Dualistic production and intralogistics (DPiL) have inherently intricate spatio-temporal coupling relationships. Meanwhile, the occurrence of disturbance propagation further amplifies such complexity, aggravating operation control difficulty. This novel problem is summarized as the dualistic production and intralogistics synchronized control problem (DPiLSynC) considering disturbance propagation. Traditional production planning and control (PPC) methods overlook the impact of disturbance propagation, significantly compromising system stability. Fortunately, digital twin (DT) technology promises rapid identification of disturbances by physical-virtual interaction. The adverse effects of disturbances may diffuse across the manufacturing system, causing operational disruptions. But analyzing the spatio-temporal characteristics of disturbance propagation is challenging due to the unclear connection-interaction mechanism of multiple DT models in manufacturing systems. This paper proposes a temporal-digital twin-network (T-DT-N) enabled Out-of-Order (OoO) synchronized control approach. Firstly, the modified OoO framework with disturbance propagation analysis is designed to reconstruct the operation logic of DPiL. Secondly, T-DT-N is developed to reveal the spatio-temporal evolution patterns of disturbance propagation by introducing the temporal network to model connection-interaction relationships of DT models. Thirdly, OoO synchronized control using propagation analysis results is proposed, involving three distinct control phases, each considering multiple factors aiming to achieve resilient DPiL operation. Computation results indicate this approach achieves shorter order makespan, shopfloor throughput time, and waiting time, while decreasing the percentage tardy. Furthermore, it exhibits stability across varying queue lengths and uncertain levels, thereby validating its application in a dynamic environment.

Digitalization initiatives have led to the widespread adoption of digital twins (DTs), particularly in manufacturing. The explosive growth of DTs makes us think about what the most abstract parts in building the DT systems are and how to promote the reconfigurability of DTs so that they can be agile to adopt the flexibility needs from Industry 4.0 manufacturing systems. However, the heterogeneous granularity and dynamic evolution of manufacturing systems present significant challenges to the reconfigurability of their corresponding DTs. To address these challenges, this paper presents MetaTwin, a reconfigurable DT framework as an integrated solution to abstract DTs into general domains. To be more specific, a novel Domain-Meta structure is first proposed, which addresses the granularity misalignment problem by decoupling the system into stable Domains and reconfigurable Meta units. Then, a dual-dimensional reconfiguration mechanism decomposes complex reconfigurations into macro and micro dimensions, achieved by composing Meta units and reconfiguring their internal pattern-driven Meta Models. A prototype-based case study involving the progressive reconfiguration from an assembly-only DT to a printing-and-assembly DT illustrates how MetaTwin supports reconfiguration scope identification, macro-level structural changes, micro-level behavioral adaptations, and runtime coordination. Furthermore, implementation-level indicators further suggest that the MetaTwin-based implementation localizes changes more effectively than a conventional implementation in the case setting.

Personalized production is widely adopted in discrete manufacturing due to its ability to meet diverse customer requirements through flexible combinations of components. However, it also poses significant challenges to production synchronization, which requires all necessary operators, tools, materials, and industrial robots to be available within the prescribed time and space. In complex multi-stage manufacturing processes, execution deviations can accumulate into operation-level uncertainty, namely the risk that an operation cannot start, proceed, or finish as scheduled because the required resources and execution conditions are no longer aligned. Such uncertainty undermines resource utilization and may trigger cascading delays across the entire process. To address this challenge, this paper proposes a spatial-temporal event compiler (STEC) framework comprising digital, knowledge, and reasoning engines for compiling and analyzing manufacturing events. Specifically, the state information of digital models is compiled into a spatial-temporal event graph, which is continuously updated through a multi-clock alignment scheme that aligns distinct planning, scheduling, and execution time scales. Within the reasoning engine, a look-around reasoning approach is developed to capture historical consistency, the current execution context, and near-future evolution trends by integrating look-backward, look-present, and look-forward perspectives, thereby enabling uncertainty assessment. Furthermore, an LLM-based synchronization mechanism is incorporated into the framework to provide status reports through multi-turn interactions, allowing managers to query the current execution status and assess potential downstream impacts for timely intervention. Finally, a case study demonstrates that STEC improves the accuracy and stability of uncertainty identification.

In the era of Industry 4.0, distributed manufacturing (DM) paradigms, exemplified by cloud manufacturing (CMfg), have garnered significant attention from both academia and industry. These paradigms aim to achieve the precise matching of geographically dispersed manufacturing resources with personalized demands. However, existing research on service configuration predominantly relies on centralized information architectures, which pose critical challenges such as single point of failure (SPoF), data privacy leakage risks, and limited scalability. To address these issues, this paper investigates how blockchain technology can be leveraged to enable trustworthy collaboration and distributed decision-making in DM systems. Specifically, we propose a Blockchain-enabled Cyber-Physical Network (BCPN) architecture and design a complete and executable business process tailored to service configuration requirements. Furthermore, a systematic solution is developed that integrates several key components, including a reputation evaluation contract, a service pre-screening contract, a proof of reputation (PoR) consensus mechanism, a distributed service configuration contract, and a backward time propagation smart contract based on breadth-first search (BFS). Simulation results validate the superiority of the proposed BCPN solution in trust establishment, privacy preservation, and system scalability, providing a novel theoretical framework and methodological pathway for the distributed governance of DM systems.

Since the promotion of sustainable investment and development, academia and practitioners have shown considerable interest in environmental, social, and governance (ESG) disclosure. Nonetheless, the authenticity of data in ESG disclosure remains an unresolved and critical issue. This study proposes a new method based on Petri Net to address the aforementioned problem. Firstly, we designed a novel Smart ESG Disclosure System (S-ESG), harnessing the power of Internet of Things (IoT) and blockchain. The seamless integration of IoT and blockchain technology not only simplifies the automatic data gathering and transmission in the ESG disclosure process, but also ensures the authenticity, consistency, and transparency of data. Secondly, within the S-ESG, a dual-layer data authenticity verification process is crafted utilizing the nested Petri net. This innovative approach facilitates event modeling and verifies the authenticity of the data associated with these events. Subsequently, based on the nested Petri net framework, this study pioneers the development of data authenticity analytics algorithms encompassing spatial-temporal analytics and authenticity score calculation. Furthermore, a simulated experiment is conducted to showcase the practical deployment of the S-ESG and assess the effectiveness of the proposed solution. This study employs a controlled variable approach coupled with a systematic sensitivity analysis, evaluating 243 variable combinations. Experimental validation identifies five key parameters that exhibit statistically significant correlations with verification accuracy, providing empirically supported deployment guidelines for production environments. The numerical results demonstrate that each algorithm achieves AUC values ranging from 0.91 to 0.94. The study concludes by outlining future endeavors, including algorithm enhancements and field experiments in manufacturing factories. Anticipated to be beneficial for academia and industry, this research aims to facilitate the application of the solution in analogous situations and spark innovative thoughts.

Straw, as a byproduct of agricultural production, is used in resource-oriented, feed-oriented, and substrate-oriented ways, while straw recycling is the first step toward its efficient utilization. In practice, straw recycling enterprises first determine the baling plan and then the transport plan. Such a sequential decision-making method increases the time delay between these two processes and the maximum completion time. However, existing studies have rarely taken this issue into account. This study proposes an integrated scheduling framework to synchronize straw baling and transportation by jointly determining the working routes of straw balers and transporters, and proposes an efficient solution algorithm to solve it. Numerical experiments using real-world datasets show that our proposed methodology provides high-quality solutions within a reasonable computation time and reduces the maximum completion time, compared with the sequential decision-making method. Expanding the fleets of straw balers and transporters shortens the maximum completion time, and more straw balers also significantly decrease waiting time.

Robotic cellular warehousing systems (RCWS) represent a new paradigm for automated logistics, yet existing multi-agent pickup and delivery (MAPD) formulations typically assume static orders and ideal communication, which rarely hold in practice. This paper studies dynamic MAPD in RCWS under communication uncertainty for the first time, where order specifications may evolve during execution and coordination decisions rely on delayed information. To capture communication effects at the system level, we model uncertainty as location-dependent stochastic transmission delays using radio maps, which reflect spatial heterogeneity induced by warehouse layout and scale. Based on this model, we develop a communication-aware token-passing framework that adapts replanning decisions to both the timing and spatial context of received updates, enabling effective coordination under non-ideal information exchange. Extensive simulations under varying order dynamics, robot fleet sizes, and communication conditions demonstrate that coordination strategies designed for static and delay-free settings can degrade significantly in uncertain environments. The results further reveal fundamental trade-offs between responsiveness and resource utilization, and show that exploiting idle resources through cooperative coordination can substantially improve system performance. These findings highlight the critical role of communication in dynamic warehouse operations and provide practical guidance for designing coordination policies and system configurations in RCWS.

Recommending maintenance plans presents significant challenges due to the low standardization of maintenance records and unclear pathways for identifying appropriate plans. While knowledge graphs have been extensively researched for integrating and evolving maintenance data, these issues hinder the accurate recommendation of maintenance solutions within large-scale maintenance knowledge systems. This paper proposes a causality and equipment structure enhanced maintenance plan matching and recommendation (CEE-MPMR) method to address these challenges. The method leverages an unsupervised SimCSE model to normalize domain vocabulary in the absence of domain lexicon, and proposes a maintenance plan reasoning method based on RotatE cc. The proposed method achieves a maintenance plan matching accuracy of 90.80%, effectively improving the precision of maintenance plan recommendations. Finally, we applied and validated the approach on real-world data from a nuclear power enterprise and integrated the algorithm into a maintenance plan recommendation system, supporting intelligent analysis and decision-making for nuclear complex equipment maintenance.

Process data, characterized by strong nonlinearity, dynamics, and complex coupling, are ubiquitous in real-world industrial production. With the rapid development of increasingly complex modern industries, traditional shallow models struggle to capture the wealth of implicit information in massive industrial data. The robust feature extraction capabilities of deep neural networks have inspired the development of a large body of deep networks-based methods in the field of process modeling. However, there remains a notable absence of an up-to-date and systematic review on data imputation and soft sensing techniques, ranging from small models to large models. To address this gap, this work conducts a comprehensive review of deep learning methodologies for data imputation and soft sensing, spanning five classic architectures and the emerging technical frameworks of Large Language Models (LLMs). We primarily give the motivation of jointly handling data imputation and soft sensing, and present a general framework of deep network for process modeling. Finally, we propose a prospective research framework that utilizes LLMs for imputation and sensing tasks. This survey brings together the latest strides in both small and large models, offering researchers an up-to-date perspective on current breakthroughs and future research opportunities.

Well-designed production sequences are essential for achieving high efficiency and low cost in automobile manufacturing. However, defective products inevitably occur in practice, and existing approaches fail to effectively support coordinative rescheduling that explicitly accounts for such defects. To address this issue, a coordinative rescheduling optimization approach is proposed based on a utilization-efficient computing strategy (UCS) for a multi-workshop automotive manufacturing system (MWAMS) considering defective products. First, a coordinative rescheduling strategy and corresponding mathematical model are developed, in which the rescheduling schemes of unprocessed and repaired automobiles are treated as decision variables. Next, an UCS that combines the advantages of distributed and multi-thread computation strategies is developed to improve computation performance for the optimization process. Finally, a practical MWAMS case study is conducted to validate the proposed approach, and the outcomes demonstrate that the coordinative rescheduling method is usable and outperforms the existing rescheduling strategies and computation strategies.

Embodied intelligence (EI) is the latest paradigm in the evolution of artificial intelligence (AI), which becomes a key enabler for advancing intelligent manufacturing in industrial scenarios. In recent years, the explosive growth of large language models (LLMs) has provided a critical pathway for realizing EI, rapidly spawning numerous groundbreaking applications. The widely application of industrial robots is the core characteristics of intelligent manufacturing, where the emergence of EI makes industrial robots becoming smarter and smarter with the capability of executing ever more complex tasks today. Consequently, research on LLMs for industrial EI has rapidly attracted attention from researchers worldwide in the past few years. To better understand the current situation in this field, this paper reviews the progress of LLMs towards industrial EI and discusses the corresponding research status and development trend from the perspectives of the capability breakthrough of single-modal LLMs, the fusion enhancement of multimodal large language models (MLLMs), and the embodiment realization of LLMs. This review aims to provide useful references for the further advancement of LLMs to support the development of industrial EI.

As the demand for personalised customisation increases, manufacturing enterprises are continually enhancing their capabilities to meet diverse market requirements. This paper investigates a two-stage assembly flowshop scheduling problem, focusing on minimising total tardiness while accounting for resource flexibility and dynamic product arrivals using multi-agent deep reinforcement learning (MADRL). To explore the operational benefits of resource flexibility in scheduling environments, a skill matrix is introduced to assess the flexibility level of the processing machines. The problem is modelled as a Markov decision process (MDP), and a Multi-Agent Proximal Policy Optimisation algorithm with an Epsilon-greedy approach (E-MAPPO) is proposed for product sequencing and processing machine allocation. Experimental results show that, across different numbers of arrival products and flexibility levels, the scheduling agent trained with the E-MAPPO effectively learns to apply suitable dispatching rules. Notably, the algorithm outperforms 18 composite dispatching rules and two other deep reinforcement learning algorithms, namely VDN and standard MAPPO. Our findings demonstrate that enhanced resource flexibility positively impacts scheduling performance, which indicates that enterprises must balance the benefits of flexibility against its implementation costs to satisfy customer demand.

Collaborative manufacturing in distributed synchronized manufacturing systems (DSMSs) increasingly relies on digital twins to support state awareness, coordinated optimization, and trustworthy data interaction. However, heterogeneous data across factories and partners result in digital twin data integration approaches that remain largely stage-isolated and insufficiently aligned with manufacturing hierarchies and lifecycle value flow, which limits the scalability and decision effectiveness in collaborative settings. To address these challenges, we propose a systematic data integration framework for collaborative manufacturing digital twins (CMDTs) that transforms multi-source heterogeneous digital twin data into traceable, standardized, and decision-oriented digital assets across distributed units. The framework advances beyond conventional stage-isolated integration by proposing an edge-fog-cloud-based architecture that aligns data integration functions with data granularity and collaboration scope, thereby enabling scalable and coordinated cross-unit data orchestration. Then, a five-stage data integration operational mechanism is established to tightly couple hierarchical architecture with lifecycle-oriented data value transformation, with blockchain embedded to enhance traceability and trust for CMDTs. Finally, three enabling methodologies targeting key bottlenecks are proposed: (i) a multi-objective edge caching model to jointly optimize low-latency data availability and load balancing; (ii) an LLM-aided data processing approach for efficient heterogeneous data standardization; (iii) a double-auction-based data sharing model for effective cross-unit collaboration. A coating manufacturing case study demonstrates that the proposed framework achieves consistent improvements in latency, load balancing, and data sharing efficiency compared with representative benchmarks. This research provides a structured and scalable foundation for digital twin data integration and data-driven optimization in collaborative manufacturing.

Welding in shipbuilding remains highly dependent on skilled labour, while expert know-how is difficult to formalise and existing automation solutions are often fragmented at the system level. This paper presents a welder-centred intelligent welding framework that links IoT-enabled data collection, a welding data and knowledge repository and an LLM-based welding copilot. At the core of the copilot, we develop WeldGPT, a welding-domain large-model module built on Qwen-7B with LoRA-based fine-tuning and retrieval-augmented prompting over structured welding manuals and experience data. Welding tasks are represented by a unified Task and Condition Description, which, together with entries retrieved from Welding Manual and Welding Knowledge and Experience, is transformed by WeldGPT into Robot-Executable Welding Procedures in terms of parameter settings and robot-oriented operation plans. A prototype Weld Copilot system is instantiated using historical data and procedures from a robotic fillet-weld production line at a medium-sized shipyard located in Zhuhai, China. Preliminary offline results indicate that WeldGPT can reproduce the main parameter-setting patterns encoded in existing procedures and provide practically useful starting points for configuring and refining robotic welding programs, supported by a web-based interface for interactive inspection and adjustment.

Large language models (LLMs) show strong text analysis and prediction abilities, enhancing the operational efficiency and demand forecasting in operations and supply chain management (OSCM). To improve the operational efficiency of cross-border e-commerce, suppliers can select appropriate multi-agency modes according to their specific circumstances. However, the influence of LLMs on online sales and the determination of viable operational management recommendations remain unexplored. To address this issue, this paper develops a Stackelberg game model for a cross-border supply chain comprising a manufacturer, a supplier, and an e-retailer. We derive and analyze the optimal order quantities and equilibrium solutions for supplier across different multi-agency modes and LLMs application scenarios in hybrid channels. Additionally, we compare different models and analyze how LLMs application levels and customer channel preferences impact supplier decisions. Our model analysis and numerical study yield the following key findings. First, in the multi-agency mode, supplier can attain higher profits through the semi-managed mode. Furthermore, the direct channel generates more profit than the agency channel. Second, LLMs can significantly boost order volumes in both channels when operating in fully-managed mode. When customers show a preference for purchasing via the agency channel, the application of LLMs in either fully-managed or semi-managed modes can enhance order volumes in both channels. Third, adopting a suitable LLMs application level can markedly boost supplier profits in the fully-managed mode, particularly when customers favor the agency channel, leading to maximum profit. This research offers valuable suggestions for LLMs application and supplier multi-agency mode choices in cross-border e-commerce.

The growing demand for customization in fast-fashion apparel manufacturing, characterized by high product variety, short lead times, unpredictable order arrivals, and varying batch sizes, has introduced significant challenges to Hybrid Flow Shop (HFS) operations. These challenges are aggravated by the inherent complexity of HFS, which features multiple stages and shared resources. The resulting variability, such as fluctuations in machine availability, processing times, and inter-stage dependencies, necessitates dynamically reallocating resources, reprioritizing jobs, and coordinating buffers among stages. In response, this study presents a Cyber-Physical Internet (CPI)-enabled multi-stage scheduling method to address the challenges of dynamic HFS operations. Firstly, a CPI-enabled HFS is constructed, establishing cyber-physical synchronization through hierarchical gateways at both the shop floor and stage levels. Subsequently, CPI routing tables are developed by multiple information tables, enabling real-time data traceability and visualization throughout the production process. Building on this foundation, this study proposes an Out-of-Order Execution (OoOE) method with dynamic Workload consideration (OoOEW) for HFS scheduling problem. The OoOEW method reduces disturbances in high-variability production environments and enhances scheduling flexibility by incorporating workload balancing across upstream and downstream stages. Finally, a case study is conducted to evaluate the proposed solution in a mass-customization production environment. The experiment results demonstrate that OoOEW significantly outperforms traditional methods, in terms of shop floor throughput and delivery time. The proposed method shows operation flexibility and enhances adaptability under dynamic manufacturing conditions in HFS.

Under the dual pressure of explosive growth in cross-border e-commerce demand and increasing timeliness requirements from overseas customers, cross-border logistics service providers are compelled to establish logistics facilities and deploy fleets across multiple regions to ensure rapid response. However, during freight transportation, the lack of effective management over these complex and heterogeneous fleets-particularly in terms of fleet composition and routing decisions-has led to high transportation costs and low operational efficiency. This study is grounded in the practical operational context of cross-border logistics in the Guangdong-Hong Kong-Macau Greater Bay Area and models a multi-level, multi-node cross-border transportation network. To minimize the overall operational cost, the problem is addressed from two interrelated decisionmaking perspectives: fleet composition at the strategic level and routing planning at the operational level. Thus, a bi-level programming model is proposed to systematically capture the hierarchical structure and the logical relationship between these two decision layers. Furthermore, the model incorporates cost differences among trucks with different functional capabilities to reflect the significant disparity in logistics cost structures between domestic and overseas operations. To address the above multi-objective mixed-integer linear programming (MILP) problem, a tailored Non-dominated Sorting Genetic Algorithm II (MNSGA-II) is developed. Several key components of the algorithm are modified and enhanced to improve its search efficiency and solution quality in handling the problem's complexity. Comparative experiments against classical algorithms demonstrate the superior solution quality and robustness of the proposed approach. The influence of cost differentials on composition and scheduling decisions is further analyzed, providing practical insights for the strategic planning of cross-border logistics systems.

Influenced by factors such as residents’ living habits, commuting patterns, and commercial activity cycles, the generation of domestic waste exhibits a distinct double-peak distribution. To meet the high demand during peak periods, collection companies typically deploy excess transportation capacity, which leads to severe resource idleness during off-peak periods, imposing significant economic and environmental burdens. To address this issue, this study develops a dynamic smart waste collection routing model aimed at minimizing the coordinated collection cost between self-owned and outsourced multi-compartment vehicles, and designs a two-phase algorithm to solve it. In the pre-optimization phase, an improved Harris Hawks Optimization algorithm integrated with multiple heuristic algorithms is employed to generate initial collection routes. In the re-optimization phase, a hybrid strategy combining periodic and continuous re-optimization is used to dynamically update collection routes. Finally, the effectiveness of the proposed model and algorithm is validated through case studies. Furthermore, a systematic sensitivity analysis is conducted to investigate the impact of key parameters, yielding practical insights for waste collection management.

With the rapid development of global logistics, the role of smart ports in optimizing sea-rail intermodal transport is becoming increasingly prominent. However, the lack of unified information sharing and real-time coordination mechanisms between heterogeneous transport networks has become the main bottleneck in improving efficiency. This paper proposes a distributed decision-making framework based on the Cyber-Physical Internet (CPI) to address the key issues in smart port operations involving sea-rail multimodality. A five-layer CPI architecture, inspired by the TCP/IP model, is developed to standardize information exchange and operational processes. Departing from the conventional ship -* yard -* train workflow, we investigate a novel sea-rail coordination strategy wherein trains serve as mobile temporary storage yards. This setup facilitates direct ship-to-train transshipment. A mathematical model and a tailored CPI gateway protocol (CPIGP) algorithm are developed to optimize container allocation, maximize train loading rates, minimize storage yard usage, and reduce operational costs. Through extensive case studies within the CPI five-layer framework, spanning both small-scale and large-scale scenarios, our CPIGP algorithm demonstrates superior performance over exact, non-black-box baselines, including fixed-rule manual and dynamic programming methods. The advantage is particularly apparent in greater efficiency and resilience under complex operating conditions. Collectively, the work advances intelligent, sustainable, and adaptive smart port operations through distributed, CPI-enabled decision-making.

With the rapid growth of B2C cross-border e-commerce, logistics service providers (LSPs) are facing significant challenges in balancing cost efficiency and operational performance due to increasingly frequent, small-batch, and highly fragmented order demands. To achieve economies of scale and improve resource utilization, fragmented orders are typically consolidated prior to logistics operations and transported in batches through cross-border haul transportation and terminal delivery stages. However, existing studies predominantly adopt static order consolidation strategies based on time, quantity, or hybrid time-quantity rules to generate transportation batches, which often fail to effectively balance transportation cost and delivery efficiency. To address this issue, this paper proposes a Cyber-physical internet (CPI)-based joint optimization approach for order consolidation and routing. First, a CPI-based B2C Logistics (CPIBCL) framework is developed. By designing a CPI routing table, the CPIBCL enables full-process visualization and traceability of logistics data packets. Second, four heuristic order consolidation rules (OCRs) are designed, and a knowledge-driven mechanism for strategy generation and dynamic updating is proposed to adapt to uncertain demand environments. Furthermore, to solve the joint optimization problem of order consolidation and two-echelon vehicle routing, an integrated solution approach is developed by combining an improved whale optimization algorithm and an adaptive large neighborhood search, enabling the coordinated optimization of OCRs weights and routing decisions. The effectiveness of the proposed approach is validated through benchmark function tests and simulation experiments. The results show that, compared with single-rule consolidation strategies, the proposed method reduces order delay rate and transportation cost by 11.37% and 3.55%, respectively, while improving truck load utilization by 2.47%, demonstrating its effectiveness and practical applicability in complex cross-border logistics scenarios.