Study Of Information Propagation And Immunization Strategy On Complex Networks

Complex networks is a tool and a method of studying complex systems,and it is the foundation of understanding the function and property of complex systems.With the development of complex network theory,the complex network theory in the study of complex systems has become one of the most active branches.Firstly,this dissertation introduces the progress of complex network theory in the field of the dissemination of information,the statistical characteristics of complex networks and some basic network models,the frequently-used communication model and some basic immunization strategies.Secondly,it conducts in-depth analysis of multi-information dissemination in the weighted network.Lastly,it studies how to restrain the spread of unhealthy information in the directed online social network.The specific research work is as follows:1.Study the propagation characteristics of two messages in the GBBV weighted network and real data sets respectively.Assume that the propagation rates of the two pieces of information are the same,the appeal are different,the attractiveness of the information itself is different from each other,in which the attractiveness of information 1 is relative high,and the attractiveness of information 2 is relative low.From the two aspects of the structure of network and the number of initial propagation source of low attraction information,we study the communication behavior of multi-information of the competitive relation in the GBBV weighted network.Study shows that when the average degree of a network is 6 to 8,the spread range of the low attractiveness information 2 is relative small,which is not much relationship with the average clustering coefficient and the average edge weight of network;When the average degree of the network is less than 8,the spread range of the information 2 increases with the increase of the average weight of the network.This paper puts forward several methods of improving the propagation range of information 2 by increasing the number of its initial propagation source,and then compares the results of these methods.We find that the effect of increasing the initial propagation nodes by descending degree is notable in sparse networks,and the propagation range of information 2 will exceed information 1 by only adding a small amount of initial propagation nodes.2.Study the propagation characteristics of two messages with a competitive relationship in Epinions online social network and compare several immune strategies.Assume that the positive information is in the spread of inferiority,and the negative information is in the dissemination superiority.Study shows that the size of Epinions social network has no effect on the propagation behavior of multi-information with a competitive relationship.When the information propagate stably,most people in the network is in the state of accepting negative information,only a small portion of the population is in the state of receiving positive information,besides a small part of the population is still in a neutral state.In order to increase the proportion of population which accept the positive information in the network and reduce the proportion of population which receive the negative information,we put forward several immune strategies,where the immune nodes are selected with descending order according to the nodes betweenness,the nodes in-degree,the nodes out degree,the nodes closeness or randomly selecting nodes respectively.Study shows that the node betweenness approach is the best one,the approach of nodes in-degree,nodes out-degree take the second place,and the methods of nodes closeness and random selection are the worst in the mentioned methods.