED ATTACKS IN COMPLEX NETWORKS

In this paper, we study a new strategy to find the influential nodes in the complex networks. This strategy is based on Structural Centrality (SC) of the node in the network. In this strategy, by using graph spectral analysis of the network, we find the hierarchy of the influential nodes in the form of central nodes in the network. The structural centrality of each node is ranked in the topology of complex networks which are modeled as the scale free networks. We have explored the structural centrality based targeted attack and compared our result with the degree based targeted attack. The robustness of the real world complex network has been measured efficiently against the degree, structural centrality based targeted attack and compared with the random attack and compared it. In the social networks, the mechanism to suppress the harmful rumors is of great importance. A rumor spreading model has been defined using the susceptible-infected-refractory (SIR) model to characterize the rumor propagation in the social networks. Inoculation strategy based on the structural centrality has been applied on the rumor spreading model for the heterogeneous networks. It is compared with the random and degree based targeted inoculations. The nodes with higher structural centrality are chosen for the inoculation in the proposed strategy. The structural centrality based targeted inoculation strategy is found to be more efficient in comparison to the random and degree based targeted inoculation strategies. One of the bottlenecks of this approach is the high complexity in computing the structural centrality of the nodes in the complex networks with very large number of nodes. Further, appearance of giant component has been studied in the network with random attacks, and degree and structural centrality based attacks. The proposed hypothesis has been verified using simulation results for e-mail network data and also for the generated scale free networks.