基于大语言模型智能体的船舶堆场数字孪生动态布局生成方法

doi:10.3901/JME.260625

摘要/Abstract

摘要： 针对船舶建造堆场布局需频繁响应坞期调整与工艺变更的动态特性，提出融合智能体推理机制的动态布局生成方法。通过将智能体技术引入堆场布局设计，旨在敏捷响应用户需求，同时精细处理设备间的复杂空间约束，展现出强大的自适应和定制化潜力。为提升模型推理过程的信息处理效率与准确性，本研究构建了一套基于智能体的任务分解-规划-评估反思推理框架。该框架将多个推理步骤串联形成推理链路，使模型能够分阶段、结构化地对复杂布局问题进行深入分析与推导。此方法不仅有效缓解了信息过载问题，还显著提升了模型在处理此类复杂任务时的准确性和效率。此外，为确保多步推理迭代中信息处理的可靠性与决策的正确性，系统制定了明确的标准和判定规则，以保障生成有效的推理序列。最后，以大型造船企业分段堆场为案例，验证方法在应对突发坞期变更场景下的布局重构能力。实验表明，该方法在保持空间利用率的前提下，显著提升布局方案对动态需求的响应速度与约束满足度，为船舶制造堆场的智能化动态调整提供了新思路。

关键词: 大语言模型智能体, 数字孪生, 布局设计, 增强检索生成, 船舶制造

Abstract: This study proposes a dynamic layout generation method integrating an agent-based reasoning mechanism to address the dynamic nature of shipyard block layouts, which must frequently adapt to docking schedule adjustments and process changes. By introducing agent technology into layout design, this approach enables an agile response to user requirements while precisely managing complex spatial constraints between equipment, demonstrating strong potential for adaptation and customization. To enhance the efficiency and accuracy of the model’s reasoning process, we constructed an agent-based framework for inference centered on task decomposition, planning, and evaluation-reflection. This framework connects multiple reasoning steps into a coherent chain, allowing the model to systematically analyze and solve complex layout problems in a phased, structured manner. This method not only mitigates information overload but also significantly improves the model's accuracy and efficiency in handling such complex tasks. Furthermore, the system establishes clear standards and decision rules to ensure the reliability of information processing and the validity of decisions throughout the multi-step reasoning iterations, guaranteeing the generation of effective inference sequences. Finally, a case study on the block storage yard of a large shipbuilding enterprise was conducted to validate the method's ability to reconfigure layouts in response to sudden docking schedule changes. The results show that the proposed method significantly improves the layout's responsiveness to dynamic demands and its constraint satisfaction rate while maintaining spatial utilization, offering a novel approach for the intelligent and dynamic adjustment of shipyard layouts.

Key words: LLM agent, digital twin, layout design, retrieval-augmented generation (RAG), shipbuilding

中图分类号:

U169.65

仝小冬, 司佳辉, 袁轶, 杨梦凯, 严哲, 鲍劲松. 基于大语言模型智能体的船舶堆场数字孪生动态布局生成方法[J]. 机械工程学报, 2026, 62(15): 126-141.

TONG Xiaodong, SI Jiahui, YUAN Yi, YANG Mengkai, YAN Zhe, BAO Jinsong. A Dynamic Layout Generation Method for Shipyard Digital Twin Based on LLM Agents[J]. Journal of Mechanical Engineering, 2026, 62(15): 126-141.

导出引用管理器 EndNote|Reference Manager|ProCite|BibTeX|RefWorks

链接本文: http://www.cjmenet.com.cn/CN/10.3901/JME.260625

http://www.cjmenet.com.cn/CN/Y2026/V62/I15/126

参考文献

[1] 张倩，单忠德，吴乐，等. 船舶分段多工位协同制造系统数学模型研究及应用[J]. 机械工程学报，2020，56(9)：191-198. ZHANG Qian，SHAN Zhongde，WU Le，et al. Research and application of mathematical model of ship sectional multi-station cooperative manufacturing system[J]. Journal of Mechanical Engineering，2020，56(9)：191-198.
[2] LIU Z，LIU H，LIU H，et al. Many-objective flexible job-shop scheduling based on a loose non-dominated sorting genetic algorithm[J]. Engineering Optimization，2024，57(6)，1423-1444.
[3] QIU Chan，TAN Jianrong，LIU Zhenyu，et al. Design theory and method of complex products：A review[J]. Chinese Journal of Mechanical Engineering，2022，35(1)：1-16.
[4] PEKDEMIR B M，ASILISKENDER B. Evolution of production spaces：A historical review for projecting smart factories[J]. Iconarp International J of Architecture and Planning，2023，11：716-33.
[5] 王无印，黄子钊，庄子龙，等. 基于深度强化学习的自动化码头堆场场桥调度方法[J]. 机械工程学报，2024，60(6)：44-57. WANG Wuyin，HUANG Zizhao，ZHUANG Zilong，at al. Yard crane scheduling method based on deep reinforcement learning for the automated container terminal[J]. Journal of Mechanical Engineering，2024，60(6)：44-57.
[6] OUYANG L，WU J，JIANG X，et al. Training language models to follow instructions with human feedback[C]// In Proceedings of the 36th International Conference on Neural Information Processing Systems (NIPS '22). NY，USA，2011，27730-27744.
[7] 宋学官，李昆鹏，胡正国，等. 高性能制造：设计与制造的协同机制[J]. 机械工程学报，2024，60(1)：2-12. SONG Xueguan，LI Kunpeng，HU Zhengguo，et al. High performance manufacturing：Collaborative mechanisms for design and manufacturing[J]. Journal of Mechanical Engineering，2024，60(1)：2-12.
[8] 张栋豪，刘振宇，郏维强，等. 知识图谱在智能制造领域的研究现状及其应用前景综述[J]. 机械工程学报，2021，57(5)：90-113. ZHANG Donghao，LIU Zhenyu，JIA Weiqiang，et al. A review on knowledge graph and its application prospects to intelligent manufacturing[J]. Journal of Mechanical Engineering，2021，57(5)：90-113.
[9] DU X，NOVOA J，PLASEK J M，et al. Enhancing early detection of cognitive decline in the elderly：A comparative study utilizing large language models in clinical notes[J]. eBioMedicine，2024，109：105401.
[10] 郏维强，刘振宇，刘达新，等. 基于模糊关联的复杂产品模块化设计方法及其应用[J]. 机械工程学报，2015，51(5)：130-142. JIA Weiqiang，LIU Zhenyu，LIU Daxin，et al. Modular design method and application for complex product based on fuzzy correlation analysis[J]. Journal of Mechanical Engineering，2015，51(5)：130-142.
[11] 程锦，叶虎强，谭建荣，等. 三维CAD技术研究进展及其发展趋势综述[J]. 机械工程学报，2023，59(23)：158-185. CHENG Jin，YE Huqiang，TAN Jianrong，et al. Review of research progress and development trends of 3D CAD technology[J]. Journal of Mechanical Engineering，2023，59(23)：158-185.
[12] 颜硕，侯泽普，郭春东. 车间布局优化对生产效率和生产质量的影响[J]. 化纤与纺织技术，2023，52(9)：94-6. YAN Shuo，HOU Zepu，GUO Chundong. The impact of workshop layout optimization on production efficiency and quality[J]. Chemical Fiber and Textile Technology，2023，52(9)：94-96.
[13] Klar M，Ruediger P，Schuermann M，et al. Explainable generative design in manufacturing for reinforcement learning based factory layout planning[J]. Journal of Manufacturing Systems，2024，72：74-92.
[14] LIND A，HÖGBERG D，SYBERFELDT A，et al. Evaluating a digital twin concept for an automatic up-to-date factory layout setup[M]. Amsterdam：IOS Press，2022.
[15] 刘惠，刘振宇，郏维强，等. 深度学习在装备剩余使用寿命预测技术中的研究现状与挑战[J]. 计算集成制造系统，2021，27(1)：34-52. LIU Hui，LIU Zhenyu，JIA Weiqiang，et al. Research status and challenges of deep learning in equipment remaining service life prediction technology[J]. Computational Integrated Manufacturing Systems，2021，27(1)：34-52.
[16] KLAR M，SCHWORM P，WU X，et al. Quantum Annealing based factory layout planning[J]. Manufacturing Letters，2022，32：59-62.
[17] ZHU L，WANG R，ZHAO Z. Factory layout optimization based on systematic layout planning and genetic algorithm[C]//International Conference on Algorithms, Microchips and Network Applications. SPIE，2022：121760.
[18] ZHAO D，YANG J，ZHOU H. Layout design of warehouse based on systematic layout planning and GA-ACO algorithm[C]//2020 Chinese Automation Congress(CAC). Shanghai，China：IEEE，2020.
[19] 彭建华，王岳，吕峰，等. 基于SLP和遗传算法的船舶管加工车间设备布局设计[J]. 船舶工程，2022，44(10)：15-21. PENG Jianhua，WANG Yue，LÜ Feng，et al. Equipment layout design for ship pipe processing workshop based on SLP and genetic algorithm[J]. Ship Engineering，2022，44(10)：15-21.
[20] KLAR M，GLATT M，AURICH J C. An implementation of a reinforcement learning based algorithm for factory layout planning[J]. Manufacturing Letters，2021，30：1-4.
[21] KLAR M，GLATT M，AURICH J C. Performance comparison of reinforcement learning and metaheuristics for factory layout planning[J]. CIRP Journal of Manufacturing Science and Technology，2023，45：10-25.
[22] BROWN T，MANN B，RYDER N，et al. Language models are few-shot learners[J]. Advances in neural information processing systems，2020，33：1877-901.
[23] 刘达新，王科，刘振宇，等. 基于数据融合与知识推理的机器人装配单元数字孪生建模方法研究[J]. 机械工程学报，2024，60(5)：36-50. LIU Daxin，WANG Ke，LIU Zhenyu，et al. Research on digital twin modeling method for robotic assembly cell based on data fusion and knowledge reasoning[J]. Journal of Mechanical Engineering，2024，60(5)：36-50.
[24] HONG Y，Zhen H，Chen P，et al. 3D-LLM：Injecting the 3D world into large language models[J]. Advances in Neural Information Processing Systems，2023，36：20482-20494.
[25] XIA X，ZHANG D，LIAO Z，et al. SceneGenAgent：Precise industrial scene generation with coding agent[C]//Proceedings of the 63rd Annual Meeting of the Association for Computational Linguistics (Volume 1：Long Papers). 2025：17847-17875.
[26] DISSELKAMP J P，KURPICK D，SCHUTTE B，et al. Use cases of generative AI in factory planning：potential and challenges[C]//NordDesign 2024. Reykjavik，Iceland，12-14 Aug 2024. DS 130，2024：196-205.
[27] WERHEID J，MELNYCHUK O，ZHOU H，et al. Designing an LLM-based copilot for manufacturing equipment selection[J]. Manufacturing Letters，2025，46：123-127.
[28] HUANG H，WANG Z，HUANG R，et al. Chat-3D v2：Bridging 3D scene and large language models with object identifiers[J]. arXiv:2312.08168，2023.
[29] FAN H，LIU X，FUH J Y H，et al. Embodied intelligence in manufacturing：Leveraging large language models for autonomous industrial robotics[J]. Journal of Intelligent Manufacturing，2024，36(2)：1141-1157.
[30] FENG W，ZHU W，FU T-J，et al. LayoutGPT：Compositional visual planning and generation with large language models[C]//In Proceedings of the 37th International Conference on Neural Information Processing Systems (NIPS '23). NY，USA，2023，802，18225-18250.