Trajectory Scheduling And Task Offloading For Unmanned Aerial Vehicle Enabled Mobile Edge Computing Networks

The vigorous development of the Internet of Things and the mobile communication technologies has spawned a large number of computing-intensive and delay-sensitive emerging applications,which pose severe challenges to the limited computing capabilities of user devices.Unmanned aerial vehicle(UAV)-enabled mobile edge computing(MEC)networks not only make up for the dependence of traditional terrestrial MEC networks on ground communication infrastructures,but also provide computing resources for user devices at the closer edge of the networks with the high flexibility and high mobility,to meet the computing needs of emerging applications.However,in the UAV-enabled MEC networks,UAV trajectory and user devices’ task offloading strategies not only affect the overall energy consumption of UAV,restricting the service time of UAV,but also affect the quality of service(Qo S)experience of user devices.How to schedule UAV trajectory flexibly and design efficient task offloading strategies has become the key to effectively deal with the challenges of UAV energy constraints and user devices’ diversity requirements.Therefore,this thesis focuses on methods of UAV trajectory scheduling and task offloading in UAV-enabled MEC networks from the perspective of UAV’s utility and user devices’ utility,respectively.The main research contents of the thesis are as follows.1)To address the challenge of UAV’s battery energy limitation in UAV-enabled MEC networks,an UAV energy optimization-oriented task offloading and UAV trajectory scheduling method is proposed.The method formulates the UAV’s total energy consumption minimization optimization problem by jointly optimizing the area-division-based task offloading decisions and the UAV trajectory,under the condition that the user devices’ exact locations information is unknown.To solve this problem,it is decomposed into two independent sub-problems,i.e.,the region partitioning sub-problem and the UAV trajectory optimization sub-problem.For the region partitioning sub-problem,an optimal transport theory-based region partitioning iterative algorithm is proposed to realize the continuous division of the target service area.The existence and closed expression of the optimal solution are theoretically proved.Meanwhile,the simulated annealing-based UAV trajectory optimization algorithm is proposed to solve UAV trajectory scheduling,so as to minimize the UAV’s total energy consumption under the user devices’ arbitrary distribution.The simulation results show that the proposed method not only reduces the UAV’s total energy consumption,but also achieves the load balance among the service regions.2)To solve the challenge of diverse requirements of user devices’ tasks in UAV-enabled MEC networks,a service satisfaction-oriented task offloading and UAV scheduling method is proposed.First of all,a novel task priority model is proposed by jointly considering the user devices’ task delay requirements and remaining energy status.A fine-grained service satisfaction model is proposed by considering the user devices’ task processing delay and energy saving,to measure the satisfaction degree of the user devices’ diverse demands in multiple perspectives.Secondly,a service satisfaction-oriented two-stage computing offloading framework is proposed.In the first stage,the user devices grouping optimization is formulated,and the K-means-based grouping algorithm is proposed to divide all the user devices into different groups.In the second stage,the overall service satisfaction maximization optimization problem is formulated by jointly optimizing the task offloading and UAV scheduling.To solve it,it is decomposed into two sub-problems,i.e.,the UAV scheduling sub-problem and task offloading sub-problem,and the genetic algorithm is used to solve them,respectively.Then a joint task offloading and UAV scheduling iterative optimization algorithm is proposed,to address the overall users’ satisfaction maximization problem by addressing the two sub-problems alternately,achieving reasonable and effective task offloading and UAV scheduling scheme.The simulation results show that the proposed method not only achieves fast convergence,but also ensures the satisfaction of lowpriority user devices while guaranteeing most high-priority user devices’ satisfaction,thereby effectively improving the overall satisfaction of user devices.In the case of 170 user devices,the total users’ satisfaction of the proposed scheme exceeds the baselines by up to 16.9%.