Research On Computing Offload And Resource Management In UAV-Aided Edge Computing

Edge computing is a technology that utilizes computing resources at the edge of the network(such as cellular base stations)to provide services to user equipment.However,traditional edge computing systems are highly dependent on ground base stations.In some complex scenarios,such as earthquakes,floods,etc.,ground base stations are easily damaged and cannot be repaired in time.During peak periods,such as major sports events,gatherings,etc.,the computing resources of the ground edge server are not enough to meet the request,resulting in the inability of user tasks to be processed in time.Compared with traditional base stations,UAVs can be flexibly used in complex areas such as mountains and oceans due to their low deployment cost,high mobility and load ability.They can also act as mobile base stations for installing edge servers to help equipment offload tasks and meet the needs of its application requirements.In this paper,resource allocation,unloading decision and UAV deployment in UAV-based edge calculation are studied as follows:In the edge computing system assisted by a single UAV,considering the limited energy consumption of the UAV,it provides services for user equipment in a periodic circular cruise mode.In order to maximize the total unloading of all users in the system,a multivariate joint iterative algorithm based on the idea of block coordinate descent is proposed.The algorithm divides the non-convex user offloading optimization problem under the constraints of energy consumption and service quality into three sub-problems.First,the sub-problems are solved by using integer programming and successive convex optimization methods,and then the overall optimization of unmanned machine trajectory,user offloading strategy and user upload power,so as to maximize the total user offloading.The simulation experiment results show that,compared with the algorithm with fixed user upload power and fixed UAV trajectory,the proposed multivariable joint iterative algorithm based on block coordinate descent has higher performance,and the unloading amount is increased by about 11%and 21.4%,respectively.In a multi-drone-assisted edge computing system,consider hovering multiple drones each at certain locations to serve user equipment.Aiming at the mixed decision-making multivariable optimization problem,the original problem with multiple variables is divided into two sub-problems,and a parameter adaptive differential evolution algorithm is proposed to solve the multi-UAV deployment sub-problem,and the greedy algorithm is used to solve the corresponding sub-problems.The user offloading decision and resource allocation sub-problems to minimize system energy consumption.Here,not only the unloading decision problem between multi-UAVs and multi-user devices is considered,but also the number of UAVs and their hovering positions are optimized at the same time to ensure that there will be no collisions between UAVs and improve task completion.rate,reducing the total energy consumption of the system.