Data Transmission Based On Social Network In D2D Communications

The high-rate data transmission is one of the research hotspots pursued by the fields of industry and academic.On the one hand,the rapid growth of demand for the multimedia service challenges the data transmission efficiency,on the other hand,the subjectivity of human beings affect the communication link between users.Furthermore,the combination of D2 D network and social network could meet the heterogeneous requirements of users.Therefore,exploiting the valuable information from social network,this thesis focuses on the issues of content priority determination and information diffusion modeling to improve the D2 D transmission performance,the concrete research contents are as follows:1.Aiming at the multi-content transmission problem in D2 D communications,a social tie-driven priority scheme,which combines social and physical D2 D networks with priority decision issue,is proposed in this thesis.Specifically,considering the time-decaying property of social tie,the priority issue is modeled as an increment of social tie strength(ISTS)maximization problem based on the social trust and social reciprocity.Additionally,due to the interplay between social tie and physical link,the ISTS-maximization decision model is transformed into another one aiming at maximizing the system capacity,which is solved eventually by the simulated annealing with low complexity.Finally,the effectiveness and the superior performance of the proposed scheme are both verified by some simulations.2.Aiming at the defect of traditional epidemic models that only focus on the network topology structure,this thesis researches a social network-based information diffusion model for D2 D communications.In particular,with consideration of social relationship and users' interest degree,the infection probability is introduced to describe the success rate of data dissemination between D2 D users.Meanwhile,the recovery rate,which relates to the survival time of information,is also defined based on the influence of users.Furthermore,the first-order Markov Chain is employed to characterize the state transition process.Simulation results show that the researched model outperforms the traditional models in terms of the diffusion efficiency and match degree with the actual network.