An efficient structure-driven multiplex network dismantling approach based on network percolation

Recent studies show that the structure and function of multiplex networks are greatly affected by a small set of critical nodes, whose breakage might lead to disastrous consequences, such as cascading failures in power grid systems and epidemic spreading across multiple channels. One potential scheme to prevent these catastrophes is to develop methods to identify and protect such a critical set, also known as the network dismantling problem. Hence, in the paper, we present an efficient structure-driven multiplex network dismantling approach, the integrated refining search (IRS) approach, to tackle the network dismantling on multiplex networks such that the critical node set can be minimized, given the fact that all protections would be carried out with the cost. We find that the studied problem can be well characterized by the percolation process and thus we can generalize the relationship-related strategy on single-layer networks to multiplex networks. In particular, we develop our method by integrating a well-designed neighborhood search mechanism and explosive percolation rules. We also conduct extensive experiments on 14 multiplex networks to verify the effectiveness of the proposed methods. Results show that our method is much more capable of finding the optimal critical node set compared to the state of the art.