Diffusion Delay Centrality: Decelerating Diffusion Processes Across Networks

Abstract

This paper presents a methodology to identify agents who should be put into isolation to decelerate a diffusion process spreading through-out a network. We refer to this measure as to Diffusion Delay Centrality (DDC). We show that DDC assigns a high rank to agents acting as the gatekeepers of the fringe of the network. When these are isolated, the spreading of diffusion from the periphery to the core of the network (or vice versa) is prevented or, at least, decelerated. We also show that the ranking of nodes obtained from the DDC is predicted by standard measures of network centrality. Specifically, the ranking can be recovered using the difference in the values of betweenness and eigenvector centrality of network agents. We suggest that the findings presented in this paper might represent a useful tool for policies intending to reduce diffusion processes.