A New Network Robustness Topology Measure based on Information Theory

TLDR: A novel methodology to measure the topological resilience and robustness of a network based on Information Theory, which can be used with any probability distribution able to represent the network’s properties is proposed.

Abstract: Industrial and Systems Engineering, University of Florida, Gainesville, FL, USA(Dated: November 3, 2014)A crucial challenge in network theory is to study how robust a network is when facing failuresor attacks. In this work, we propose a novel methodology to measure the topological resilience androbustness of a network based on Information Theory quanti ers. This measure can be used withany probability distribution able to represent the network’s properties. In particular, we analyzethe eciency in capturing small perturbations in the network’s topology when using the degree anddistance distributions. Theoretical examples and real networks are used to study the performanceof this methodology. Although both cases show to be able to detect any single topological change,the distance distribution seems to be more consistent to reect the network structural deviations. Inall cases, the novel resilience and robustness measures computed by using the distance distributionreect better the consequences of the failures, outperforming other methods.