Graph partitioning and multi-criteria TOPSIS analysis for assessing vulnerabilities of critical infrastructure networks

The identification of vulnerable nodes in complex networks has great practical importance in critical infrastructure networks. Because the centrality measures considered individually cannot accurately reflect the centrality of a node in a network, this study combines four relevant local and global centrality metrics in a TOPSIS multi-criteria analysis to evaluate the criticality of the nodes in a network. The results performed on various networks showed a good ranking performance of the proposed method, underlining the importance of aggregating several measures of centrality in the analysis, to mitigate the risks presented by cyber threats.