Identifying key nodes in interdependent networks based on Supra-Laplacian energy

Abstract

Identification of fundamental agents in interdependent networks with heterogeneous nodes is a key and challenging topic. It is crucial to understand the topology and dynamic processes of multilayer networks. Nodes in interdependent networks have strong coupling effect which implies that the whole is not the same that the sum of its parts, so it will reduce the collective influence of nodes. Most methods of influential node identification have some defects in that they do not consider coupling between nodes. In this paper, we use Supra-matrix to represent interdependent networks, and propose a novel nodes identification method (SLE) based on Supra-Laplacian Laplacian energy. This method selects a set of influential nodes with minimal coupling effect to maximize the collective influence. The monolayer networks and multilayer networks are used to evaluate the performance of the SLE. The experimental results show that the proposed approach is a good alternative index to identify real important nodes. Compared with monolayer network, SLE algorithm is sensitive in interdependent networks in terms of connected component number, size of giant component, shortest distance and so on.