Journal of Manufacturing Systems最新文献_第3页

Design of laser-assisted processing digital twin system 激光辅助加工数字孪生系统的设计

IF 14.2 1区工程技术 Q1 ENGINEERING, INDUSTRIAL

Journal of Manufacturing Systems

Pub Date : 2026-01-13 DOI: 10.1016/j.jmsy.2026.01.001

Xuefeng Wu , Tiankuo Yu , Xianli Liu , Caixu Yue , Xulu Jiang , Jiayu Li , Zhonghua Li

The machining of high-performance, difficult-to-cut materials poses a critical challenge in advanced manufacturing. While laser-assisted machining (LAM) has emerged as a viable solution, its effectiveness is often compromised in practice by insufficient synchronization between laser heating and cutting operations, leading to processing defects. To address this limitation, this study develops an intelligent collaborative LAM system based on digital twin technology. By integrating Unity3D with deep learning techniques, a systematic architecture suitable for laser-thermal-assisted machining is constructed. A dynamic multi-physics field coupling model is established to achieve real-time control of laser incidence posture along with simultaneous monitoring and prediction of the temperature distribution in the machining region. This integrated system exhibits enhanced laser positioning agility, reduced thermal fluctuations, improved laser energy utilization efficiency, and consistent processing quality. Experimental validation conducted on forged superalloy Inconel 718 and SiCp/Al composites demonstrates remarkable improvements in both surface integrity and dimensional accuracy. Moreover, machine learning-based reliability assessment reliability assessment confirms only minor deviations in experimental outcomes, thereby providing a robust intelligent process assurance mechanism for machining difficult-to-process components such as aero-engine blades.

高性能、难切削材料的加工对先进制造提出了严峻的挑战。虽然激光辅助加工（LAM）已经成为一种可行的解决方案，但在实践中，由于激光加热和切割操作之间的不同步，其有效性经常受到影响，导致加工缺陷。为了解决这一限制，本研究开发了一个基于数字孪生技术的智能协同LAM系统。将Unity3D与深度学习技术相结合，构建了适合激光热辅助加工的系统架构。建立了动态多物理场耦合模型，实现了激光入射姿态的实时控制，同时实现了加工区域温度分布的监测和预测。该集成系统具有增强的激光定位灵活性，减少热波动，提高激光能量利用效率和一致的加工质量。对锻造高温合金Inconel 718和SiCp/Al复合材料进行的实验验证表明，表面完整性和尺寸精度都有显著提高。此外，基于机器学习的可靠性评估可靠性评估仅确认实验结果中的微小偏差，从而为航空发动机叶片等难加工部件的加工提供强大的智能过程保证机制。

{"title":"Design of laser-assisted processing digital twin system","authors":"Xuefeng Wu , Tiankuo Yu , Xianli Liu , Caixu Yue , Xulu Jiang , Jiayu Li , Zhonghua Li","doi":"10.1016/j.jmsy.2026.01.001","DOIUrl":"10.1016/j.jmsy.2026.01.001","url":null,"abstract":"<div><div>The machining of high-performance, difficult-to-cut materials poses a critical challenge in advanced manufacturing. While laser-assisted machining (LAM) has emerged as a viable solution, its effectiveness is often compromised in practice by insufficient synchronization between laser heating and cutting operations, leading to processing defects. To address this limitation, this study develops an intelligent collaborative LAM system based on digital twin technology. By integrating Unity3D with deep learning techniques, a systematic architecture suitable for laser-thermal-assisted machining is constructed. A dynamic multi-physics field coupling model is established to achieve real-time control of laser incidence posture along with simultaneous monitoring and prediction of the temperature distribution in the machining region. This integrated system exhibits enhanced laser positioning agility, reduced thermal fluctuations, improved laser energy utilization efficiency, and consistent processing quality. Experimental validation conducted on forged superalloy Inconel 718 and SiCp/Al composites demonstrates remarkable improvements in both surface integrity and dimensional accuracy. Moreover, machine learning-based reliability assessment reliability assessment confirms only minor deviations in experimental outcomes, thereby providing a robust intelligent process assurance mechanism for machining difficult-to-process components such as aero-engine blades.</div></div>","PeriodicalId":16227,"journal":{"name":"Journal of Manufacturing Systems","volume":"85 ","pages":"Pages 155-174"},"PeriodicalIF":14.2,"publicationDate":"2026-01-13","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"145978548","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":1,"RegionCategory":"工程技术","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}

引用次数: 0

A literature review on deep reinforcement learning for machine scheduling problems 机器调度问题中深度强化学习的文献综述

IF 14.2 1区工程技术 Q1 ENGINEERING, INDUSTRIAL

Journal of Manufacturing Systems

Pub Date : 2026-01-12 DOI: 10.1016/j.jmsy.2025.12.017

Constantin Waubert de Puiseau , Furkan Ercan , Jannik Peters , Marvin Brune , Hasan Tercan , Christopher Prinz , Tobias Meisen , Bernd Kuhlenkötter

Machine scheduling represents a core challenge in industrial production systems due to its inherently complex combinatorial nature and its critical role in enhancing operational efficiency. With recent advances in artificial intelligence, deep reinforcement learning (DRL) has gained increasing attention as an innovative tool to address scheduling tasks with a self-adapting, data-driven approach. This survey presents a comprehensive review of 143 publications between 2018 and February 2025 that apply DRL to machine scheduling problems. We develop a structured framework to classify and compare problem settings, algorithmic designs, and evaluation methodologies. Key aspects such as action and observation space design, reward functions, neural network architectures, and experimental benchmarks are systematically analyzed. The review identifies current trends, outlines promising patterns, and highlights open research opportunities for DRL-based scheduling solutions. The goal of this survey is to make the rapidly evolving research landscape more accessible to both academics and practitioners and to identify the next steps in research and application. To facilitate reproducible research and customized analysis, we publish the dataset underpinning this review, which includes 61 annotated features per publication, allowing for customizable filtering and further in-depth exploration of niches within the field. This dataset is publicly accessible online.

机器调度是工业生产系统的核心挑战，因为它本身具有复杂的组合性质，在提高运行效率方面具有关键作用。随着人工智能的最新进展，深度强化学习（DRL）作为一种创新工具，通过自适应、数据驱动的方法来解决调度任务，受到越来越多的关注。本调查对2018年至2025年2月期间将DRL应用于机器调度问题的143份出版物进行了全面审查。我们开发了一个结构化的框架来分类和比较问题设置、算法设计和评估方法。系统分析了行动和观察空间设计、奖励函数、神经网络架构和实验基准等关键方面。该综述确定了当前的趋势，概述了有希望的模式，并强调了基于drl的调度解决方案的开放研究机会。这项调查的目的是使快速发展的研究领域更容易为学者和从业者所了解，并确定研究和应用的下一步。为了促进可重复的研究和定制分析，我们发布了支持本综述的数据集，其中包括每篇出版物的61个注释功能，允许自定义过滤和进一步深入探索该领域的利基。此数据集可在线公开访问。

{"title":"A literature review on deep reinforcement learning for machine scheduling problems","authors":"Constantin Waubert de Puiseau , Furkan Ercan , Jannik Peters , Marvin Brune , Hasan Tercan , Christopher Prinz , Tobias Meisen , Bernd Kuhlenkötter","doi":"10.1016/j.jmsy.2025.12.017","DOIUrl":"10.1016/j.jmsy.2025.12.017","url":null,"abstract":"<div><div>Machine scheduling represents a core challenge in industrial production systems due to its inherently complex combinatorial nature and its critical role in enhancing operational efficiency. With recent advances in artificial intelligence, deep reinforcement learning (DRL) has gained increasing attention as an innovative tool to address scheduling tasks with a self-adapting, data-driven approach. This survey presents a comprehensive review of 143 publications between 2018 and February 2025 that apply DRL to machine scheduling problems. We develop a structured framework to classify and compare problem settings, algorithmic designs, and evaluation methodologies. Key aspects such as action and observation space design, reward functions, neural network architectures, and experimental benchmarks are systematically analyzed. The review identifies current trends, outlines promising patterns, and highlights open research opportunities for DRL-based scheduling solutions. The goal of this survey is to make the rapidly evolving research landscape more accessible to both academics and practitioners and to identify the next steps in research and application. To facilitate reproducible research and customized analysis, we publish the dataset underpinning this review, which includes 61 annotated features per publication, allowing for customizable filtering and further in-depth exploration of niches within the field. This dataset is publicly accessible <span><span>online</span><svg><path></path></svg></span>.</div></div>","PeriodicalId":16227,"journal":{"name":"Journal of Manufacturing Systems","volume":"85 ","pages":"Pages 96-126"},"PeriodicalIF":14.2,"publicationDate":"2026-01-12","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"145979110","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":1,"RegionCategory":"工程技术","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}

引用次数: 0

Physics-informed machine learning across manufacturing processes: Recent advances, challenges, and directions 跨制造过程的物理信息机器学习：最新进展，挑战和方向

IF 14.2 1区工程技术 Q1 ENGINEERING, INDUSTRIAL

Journal of Manufacturing Systems

Pub Date : 2026-01-12 DOI: 10.1016/j.jmsy.2026.01.002

Donghyun Ra , Jaeryun Lee , Minwoo Lee , Seongmin Kwak , Seungchul Lee , Sooyoung Lee

Data-driven methods have shown remarkable promise across manufacturing disciplines. However, their application to manufacturing processes, where physical behaviors dominate, faces several limitations including the need for large datasets, the risk of physically implausible results, and limited interpretability. Physics-informed machine learning (PIML), which incorporates governing physical laws into deep neural networks, has emerged as a promising direction for overcoming these limitations by enabling physically consistent and data-efficient modeling. In this article, we systematically analyze PIML applications from a manufacturing-process perspective. We organize the manufacturing landscape into five representative categories based on their dominant physical mechanisms and the availability of relevant literature: mechanical processes, chemical processes, thermal-driven processes, additive manufacturing, and semiconductor fabrication. We comprehensively review and discuss domain-dependent characteristics, including governing physics, data regimes, and modeling objectives across manufacturing processes. We further discuss fundamental challenges of PIML approaches, along with practical considerations for real-world deployment. This review aims to provide a foundation for accelerating physics-informed learning and its advancement within manufacturing research and practice.

数据驱动的方法在制造学科中显示出显著的前景。然而，在物理行为占主导地位的制造过程中，它们的应用面临着一些限制，包括对大型数据集的需求、物理上不可信结果的风险以及有限的可解释性。基于物理的机器学习（PIML）将控制物理定律整合到深度神经网络中，通过实现物理一致性和数据高效建模，已经成为克服这些限制的一个有希望的方向。本文从制造过程的角度系统地分析了PIML的应用。我们根据其主要的物理机制和相关文献的可用性将制造业景观分为五个代表性类别：机械过程，化学过程，热驱动过程，增材制造和半导体制造。我们全面地回顾和讨论领域相关的特征，包括管理物理、数据制度和跨制造过程的建模目标。我们进一步讨论了PIML方法的基本挑战，以及实际部署的实际考虑。本综述旨在为加速物理知识学习及其在制造研究和实践中的进步提供基础。

{"title":"Physics-informed machine learning across manufacturing processes: Recent advances, challenges, and directions","authors":"Donghyun Ra , Jaeryun Lee , Minwoo Lee , Seongmin Kwak , Seungchul Lee , Sooyoung Lee","doi":"10.1016/j.jmsy.2026.01.002","DOIUrl":"10.1016/j.jmsy.2026.01.002","url":null,"abstract":"<div><div>Data-driven methods have shown remarkable promise across manufacturing disciplines. However, their application to manufacturing processes, where physical behaviors dominate, faces several limitations including the need for large datasets, the risk of physically implausible results, and limited interpretability. Physics-informed machine learning (PIML), which incorporates governing physical laws into deep neural networks, has emerged as a promising direction for overcoming these limitations by enabling physically consistent and data-efficient modeling. In this article, we systematically analyze PIML applications from a manufacturing-process perspective. We organize the manufacturing landscape into five representative categories based on their dominant physical mechanisms and the availability of relevant literature: mechanical processes, chemical processes, thermal-driven processes, additive manufacturing, and semiconductor fabrication. We comprehensively review and discuss domain-dependent characteristics, including governing physics, data regimes, and modeling objectives across manufacturing processes. We further discuss fundamental challenges of PIML approaches, along with practical considerations for real-world deployment. This review aims to provide a foundation for accelerating physics-informed learning and its advancement within manufacturing research and practice.</div></div>","PeriodicalId":16227,"journal":{"name":"Journal of Manufacturing Systems","volume":"85 ","pages":"Pages 72-95"},"PeriodicalIF":14.2,"publicationDate":"2026-01-12","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"145979111","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":1,"RegionCategory":"工程技术","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}

引用次数: 0

Hybrid digital twins for smart manufacturing: Architectures, fusion paradigm, and implementation challenges 智能制造的混合数字孪生：架构、融合范式和实施挑战

IF 14.2 1区工程技术 Q1 ENGINEERING, INDUSTRIAL

Journal of Manufacturing Systems

Pub Date : 2026-01-09 DOI: 10.1016/j.jmsy.2025.12.029

Xi Zhang , Yiqun Kou , Xin Zhang , Qi Shi , Youmin Hu , Huapeng Wu , Shimin Liu , Pai Zheng

As a high-fidelity representation of physical objects, the digital twin (DT) emerges as a crucial enabling tool supporting intelligent monitoring, prediction, and decision-making for smart manufacturing. To achieve reliable, accurate, and explainable DT modeling under dynamic conditions, it is necessary to integrate multiple models, including first-principles knowledge, data-driven algorithms, and simulation. Furthermore, with the emergence of state-of-the-art artificial intelligence (AI) technologies, such as Generative AI and Large Language Models, new drivers for DT modeling can be provided. However, the specific paradigm for hybridizing these models varies significantly depending on the application scenario, the object, and the critical requirements. This diversity poses a significant challenge for systematically selecting and combining modeling techniques in smart manufacturing. This review addresses this gap by providing a systematic exploration of the Hybrid Digital Twin (HDT) modeling paradigm, which focuses on the integration of multiple heterogeneous models. Therefore, this paper aims to: (1) clarify the architecture and core characteristics of HDT; (2) categorize critical technologies and fusion paradigms for HDT implementation; and (3) outline potential future research directions. It is hoped that this paper will serve as a systematic reference for researchers and engineers seeking to apply HDT to build more accurate, reliable, and adaptive DT applications.

作为物理对象的高保真表示，数字孪生（DT）成为支持智能制造智能监控、预测和决策的关键支持工具。为了在动态条件下实现可靠、准确和可解释的DT建模，需要集成多个模型，包括第一性原理知识、数据驱动算法和仿真。此外，随着最先进的人工智能（AI）技术的出现，如生成式AI和大型语言模型，可以为DT建模提供新的驱动因素。然而，混合这些模型的具体范例根据应用程序场景、对象和关键需求而有很大的不同。这种多样性对智能制造中建模技术的系统选择和组合提出了重大挑战。本文通过对混合数字孪生（HDT）建模范式的系统探索来解决这一差距，该范式侧重于多个异构模型的集成。因此，本文旨在：(1)阐明HDT的体系结构和核心特征；(2)对HDT实施的关键技术和融合范式进行分类；(3)概述了未来可能的研究方向。希望本文能够为寻求应用HDT构建更准确、可靠和自适应的DT应用的研究人员和工程师提供系统的参考。

{"title":"Hybrid digital twins for smart manufacturing: Architectures, fusion paradigm, and implementation challenges","authors":"Xi Zhang , Yiqun Kou , Xin Zhang , Qi Shi , Youmin Hu , Huapeng Wu , Shimin Liu , Pai Zheng","doi":"10.1016/j.jmsy.2025.12.029","DOIUrl":"10.1016/j.jmsy.2025.12.029","url":null,"abstract":"<div><div>As a high-fidelity representation of physical objects, the digital twin (DT) emerges as a crucial enabling tool supporting intelligent monitoring, prediction, and decision-making for smart manufacturing. To achieve reliable, accurate, and explainable DT modeling under dynamic conditions, it is necessary to integrate multiple models, including first-principles knowledge, data-driven algorithms, and simulation. Furthermore, with the emergence of state-of-the-art artificial intelligence (AI) technologies, such as Generative AI and Large Language Models, new drivers for DT modeling can be provided. However, the specific paradigm for hybridizing these models varies significantly depending on the application scenario, the object, and the critical requirements. This diversity poses a significant challenge for systematically selecting and combining modeling techniques in smart manufacturing. This review addresses this gap by providing a systematic exploration of the Hybrid Digital Twin (HDT) modeling paradigm, which focuses on the integration of multiple heterogeneous models. Therefore, this paper aims to: (1) clarify the architecture and core characteristics of HDT; (2) categorize critical technologies and fusion paradigms for HDT implementation; and (3) outline potential future research directions. It is hoped that this paper will serve as a systematic reference for researchers and engineers seeking to apply HDT to build more accurate, reliable, and adaptive DT applications.</div></div>","PeriodicalId":16227,"journal":{"name":"Journal of Manufacturing Systems","volume":"85 ","pages":"Pages 51-71"},"PeriodicalIF":14.2,"publicationDate":"2026-01-09","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"145928459","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":1,"RegionCategory":"工程技术","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}

引用次数: 0

An inspection path planning approach based on hierarchical reinforcement learning 一种基于分层强化学习的检测路径规划方法

IF 14.2 1区工程技术 Q1 ENGINEERING, INDUSTRIAL

Journal of Manufacturing Systems

Pub Date : 2026-01-08 DOI: 10.1016/j.jmsy.2025.12.027

Kuanhong Yan , Daxin Liu , Zhenyu Liu , Jianrong Tan

As manufacturing scales up and demand for customization grows, computer-aided inspection technologies face increasing pressure to enhance efficiency and capability. Inspection path planning (IPP) for touch-trigger probes on coordinate measuring machines (CMMs) still suffers from heavy reliance on manual rules and poor coordination among planning stages. To address these limitations, this paper proposes a novel approach for IPP based on a hybrid planning framework, called Inspection Path Optimization via Deep Reinforcement Learning & Heuristic Search (InPO-DRLHS). InPO-DRLHS leverages hierarchical reinforcement learning (HRL) to generate collision-free paths (CFPs) while accounting for probe orientation changes, and integrates this architecture with heuristic-based optimization of measurement sequence to improve global planning efficiency. Specifically, within the HRL module: the lower-level agent employs an improved deep Q-learning approach to generate optimal CFPs under fixed probe orientations; the higher-level agent predicts optimal probe rotation positions from a voxelized scene representation, guiding orientation adjustments for the lower-level planning. Furthermore, the value function learned by the HRL module serves as a surrogate heuristic estimate within the heuristic algorithm during sequence planning, enabling path quality assessment without explicit CFP generation. Experimental results show that the proposed method achieves over 90% success rate in 3D CFP planning and reduces total measurement time by 21.7% compared to conventional approaches, indicating substantial gains in inspection efficiency.

随着制造业规模的扩大和定制需求的增长，计算机辅助检测技术在提高效率和能力方面面临着越来越大的压力。三坐标测量机上触摸触发探头的检测路径规划仍然严重依赖人工规则，规划阶段之间的协调性差。为了解决这些限制，本文提出了一种基于混合规划框架的IPP新方法，称为通过深度强化学习和启发式搜索的检查路径优化（InPO-DRLHS）。InPO-DRLHS利用分层强化学习（HRL）在考虑探针方向变化的情况下生成无碰撞路径（CFPs），并将该架构与基于启发式的测量序列优化相结合，以提高全局规划效率。具体而言，在HRL模块中：底层智能体采用改进的深度q -学习方法在固定探针方向下生成最优cfp；高级代理从体素化场景表示中预测最佳探针旋转位置，指导低级规划的方向调整。此外，HRL模块学习的值函数在序列规划过程中充当启发式算法中的代理启发式估计，使路径质量评估无需显式生成CFP。实验结果表明，该方法在三维CFP规划中实现了90%以上的成功率，与传统方法相比，总测量时间缩短了21.7%，检测效率大幅提高。

{"title":"An inspection path planning approach based on hierarchical reinforcement learning","authors":"Kuanhong Yan , Daxin Liu , Zhenyu Liu , Jianrong Tan","doi":"10.1016/j.jmsy.2025.12.027","DOIUrl":"10.1016/j.jmsy.2025.12.027","url":null,"abstract":"<div><div>As manufacturing scales up and demand for customization grows, computer-aided inspection technologies face increasing pressure to enhance efficiency and capability. Inspection path planning (IPP) for touch-trigger probes on coordinate measuring machines (CMMs) still suffers from heavy reliance on manual rules and poor coordination among planning stages. To address these limitations, this paper proposes a novel approach for IPP based on a hybrid planning framework, called <strong>Inspection Path Optimization via Deep Reinforcement Learning & Heuristic Search (InPO-DRLHS)</strong>. InPO-DRLHS leverages hierarchical reinforcement learning (HRL) to generate collision-free paths (CFPs) while accounting for probe orientation changes, and integrates this architecture with heuristic-based optimization of measurement sequence to improve global planning efficiency. Specifically, within the HRL module: the lower-level agent employs an improved deep Q-learning approach to generate optimal CFPs under fixed probe orientations; the higher-level agent predicts optimal probe rotation positions from a voxelized scene representation, guiding orientation adjustments for the lower-level planning. Furthermore, the value function learned by the HRL module serves as a surrogate heuristic estimate within the heuristic algorithm during sequence planning, enabling path quality assessment without explicit CFP generation. Experimental results show that the proposed method achieves over 90% success rate in 3D CFP planning and reduces total measurement time by 21.7% compared to conventional approaches, indicating substantial gains in inspection efficiency.</div></div>","PeriodicalId":16227,"journal":{"name":"Journal of Manufacturing Systems","volume":"85 ","pages":"Pages 34-50"},"PeriodicalIF":14.2,"publicationDate":"2026-01-08","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"145928458","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":1,"RegionCategory":"工程技术","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}

引用次数: 0

A collaborative process parameter recommender system for fleets of networked manufacturing machines — with application to 3D printing 一个协作过程参数推荐系统的车队网络化制造机器-应用于3D打印

IF 14.2 1区工程技术 Q1 ENGINEERING, INDUSTRIAL

Journal of Manufacturing Systems

Pub Date : 2026-01-02 DOI: 10.1016/j.jmsy.2025.12.028

Sicong Guo , Weishi Wang , Chenhuan Jiang , Mohamed Elidrisi , Myungjin Lee , Harsha V. Madhyastha , Raed Al Kontar , Chinedum E. Okwudire

Fleets of networked manufacturing machines of the same type, that are collocated or geographically distributed, are growing in popularity. An excellent example is the rise of 3D print farms, which consist of multiple networked 3D printers operating in parallel, enabling faster production and efficient mass customization. However, optimizing process parameters across a fleet of manufacturing machines, even of the same type, remains a challenge due to machine-to-machine variability. Traditional trial-and-error approaches are inefficient, requiring extensive testing to determine optimal process parameters for an entire fleet. In this work, we introduce a machine learning-based collaborative recommender system that optimizes process parameters for each machine in a fleet by modeling the problem as a sequential matrix completion task. Our approach leverages spectral clustering and alternating least squares to iteratively refine parameter predictions, enabling real-time collaboration among the machines in a fleet while minimizing the number of experimental trials. We validate our method using a mini 3D print farm consisting of ten 3D printers for which we optimize acceleration and speed settings to minimize surface roughness and printing time, thus maximizing print quality and productivity. Our approach achieves significantly faster convergence to optimal process parameters relative to a comparable non-collaborative technique.

同一类型的联网制造机器的机群，无论是在同一地点还是在不同地理位置，都越来越受欢迎。3D打印农场的兴起就是一个很好的例子，它由多台并行运行的联网3D打印机组成，从而实现了更快的生产和高效的大规模定制。然而，由于机器对机器的可变性，在一组制造机器（即使是同一类型）之间优化工艺参数仍然是一个挑战。传统的试错法效率低下，需要大量的测试来确定整个车队的最佳工艺参数。在这项工作中，我们引入了一个基于机器学习的协作推荐系统，该系统通过将问题建模为顺序矩阵完成任务来优化车队中每台机器的过程参数。我们的方法利用光谱聚类和交替最小二乘来迭代地改进参数预测，实现车队中机器之间的实时协作，同时最大限度地减少实验试验的数量。我们使用一个由10台3D打印机组成的迷你3D打印农场来验证我们的方法，我们优化了加速和速度设置，以最大限度地减少表面粗糙度和打印时间，从而最大限度地提高了打印质量和生产率。我们的方法实现了显著更快的收敛到最优工艺参数相对于一个可比的非协作技术。

{"title":"A collaborative process parameter recommender system for fleets of networked manufacturing machines — with application to 3D printing","authors":"Sicong Guo , Weishi Wang , Chenhuan Jiang , Mohamed Elidrisi , Myungjin Lee , Harsha V. Madhyastha , Raed Al Kontar , Chinedum E. Okwudire","doi":"10.1016/j.jmsy.2025.12.028","DOIUrl":"10.1016/j.jmsy.2025.12.028","url":null,"abstract":"<div><div>Fleets of networked manufacturing machines of the same type, that are collocated or geographically distributed, are growing in popularity. An excellent example is the rise of 3D print farms, which consist of multiple networked 3D printers operating in parallel, enabling faster production and efficient mass customization. However, optimizing process parameters across a fleet of manufacturing machines, even of the same type, remains a challenge due to machine-to-machine variability. Traditional trial-and-error approaches are inefficient, requiring extensive testing to determine optimal process parameters for an entire fleet. In this work, we introduce a machine learning-based collaborative recommender system that optimizes process parameters for each machine in a fleet by modeling the problem as a sequential matrix completion task. Our approach leverages spectral clustering and alternating least squares to iteratively refine parameter predictions, enabling real-time collaboration among the machines in a fleet while minimizing the number of experimental trials. We validate our method using a mini 3D print farm consisting of ten 3D printers for which we optimize acceleration and speed settings to minimize surface roughness and printing time, thus maximizing print quality and productivity. Our approach achieves significantly faster convergence to optimal process parameters relative to a comparable non-collaborative technique.</div></div>","PeriodicalId":16227,"journal":{"name":"Journal of Manufacturing Systems","volume":"85 ","pages":"Pages 22-33"},"PeriodicalIF":14.2,"publicationDate":"2026-01-02","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"145882953","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":1,"RegionCategory":"工程技术","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}

引用次数: 0

Predicting transfer times across production lines using data pooling 使用数据池预测跨生产线的传输时间

IF 14.2 1区工程技术 Q1 ENGINEERING, INDUSTRIAL

Journal of Manufacturing Systems

Pub Date : 2026-01-02 DOI: 10.1016/j.jmsy.2025.12.025

Seohyun Choi , Young Ha Joo , Hoonseok Park , Younkook Kang , Ri Choe , Jae-Yoon Jung

Reliable transfer time prediction is crucial for productivity in automated material handling systems within modern manufacturing environments. However, the complexity and dynamic behavior of manufacturing and logistics systems make accurate transfer time estimation highly challenging. This study proposes a transfer time prediction approach for automated material handling systems operating across production lines. To forecast inter-building and inter-floor transfer times, this study proposes a hybrid method, called time-series residual regression, that integrates linear time-series analysis with nonlinear machine learning. The framework further employs three data pooling strategies to effectively capture device heterogeneity and improve forecasting robustness. The hybrid method was validated using transfer records from inter-building stockers and inter-floor lifters in a Korean semiconductor fab. The experimental results show that the proposed model delivers superior performance, achieving R-squared values of 64.01 % for inter-building transfers and 72.00 % for inter-floor transfers.

在现代制造环境中，可靠的传递时间预测对于自动化物料处理系统的生产率至关重要。然而，制造和物流系统的复杂性和动态行为使得准确的传递时间估计极具挑战性。本研究提出一种跨生产线自动化物料搬运系统的转移时间预测方法。为了预测建筑物间和楼层间的转移时间，本研究提出了一种称为时间序列残差回归的混合方法，该方法将线性时间序列分析与非线性机器学习相结合。该框架进一步采用三种数据池策略来有效捕获设备异质性并提高预测的鲁棒性。使用韩国半导体工厂的楼间仓库和楼间升降机的转移记录验证了混合方法。实验结果表明，所提出的模型具有较好的性能，楼间传输的r平方值为64.01 %，楼间传输的r平方值为72.00 %。

{"title":"Predicting transfer times across production lines using data pooling","authors":"Seohyun Choi , Young Ha Joo , Hoonseok Park , Younkook Kang , Ri Choe , Jae-Yoon Jung","doi":"10.1016/j.jmsy.2025.12.025","DOIUrl":"10.1016/j.jmsy.2025.12.025","url":null,"abstract":"<div><div>Reliable transfer time prediction is crucial for productivity in automated material handling systems within modern manufacturing environments. However, the complexity and dynamic behavior of manufacturing and logistics systems make accurate transfer time estimation highly challenging. This study proposes a transfer time prediction approach for automated material handling systems operating across production lines. To forecast inter-building and inter-floor transfer times, this study proposes a hybrid method, called time-series residual regression, that integrates linear time-series analysis with nonlinear machine learning. The framework further employs three data pooling strategies to effectively capture device heterogeneity and improve forecasting robustness. The hybrid method was validated using transfer records from inter-building stockers and inter-floor lifters in a Korean semiconductor fab. The experimental results show that the proposed model delivers superior performance, achieving R-squared values of 64.01 % for inter-building transfers and 72.00 % for inter-floor transfers.</div></div>","PeriodicalId":16227,"journal":{"name":"Journal of Manufacturing Systems","volume":"84 ","pages":"Pages 629-645"},"PeriodicalIF":14.2,"publicationDate":"2026-01-02","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"145880098","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":1,"RegionCategory":"工程技术","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}

引用次数: 0

A hybrid multiobjective particle swarm optimization with Q-learning-driven local search for distributed heterogeneous hybrid flow-shop scheduling problem with worker–machine–environment collaboration 基于q -学习驱动局部搜索的混合多目标粒子群算法求解工人-机器-环境协同的分布式异构混合流水车间调度问题

IF 14.2 1区工程技术 Q1 ENGINEERING, INDUSTRIAL

Journal of Manufacturing Systems

Pub Date : 2025-12-31 DOI: 10.1016/j.jmsy.2025.12.024

Wenqiang Zhang , Sen Yang , Mingzhe Li , Yashuang Mu , Guohui Zhang , Mitsuo Gen , Peng Li

The distributed heterogeneous hybrid flow-shop scheduling problem (DHHFSP) involves geographically dispersed factories, heterogeneous machines, and skilled workers, leading to complex multiobjective scheduling challenges. Existing studies usually ignore the critical role of workers in distributed manufacturing, and conventional multiobjective optimization algorithms struggle to balance convergence and solution diversity. To address these gaps, this paper develops a worker–machine–environment collaborative model for DHHFSP that simultaneously minimizes makespan, total energy consumption, and total worker cost, and proposes a hybrid multiobjective particle swarm optimization algorithm with Q-learning-driven local search (HMOPSO-QLS). The proposed approach features a hybrid framework that combines multidirectional particle swarm global search with reinforcement learning-based local search, a Pareto front oriented four directional swarm decomposition and update mechanism where three boundary exploration sub-swarms and one central exploration sub-swarm cooperatively guide particle evolution, and a two level local search scheme that integrates domain knowledge based inter-factory load balancing with Q-learning based adaptive intra-factory variable neighborhood search. Comprehensive experiments demonstrate that HMOPSO-QLS significantly outperforms classical multiobjective optimization algorithms in solving DHHFSP in terms of convergence and solution diversity. In distributed human–machine collaborative manufacturing, the proposed framework and algorithm support more effective configuration of factories, machines, and workers, providing robust schedules that are directly applicable to practical production decision-making and cross-factory coordination.

分布式异构混合流水车间调度问题（DHHFSP）涉及地理上分散的工厂、异构机器和熟练工人，导致复杂的多目标调度挑战。现有的研究往往忽略了工人在分布式制造中的关键作用，传统的多目标优化算法难以平衡收敛性和解的多样性。为了解决这些问题，本文开发了DHHFSP的工人-机器-环境协同模型，同时最小化完工时间、总能耗和总工人成本，并提出了一种基于q -学习驱动的局部搜索（hmpso - qls）的混合多目标粒子群优化算法。该方法采用多向粒子群全局搜索与基于强化学习的局部搜索相结合的混合框架，采用三个边界探索子群与一个中心探索子群协同引导粒子进化的Pareto前沿四向群体分解与更新机制；结合基于领域知识的工厂间负载均衡和基于q学习的自适应工厂内变量邻域搜索的两级局部搜索方案。综合实验表明，hmpso - qls在求解DHHFSP的收敛性和解的多样性方面明显优于经典多目标优化算法。在分布式人机协同制造中，提出的框架和算法支持更有效的工厂、机器和工人配置，提供直接适用于实际生产决策和跨工厂协调的鲁棒调度。

{"title":"A hybrid multiobjective particle swarm optimization with Q-learning-driven local search for distributed heterogeneous hybrid flow-shop scheduling problem with worker–machine–environment collaboration","authors":"Wenqiang Zhang , Sen Yang , Mingzhe Li , Yashuang Mu , Guohui Zhang , Mitsuo Gen , Peng Li","doi":"10.1016/j.jmsy.2025.12.024","DOIUrl":"10.1016/j.jmsy.2025.12.024","url":null,"abstract":"<div><div>The distributed heterogeneous hybrid flow-shop scheduling problem (DHHFSP) involves geographically dispersed factories, heterogeneous machines, and skilled workers, leading to complex multiobjective scheduling challenges. Existing studies usually ignore the critical role of workers in distributed manufacturing, and conventional multiobjective optimization algorithms struggle to balance convergence and solution diversity. To address these gaps, this paper develops a worker–machine–environment collaborative model for DHHFSP that simultaneously minimizes makespan, total energy consumption, and total worker cost, and proposes a hybrid multiobjective particle swarm optimization algorithm with Q-learning-driven local search (HMOPSO-QLS). The proposed approach features a hybrid framework that combines multidirectional particle swarm global search with reinforcement learning-based local search, a Pareto front oriented four directional swarm decomposition and update mechanism where three boundary exploration sub-swarms and one central exploration sub-swarm cooperatively guide particle evolution, and a two level local search scheme that integrates domain knowledge based inter-factory load balancing with Q-learning based adaptive intra-factory variable neighborhood search. Comprehensive experiments demonstrate that HMOPSO-QLS significantly outperforms classical multiobjective optimization algorithms in solving DHHFSP in terms of convergence and solution diversity. In distributed human–machine collaborative manufacturing, the proposed framework and algorithm support more effective configuration of factories, machines, and workers, providing robust schedules that are directly applicable to practical production decision-making and cross-factory coordination.</div></div>","PeriodicalId":16227,"journal":{"name":"Journal of Manufacturing Systems","volume":"85 ","pages":"Pages 1-21"},"PeriodicalIF":14.2,"publicationDate":"2025-12-31","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"145852447","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":1,"RegionCategory":"工程技术","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}

引用次数: 0

A knowledge-enhanced discrete artificial bee colony algorithm for flexible job shop scheduling problem with transport robots 运输机器人柔性作业车间调度问题的知识增强离散人工蜂群算法

IF 14.2 1区工程技术 Q1 ENGINEERING, INDUSTRIAL

Journal of Manufacturing Systems

Pub Date : 2025-12-30 DOI: 10.1016/j.jmsy.2025.12.023

Youjie Yao , Kai Ye , Chunjiang Zhang , Xinyu Li , Liang Gao

The flexible job shop scheduling problem with transport robots (FJSPTR) has garnered significant attention due to the widespread integration of transport robots in manufacturing systems. Owing to its dual-resource scheduling characteristics, metaheuristic algorithms have become the dominant approach for solving the FJSPTR, as they can effectively handle its inherent complexity. However, existing algorithms often lack domain-specific knowledge, leading to suboptimal solution quality and limited robustness. To address these challenges, this study proposes a knowledge-enhanced discrete artificial bee colony (KEDABC) algorithm aimed at minimizing makespan. First, three tailored neighborhood structures are designed, and two feasibility theorems are introduced to ensure the validity of the generated neighborhood solutions. Second, these neighborhood structures are integrated with tabu search to form a knowledge-driven search approach, enabling a more comprehensive and detailed search of the neighborhoods. Third, to integrate domain knowledge into the discrete artificial bee colony algorithm, a novel task-based solution representation is proposed, which provides a complete mapping from the encoding to the solution space. In addition, a machine assignment-based local search strategy is developed to further improve algorithm performance. Finally, extensive experiments are conducted to benchmark the proposed algorithm against state-of-the-art methods using two benchmark datasets. The comparison results demonstrate that the KEDABC algorithm consistently achieves optimal or near-optimal solutions, outperforming other methods in both solution quality and computational efficiency.

由于运输机器人在制造系统中的广泛集成，运输机器人柔性作业车间调度问题（FJSPTR）受到了广泛关注。由于FJSPTR的双资源调度特性，元启发式算法能够有效地处理其固有的复杂性，已成为解决该问题的主要方法。然而，现有的算法往往缺乏特定领域的知识，导致求解质量不佳，鲁棒性有限。为了解决这些挑战，本研究提出了一种知识增强离散人工蜂群（KEDABC）算法，旨在最小化最大完工时间。首先，设计了三种定制的邻域结构，并引入了两个可行性定理来保证生成的邻域解的有效性。其次，将这些邻域结构与禁忌搜索相结合，形成知识驱动搜索方法，对邻域进行更全面、更细致的搜索。第三，为了将领域知识整合到离散人工蜂群算法中，提出了一种新的基于任务的解表示，提供了从编码到解空间的完整映射。此外，为了进一步提高算法性能，提出了一种基于机器分配的局部搜索策略。最后，使用两个基准数据集对所提出的算法进行了广泛的实验，以比较最先进的方法。对比结果表明，KEDABC算法始终能得到最优或近最优解，在解质量和计算效率方面都优于其他方法。

{"title":"A knowledge-enhanced discrete artificial bee colony algorithm for flexible job shop scheduling problem with transport robots","authors":"Youjie Yao , Kai Ye , Chunjiang Zhang , Xinyu Li , Liang Gao","doi":"10.1016/j.jmsy.2025.12.023","DOIUrl":"10.1016/j.jmsy.2025.12.023","url":null,"abstract":"<div><div>The flexible job shop scheduling problem with transport robots (FJSPTR) has garnered significant attention due to the widespread integration of transport robots in manufacturing systems. Owing to its dual-resource scheduling characteristics, metaheuristic algorithms have become the dominant approach for solving the FJSPTR, as they can effectively handle its inherent complexity. However, existing algorithms often lack domain-specific knowledge, leading to suboptimal solution quality and limited robustness. To address these challenges, this study proposes a knowledge-enhanced discrete artificial bee colony (KEDABC) algorithm aimed at minimizing makespan. First, three tailored neighborhood structures are designed, and two feasibility theorems are introduced to ensure the validity of the generated neighborhood solutions. Second, these neighborhood structures are integrated with tabu search to form a knowledge-driven search approach, enabling a more comprehensive and detailed search of the neighborhoods. Third, to integrate domain knowledge into the discrete artificial bee colony algorithm, a novel task-based solution representation is proposed, which provides a complete mapping from the encoding to the solution space. In addition, a machine assignment-based local search strategy is developed to further improve algorithm performance. Finally, extensive experiments are conducted to benchmark the proposed algorithm against state-of-the-art methods using two benchmark datasets. The comparison results demonstrate that the KEDABC algorithm consistently achieves optimal or near-optimal solutions, outperforming other methods in both solution quality and computational efficiency.</div></div>","PeriodicalId":16227,"journal":{"name":"Journal of Manufacturing Systems","volume":"84 ","pages":"Pages 614-628"},"PeriodicalIF":14.2,"publicationDate":"2025-12-30","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"145879994","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":1,"RegionCategory":"工程技术","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}

引用次数: 0

Small-sample machining quality prediction via a fuzzy broad learning system enhanced by prior knowledge 基于先验知识增强模糊广义学习系统的小样本加工质量预测

IF 14.2 1区工程技术 Q1 ENGINEERING, INDUSTRIAL

Journal of Manufacturing Systems

Pub Date : 2025-12-29 DOI: 10.1016/j.jmsy.2025.12.021

Zewen Hu , Yu Shen , Shuyue Zhang , Hongcai Chen , Kanjian Zhang , Haikun Wei

Surface roughness is a critical indicator of machined workpiece quality, and accurately modeling its relationship with process parameters is essential for process optimization and intelligent decision-making. Fuzzy broad learning system (FBLS) has demonstrated considerable advantages in nonlinear predictive modeling; however, its performance under small-sample conditions may be limited due to an incomplete rule base and the lack of explicit physical mechanisms. To address this challenge, this article proposes a knowledge-enhanced fuzzy broad learning system (KEFBLS) that integrates dual sources of prior knowledge — expert-knowledge-guided fuzzy partition and physics-based fuzzy rule consequents — to improve predictive accuracy and generalization ability. The effectiveness of KEFBLS is validated on both real-world robotic grinding experiments and a publicly available machining dataset, achieving average prediction errors of only 10.3% and 4.7%, respectively, representing over 20% accuracy improvement over the FBLS baseline. These results highlight the significance of combining domain-specific prior knowledge with data-driven learning, enabling robust performance under limited-data conditions. Overall, KEFBLS provides a unified knowledge- and data-driven framework for surface roughness prediction, with potential applicability to other manufacturing processes where labeled data are scarce.

表面粗糙度是加工工件质量的重要指标，准确建模其与工艺参数的关系对工艺优化和智能决策至关重要。模糊广义学习系统（FBLS）在非线性预测建模方面显示出相当大的优势；然而，由于不完整的规则库和缺乏明确的物理机制，其在小样本条件下的性能可能受到限制。为了解决这一挑战，本文提出了一种知识增强模糊广义学习系统（KEFBLS），该系统集成了先验知识的双重来源-专家知识引导的模糊划分和基于物理的模糊规则结果-以提高预测精度和泛化能力。KEFBLS的有效性在实际机器人磨削实验和公开可用的加工数据集上得到了验证，平均预测误差分别仅为10.3%和4.7%，比FBLS基线精度提高了20%以上。这些结果强调了将特定领域的先验知识与数据驱动学习相结合的重要性，从而在有限数据条件下实现稳健的性能。总体而言，KEFBLS为表面粗糙度预测提供了统一的知识和数据驱动框架，具有潜在的适用性，可用于标记数据稀缺的其他制造过程。

{"title":"Small-sample machining quality prediction via a fuzzy broad learning system enhanced by prior knowledge","authors":"Zewen Hu , Yu Shen , Shuyue Zhang , Hongcai Chen , Kanjian Zhang , Haikun Wei","doi":"10.1016/j.jmsy.2025.12.021","DOIUrl":"10.1016/j.jmsy.2025.12.021","url":null,"abstract":"<div><div>Surface roughness is a critical indicator of machined workpiece quality, and accurately modeling its relationship with process parameters is essential for process optimization and intelligent decision-making. Fuzzy broad learning system (FBLS) has demonstrated considerable advantages in nonlinear predictive modeling; however, its performance under small-sample conditions may be limited due to an incomplete rule base and the lack of explicit physical mechanisms. To address this challenge, this article proposes a knowledge-enhanced fuzzy broad learning system (KEFBLS) that integrates dual sources of prior knowledge — expert-knowledge-guided fuzzy partition and physics-based fuzzy rule consequents — to improve predictive accuracy and generalization ability. The effectiveness of KEFBLS is validated on both real-world robotic grinding experiments and a publicly available machining dataset, achieving average prediction errors of only 10.3% and 4.7%, respectively, representing over 20% accuracy improvement over the FBLS baseline. These results highlight the significance of combining domain-specific prior knowledge with data-driven learning, enabling robust performance under limited-data conditions. Overall, KEFBLS provides a unified knowledge- and data-driven framework for surface roughness prediction, with potential applicability to other manufacturing processes where labeled data are scarce.</div></div>","PeriodicalId":16227,"journal":{"name":"Journal of Manufacturing Systems","volume":"84 ","pages":"Pages 601-613"},"PeriodicalIF":14.2,"publicationDate":"2025-12-29","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"145879995","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":1,"RegionCategory":"工程技术","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}

引用次数: 0