Generalized individual-based epidemic model for vulnerability assessment of correlated scale-free complex networks

TL;DR

This paper introduces a generalized SIR epidemic model for any network type, providing a new way to assess network vulnerability to epidemics based on spectral properties and validated through simulations.

Contribution

The paper develops a novel continuous-time Markov chain based SIR model applicable to all network types, enabling comprehensive vulnerability assessment.

Findings

Epidemic threshold inversely proportional to spectral radius.

Disassortative scale-free networks are more vulnerable.

Model validated with extensive numerical simulations.

Abstract

Many complex networks exhibit vulnerability to spreading of epidemics, and such vulnerability relates to the viral strain as well as to the network characteristics. For instance, the structure of the network plays an important role in spreading of epidemics. Additionally, properties of previous epidemic models require prior knowledge of the complex network structure, which means the models are limited to only well-known network structures. In this paper, we propose a new epidemiological SIR model based on the continuous time Markov chain, which is generalized to any type of network. The new model is capable of evaluating the states of every individual in the network. Through mathematical analysis, we prove an epidemic threshold exists below which an epidemic does not propagate in the network. We also show that the new epidemic threshold is inversely proportional to the spectral radius of the network. In particular, we employ the new epidemic model as a novel measure to assess the vulnerability of networks to the spread of epidemics. The new measure considers all possible effective infection rates that an epidemic might possess. Next, we apply the measure to correlated networks to evaluate the vulnerability of disassortative and assortative scalefree networks. Ultimately, we verify the accuracy of the theoretical epidemic threshold through extensive numerical simulations. Within the set of tested networks, the numerical results show that disassortative scale-free networks are more vulnerable to spreading of epidemics than assortative scale-free networks.