Research On Optimization Strategy Of Mobile Edge Computing For Terminal Multi-service Model

With the development of delay sensitive applications such as virtual reality,augmented reality and digital holography in smart city,traditional cloud computing is difficult to meet the real-time requirements of computing and storage due to the long transmission delay.By deploying the server to the edge of the wireless access network which is closer to the terminal layer,edge computing can not only effectively reduce the transmission delay,but also provide the terminal with additional computing power.These changes in network structure provide broad application potential for new services with large amount of computation and severe delay,but present new challenges to resource deployment: because of the unbalanced spatial distribution of base stations and terminals,and the different computing ability of edge servers,we need to find the strategies of terminal access selection,service offload decision,and system resource deployment in different scenarios.The main contents and contributions of this thesis are as follows:1)For the single cell-multiple terminals-multiple services edge computing scenario,we study the strategy of service offloading and resource deployment,for the purpose of obtaining the solution of the joint optimization problem.Through constructing the utility index minimization problem which comprehensively considers the service processing delay and the weighted energy consumption of the system,we find the optimal service offloading decision and resource deployment strategy under the constraints of computing resources,by jointly optimizing the service offloading decision,the transmission power and computing resource allocation strategy of terminal and the base station.Considering that the problem is a mixed integer nonlinear programming problem(MINLP)belonging to NP-hard problems and difficult to solve using conventional algorithms,we use the successive convex approximation(SCA)algorithm to find the sub-optimal solution of the problem and verify the approximate algorithm can effectively improve the system performance by designing simulation experiments.2)On the basis of single cell scenario,we consider more ubiquitous multiple cellsmultiple terminals-mutiple services edge computing scenarios.By constructing the system weighted energy minimization problem,we jointly optimize the terminal access node selection,service offloading decision and resource deployment strategy.In order to reduce the complexity of the solution,the original MINLP problem is decoupled into: terminal access selection and service offloading sub-problem,and system resource deployment subproblem.We propose an iterative solution framework to solve the problem,by using adaptive genetic algorithm and convex optimization method.The simulation results show that the iterative algorithm can effectively reduce the weighted energy consumption of the system.Through the research of two typical mobile edge computing scenarios,single cell and multiple cell,we obtain an efficient terminal access selection strategy,moreover solve an efficient service offloading decision and system resource deployment strategy.Our study provides useful exploration for the future integrated design of 6G communication,computing,storage,and control,and also lays a good preliminary foundation for future research on timeliness and energy consumption of new services.