Group User Behavior Aware Collaborative Caching In Mobile Networks

Caching content on the edge of the network has recently been considered as an efficient way to reduce the transmission cost and improve the users' experience in emerging mobile networks.With the development of the mobile network,there are more advantages of the structure that can be taken to further increase the efficiency of the service.At the same time,a rise of Sharing Economy has recently brings to more and more attention of the society,people begin to realize that in the process of resource scheduling and configuration,the description of user groups with similar interests or behaviors can effectively improve the utilization efficiency of resources,and thus provide a new method for the resource sharing in the mobile network.To solve the problem that the caching is not accurate enough and to improve the low transmission efficiency in the existing mobile network caching strategy,we proposed a Group User Behavior Aware Collaborative Caching Method(GUCC)in this paper,and the main contribution is showed below:1.We make an analysis of the Usage Detail Records(UDR)and explore the heterogeneity,predictability and the group behavior of users during the access of location and content.2.We design a collaborative relationship model based on group user behavior.We use content and location similarity network to depict the group behavior.Weighted Similarity Network Fusion(WSNF)algorithm is raised to solve the fusion problem of largescale similarity network.And the user groups are formed based on the fused network.The collaborative relationship among the base stations is established according to the user groups they serve.3.We propose a caching placement scheme that based on the collaborative relationship among base stations.In the scheme,both the cost of the transmission among base stations and the hit ratio of the requests are considered.A multi-objective optimization problem is solved to maximize the hit ratio and minimize the transmission cost at the same time.We analyze and solve the problem according to the Game Theory.4.To solve the cache updating problem during the real application,we propose a cache updating method driven by the base station cooperative relationship.In detail,we propose an articulate point discovering method based on the distribution of users'access to the base stations.In this method,we decide the updating time according to the access behavior on the articulate point to finally update the cache.The numerical results show that the proposed method achieves performance gain in terms of both hit rate and transmission cost.