Research On Task Offloading Strategies In Mobile Edge Computing

With the widespread deployment of the 5th Generation Communication System(5G)around the world,new services with latency-sensitive and computation-intensive characteristics are emerging,such as unmanned driving,virtual reality,augmented reality,4K/8K ultra-high definition video,and so on.However,mobile devices cannot effectively meet the demand for computing resources of these new applications due to their size and battery capacity limitations.In addition,traditional Cloud computing servers are usually deployed far away from mobile users,which can bring large latency to application services and lead to reduced user service experience.Mobile Edge Computing(MEC),as one of the key technologies of 5G,deploys caching resources and computing resources in Cloud computing closer to the edge of the network on the user side,which can solve the problem of insufficient computing power of mobile devices or large delay caused by long-distance transmission to a certain extent,but compared to cloud computing,MEC computing resources are insufficient.Therefore,how to improve the efficiency of computing resource utilization,optimize task offloading decisions,and improve the quality of service(Qo S)of users is still an important challenge faced in the development of MEC technology.In this paper,we conduct an in-depth study of task offloading problems in different edge computing scenarios from the perspectives of users and edge service providers,respectively.The main research contents are summarized as follows.(1)Research on efficient task offloading strategy for multiple users in single MEC server scenario.To address the problem of low task execution success rate in multiuser single MEC server networks,an edge computing-based wireless access network model is proposed to construct an optimization problem with the goal of maximizing task execution success rate by considering constraints such as task delay,user device transmission power,and available computational resources of MEC servers.To solve this optimization problem,a priority-based task offloading algorithm is proposed to realize the joint optimization of task offloading strategy and computational resources.Simulation results show that the proposed algorithm can offload user tasks efficiently and effectively improve the execution success rate of tasks.(2)Research on task offloading strategy based on dynamic pricing of services in a multi-MEC server scenario.To address the problem of unreasonable allocation of computing resources for MEC servers,a computational offloading framework for multi-MEC server collaboration is established,and an optimization problem with the objective of minimizing system energy consumption and maximizing economic benefits of ESP is constructed under the constraint of satisfying users’ Qo S requirements.To solve this optimization problem,a task offloading algorithm based on dynamic pricing of services is proposed to optimize both MEC server computational resource prices and task offloading decisions.Simulation results show that the proposed algorithm can improve the efficiency of resource utilization and increase the economic benifit of ESP while effectively reducing system energy consumption.(3)Research on task offloading strategies assisted by Intelligent Reflective Rurfaces(IRS)in edge computing scenario.To address the problem of poor wireless communication link performance or communication interruption caused by obstacle blockage,IRS technology is used to improve communication quality,establish an IRSassisted UAV communication system model,and construct an optimization problem with the objective of minimizing system energy consumption.To solve this optimization problem,a task offloading algorithm based on the improved block coordinate descent method is proposed to jointly optimize the task offloading coefficient,user device transmission power,IRS phase shift matrix and UAV trajectory.Simulation results show that the proposed algorithm can improve the communication link quality,effectively reduce the system energy consumption.