Research On Performance Optimization Of Uav-assisted Mobile Edge Computing System

With the rapid development of the Internet of Things technology,terminal devices have generated a large number of compute-intensive and delay sensitive tasks.However,terminal batteries and computing resources are often limited,which makes their performance unable to meet the needs of task processing.Mobile Edge Computing(MEC)technology provides an efficient and feasible solution to this problem.Because Unmanned Aerial Vehicle(UAV)has the characteristics of mobility,flexibility,and high probability of Line of Sight(LoS)link,UAV-assisted communication system can be applied to various scenarios to support communication services,and has a good development prospect.Based on this,this thesis mainly investigates the performance optimization of UAV-assisted mobile edge computing system by resource allocation and UAV flight trajectory planning in terms of energy consumption and delay.The main research contents and innovative work of this thesis are as follows:(1)An energy consumption iterative optimization algorithm based on linearization is proposed.In the UAV-assisted MEC system,due to the limited computing capacity of the terminal device,computing tasks can be unloaded to the UAV for processing partially.In order to comprehensively balance the energy consumption of the UAV and terminal equipments,a weight factor is proposed,and the goal of minimizing the total weighted energy consumption is achieved by optimizing the trajectory of the UAV,the allocation of system computing resources and the unloading time of the terminal equipments.On this basis,a new trajectory initialization method based on spline interpolation is proposed.Simulation results verify the effectiveness of the proposed iterative optimization algorithm,and show that it can achieve lower total energy consumption than the benchmark scheme.(2)A multi-stage delay optimization algorithm based on Lyapunov method is proposed.In this system,the UAV and the central cloud can collaborate to provide computing services to terminal devices.To solve the waiting delay problem of the terminal equipments in the system,by controlling the energy consumption queue,the original long-term stochastic optimization problem is decomposed into a series of deterministic online optimization subproblems,and then the closed-form solutions of the corresponding problems are obtained by using Lagrangian duality theory,alternative optimization and successive convex approximation(SCA)technique.Finally,simulation results and performance analysis show the convergence and effectiveness of the proposed algorithm.