NSKSD: Interdependent Network Dismantling via Nonlinear-Metric

Abstract

Most networks are not isolated but interdependent in real applications. Studies on the dismantling of interdependent networks have motivated crucial and significant improvements to our understanding of fundamental network behaviors, i.e., function, robustness, structure characters, etc. The popular way to deal with such a dismantling problem is to extract the influential spreaders based on centrality measures. This brief proposes the node significance (NS) for individual node measures based on the novel sigmoid-like similarity calculation. A new index KSD is proposed to identify the top influential nodes, combining the different node centralities’ effects. Notably, we put forward two types of significance, i.e., the overlapping node significance (ONS) and the overlapping KSD (OKSD); both are set as benchmarks for computing the spreading capability of each node. Simulation results validate that our proposed NSKSD is much better than the state-of-the-art methods in terms of the efficient and effective dismantling of interdependent networks.