Contagion dynamics on a compound model

Abstract

In the real world, a single homogeneous or heterogeneous network is very rare, so we consider a compound model which contains two different types of subnetworks. The compound model may be applied to analyze the contagion dynamics of two different communities where they have a probability to communicate. The global infection threshold is obtained, and many factors affecting spread of epidemics are analyzed, including subnetwork density, subnetwork size, connecting probability and the maximum degree. An interesting phenomenon is that we find the transition points of the global threshold from increasing to decreasing, which was not mentioned in previous work. Therefore, we give many measures to improve infection threshold and prevent the spread of epidemics. The experimental results confirm the conclusions of epidemic analysis.