Research On Computing Offloading Algorithm In Mobile Edge Networks

In recent years,with the rapid development of mobile communication and Internet technology,various applications have appeared,such as smart city,Internet of vehicles and telemedicine,which brings a great challenge to the computation and storage capacity of current mobile communication networks.In order to solve these problems,mobile edge computing(MEC)has emerged as one of the key technologies in the fifth-generation mobile communication(5G).MEC can provide computing offloading service for all the users by sinking computation resources and storage resources to the edge of the network.Compared with mobile colud computing(MCC),it can not only avoid the long-distance data transmission by offloading the tasks to MEC servers for execution,but also greatly reduce the task execution latency and provide higher quality of service.Therefore,it is necessary to achieve a reasonable and effective computing offloading strategy for each user in mobile edge networks.This thesis mainly studies the computing offloading problem in mobile edge networks.Firstly,to solve the problem of computing offloading in multi-cell and multi-user mobile edge networks,the offloading decisions of mobile terminals,the interference and the allocation of computation resources are taken into consideration.Thus,a computing offloading and resource allocation algorithm based on deep reinforcement learning is designed.In this algorithm,deep reinforcement learning is used to handle the offloading decision problem,genetic algorithm is utilized to deal with the resource allocation problem and the optimal solution is obtained by solving the two problems alternately.Simulation results prove that compared with traditional algorithms,the proposed algorithm can effectively reduce the total cost of the system.Secondly,in order to handle the computing offloading problem in a MEC and Device-to-Device(D2D)hybrid network,the usage of idle computation resources,the different priorities of tasks and the potential waiting delay are considered.A computing offloading algorithm based on sequential game is designed.Simulation results prove that compared with several other algorithms,the proposed algorithm has a faster speed of convergence and a lower energy consumption of task execution.Lastly,an actual mobile edge network test environment is constructed,which consists of a MEC server,a small cell and several mobile terminals.A terminal control system based on MEC is designed and implemented.The system can not only take an access control of communication and Internet for all the users,but also realize the functions of task offloading and website access authorization.Test results prove the effectiveness and practicability of the system.