Modeling And Analyzing Virus Propagation On Evolution Of Network

Recently complex network has become a new hotspot in domestic and overseas academic research, it has widespread application in many fields including sociology, engineering, medicine, politics, economics, management, Internet, etc. The development of complex network has caused attention of scholars in different research areas widely. Any kind of complex system can be abstracted to a network composed by interactional individuals. Because of the complexity of real system, the core problems of complex network research are focused on the way to construct evolution network of a more realistic system and understand the relationship between network structure and dynamics. Generation mechanism and virus process are researched using knowledge of complex network and dynamics.Firstly, local world selection mechanism and triangle coupling mechanism are introduced based on traditional BBV weighted network model, three kinds of modified weighted network models (GBBVIII) have been built, in which the theoretical analysis and simulation of the two mechanisms mentioned above are emphasized. Result of simulation analysis indicated as follows:Degree distribution, intensity distribution of node and edge weight distribution of GBBVIII model are compliant with power-law distribution, which, like BBV model, corresponds to theoretical derivation and massive network features. But, with consider of these two mechanisms, GBBVIII model is more coincident with real network than BBV model; Although both of BBV and GBBVIII model have a linear function of the degree and intensity of node, GBBVIII model is equipped with the ability to adjust the two additional mechanisms to accelerate the increase speed of node intensity exceeding that of node degree, which contributes to a more coincident with network in real life, such as railway network; GBBVIII model is more sound and efficient in tuning clustering coefficient than that of BBV model. Virus spread process on GBBVIII model has been investigated. Virus spreading characteristic has been inspected in sight of the average intensity of infected node, selection of the initial infected node, local world scale and immune mechanism.Secondly, most virus spread models are based on some specific network such as homogeneous network, and most are investigated in static network without considering of the reality in which network nodes are growing and perishing continuously、the network is divisible and scale free、nodes in the network are self-adaptive (feedback mechanism) and network viruses are emerging in endlessly and it is impossible for recovered computers from infection to be immune permanently, they may be infected again. Taking into account of the four factors above, this article establish a dynamic self-adaptive network and then a virus spread model is constructed. Influence on virus spread caused by adding speed and self-adaptability (feedback mechanism) of nodes and clustering coefficient has been studied. Simulation results indicate:(1) virus spreads fast and widely when dynamic self-adaptive network has a large-scale fluctuating clustering coefficient;(2) The higher the network node adding speed is, the faster the virus spreads and the shorter the immune time is. Consequently, for Internet administrators, it is difficult to control the velocity because of the uncertainty of exiting speed. By means of controlling the number of computers adding into the Internet each day we may control the spread of virus, which provides a new idea on health development of the Internet;(3) For dynamic network, self-adaptability (feedback mechanism) is an effective mechanism to control the spread of virus.