Research On Resource Allocation Of Mobile Edge Computing Based On UAV

Due to the extreme sensitivity to latency and energy consumption,many computation and data-intensive tasks are difficult to be implemented on user terminals and cannot meet the needs of the rapid development of user equipment.Mobile edge computing can provide low-latency,low-energy computing services,providing an effective solution to these problems.However,the location of mobile edge servers is usually fixed and cannot be changed flexibly according to the needs of mobile users,which limits the capability of mobile edge servers.Unmanned aerial vehicle has good mobility and flexibility,and arranging mobile edge servers on UAVs has two advantages: flexible deployment to shorten the transmission distance and help edge servers to cope with the challenges of transient computing surge;and the ability to provide better line-of-sight links to mobile users with higher probability.However,the limited endurance of UAVs and the mobile edge servers they carry can affect the performance of mobile edge computing services and easily lead to incomplete mobile edge services within the time limit problem.Therefore,this paper addresses the problem of UAVs serving ground terminals when the task volume of multiple terminals is too large,resulting in UAVs failing to meet user service quality.By designing a reasonable UAV deployment location and also studying the optimization problem of UAV resource allocation,the terminal tasks can be executed more effectively.Based on this,this paper addresses the resource allocation problem in the field of UAV-assisted edge computing,and the main work is as follows.（1）In terms of UAV server deployment.This paper constructs a mathematical model of multiple UAVs serving multi-faceted terminals with UAV deployment cost and UAV energy consumption as indicators,proposes an improved mean drift clustering algorithm,and optimizes the deployment location and number of UAVs by searching the optimal bandwidth value of the algorithm.Simulation results show that the UAV deployment algorithm proposed in this paper has better coverage and lower system energy consumption compared with the benchmark algorithm.（2）In terms of single UAV resource allocation.This paper takes a single UAV serving multiple ground terminals as a scenario,constructs an energy consumption model under the time delay constraint,constructs an objective function with the goal of optimizing energy consumption,considers issues such as UAV bandwidth and the way of computing resources allocation,and finds the optimal unloading location of the terminal through genetic algorithm to minimize the energy consumption of a single UAV serving ground terminals.Simulation experiments prove that the resource allocation strategy based on genetic algorithm proposed in this paper has faster convergence speed and fewer iterations compared with the benchmark algorithm,and the objective function value is better.（3）In terms of multi-UAV resource allocation.In this paper,a multi-UAV multi-terminal resource service model under time delay and load constraints is constructed for task unloading strategy,UAV task migration strategy,and resource allocation strategy of UAVs.In the multi-UAV multi-terminal resource service model,a resource allocation strategy based on non-dominated ranking genetic algorithm is designed to achieve the goals of optimal user service quality and UAV load balancing in multi-UAV scenarios.Simulation experiments prove that the proposed algorithm in this paper has better performance compared with the benchmark algorithm in terms of task latency and server utilization.