Research Of Caching Strategies And User Request Prediction Algorithm In Mobile Edge Networks

Nowadays,the exchange of information between people and people,people and things,things and things is more and more frequent,and the amount of data transmission is getting larger and larger,which makes the data transmission in wireless communication networks explode.The sharp increase in users' data has led to great pressure on network bandwidth and aggravation of user service delays.How to reduce network bandwidth pressure,reduce user service delay,and improve wireless network system capacity is a very practical issue.The introduction of edge caching technology in the wireless edge network allows the content requested by users to be closer to users,which can reduce the bandwidth pressure of the core network and improve the user service delay effectively.This paper introduces edge caching technology into mobile edge networks.The main research and contributions are as follows:Firstly,this paper deeply studies the energy consumption of a two-tier heterogeneous networks under two different caching strategies.The energy consumption of each network components was modelled,and the relationship between the energy consumption of heterogeneous networks and the distribution density of small cell,user load per cell and caching capacity was investigated.The simulation results show that edge caching can effectively reduce the energy consumption of heterogeneous networks and the performance of the two caching strategies are different under different density of small base station and user load.Specifically,the energy consumption of both the network and per user under the two caching strategies are reduced with the increase of the density of the small base station and the user load per cell,respectively.Furthermore,the influence of small base station's caching capacity can not be neglected in total energy consumption.Then,this paper proposes an improved recommendation algorithm that takes into account user preferences,request history,and similar user request data for selecting user cached content.Firstly,the user-content scoring matrix is complemented by the improved collaborative filtering algorithm,and then users are ranked according to users' request frequency.Finally,the content rating is comprehensively evaluated according to the user level,the user rating and the access history of the content.We can select some of the highest-rated content as the content to be cached,and then actively cache the content to egde node at the appropriate time.The simulation results show that the introduction of such a recommendation algorithm in the edge network can effectively improve the hit probability of the user request.