Joint Optimization Strategy Of Task Offloading And Resource Allocation In Mobile-Edge Computing

With the popularity of mobile devices(MD)and the rapid development of new applications for 5G networks,the requirements for large-scale device connectivity and computing requirements have increased.New types of mobile applications such as e-health,face recognition,augmented reality,and virtual reality are not only computationally intensive,but also place higher demands on latency.The development of mobile devices has limitations such as battery life,processing power,and storage capacity,and they cannot meet the needs of the above applications.Mobile-Edge Computing(MEC)provides a solution for this.Mobile-edge computing sinks computing and storage resources to the edge of the mobile network,and MD can offload compute-intensive or delay-sensitive tasks to the MEC servers for calculation,thereby reducing latency and energy consumption,and extending battery life.Nonetheless,offloading will bring additional overhead.Compared to mobile cloud computing(MCC)systems,MEC still has the problem of limited computing resources.Therefore,research on task offloading strategy and resource allocation is very necessary.Firstly,this thesis summarizes the theory of mobile-edge computing,and analyzes the scenarios and advantages of unmanned aerial vehicle(UAV)-assisted edge computing.The technical theories of task offloading and resource allocation in the MEC system are explored separately,and the development status of the joint optimization strategy of the two is introduced.Secondly,according to Kang Cheng et al.'s research on multi-user and multi-MEC base station systems in 2018,this thesis further expands it into a multi-access edge base station network scenario,and a total capacity of mobile devices meeting the task delay constraint was formulated.The problem of minimum energy consumption is jointly optimized for three variables: computing offload,wireless resource allocation and computing resource allocation.Due to the coupling between variables,it cannot be decomposed into sub-problems.Therefore,in order to solve the mixed integer non-linear programming(MINLP)and non-convex problem,this thesis transforms them through the combination of variables,and proposes two solutions to obtain the optimal solution and the suboptimal solution.Simulation results show that compared with the local execution scheme and non-joint optimization strategy,this scheme can achieve the lowest energy consumption,and the suboptimal solution greatly reduces the algorithm running time on the basis of ensuring performance,which is more conducive to the deployment of actual scenarios.Thirdly,for the UAV-assisted edge computing system,according to Jingyu Xiong et al.'s research on single UAV and multi-MD system in 2019,a two-layer UAV scene model is proposed to jointly optimize the task offloading decision,bit allocation and UAV flight trajectory during the transmission process,aiming to minimize the overall performance of the mobile device.Through the iterative algorithm,the above non-convex MINLP problem is solved.Simulation results show that this scheme can achieve better performance than other benchmark strategy and non-joint optimization strategy.