The 15th Research Institute, China Electronics Technology Group Corporation, Beijing, China
Email: mjjblcu@126.com (J.J.M.); 419061969@qq.com (X.H.L.); 17310556670@163.com (Y.X.K.)
*Corresponding author
Manuscript received March 9, 2026; revised March 30, 2026; accepted April 9, 2026; published May 20, 2026
Abstract—The vigorous development of the low-altitude economy has led to a core contradiction in the field of unmanned logistics delivery: dynamically changing delivery demands versus limited resource allocation. To address this issue, a dynamic scheduling algorithm for unmanned delivery resources based on hierarchical attention is proposed, aiming to solve the problems of agile cooperative scheduling and dynamic adaptive expansion of unmanned delivery resources in an environment with variable delivery demands and achieve the overall optimization of delivery efficiency. The unmanned delivery resource scheduling task is modeled as a Markov Decision Process (MDP) with the goal of global delivery optimization. A multi-agent reinforcement learning framework is adopted to realize cross-domain cooperative decision-making of unmanned delivery resources, and a hierarchical attention mechanism is designed to support the unified representation of variable-scale cross-domain heterogeneous unmanned clusters. The prioritized experience replay technology is applied and a multi-scale reward function is established to solve the problem of reward sparsity in the multi-agent cooperative environment, which effectively improves the stability of the algorithm training process and the accuracy of the results. Experimental results show that the hierarchical attention mechanism has good multi-dimensional cooperative perception ability of environmental information; in the scenario where unmanned delivery resources can be newly added and expanded, the final task completion rate reaches 0.81, which is 12.5% higher than that of the strategy without expansion. The proposed algorithm provides an efficient solution for dynamic resource scheduling in unmanned logistics while laying a scalable and adaptable technical foundation for broader applications in intelligent transportation systems and smart city logistics.
Keywords—unmanned delivery resource scheduling, hierarchical attention mechanism, multi-agent reinforcement learning, adaptive expansion, dynamic scheduling
Cite: Jingjing Miao, Xinghua Li, and Yaoxi Kang, "Dynamic Scheduling Algorithm for Unmanned Delivery Resources Based on Hierarchical Attention," International Journal of Engineering and Technology, vol. 18, no. 2, pp. 37-44, 2026.
Copyright © 2026 by the authors. This is an open access article distributed under the Creative Commons Attribution License which permits unrestricted use, distribution, and reproduction in any medium, provided the original work is properly cited (
CC BY 4.0).
