Mobile Edge Caching Algorithm Design Based On User Social-ties

The growth in the number of mobile devices around the world has led to a proliferation of mobile data traffic which has made the current cellular network can't afford such a large number of users access cellular networks at the same time.According to the traffic statistics,the popularity of content follows a Zipf distribution,which means 80% of users request 20% of files.Facing the problem of network repetitive transmission caused by the same content requests among different users,it has been proposed in some papers to cache the content in the cellular base stations to reduce the duplication of file transmission.With the standardization of D2D(Device-to-Device)communication technology,users can share content through D2 D communication,thus making mobile edge caching become a research hotspot.Although some achievements on mobile edge caching have been made,there still exist some problems to be solved,including the influence of social ties on user communication,how to choose users as cache nodes and the node space waste problem caused by content duplication caching when a user communicates with multiple users.All of these problems are the focus of our work.An important user selection algorithm based on physical location and user social ties is studied.In order to address the impact of social networks on content transmission between users,a social network layer model on the physical layer is introduced and a new physical-social network joint mode has been created.Based on this model,we define the social throughput of a user to measure the importance of a user.We further formulate the important user selection problem as a 0-1 integer programming problem to maximize social throughput and propose an important user selection algorithm to obtain the approximate solution.The simulation results shows that the important user selection algorithm reduces the system average delay by 3% compared with the random selection of the user and degree centrality selection of the user.With the joint model of physical-social network and the content cache node,we further investigate the content caching problem in a scenario where the user can communicate with one user or a number of important users.The content caching problem is then described as a nonlinear integer programming problem to maximize the gain of the system average delay.We prove that the content caching problem is a NP-hard problem,the constraint condition is a partition matroid matrix and the objective function is a monotone submodular function.Finally,a content caching algorithm is proposed and the impact of important users,cache space,content popularity parameters on the system average delay is analyzed.Compared with the most popular content caching and random caching,the proposed content caching algorithm joint important user selection algorithm reduces system average delay by 5% and 10%.