Research On Network Planning And Resource Optimization In Mobile Edge Computing Networks

With the rapid development of mobile phone chip technology,intelligent terminals have gradually transformed into powerful devices with communication and entertainment functions.In traditional Mobile Cloud Computing(MCC),since the server is far from the User Equipment(UE),the real-time application of UE cannot be satisfied.Therefore,people have to move the lower layer of the server to the edge of the network to enable it to provide services near the UEs.Therefore,Mobile Edge Computing(MEC)technology is proposed.It provides UEs with computing offloading,caching and other functions at the edge of mobile network,so as to make up for the deficiency of MCC.However,MEC servers are usually deployed in the access network and binded with Base Station(BS),which is not deployed in the core network like MCC server.Therefore,how to deploy MEC servers has become a key issue.At the same time,the resources of MEC servers are usually limited.How to optimize the resources in MEC servers has also become a research hotspot of MEC networks.To solve the above problems,this thesis studies the network planning and resource optimization methods in MEC networks.The main works are as follows:(1)For multi-user MEC networks under the requirements of homogeneous services,the deployment method of MEC server is studied.Firstly,this thesis proposes a fast estimation method of network scale based on a greedy algorithm,and obtains the feasible range of the number of MEC servers in the network.Then,a MEC server location optimization algorithm based on improved Particle Swarm Optimization(PSO)is proposed.When the number of MEC servers is fixed,the algorithm optimizes the location of MEC servers,and the goal is to maximize the service rate of users in the network.Finally,an optimization algorithm for minimizing the number of MEC servers based on bisection search is proposed.The algorithm optimizes the number of MEC servers through bisection search and combined with the MEC server location optimization algorithm,and finds out the minimum number and locations of MEC servers to meet the user needs.The simulation results show that compared to the location optimization algorithm based on traditional PSO,random weight PSO and linear decreasing weight PSO,the improved PSO algorithm proposed in this thesis can meet the needs of users with fewer MEC servers and provide computing and communication services for more users.(2)For multi-user MEC networks under heterogeneous business requirements,the methods of user access control and resource optimization in MEC networks are studied.Firstly,this thesis proposes a user access control method for MEC networks based on the backtracking algorithm.Based on the traditional Signal-to-Noise Ratio(SNR)access method,this algorithm comprehensively considers the limitations of communication and computing resources,and realizes the effective access of users with heterogeneous service needs.Then,a resource optimization algorithm based on the improved Genetic Algorithm(GA)is proposed.The algorithm jointly optimizes the transmission power of BS,the number of subcarriers,the computing resources and cache resources of MEC servers.The goal is to minimize the system cost under the constraints of user communication,computing offloading and cache requirements.Simulation results show that the proposed method can meet the needs of users with lower network construction cost compared with the algorithms based on traditional GA and maximum SNR.