Epidemic Spreading And Controlling In Complex Community Networks

The rampancy of computer viruses on the Internet and the propagation of epidemic inpopulation has brought great loss to human beings, and the fast-developing theory of complexnetworks can provide solutions on how to controlling the spreading of virus by investigating theirpropagation behavior. With the gradual deepening of research on complex networks, the communitystructure is found to exist in lots of real-world networks in meso-scale which affect both trafficdynamics and epidemic spreading and some epidemic will spread only when there is a kind oftraffic dynamics on the network. In this dissertation, some strategies are proposed to enhancecapacity of network and control spreading of epidemic. The main research work and contributionsof the dissertation are as follows:1. The impact of community structure on epidemic spreading in homogeneous networks andcorresponding immunization strategy are studied based on mean-field theory. The epidemicspreading in homogeneous networks is thought to be related to the average degree in the traditionalresearch. It is found that community structure has influence on epidemic spreading and pronouncecommunity structure will reduce the speed of epidemic spreading in homogeneous networks withsame average degree. Target immunization strategies based on shortest-path betweenness andrandom-walk betweenness are proposed to control the epidemic spreading in homogeneousnetworks.2. The traffic dynamics in community network and the impact of community structure on traffictransportation are investigated in this dissertation. It is found that networks with pronouncedcommunity structure are less efficient in terms of packet delivery in both shortest path routingstrategy and efficient path routing strategy. An optimal routing strategy based on communitystructure is proposed which can reduce the betweenness centrality of the nodes by minimizing thenumber of the communities through which the routing path passes. Simulations show that the newstrategy can enhance the packet delivery capability with the small-world character, and the moreaccurate the community is identified, the more efficient the new strategy is.3. The traffic-driven epidemic spreading and corresponding controlling strategy in communitynetworks are discussed. The speed of epidemic spreading will increase with the enhancement oftraffic flow. The speed of epidemic spreading and the critical epidemic threshold have a high correlation with the nodes average routing betweenness in homogeneous networks while with theratio between the first and second moments of the nodes routing betweenness distribution in scalefree networks. Different from the traditional epidemic spreading models, pronounce communitystructure will improve the speed of epidemic spreading in the traffic-driven epidemic spreadingmodel. Moreover, a controlling strategy based on community structure is presented which cansuppress epidemic spreading in scale-free networks. The more accurate the community is identifiedand the stronger community structure the network has, the more efficient the new strategy is.4. Finally, the influence of network topology on epidemic spreading is studied. While thewell-known immunization strategies may break the integrality of the network, the network topologycan be changed by adding or deleting some edges so as to suppress the epidemic spreading. Inscale-free networks, kicking out edges according to the product of the degrees, the product ofshortest path betweenness, and the random walk betweenness of two nodes of the edge can raise thecritical spreading threshold. And the critical spreading threshold will firstly increase and thendescend while removing edges continuously. On the contrary, adding those higher weight edges isbeneficial to suppress the epidemic spreading and enhance the critical spreading threshold.Moreover, the more pronounced community structure the network has, the better the newcontrolling strategy works.