Research On Mobile Edge Computing Task Offloading Strategy Based On Multi-UAV Collaboratio

The Internet of Things,with its iteration update of communication technology,has enabled a variety of applications to become part of people’s everyday lives.However,the computing resources and power storage of most Internet of Things devices are very limited,which hinders the development of the Internet of Things.Therefore,mobile edge computing came into being.However,mobile edge computing relies too much on the ground base stations with high cost and fixed location.At this point,it is of great research significance and application value to combine mobile edge computing with UAVs with strong mobility,low cost and flexible deployment.However,in the scenario of UAV assisted mobile edge computing,there are many problems to be solved.Most researches on UAV assisted mobile edge computing do not consider the cooperation between UAVs.Some UAVs are idle and some UAVs are overloaded.Redundant tasks can only be transmitted to remote base stations or clouds for calculation.At the same time,few studies have considered the transmission queuing problem caused by multiple users unloading tasks to UAVs at the same time in a short time.And in some scenarios,there are some idle users with large computing resources.How to effectively use these idle users’ computing resources is also a problem worth studying.This paper undertakes two aspects of research in light of these issues:(1)Aiming at the problem of UAV collaboration and transmission queuing,a strategy for offloading mobile edge computing,based on non-orthogonal multiple access collaboration with multiple UAVs,is presented in this paper.First,in the network model of multiple UAVassisted mobile edge computing,the problem of uneven load is solved by enabling UAVs to transmit tasks to each other,and the transmission queue problem is solved by using nonorthogonal multiple access technology.Contemplating the restrictions of time-delay energy consumption,an optimization problem of unloading strategy is constructed to minimize the system’s total energy expenditure is essential.This paper introduces a deep Q network-based algorithm for task uninstallation to address this optimization issue.Simulation results demonstrate that the proposed scheme and the task offloading decision can significantly decrease the system’s overall energy consumption.(2)In scenarios where some users have large computing resources,how to utilize the idle computing resources of users is also a work worth studying.Therefore,in the second part of this paper,D2 D communication technology is introduced into the multi-UAV cooperative mobile edge computing model to enable users to transfer tasks to each other and improve the utilization of users’ computing resources.At the same time,considering that the first part does not optimize the unmanned aerial vehicle deployment location,under the constraints of time delay and energy consumption,a joint optimization uninstallation strategy and the optimization of the unmanned aerial vehicle deployment location are proposed.To reduce the total energy expenditure of the system,an iterative algorithm based on deep double Q network and differential evolution is presented to solve the optimization problem.Simulation outcomes demonstrate that the suggested plan can effectively augment the idle resource utilization of users and decrease the system’s total energy expenditure.