Joint Optimization Of Cache Placement And Resource Allocation In Cellular Networks With Multiple Time Scales

With the popularization of mobile devices and the growing of multimedia applications,mobile data traffic and content diversity are increasing rapidly,which brings great burden to the traffic offloading of backhaul link.In order to meet the huge demand of content traffic in mobile cellular networks,edge caching is proposed as one of the most promising methods.Caching popular content in the edge node can bring the content closer to users,which can effectively reduce the repeated transmission of the same content in the backhaul link,the burden of the backhaul link and the response time of the content request.Meanwhile,system performance such as spectrum efficiency,energy efficiency,transmission delay,backhaul cost and throughput can be optimized.This work is supported by the National Natural Science Foundation"Research on Multilevel Cooperative Caching in UAV-Assisted Cellular Networks".Caching placement and resource allocation are critical to network performance in dynamic cellular networks.In order to improve the utilization of cache space and the efficiency of resource allocation,this thesis jointly optimizes caching placement and resource allocation in a long-term perspective to improve the system performance.In this thesis,two scenarios of base station cooperative caching and base station multihop transmission are considered,and the corresponding joint optimization algorithm of caching placement and resource allocation is proposed respectively.The main work of this thesis is as follows:1)This thesis summarizes the current research about edge caching and resource allocation in cellular networks.Firstly,the classification and main research scenarios of edge caching in mobile network are summarized.Then we introduce the research status of caching placement and resource allocation algorithms in cellular networks including research progress and main problems.Finally,challenges in the joint optimization of caching placement and resource allocation in base station cooperative cache scenario are summarized.2)Considering the dynamic change of user location and content popularity,this thesis proposes a joint dynamic optimization algorithm of caching placement and user association in base station cooperative caching scenario.Firstly,an optimization problem is established to minimize the long-term content delivery delay in dynamic scenarios.To solve this problem,this thesis decomposes it into two sub-problems,the user association sub-problem in a short time scale and the caching placement in a long time scale.Specifically,we propose a low complexity distributed Belief Propagation(BP)based user association algorithm for a given caching placement in the short time scale.Then we model the other subproblem as a Markov decision process(MDP)and develop a deep deterministic policy gradient(DDPG)based caching placement algorithm which involves the short time-scale user association decisions in the long time scale.3)In the multi-hop transmission scenario of base stations,considering content cooperation and sharing between base stations,this thesis proposes a joint optimization algorithm of caching placement and user association and bandwidth allocation.Firstly,this thesis establishes an optimization problem to minimize the long-term content delivery delay.This thesis designs two time scales to implement content caching placement,user association and bandwidth allocation respectively.In the long time scale,the cooperative caching problem is modeled as a Markov decision process,and a Twin Delayed Deterministic Policy Gradient(TD3)based caching placement algorithm is proposed.In the short time scale,this thesis proposes a switch matching based user association algorithm and a convex optimization based on bandwidth allocation algorithm respectively.