目的 提高四向穿梭车立体仓储系统（FWSS/RS）在出库和入库作业任务并存的复合作业场景下的运行效率和资源利用率。方法 针对现有研究主要聚焦出库或入库的单一作业模式,对穿梭车、提升机等设备协同调度考虑不足的问题,面向出库和入库作业任务并存的复合作业场景,提出一种基于改进独立深度Q网络（IDQN）的任务分配方法。首先,构建以总完成时间最小化为目标的FWSS/RS复合作业任务分配模型。其次,将模型转化为马尔可夫决策过程（MDP）,设计状态空间、动作空间、奖励函数等MDP要素,采用改进IDQN算法求解。结果 开展实验研究,与拍卖算法、遗传算法等方法进行对比,分析任务规模、运输资源配置等因素对任务分配方法性能的影响,验证了所提IDQN方法的有效性和优势。结论 基于改进IDQN的复合作业任务分配方法能够显著提升四向穿梭车与提升机的协同调度水平,可有效支撑实际物流场景下的作业任务分配需求。
Abstract
The work aims to improve the operational efficiency and resource utilization rate of the four-way shuttle storage and retrieval system (FWSS/RS) in the composite operation scenario where both outbound and inbound operation tasks coexist. With focus on single inbound or outbound modes and neglect collaborative scheduling between shuttles and lifts, an improved independent deep Q-network (IDQN)-based task allocation method was proposed for mixed inbound-outbound scenarios. A task allocation model for FWSS/RS composite operations was constructed with the objective of minimizing makespan. Furthermore, the model was transformed into a Markov decision process (MDP), where the MDP elements like state space, action space, and reward function were designed, and the improved IDQN algorithm was employed to derive optimal solutions. Finally, experiments were conducted in conjunction with engineering practice, and comparisons were made with the auction algorithm and the genetic algorithm. Meanwhile, the effects of task scale, transport resource configuration, and other factors on the performance of the proposed method were analyzed, verifying its effectiveness and advantages. The composite operation task allocation method based on the improved IDQN can significantly enhance the collaborative scheduling level of four-way shuttles and lifts, and can effectively support the operation task allocation requirements in actual logistics scenarios.
关键词
四向穿梭车立体仓储系统 /
复合作业 /
任务分配 /
马尔可夫决策过程 /
独立深度Q网络
Key words
four-way shuttle storage and retrieval system /
composite operation /
task allocation /
Markov decision process /
independent deep q-network
{{custom_sec.title}}
{{custom_sec.title}}
{{custom_sec.content}}
参考文献
[1] KUHN H.The Hungarian Method for the Assignment Problem[J]. Naval Research Logistics, 2010(52): 7-21.
[2] MARTIN J, FREJO J, GARCIA R, et al.Multi-robot Task Allocation Problem with Multiple Nonlinear Criteria Using Branch and Bound and Genetic Algorithms[J]. Intelligent Service Robotics, 2021(14): 707-727.
[3] QIU Z, LONG J, YU Y, et al.Integrated Task Assignment and Path Planning for Multi-type Robots in an Intelligent Warehouse System[J]. Transportation Research Part E-Logistics and Transportation Review, 2025(194): 103883.
[4] BAROUDI U, ALSHABOTI M, KOUBAA A, et al.Dynamic Multi-objective Auction-based (DYMO-auction) Task Allocation[J]. Applied Sciences, 2020(10): 3264.
[5] LI J, LIU K, WANG H.Task Assignment Optimization of Multi-logistics Robot Based on Improved Auction Algorithm[C]//Proceedings of the 9th International Conference on Logistics, Informatics and Service Sciences. Singapore: Springer, 2020: 41-54
[6] MA T, LU P, DENG F, et al.Air-ground Collaborative Multi-target Detection Task Assignment and Path Planning Optimization[J]. Drones, 2024, 8: 110.
[7] ASMA A, SADOK B.PSO-based Dynamic Distributed Algorithm for Automatic Task Clustering in a Robotic Swarm[J]. Procedia Compute Sciences, 2019(159): 1103-1112.
[8] GUO H, MIAO Z, PAN Q.An Effective Collaboration Evolutionary Algorithm for Multi-robot Task Allocation and Scheduling in a Smart Farm[J]. Knowledge-Based Systems, 2024(289): 111474.
[9] CHEN R, YANG J, YU Y, et al.Retrieval Request Scheduling in a Shuttle-based Storage and Retrieval System with Two Lifts[J]. Transportation Research Part E-Logistics and Transportation Review, 2023(174): 103117.
[10] EKREN B, ARSLAN B.A Reinforcement Learning Approach for Transaction Scheduling in a Shuttle-based Storage and Retrieval System[J]. International Transactions in Operational Research, 2024(31): 274-295.
[11] WU S, CHIU W, WU C.Deep Reinforcement Learning for Task Assignment and Shelf Reallocation in Smart Warehouses[J]. IEEE Access, 2024(12): 58915-58926.
[12] AGRAWAL A, HARIHARAN S, BEDI A, et al.DC-MRTA: Decentralized Multi-robot Task Allocation and Navigation in Complex Environments[C]//Proceedings of the 2022 IEEE/RSJ International Conference on Intelligent Robots and Systems, Kyoto: IEEE, 2022: 11711-11718.
[13] HERSI A, UDAYAN J.Efficient and Robust Multirobot Navigation and Task Allocation Using Soft Actor Critic[J]. Procedia Computer Science, 2024(235): 484-495.
[14] 武照云, 赵彬彬, 张中伟, 等. 基于改进DQN的多深度四向穿梭车仓储系统货位分配研究[J]. 现代电子技术, 2026, 49(2): 178-186.
WU Z Y, ZHAO B B, ZHANG Z W, et al.Research on Improved DQN Based Location Allocation for MFSWS[J]. Modern Electronic Technology, 2026, 49(2): 178-186.
[15] 张中伟, 高增恩, 王菁锐, 等. 智能制造车间分布式AGV任务分配研究[J]. 包装工程, 2025, 46(7): 142-149.
ZHANG Z W, GAO Z E, WANG J R, et al.Research on Distributed AGV Task Allocation in Intelligent Manufacturing Workshop[J]. Packaging Engineering, 2025, 46(7): 142-149.
[16] EIJYNE T, G R, SG P. Development of a Task-oriented, Auction-based Task Allocation Framework for a Heterogeneous Multirobot System[J]. Sadhana, 2020, 45(1): 115.
[17] WATANABE M, IDA K, GEN M.A Genetic Algorithm with Modified Crossover Operator and Search Area Adaptation for the Job-shop Scheduling Problem[J]. Computers & Industrial Engineering, 2005, 48(4): 743-752.
基金
河南省科技研发计划联合基金项目（242103810064）; 河南省科技副总资助项目（HNFZ20250128）; 河南工业大学校企联合项目（243H2024JD243,414H2024JD414）
PDF(2074 KB)
渝公网安备 50010702501716号 渝ICP备15012534号-2
