To accurately represent disease spread, epidemiological models must account for the complex network topology and contact heterogeneity. Traditionally, most studies have used random heterogeneous networks, which ignore correlations between the nodes' degrees. Yet, many real-world networks exhibit degree assortativity—the tendency for nodes with similar degrees to connect. Here we explore the effect degree assortativity (or disassortativity) has on long-term dynamics and disease extinction in the realm of the susceptible-infected-susceptible model on heterogeneous networks. We derive analytical results for the mean time to extinction (MTE) in assortative networks with weak heterogeneity, and show that increased assortativity reduces the MTE and that assortativity and degree heterogeneity are interchangeable with regard to their impact on the MTE. Our analytical results are verified using the weighted ensemble numerical method, on both synthetic and real-world networks. Notably, this method allows us to go beyond the capabilities of traditional numerical tools, enabling us to study rare events in large assortative networks, which were previously inaccessible.

Close

PrevNext

View figure in article

• Stochastic processes
• Hamiltonian systems
• Epidemic spreading
• Large deviation & rare event statistics
• Master equation
• Monte Carlo methods
• SIS model
• Semiclassical methods