Research On Optimization Of MEC System Offloading Computation Based On Deep Learning

With the rapid development of 5g and the proliferation of mobile devices,traditional mobile cloud computing(MCC)has been unable to meet the requirements of low delay service for computing intensive tasks.At the same time,mobile devices are limited by computing power and battery capacity and can not bring users a satisfactory experience.In order to solve the above problems,mobile edge computing(MEC)transfers the powerful cloud computing power to the network edge by deploying high-performance servers near mobile users.Users can uninstall tasks nearby,which can significantly reduce the computing delay of tasks.However,there are still some difficulties in how to realize low latency offload calculation of a large number of user tasks based on limited MEC resources.Therefore,this paper focuses on the optimization of unloading calculation of multi-user MEC system,and discusses the methods to improve the amount of security data and reduce user delay under the constraints of power and energy consumption.The main research contents and achievements of this paper are as follows:(1)This paper studies the low delay MEC unloading calculation problem of a large number of user tasks under MEC resource constraints,establishes the joint optimization problem of multi-user unloading and computing scheduling under end-to-side cooperation,and proposes a multi-user edge computing optimization scheme based on deep reinforcement learning(DRL).The scheme includes two sub algorithms: one is task scheduling algorithm,which realizes the optimal scheduling of computing tasks in MEC server;The second is the SRLA unloading decision algorithm,which solves the user side unloading decision and the selection of edge server.The results show that the proposed scheme can effectively reduce the total processing delay of tasks and improve the service efficiency of limited resources to a large number of concurrent user computing tasks.(2)The problem of multi-user unloading in secure MEC scenario is studied.The local computing model and edge computing model are established,and the optimization problem is established to maximize the amount of user safe unloading computing data.In this scheme,the user’s security data includes the amount of local computing and the amount of tasks safely unloaded to the edge server.Our objective function is solved by optimizing the user’s transmission power and local computing.In order to solve the above problems,this paper divides the objective function into two subproblems:(1)local computation optimization subproblem;(2)User transmit power optimization subproblem.Each subproblem obtains the optimal solution by giving another variable.Two subproblems are solved respectively to obtain the optimal solution of the original problem,and then an optimal partial task safe unloading scheme is proposed.The results verify the performance of the proposed scheme,and show that the proposed scheme can achieve better performance than the binary unloading scheme.In short,this paper makes an in-depth study on the unloading calculation of a large number of user tasks in MEC environment.By optimizing the unloading sequence of multiple users,the user delay is effectively reduced,and by optimizing the transmission power and local calculation,the user’s security data is effectively improved in the case of eavesdropping.