随着工业4.0时代的到来，智能制造竞争日益激烈，仓储物流的繁忙程度也随之增加，在许多大型电商平台的自动化仓储中，成百上千个AGV在仓库中自动导航，将货物从一个指定地点运送到另一个指定地点。智能仓储优化算法主要集中在仓储布局、货位分配和AGV路径规划这三部分，当AGV规模庞大时，由于AGV会产生阻塞、冲突等问题，导致仓储吞吐量降低，严重影响了智能仓储的运行效率。主要工作如下：

（1）针对智能仓储布局不合理导致仓储通道利用率低、AGV阻塞率高、仓储吞吐量小等问题，以最大化仓储吞吐量为目标，提出一种基于DNN-MAP-Elites的智能仓储布局优化及评价算法。该算法保留了MAP-Elites算法布局通道多样性，融合ε-支配准则和深度代理模型，降低算法计算资源代价，设计以通道利用率、阻塞率、吞吐量为核心指标的仓储布局质量评价指标体系，分别对DNN-MAP-Elites布局方法、传统型布局方法、Flying-V型布局方法、鱼骨型布局方法进行AGV路径规划性能测试。实验结果表明，本文方法在小尺寸仓库阻塞率相较Flying-V型和鱼骨型布局方法低10%，吞吐量相较Flying-V型和鱼骨型布局方法提升33%。在大尺寸仓库阻塞率相较Flying-V型和鱼骨型布局方法低13%，吞吐量相较其他布局方法提升27%。

（2）针对智能仓储货位分配相邻货架通道阻塞严重、AGV运输效率低、仓储吞吐量低等问题，提出一种基于FD-CGA的货位分配算法。该算法首先通过FP-Growth算法对货物关联规则进行分析，其次，根据基于路径规划算法计算各个货架到达工作站点的最佳出库路径，最后，使用改进的混沌遗传算法利用计算完成的数据给出货位分配结果，从而将货架到工作站点的路径与货物的关联规则以及货位分配相结合，实现货位的合理分配。在本文第一部分生成的最优布局上进行货位分配实验。实验结果表明，在随机订单情况下，FD-CGA算法相较于NSGA-II算法和贪心算法任务完成耗时下降47%。在高支持度、高置信度货物，低支持度、高置信度，高支持度、低置信度货物订单实验中，FD-CGA算法相对于NSGA-II算法与贪心算法，冲突次数与阻塞次数显著减少，任务完成时间降低17%。

（3）为验证以上方法的有效性，设计一个智能仓储管理系统。该系统具有用户及权限管理、出入库管理、货位管理、布局优化的功能。其中布局优化功能基于DNN-MAP-Elites布局优化算法，货位管理功能可以使用FD-CGA货位分配算法对货位进行分配。通过对系统进行功能测试与性能测试，测试结果表明，该系统具有易用性、高效性、可靠性，能够满足智能仓储管理的预期需求。

With the advent of the Industry 4.0 era, the competition in intelligent manufacturing has become increasingly fierce, and the busyness of warehouse logistics has correspondingly increased. In the automated warehouses of many large e-commerce platforms, hundreds to thousands of AGVs automatically navigate within the facilities, transporting goods from one designated location to another. Intelligent warehouse optimization algorithms primarily focus on three aspects: warehouse layout, storage allocation assignment, and AGV path planning. However, when AGV fleets scale up, issues such as congestion and conflicts arise, reducing warehouse throughput and severely impacting operational efficiency. The main contributions of this work are as follows:

To address the issues of low storage aisle utilization, high AGV blockage rates, and limited warehousing throughput caused by irrational intelligent warehousing layouts, this study proposes a DNN-MAP-Elites-based intelligent warehousing layout optimization and evaluation algorithm aiming to maximize warehousing throughput. The algorithm preserves the layout channel diversity of the MAP-Elites algorithm, integrates the ε-dominance criterion with a deep neural network surrogate model to reduce computational resource costs, and designs a warehousing layout quality evaluation model focusing on core metrics including aisle utilization, blockage rate, and throughput. Performance tests of AGV path planning are conducted for the DNN-MAP-Elites layout method, conventional layouts, Flying-V layouts, and fishbone layouts. Experimental results demonstrate that in small-scale warehouses, DNN-MAP-Elites achieves blockage rate of 10%, 10% lower than Flying-V and fishbone layouts, while increasing throughput by 33% compared to other methods. In large-scale warehouses, it maintains blockage rate of 10.4%, 13% lower than Flying-V and fishbone layouts, and enhances throughput by 27%.

To address the issues of severe congestion in adjacent shelf aisles during intelligent warehouse storage location allocation, low AGV transportation efficiency, and low warehouse throughput, a storage location allocation algorithm based on FD-CGA is proposed. The algorithm first analyzes goods association rules using the FP-Growth algorithm. Second, it calculates optimal outbound paths for each shelf based on path planning algorithms. Finally, the improved chaotic genetic algorithm utilizes the computed data to generate goods allocation results, integrating shelf outbound paths, product association rules, and storage location allocation to achieve rational goods placement. Storage location assignment experiments were conducted based on the optimal layout generated in the first part of this paper. The experimental results show that under random order conditions, the FD-CGA algorithm reduces task completion time by 47% compared to the NSGA-II algorithm and the greedy algorithm. In experiments involving orders of high-support, high-confidence goods; low-support, high-confidence goods; and high-support, low-confidence goods, the FD-CGA algorithm significantly reduces the number of conflicts and blockages and decreases task completion time by 17% compared to the NSGA-II algorithm and the greedy algorithm.

To evaluate the effectiveness of the above methods, An intelligent warehouse management system is designed. This system includes user and permission management, inbound and outbound management, storage location management, and layout optimization functions. The layout optimization function is based on the DNN-MAP-Elites layout optimization algorithm, while the storage location management function employs the FD-CGA storage location allocation algorithm for slot assignment. Through functional and performance testing, the results demonstrate that the system exhibits usability, efficiency, and reliability, meeting the expected requirements of intelligent warehouse management.

[1]杨伊静. 抢占关键技术制高点 力推制造业转型升级——工信部等多部门印发《“十四五”智能制造发展规划》[J]. 中国科技产业, 2022, 391(01): 40-41.

[2]司明, 邬伯藩, 胡灿等. 智能仓储交通信号与多AGV路径规划协同控制方法[J]. 计算机工程与应用, 2024, 60(11): 290-297.

[3]Zhang X, Mo T, Zhang Y. Optimization of Storage Location Assignment for Non-Traditional Layout Warehouses Based on the Firework Algorithm[J]. Sustainability, 2023, 15(13): 10242.

[4]丁海毅, 佘世刚, 强运哲等. 大型智能仓储多AGV无冲突路径规划研究[J]. 机床与液压, 2025, 53(03): 74-80.

[5]Mohamud I H, Kafi M A, Shahron S A, et al. The Role of Warehouse Layout and Operations in Warehouse Efficiency: A Literature Review[J]. Journal Européen des Systèmes Automatisés, 2023, 56(1).

[6]侯智, 孟祥超. 基于SLP与遗传算法的仓储布局优化[J]. 组合机床与自动化加工技术, 2020, (05): 159-163.

[7]祝凌瑶, 周丽. 现代仓储中心存储布局的优化研究[J]. 工程数学学报, 2022, 39(06): 862-874.

[8]Shi X, Deng F, Lu S, et al. A bi-level optimization approach for joint rack sequencing and storage assignment in robotic mobile fulfillment systems[J]. Science China Information Sciences, 2023, 66(11): 212202.

[9]李丽, 刘保国, 武照云等. 基于改进人工蜂群算法的四向穿梭车仓储系统货位优化研究[J]. 包装工程, 2024, 45(19): 265-274.

[10]邱雄飞, 张桦, 赵润泽. 基于PSO-GA算法的后方仓库货位分配优化[J]. 信息工程大学学报, 2024, 25(04): 423-427.

[11]Chen Z, Alonso-Mora J, Bai X, et al. Integrated task assignment and path planning for capacitated multi-agent pickup and delivery[J]. IEEE Robotics and Automation Letters, 2021, 6(3): 5816-5823.

[12]宋作玲, 殷祥栋. 基于A*算法的AGV路径规划研究文献综述[J]. 物流科技, 2025, 48(03): 41-43.

[13]杨振, 李俊丽, 杨立炜等. 安全性A*融合DWA的分布式多移动机器人路径规划方法[J]. 控制工程, 2024, 31(12): 2284-2295.

[14]张开来. 面向智能仓储的仓位布局优化与多AGV动态调度方法研究[D]. 湖南大学, 2022.

[15]张洪琳. 多载位移动机器人拣选系统订单策略与路径协同优化研究[D]. 山东大学, 2023.

[16]Zhang Y, Fontaine M C, Bhatt V, et al. Multi-robot coordination and layout design for automated warehousing[C]//Proceedings of the International Symposium on Combinatorial Search. 2024, 17: 305-306.

[17]Alkazzi J M, Okumura K. A Comprehensive Review on Leveraging Machine Learning for Multi-Agent Path Finding[J]. IEEE Access, 2024, 12: 57390-57409.

[18]杨丽. 煤矿智能仓储系统研究与设计[J]. 中国煤炭, 2022, 48(02): 48-54.

[19]曹霞, 曹民. 基于改进克隆选择算法的仓储布局优化设计[J]. 控制工程, 2020, 27(02): 329-334.

[20]Pinto A R F, Nagano M S, Boz E. A classification approach to order picking systems and policies: Integrating automation and optimization for future research[J]. Results in Control and Optimization, 2023, 12: 100281.

[21]Lamballais T, Roy D, De Koster M B M. Estimating performance in a Robotic Mobile Fulfillment System[J]. European Journal of Operational Research, 2017, 256(3): 976-990.

[22]Xu T, Yang P, Guo H. Energy efficiency analysis on robotic mobile fulfillment system[C]//2019 IEEE 6th International Conference on Industrial Engineering and Applications (ICIEA). IEEE, 2019: 145-149.

[23]闫青, 鲁建厦, 江伟光等. 考虑双端口布局的紧致化仓储系统堆垛机路径优化[J]. 上海交通大学学报, 2022, 56(07): 858-867.

[24]邓旭东, 张马萍, 吴应强等. 三维紧致化存储系统中货架尺寸的优化研究[J]. 包装工程, 2019, 40(21): 173-178.

[25]Xu X, Zhao X, Zou B, et al. Travel time modelsfor a threedimensional compact AS/RS consideringdifferent I/O point policies[J]. International Journalof Production Research, 2019, 58(18): 5432-5455.

[26]刘建胜, 张有功, 熊峰等. Flying-V型仓储布局货位分配优化方法研究[J]. 运筹与管理, 2019, 28(11): 27-33.

[27]Yang X, Liu X, Feng L, et al. Non-traditional layout design for robotic mobile fulfillment system with multiple workstations[J]. Algorithms, 2021, 14(7): 203.

[28]Wan Y, Liu Y. Integrating optimized fishbone warehouse layout, storage location assignment and picker routing[J]. IAENG International Journal of Computer Science, 2022, 49(3): 957-974.

[29]Zhou L, Li Z, Shi N, et al. Performance analysis of three intelligent algorithms on route selection of fishbone layout[J]. Sustainability, 2019, 11(4): 1148.

[30]Cardona L F, Rivera L, Martinez H J. Analytical study of the fishbone warehouse layout[J]. International Journal of Logistics Research and Applications, 2012, 15(6): 365-388.

[31]Zhang Z Y, Liang Y, Hou Y P, et al. Designing a warehouse internal layout using a parabolic aisles based method[J]. Advances in Production Engineering & Management, 2021, 16(2): 223-239.

[32]Wang Y, Man R, Zhao W, et al. Storage assignment optimization for fishbone robotic mobile fulfillment systems[J]. Complex & Intelligent Systems, 2022, 8(6): 4587-4602.

[33]赵巍, 连泰湖, 张雷等. 基于数字孪生的自动化立体仓库货位分配优化方法[J]. 航空制造技术, 2023, 66(06): 66-73.

[34]曹春玲, 邵杨, 何龙龙等. 面向煤炭企业自动化立体仓库的改进SAGA货位优化[J]. 煤炭技术, 2023, 42(10): 190-194.

[35]范贤光, 吴俊涛, 尹艺玲等. 基于改进鱼群算法的双向式立体库货位优化研究[J]. 计算机仿真, 2023, 40(05): 353-357+373.

[36]郭科仁. 基于多目标粒子群算法的G电商企业仓储中心零拣区货位优化研究[D]. 福州大学, 2023.

[37]Tinelli L M, Vivaldini K C T, Becker M. Intelligent warehouse product position optimization by applying a multi-criteria tool[C]//International Workshop on Robotics in Smart Manufacturing. Berlin, Heidelberg: Springer Berlin Heidelberg, 2013: 137-145.

[38]Arnaout J P. Worm optimization for the multiple level warehouse layout problem[J]. Annals of Operations Research, 2018, 269(1): 29-51.

[39]Lanza G, Passacantando M, Scutellà M G. Assigning and sequencing storage locations under a two level storage policy: Optimization model and matheuristic approaches[J]. Omega, 2022, 108: 102565.

[40]Meng Wang, Ren Qian Zhang, Kun Fan. Improving order-picking operation through efficient storage location assignment: A new approach[J]. Computers & Industrial Engineering, 2020, 139(C): 106186.

[41]徐云龙. 面向边际效益的智能仓储多负载AGV任务调度与路径规划算法研究[D]. 西安电子科技大学, 2022.

[42]郭涵. 基于偏好的质量多样性算法[D]. 电子科技大学, 2024.

[43]詹红有, 梁峻源, 肖宁聪. 基于主动学习代理模型的结构时空相关可靠性分析方法[J]. 电子科技大学学报, 2025, 54(01): 84-90.

[44]Bhatt V, Tjanaka B, Fontaine M, et al. Deep surrogate assisted generation of environments[J]. Advances in Neural Information Processing Systems, 2022, 35: 37762-37777.

[45]Zhang Y, Fontaine M C, Bhatt V, et al. Multi-robot coordination and layout design for automated warehousing[C]//Proceedings of the International Symposium on Combinatorial Search. 2024, 17: 305-306.

[46]王凤英, 陈莹, 袁帅等. 自注意力机制结合DDPG的机器人路径规划研究[J]. 计算机工程与应用, 2024, 60(19): 158-166.

[47]黄博. 基于生成对抗网络的遥感图像融合模型研究[D]. 吉林大学, 2024.

[48]Vaswani A, Shazeer N, Parmar N, et al. Attention is all you need[J]. Advances in neural information processing systems, 2017, 30.

[49]王继禾, 吴颖, 迟恒喆等. 基于Transformer结构增强的神经网络架构搜索性能预测器[J]. 计算机学报, 2024, 47(07): 1469-1484.

[50]郭丹, 姚沈涛, 王辉等. 嵌入局部聚类描述符的视频问答Transformer模型[J]. 计算机学报, 2023, 46(04): 671-689.

[51]Gao Y ,Qiu S ,Liu M , et al. Fault Warning of Satellite Momentum Wheels With a Lightweight Transformer Improved by FastDTW[J]. IEEE/CAA Journal of Automatica Sinica, 2025, 12(03): 539-549.

[52]马瑞. 基于自注意力神经网络的室内可见光定位方法[J]. 光学技术, 2025, 51(01): 108-115.

[53]段雄, 范小宁. 基于动态RBF代理模型和进化算法的起重机主梁优化[J]. 机械设计, 2025, 42(03): 86-94.

[54]张凯, 陈旭, 刘丕养等. 基于深度学习贝叶斯模型平均代理的油藏自动历史拟合研究[J]. 钻采工艺, 2025, 48(01): 147-156.

[55]詹红有, 梁峻源, 肖宁聪. 基于主动学习代理模型的结构时空相关可靠性分析方法[J]. 电子科技大学学报, 2025, 54(01): 84-90.

[56]Kilimtzidis S, Kostopoulos V. Static aeroelastic optimization of high-aspect-ratio composite aircraft wings via surrogate modeling[J]. Aerospace, 2023, 10(3): 251.

[57]Meng D, Yang H, Yang S, et al. Kriging-assisted hybrid reliability design and optimization of offshore wind turbine support structure based on a portfolio allocation strategy[J]. Ocean Engineering, 2024, 295: 116842.

[58]Yu Y, Zhao M, Li Z, et al. Optimal design of asymmetrically arranged moorings in a floating production system based on improved particle swarm optimization and RBF surrogate model[J]. Marine Structures, 2024, 94: 103576.

[59]Wu Z, Fang G. The Research on the Improvement of FP-growth Algorithm[J]. Artificial Intelligence Technology Research, 2024, 2(1).

[60]Pan H, Liu R. Mining research on correlation factors of residential electricity stability based on improved FP-growth algorithm[J]. Management of Environmental Quality: An International Journal, 2024, 35(3): 547-566.

[61]Alhijawi B, Awajan A. Genetic algorithms: Theory, genetic operators, solutions, and applications[J]. Evolutionary Intelligence, 2024, 17(3): 1245-1256.

[62]Wang Y C, Chen T. Adapted techniques of explainable artificial intelligence for explaining genetic algorithms on the example of job scheduling[J]. Expert Systems with Applications, 2024, 237: 121369.

[63]李梦凡, 张育芝, 韩翔等. 基于多智能体深度强化学习的水声网络资源分配[J]. 电讯技术, 2025, 65(02): 283-292.

[64]魏东岩, 巨柳荫, 纪新春等. 面向复杂场景的地图轻量化与匹配定位方法[J]. 中国惯性技术学报, 2025, 33(02): 114-123.

[65]林佳俐, 李永强, 赵硕等. 结合先验知识的SAC神经纤维追踪算法及应用[J]. 小型微型计算机系统, 2024, 45(07): 1719-1727.

[66]Li J, Tinka A, Kiesel S, et al. Lifelong multi-agent path finding in large-scale warehouses[C]//Proceedings of the AAAI Conference on Artificial Intelligence. 2021, 35(13): 11272-11281.

[67]Liu Z, Lin Y, Cao Y, et al. Swin transformer: Hierarchical vision transformer using shifted windows[C]//Proceedings of the IEEE/CVF international conference on computer vision. 2021: 10012-10022.

[68]Chen J, Kao S, He H, et al. Run, don't walk: chasing higher FLOPS for faster neural networks[C]//Proceedings of the IEEE/CVF conference on computer vision and pattern recognition. 2023: 12021-12031.

[69]蔺一帅, 李青山, 陆鹏浩等. 智能仓储货位规划与AGV路径规划协同优化算法[J]. 软件学报, 2020, 31(09): 2770-2784.

[70]Duan J, Zhang L, Wang Y, et al. Distributional soft actor-critic with three refinements[J]. IEEE Transactions on Pattern Analysis and Machine Intelligence, 2025, 45(12): 12345-12356.

[71]Zhao L, Gao J, Deng D, et al. SSIR: Spatial shuffle multi-head self-attention for single image super-resolution[J]. Pattern Recognition, 2024, 148: 110195.