부산대학교 도서관

We investigate the effects of heterogeneous and clustered contact patterns on the timescale and final size of infectious disease epidemics. The abundance of transitive relationships (the number of 3 cliques) in a network and the variance of the degree distribution are shown to have large effects on the number ultimately infected and how quickly the epidemic propagates. The network model is based on a simple generalization of the configuration model, and epidemic dynamics are modeled with a low dimensional system of ordinary differential equations. Because of the simplicity of this model, we are able to explore a large parameter space and characterize dynamics over a wide range of network topologies. We find that the interaction between clustering and the degree distribution is complex, and that clustering always slows down an epidemic, but that simultaneously increasing clustering and variance of the degree distribution can potentially increase final epidemic size. In contrast to solutions for unclustered configuration model networks, we find that bond percolation solutions for the final epidemic size are potentially biased if they do not take variable infectious periods into account.
Comment: 17 pages, 4 figures