Research On Task Offloading Strategy In UAV Edge Computing Environmen

The insufficient computing and storage capabilities of current smart terminals affect the user experience,while the emergence of mobile edge computing(MEC)brings opportunities to solve the above problems.Among them,high-mobility unmanned aerial vehicles(UAVs)can be combined with MEC as a mobile communication service platform to provide computing and communication support in disaster areas.Therefore,this thesis focuses on the following two aspects of research for different situations that occur in natural disaster scenarios:(1)For the situation that ground base station(BS)are damaged or invalid due to natural disasters and cannot provide computing offloading services for UAV users,this thesis proposes a task offloading strategy suitable for UAV network edge computing environment.This strategy introduces multiple high-altitude balloons equipped with MEC servers as air auxiliary nodes,and optimizes the task offloading decision and trajectory planning problem of multiple UAV users under the premise of task execution time constraints,and minimizes UAV users’ total energy consumption as the goal to establish a nonconvex optimization problem.To effectively solve this problem,this thesis uses a block coordinate descent algorithm to decompose the original problem into two subproblems,and uses a successive convex approximation method to transform the subproblems into convex problems for processing.Numerical simulation experiments verify that the proposed strategy has significant advantages in reducing total energy consumption of UAV users compared with other comparison schemes.(2)In order to solve the problem that the MEC system cannot operate normally or has low efficiency due to the communication blocking between BS and mobile user devices in the ground network environment,this thesis proposes a UAV-assisted digital twin edge network intelligent offloading scheme.This scheme adds BS and resource devices as two types of edge servers in MEC system,and uses digital twin technology to adapt to uncertainty of MEC environment.On this basis,this thesis minimizes total energy consumption of MEC system as optimization goal to establish a mixed integer programming problem,and uses an alternating iterative optimization algorithm to decompose it into three subproblems.For each subproblem,this thesis respectively adopts double deep Q-network algorithm,closed-loop expression and iterative algorithm for solving.Numerical results show that proposed scheme significantly reduces energy consumption of MEC system compared with baseline algorithm.