Research On Virus Propagation Model And Immunization Strategy Of Complex Networks

Currently, the research on complex network has become mature; and it has permeated various disciplines. The theory research is not only limited to Mathematics but also life discipline and engineering discipline and so on. Simultaneously it has attained a series of achievements in computer network control, society network analysis and biological network. With the study of complex network, the transmission mechanism has become one of the important branches. Propagation exists widely in nature and human life, such as virus propagation, rumor propagation. While with the closer communication among people, virus outbreak in the network and disease propagation in social network will cause enormous impact on human life and economic. All these problems prompt researchers to study propagation law. The research on complex network includes virus propagation model and network immunization strategies.With the large-scale outbreak of epidemic in social network, computer virus are devouring the Internet and other networks. Because email is one of the most frequent application in network, so virus propagation attaching to email has appeared and caused a series of losses. Traditional virus propagation relies mainly on individual contact, including Susceptible-Infected (SI) model, Susceptible-Infected-Susceptible (SIS) model, and Susceptible-Infected-Removed (SIR) model and so on. However, email virus propagation is effected not only by individual contact but also other factors. For example, background knowledge of users, the ability of self-protection and trust level among users. Therefore, traditional epidemic propagation model is no longer suitable for describing such kind of virus propagation. At the same time, according to the topology of the network, some researchers have proposed random immunization, acquaintance immunization, target immunization and some improved immunization strategies. So the establishment of a reliable model and related immunization strategies are considered as a challenging problem. This dissertation proposed an improved propagation by analyzing the spreading way and characteristics of email virus propagation. Also, we propose a dynamic immunization algorithm based on importance of the nodes for interactive email virus propagation model. The main work is summarized as follows:1) Research on virus propagation model in complex networkCurrently, the existed virus propagation model is just based on individual contact. By analyzing the spreading way and characteristics of email virus propagation, we discover background knowledge of users, the ability of self-protection and trust level among users also effect virus propagation. Therefore, this dissertation presents a novel interactive email virus propagation model based on users’ ability of self-protection and trust level. Our simulation experimental results show that the proposed model can accurately describe the process of spreading of virus, and the number of infected individuals in the network is also quite low. At the same time, we research the effect of different factors on virus propagation in the novel model, which provides a novel method for network immunization.2) Research on network immunization strategyMost of immunization strategies only make use of local degree information or betweenness of nodes. This method is not suitable for large-scale network and ignores of the influence of the nearest neighbors and the next neighbors on virus propagation. Therefore, considering the effect of the nearest neighbors and the next nearest neighbors, we present an immunization strategy by local central value to compute the importance of nodes. Also we analyze and compare advantages and disadvantages between target immunization and the novel immunization strategy based on the importance of nodes proposed in this thesis. The experimental results prove dynamic immunization can reduce effectively the speed of outbreak, and has better immunization efficiency.