关键词: 末端配送, 任务分配, 强化学习, 无人机, 多目标优化

Abstract: Abstract: The difficulty, long delivery time, and high cost of last mile delivery in rural areas make efficient and accurate end-to-end delivery scheduling solutions particularly important. Aiming at the task allocation problem of multiple logistics drones in rural distribution scenarios, a multi-objective drone task allocation model is established by comprehensively considering the payload capacity of drones and the maximum flight distance of drones, with the goal of minimizing the flight distance, dispatch quantity, and not violating time windows of drones. Based on reinforcement learning, to address the problem of high dimensionality in task allocation, an encoder and attention mechanism are introduced to effectively simplify the state space. Combining the global local search strategy to explore the solution space while avoiding getting stuck in local optima, thereby improving the quality of the solution. Finally, further analysis was conducted on the parameter weight settings, and the optimal combination of weight coefficients for each sub objective function was obtained through experiments. The simulation results show that the final path length obtained by the proposed algorithm is reduced by 8.35%, 9.88%, 10.29%, and 12.48% compared to the hybrid Q-learning network method (HQM), adaptive large neighborhood search algorithm (ALNS), Q-learning algorithm (Q-learning), and genetic algorithm (GA), respectively.

Key words: End of line delivery, Task allocation, Reinforcement learning, unmanned aerial vehicle(UAV), Multi objective optimization