| With the rapid increase of complex applications such as autonomous driving,wearable devices,virtual reality,and augmented reality,the computing demand of Internet of Things devices(Io TDs)has reached an unprecedented level,which poses higher requirements for network bandwidth and transmission latency.However,limited computing power and energy make it difficult for Io TDs to provide satisfactory services.Although the quality of service can be improved by uploading computing tasks to servers,this may result in high latency and network congestion.Especially in some special emergency scenarios such as earthquakes,fires,and tsunamis,the lack of communication facilities puts greater pressure on the network.In recent years,the Mobile Edge Computing(MEC)technology assisted by Unmanned Aerial Vehicles(UAVs)has been widely used.It makes use of the strong mobility,high flexibility,and ease of maintenance of UAVs to help resource-limited Io TDs to process computing-intensive and latency-sensitive tasks,mitigate the pressure of massive data transmission on the network,and improve computing efficiency.Based on the above background,this thesis focuses on the task area allocation mechanism and multi-resource joint configuration method in the UAV-assisted mobile edge computing system,mainly including the following two aspects:Firstly,this thesis studied the multi-resource joint configuration in a single UAVassisted mobile edge computing system.In the case where there are fewer ground Io TDs and they are close to each other,these devices are grouped into a ground task area.UAVs carry edge computing devices to help Io TDs in the area process computing tasks.In order to fully utilize the large-capacity memory and highperformance processor of UAVs,different objective functions were set from the perspectives of UAVs and ground Io TDs respectively,and a multi-objective optimization problem was established.By jointly optimizing the flight trajectory of UAVs,the allocation of computing resources,and dynamic time slot division,UAVs and ground task areas can achieve their goals.Due to the existence of integer variables and non-linear functions,the optimization problem established is a multiobjective mixed-integer non-linear programming problem.By using linear weighting to transform the problem into a single-objective problem,using the idea of block coordinate descent to decouple the problem into three sub-problems for iterative solving,and respectively using linear programming,Lagrange multiplier method,and continuous convex approximation to solve the linear,nonlinear,and non-convex three sub-problems,a complete and effective algorithm was finally designed to obtain the approximate solution of the original problem.The simulation results prove the effectiveness of the proposed algorithm.Finally,based on the above research,considering the existence of multiple UAVs and multiple ground task areas composed of multiple ground Io TDs in remote and sparsely populated areas,each UAV belongs to a different airborne service provider,and each ground task area belongs to a different ground network operator.This thesis further studied the task area allocation and multi-resource joint configuration methods in a multi-UAV-assisted mobile edge computing system.A mobile edge computing system with multiple UAVs and multiple task areas was established,and two benefit functions were set for airborne service providers and ground network operators,respectively.By optimizing the joint allocation of multiple resources and task area assignment,the benefits of both parties were maximized.In order to solve the scheduling and task area assignment problems of UAVs,integer-type correlation variables were introduced,and a mixed-integer non-linear programming problem was established.To solve this mixed-integer non-linear programming problem,the assignment of task areas and the scheduling of UAVs were transformed into a matching problem,and a one-to-one stable matching algorithm was designed to solve the problem,achieving weak Pareto optimality.The simulation results have demonstrated the effectiveness of the proposed algorithm. |
