Research The Dynamics Behavior And Immune Control Strategies Of Epidemic Spreading On Complex Networks

Throughout the history of the development of human society, it is a history of struggle of fight for people with all kinds of viruses. From early measles and smallpox to the pneumonia and A(H1N1) influenza in recent years, Each large-scale spread of epidemics gave life and property of the people to bring a huge disaster. In addition, with the development of information technology, the network has gained wide popularity and application; varieties of computer viruses have quickly spread through the Internet. Its widespread formed a great threat to the whole computer network security. Therefore, process modeling the spread of the virus based on understanding the characteristics and laws of the virus spreading. Then design effective prevention and control strategies based on predicting viral trends and analysis of the epidemic spreading causes and key factors that has important theoretical and practical significance.In the past, people mainly studied the network topological and network characteristics effects on virus spreading behavior, and we generally studied it on the particular nature of the network. By understanding the current related works, we improved the classic SIRS virus propagation model and analysis of the propagation behavior in the homogeneous network and heterogeneous network. By using the theoretical analyses and numerical simulations, we gained the epidemic threshold of the model. In addition, we studied the application of the propagation behavior in advertising with the rumors on the Twitter social network. Detailed, the contents and main results of the thesis are summarized as follows.1. The infectious SIRS model with artificial immunization is proposed based on SIRS model and scale-free nature on complex networks. We have presented the state transition diagram of the model. For this improved model, we studied the dynamic behavior of the model in the homogeneous network and heterogeneous network through a mean-field theory. We found out that the threshold value in the homogeneous network and the threshold value in the heterogeneous network. We conclude that the threshold value λc of spreading velocity in the homogeneous network and heterogeneous network is determined by the network topology, which has no connection with the individual characteristics.2. We study the spreading of disease in the special complex networks through two different artificial immunization strategies based on the improved model, and simulate to verify different impact of each strategy on disease spreading. We found that the target immune could not use in the homogeneous network, but the random artificial immunization can effectively reduce infected persons. The greater infection rates, the fewer the number of infections. The degree distribution of the heterogeneous network approximate power-law distribution. In the heterogeneous network applies random immunization and target immunization. We found that the number of infections when used the target immunization is less than individuals when used the random immunization. The result shows that artificial immunization can effectively reduce infection rates and improve the system spreading threshold, so as to effectively control the disease spreading on complex networks.3. Due to the reality of the networks are not completely random, are not only with a single small-world or scale-free property, but it may also have random property, small-world property and scale-free property. In order to reflect this property of the real network, we introduced a heterogeneous network model. We construct the heterogeneous complex network based on ER random network, WS small-world network and BA scale-free network. Then we using evolutionary algorithm of the heterogeneous network model generate a network topology diagram. By analyzing the network data, we obtained the characteristic network data and draw the corresponding degree distribution graph. Analyzing the spreading behavior of the infectious SIRS model with artificial immunization based on the heterogeneous network. Then we do simulation experiments by using the A(H1N1) type influenza virus in the heterogeneous network. The result shows that artificial immunization can greatly reduce the number of infections less than the network has only a single property, and can obtain more pronounced effect.4. Twitter social network as a newly booming communication tool, information disseminate fast through social connections. It is a prevalent phenomenon that the funs forwarded interesting messages, so advertising embedded in messages is widely accepted. Firstly, we preprocess the reference data2012by SQL Server through data analysis. We conclude that the Twitter social network has a scale-free property. Then we draw the degree distribution graph. We analyzed the advertising on the Twitter social network based on the information diffusion SIR model. We conclude that the number of persons who eventually knows the information related to the degree distribution of the Twitter social network. The degree distribution of the Twitter social network approximate power-law distribution, there are part of a great degree of a node in the network. The result shows that if the company to hire such people who have a great of funs advertising on social networks, the effect better than hire only the general people who have a few of funs, and advertising can cover the entire social network in a shorter time. Accordingly, we can provide enterprises with advertising recommendations, in order to pay small cost to obtain a lager advertising effect.