Social Network Communication Model And Its Algorithm Research

The dissemination problem of social network is one of the important tasks in the study of social networks.It has been widely used in biological evolution,viral infections and marketing.Using the method of the networks,we study on several dissemination models of social networks and predict the behavior of participants during the process of transmission.Then,we give several approximate algorithms.The first chapter defines the probability that nodes become a mutation in each moment.We name the probability for the vertex probability.We prove the convergence of vertex probability which converges to the fixation probability when time goes to infinity.Then we give the solving method of the fixation probability on the weighted directed graph and the expectation number of the variation in every moment.Study the social network from its own structure.Through analyzing the conditions of occuring giant branch in the Poisson random networks,we study the effect of giant branch on the spread of variation and the size of variation dissemination.We study the properties of temporal network and research the influence of advantage variation on the participants.The second chapter introduces the linear threshold model and independent cascade model and studys the computational complexity of the maximization of influence in these models.We prove that the influence function of linear threshold model and independent cascade model is submodular and obtain a approximate algorithm of the maximization of influence,named greed climbing method.We have proved the validity of the algorithm.The third chapter studys the threshold model in the case of single and multiple products.Focusing on multiple products spread model in the root tree network.We discuss the possibility and inevitability of a product adopted by all participants.We give the sufficient condition in which a kind of products is adopted by all participants and give a polynomial time algorithm to confirm if a certain product would be adopted by all participants.