Dynamical Immunization of Data Network Controllability Processes Against Centrality Attacks on Temporal Networks

Abstract

Abstract The main goal of controllability network methods on temporal networks is to control all nodes with the minimum number of control nodes. Real-world temporal networks are faced with many errors and attacks that cause the network structure to be changed in some way so that the controllability processes are disturbed and the network is out of fully controllability mode. One of the most important attacks on temporal networks is centrality attacks. In this paper, the types of centrality attacks and their destructive effects on the controllability of temporal networks have been investigated. To deal with centrality attacks, strategies have been proposed on dynamic graph by adding new control nodes or adding new links to the network so that the network is protected against centrality attacks. The results of simulation and comparing them with conventional methods demonstrate that the proposed node addition strategy has performed better than other methods and the improvement rate in terms of execution time is 60%. On the other hand, the proposed immunization strategy by adding links has kept the network controllable with a smaller number of links (38%) and less execution time (52%) compared to other methods.