无人机辅助移动边缘计算中的计算卸载与轨迹优化策略OA

The unmanned aerial vehicle(UAV)-assisted mobile edge computing(MEC)system,known for its efficient ser-vice capabilities,provides temporary computing and communication support for ground users in special scenarios such as remote areas,emergency rescue and post-disaster reconstruction.However,obstacles such as tall buildings in the real envi-ronment cause significant interference to the flight trajectory and task execution of UAV,which not only threaten flight safety,but also reduce the efficiency of task execution.In order to solve the above problem,a strategy of jointly optimizing the computation offloading and UAV flight trajectory is proposed.A model of UAV-assisted mobile edge computing system considering the existence of obstacles is constructed,and a weight self-adaptive adjustment mechanism is introduced,with the objective of minimizing the weighted value of the total system delay and energy consumption.For this optimization objective,a deep reinforcement learning algorithm is designed,the algorithm integrates the prioritized experience replay(PER)mechanism with the twin delayed deep deterministic policy gradient(TD3)algorithm.The performance of the proposed strategy is analyzed by simulation experiments.Simulation results indicate that,compared with the three bench-mark algorithms of TD3,DDPG,and AC,when the total amount of computing tasks,bandwidth,computing capabilities of user equipment,and the number of users are changed,the weighted value of the total system delay and energy consumption of the proposed algorithm is reduced by an average of 5.31%,8.86%,and 43.30%respectively,and it can optimize a smoother flight trajectory with better obstacle avoidance capabilities.