Research On Mobile Edge Computing Resource Scheduling Algorithm For UAV

With the increase in military application requirements and the improvement of UAV performance and technology,the application range of UAVs in the military field is increasing rapidly.UAVs are widely used in reconnaissance operations because of their good maneuverability,flexibility,and ease of deployment.In the process of multi-UAVs cooperative reconnaissance,the battery capacity and computing resources of the UAV are limited,and too much energy consumption and delay will be generated when dealing with tasks.The use of cloud computing technology will not only cause a waste of bandwidth resources;data transmission delays and energy consumption are too large;real-time cannot be better satisfied;and other issues Mobile edge computing(MEC)provides a solution to the above problems by providing computing services for drones at the data source.However,the edge server is usually deployed near the base station at a fixed position on the ground,making it impossible to effectively provide communication and computing services for reconnaissance drones.Therefore,this paper studies a MEC system architecture consisting of a single unmanned helicopter(UH)and multiple reconnaissance drones.The unmanned helicopter is used as the air base station and is equipped with an edge server to provide communication and computing services for the drone.Based on the MEC system architecture,the task computing offloading strategy and resource scheduling problem of multi-UAVs are studied to reduce the delay and energy consumption of task processing,thereby reducing the battery life loss of UAVs and satisfying the user service quality experience.The main research work of this paper is as follows:(1)In the aspect of UAV task computation offloading,the MEC system model for UAV is constructed,and the objective function is constructed to minimize the energy consumption and delay of system task execution.Compared with particle swarm optimization(PSO)and genetic algorithms(GA),the bat algorithm(BA)has the advantages of having fewer adjustment parameters,high precision,high efficiency,and good stability.However,there are also shortcomings,such as the easy fall into local optimum and slow convergence speed in the later stage.Therefore,this paper improves the global search ability and fast convergence ability of the bat algorithm by introducing a linear decreasing inertia weight strategy.The simulation experiment takes the objective function value,the number of optimal solutions,and the total running time of the algorithm as the evaluation indexes.The simulation results show that compared with the benchmark algorithm,the improved bat algorithm(IBA)is more accurate,stable,and efficient in solving the computation offloading problem of multi-UAVs’ task.In terms of the weighted sum of energy consumption and delay of task execution,the IBA is 4.6%,10%,and 3.7% lower than the BA,PSO,and IPSO,respectively.(2)In terms of system resource scheduling,the MEC system model and system task queue model are constructed,and the objective function is constructed to minimize the average execution energy consumption of system tasks.Aiming at the mixed integer nonlinear programming problem caused by the coupling relationship between variables such as offloading strategy,communication resources,and computing resources in the resource scheduling problem,a system resource scheduling algorithm based on Lyapunov optimization is adopted.Firstly,the task queue is analyzed,and the Lyapunov drift plus penalty function is established.Secondly,the optimization problem is transformed into a queue stability control problem based on Lyapunov optimization.Then it is decomposed into three sub-problems: computing offloading strategy and UAV user local computing resource allocation,transmission power and bandwidth resource allocation,and MEC server computing resource allocation,and solved separately.Finally,a joint optimization scheme is designed.The simulation experiment uses task queue length and task execution energy consumption as evaluation indexes.The simulation results show that the performance of the proposed algorithm is better than that of the benchmark algorithm,and the average task execution energy consumption of the system is minimized under the premise of ensuring the stability of the task queue.