Research On Epidemic Spreading Based On Network Structure

The correlation between structure and propagation behavior is a significant aspect when referring to virus propagation in complex networks. Thus researches on network construct and evolution process of the virus propagation will be great helpful to reveal the determinant of structural dimension about different kinds of virus outbreak. The main work is summarized as follows:First, individuals may respond to epidemic by adjusting their behavior of contact with sick individuals to reduce the risk of infection. In this paper, an adaptation mechanism is proposed where healthy individuals may temporarily protect their contacts with unfamiliar individuals to reduce the risk of infection when look for healthy neighbors, and allowing restoration once ensuring the neighbors’ health. A mean-field description of this system is developed to predict the epidemic threshold, and approximate the network dynamic using Monte Carlo simulation. Simulation results demonstrate that the epidemic threshold depended on part of parameters, the link protection and rewiring edges become central to resist disease in the different parameter regimes, respectively. Additionally, the link protection mechanism can greatly enhance the community structure.Second, the interrelationship between network parameters and propagation size is discussed. Four kinds of tunable network topologies are involved: the tunable networks of clustering coefficient, average path length, assortativity coefficient and power exponent. The paper mainly researches the establishment process about the tunable network of clustering coefficient, and a network model with wider scope of tunable cluster coefficient is proposed. According to the simulation results, a negative correlation is presented between cluster coefficient and infection density; and there’s no obvious relation between power exponent and infection density.Finally, four kinds of network parameters are considered to research the relationship between network parameters and propagation velocity. They are cluster coefficient, average path length, assortativity coefficient, and power exponent. From simulation results, a network with low clustering shows a better propagation velocity to virus propagation, while one with low assortativity coefficient shows fragility to propagation velocity. What’s more, the power exponent does not appear to have a significant influence on propagation velocity.