Exploiting the weakness of preferential attachment networks

We address the general problem of how to attack and destroy a network by node removal given limited or no prior information about the edges. Networks have been used to describe many kinds of systems. In general, nodes represent systems components and edges the interactions between them. How the edges are arranged in a network has great importance because quantities of interest depend on edge placement, e.g. connectivity distribution, clustering coefficient, resilence to node and edge removal, spreading processes, and small-world effects. The rules controlling edge placement define the network structure and they can be exploited by agents that wish to attack weaknesses of the networks. In our study, we consider a family of strategies in which nodes are randomly chosen, but not removed. Instead a random acquaintance (i.e., a first neighbour) of the chosen node is removed from the network.
Our approach is a generalization of the strategy introduced by Cohen et. al. [Phys. Rev. Lett., 91 (2003)], in which the acquaintance of a randomly chosen node is promptly removed from the network as soon as it was chosen. Instead of the immediate removal, a given node needs to be pointed by other randomly chosen nodes more than once before being removed. As a result, we observe that our approach leads the network to be destroyed more quickly, i.e., it's necessary to remove a lower number of nodes, in comparison to the original strategy.