Research On Information Dissemination Problem Of Credibilistic Social Network

Information dissemination is the main way for social network users to communicate and learn.Throughout a social network,it is important for a decision maker to deal with uncertain factors in information data and find a path of Information dissemination for the shortest time and maximum flow.Therefore,on the basis of the new theoretical analysis method,the uncertain factors in social networks are described effectively,and the problem of reasonably studing information dissemination under certain constraints has gradually become a research hotspot in the field of social network analysis.This thesis proposes credibilistic social network analytical method,and is used to deal with information dissemination problem in social networks.On the one hand,three central analysis methods of degree centrality,closeness centrality and ?-betweenness centrality are given based on the credibilistic graph.On the other hand,using the fuzzy variables with credibility distributions to describe the information dissemination time and flow,and establishing several new types of credibilistic social network information dissemination models based on the expected value,value-at-risk,conditional value-at-risk and expected absolute semi-deviation methods.More specifically,according to different objective criterias and the influence factors of information dissemination chosen by decision maker,four models are constrcted,they are the information dissemination expected value model with credibility constraints,the credibilistic social network information dissemination value-at-risk model,the credibilistic social network information dissemination conditional value-at-risk model,the credibilistic social network information dissemination expected absolute semi-deviation model.At the same time,according to the characteristics of different models,the convertible models are transformed into equivalent linear programming models.The models that can not be transformed are solved by a hybrid intelligent algorithm including approximation method,neural network and genetic algorithm.Finally,the numerical examples of credibilistic social network are given to verify the feasibility and effectiveness of the proposed credibilistic optimization models and solution methods.The innovation points of this paper mainly include the following aspects:(i)The concept of credibilistic graph and network and three new central analysis indexes of degree centricity,closeness centrality and ?-betweenness centrality are proposed.(ii)According to the time and flow in the process of social network information dissemination,two new expected value models of information dissemination with credibility constraints are established and equivalent transformations are carried out.(iii)Using value-at-risk,conditional value-at-risk and expected absolute semi-deviation risk evaluation methods,several new risk models of social network information dissemination with credibility constraints are established and equivalent transformations are carried out.