Keywords: Decision making, Production planning, Smart system
Abstract: As global competition intensifies; automotive companies increasingly adopt quality improvement techniques like Lean Six Sigma (LSS). However, many face challenges in implementing LSS effectively, underscoring the need to identify and prioritize barriers to successful integration. This study aims to pinpoint and rank the primary barriers to LSS implementation within the automotive industry, providing a roadmap for overcoming these challenges to enhance operational performance. The study begins with an extensive literature review of peer-reviewed articles focusing on LSS barriers, supplemented by insights from ten industry experts. Fifteen barriers were identified and grouped into three primary categories: technological, organizational, and socio-cultural. The Best-Worst Method (BWM), a multi-criteria decision-making technique, was employed to prioritize these barriers based on expert input. This research highlights 15 critical barriers that hinder effective LSS implementation in the automotive sector. The organizational barrier, particularly the inadequacy of resources, emerged as the most significant, followed by technological and socio-cultural barriers. The study also proposes strategies to address these barriers, helping practitioners implement LSS more effectively. The barriers were identified through literature review and expert insights. Future studies could expand on this by conducting broader surveys to include additional industry-specific barriers. This paper’s main contribution is the identification and prioritization of barriers to LSS in the automotive industry using the BWM, offering actionable insights and strategies to facilitate smoother LSS integration and operational excellence.

Keywords: – Lean six sigma (LSS), Barriers, Best-worst Method (BWM), Automotive industry, Decision Making, Production planning, Smart system.

Keywords: Logistic planning, Bio manufacturing
Abstract: This paper presents a novel approach to asset tracking and environmental conditio monitoring, supporting the advancement of Industry 4.0 through a fully digitalized system tailored to medium and large-scale manufacturing and logistics operations. The research addresses a key gap in current logistics practices—namely, the absence of real-time, scalable monitoring of chemical raw materials, particularly fertilizers, during storage and transportation. Existing methods often fail to account for environmental factors that degrade product quality, resulting in inefficiencies and financial losses. To overcome these limitations, we propose an innovative system that combines RFID technology with embedded measuring micro-stations installed in “Big Bags.” These sensors continuously track critical environmental parameters such as temperature, humidity, pressure, etc. A key novelty of this solution is its automated inference engine, which leverages collected sensor data and expert knowledge to support a prototype advisory and diagnostic system for remote management of storage conditions. The motivation behind this research stems from the growing global demand for sustainable and traceable logistics, particularly considering the current chemical fertilizer shortages and the push toward smarter agricultural supply chains. While similar technologies exist in limited or niche contexts, this is the first system of its kind introduced to the Polish market, designed to scale for monitoring 10,000 big bags across 30 or more locations simultaneously. The combination of scalability, autonomy, and intelligent analytics positions this system as a transformative tool in the fields of smart logistics, supply chain optimization, and sustainable urban development.

Keywords: Logistic planning, Process planning
Abstract: Contemporary supply chains have evolved into highly complex systems characterized by multifaceted interactions between entities within and across firms. The linear and isolated view of the supply chain often fails to capture the operational inter-dependencies when addressing a supply chain problem. Our study proposes to view supply chains through the lens of a multilayer network perspective. First, we propose the principal layer, the so-called direct supplier-buyer network, encompassing the focal firm, its immediate supplier, and buying firms. Second, we extend to a deep-tier supply network layer, capturing firms with indirect relationships. Firms do not typically have good visibility for deep-tier suppliers and sometimes suffer a significant impact of the ripple effect. Third, to facilitate more effective management of material dependencies, we introduce the product-integrated network mapping the products and required materials. Finally, we propose a process-integrated network to represent how materials are transformed into final products. The four-layer network framework, therefore, offers a unified, integrated, and interoperable approach to better manage supply chain operations. We also present a case study from a leading European manufacturing firm and highlight how the presentation of a four-layer supply network supports digital transformation and enhances supply chain resilience.

Keywords: Mathematical programming, Optimization, Scheduling
Abstract: Many issues related to optimization practice belong to the class of NP-hard problems. Due to the large size of practical examples, metaheuristics are mainly used to solve them. However, calculations performed on classical computers, due to the computation time and values of the solutions determined, do not meet the expectations of many practitioners. In turn, quantum computers currently have too few qubits to solve even medium-sized examples. In this paper, we present a new approach to using neighborhoods with an exponential number of elements in local search algorithms. A hybrid CPU/QPU algorithm, in which the search of the neighborhood is performed on a quantum computer, allows for more efficient use of the currently available power of quantum computers.

Keywords: Mathematical programming, Smart maintenance
Abstract: This research develops a mathematical model to calculate the coverage of spherical objects by sensors, aiming to support a simulation environment for smart farming services. The proposed model represents plants or produce as spheres and calculates their coverage by directional, static sensors while considering overlaps between objects. This model serves as a key function or a reference value for simulations in smart farming, providing essential support for evaluating and optimizing sensor allocation performance prior to implementation.

Keywords: Production planning, Optimization, Mathematical programming
Abstract: In the context of contemporary manufacturing systems characterized by a high saturation of modern manufacturing technologies such as CNC (Computerized Numerical Control), 3D printing, and advanced transportation or information technology, the decision support processes for production necessitate a novel approach. Firstly, production resources should be treated in a multidimensional manner, i.e. not merely as means of production but also as software, service, competence, etc. Secondly, recognizing the dynamic nature of today's market, a proactive approach is recommended for planning, scheduling, and resource allocation processes. This proactive stance enables precise decision-making even in unconventional or rapidly changing situations. The paper proposes a mathematical model designed for proactive decision support in the planning, scheduling, and distribution of production process loads. The presented model addresses various critical questions, including: Is it possible to fulfill a given set of orders? What is the most cost-effective or time-efficient way to fulfill a given set of orders? Can a specified set of orders be fulfilled in the face of selected resource unavailability? The proposed model has been implemented in a mathematical programming environment and applied to an actual rolling bearing production process.

Keywords: Production planning, Optimization, Mathematical programming
Abstract: The paper presents a method of optimizing production process using audits based on TOGAF and mathematical programming models. Traditional methods, such as ERP/MRP and mathematical modeling, often overlook organizational aspects, limiting their effectiveness in complex production environments. The proposed approach integrates organizational insights from TOGAF with mathematical programming, enabling the identification and optimization of selected areas of the production system. In the case study of a medium-sized kitchen furniture manufacturer, issues such as production delays, insufficient automation, and seasonal employment were analyzed. The study shows how the proposed method can be used to optimize the production system in the presented manufacturer.

Keywords: Process engineering, Intelligent manufacturing, Digital manufacturing
Abstract: The article presents an analysis of the configuration of a reconfigurable production system in terms of the degree of use of technological machines included in the system. It was assumed that the subject of the design was a reconfigurable production system intended for the production of shaft-class parts. As part of the work carried out, based on input data, the number of necessary technological machines (machine tools) was defined, system configurations were selected to achieve the required level of productivity, and an analysis of the production process was made for selected solutions (configurations), enabling production to be carried out at the assumed level. The assessment of the production process was based on the results of simulations carried out using Enterprise Dynamics software.

Keywords: Decision making, Scheduling, Optimization
Abstract: The product-as-a-service (PaaS) market, which combines a product delivery contract with accompanying service contracts for its repair or replacement, highlights the role of associated risk. The risk of failure of leased equipment is perceived as both a reduction in the efficiency of its user and an additional cost of the provider's service. The provider's problem considered in this context comes down to planning an offer for renting leased equipment that maximizes the revenues from its rental and at the same time minimizes its maintenance costs subject to demand constraints. The implementation of this problem, reconstructed within the framework of declarative modeling, made it possible to conduct computational experiments, the results of which confirmed its usefulness in situations occurring in practice.

Keywords: Computer Aided Design, Computer aided manufacturing, Intelligent manufacturing
Abstract: The increasing integration of advanced technologies and artificial intelligence (AI) into manufacturing and business processes drives the adoption of low-code development platforms (LCDPs). This paper investigates key stakeholder expectations regarding AI support within these LCDPs, specifically within the context of manufacturing, and identifies critical gaps between these expectations and current platform offerings. I assess the perceived importance and impact of various AI features by surveying Top Management, Process Owners, Process Participants, Customers, and IT Departments in manufacturing settings. The findings reveal significant disparities in priorities across stakeholder groups within this sector.

Keywords: Scheduling, Internet of Things (IoT), Human interaction
Abstract: This paper addresses the modelling and fault-tolerant control mechanisms for multi-AGVs (Automated Guided Vehicles) systems, specifically examining a flexible transportation network that delivers products to transfer stations within a manufacturing facility's high-storage warehouse. The primary contribution lies in developing a mathematical framework for multiple AGVs and creating an algorithm to determine optimal delivery timing sequences from the outlet. The proposed methodology addresses three critical challenges: synchronization, concurrency (inherent in multi-vehicle operations), and modeling imprecision. The latter is handled through max-plus algebraic principles. The research further incorporates fault-tolerant control mechanisms, resulting in an integrated control framework that merges predictive control with max-plus algebra concepts. The study concludes with demonstrative example that showcases the system's performance under fault conditions, with comparative analysis against non-fault-tolerant control approaches.

Keywords: Optimization, Logistic planning, Mathematical programming
Abstract: The air freight supply chain plays a crucial role in global logistics, with freight forwarders serving as intermediaries between the airside and customers. Distribution strategies can follow either a single-echelon or multi-echelon structure, with the latter incorporating intermediate facilities, such as satellites, to enhance cost efficiency and service quality. One of the key challenges in this domain is the Vehicle Routing Problem (VRP), a combinatorial optimization problem that seeks to determine optimal routes for a fleet of vehicles serving a set of customers. This study focuses on the three-echelon capacitated VRP with time windows (3E-CVRPTW) and presents a mixed-integer linear programming (MILP) formulation using an arc-based approach. Using a case study from a freight forwarding company in Sri Lanka, we applied a clustering-based heuristic to solve the 3E-CVRPTW. The findings highlight the effectiveness of multi-echelon routing strategies and contribute to the development of advanced solution approaches for VRPs in air freight distribution networks.

Keywords: Logistic planning, Mathematical programming, Optimization
Abstract: The global fashion and textile industry, while contributing significantly to economic growth, faces mounting criticism for its linear supply chain practices and associated environmental impacts. Despite growing interest in Circular Economy (CE) principles, there remains a lack of integrated frameworks that leverage emerging technologies to operationalize circularity in practice. Addressing this gap, this study reviews state-of-the-art technologies—including Artificial Intelligence (AI), Machine Learning (ML), and Deep Learning—and explores their potential to enhance recycling processes, material traceability, and waste management within the textile supply chain. Using a conceptual framework development approach, the study illustrates strategic pathways for technology integration to optimize resource efficiency, reduce waste, and promote circular supply chain management. It also highlights the critical need for supportive regulatory frameworks, targeted policies, and stakeholder collaboration to drive scalable, sustainable transformation. Future research directions are proposed to advance empirical validation and cross-disciplinary innovations supporting smart, circular, and sustainable textile supply chains.

Keywords: Decision making, Computational intelligence, Intelligent manufacturing
Abstract: This paper presents an extension of the weighted aggregated sum product assessment (WASPAS) method incorporating Ordered Fuzzy Numbers (OFNs) to enhance decision-making processes in discrete event systems for enterprise operational management. The proposed framework addresses the challenge of uncertainty in multi-criteria decision-making (MCDM) by integrating proper OFNs, ensuring the interpretability of results and avoiding improper OFNs that arise from arithmetic operations. The study introduces a modified OFN-WASPAS procedure that effectively manages decision criteria with dynamic trends while preserving computational efficiency. A practical implementation of the method in an automated warehouse system is provided. The selection of the parameter controlling the impact of trends in assessments enables flexibility in modeling decision-making scenarios. However, proper parameter selection remains a crucial aspect for further research to optimize the applicability of the proposed method.

Keywords: Assembly planning, Facility layout, Production planning
Abstract: This paper introduces Generative Assembly Line Design (GALD), a novel framework leveraging artificial generative intelligence to optimize task assignment, equipment selection, and line balancing in assembly line design. GALD integrates Variational Autoencoders (VAEs) and Generative Adversarial Networks (GANs) to address the complex interdependencies of assembly line design challenges. The framework minimizes total lifetime costs, encompassing worker wages, material handling, and equipment expenses, while adhering to stringent ergonomic and safety constraints. The methodology emphasizes a unified optimization approach rather than isolated problem-solving modules, with explicit consideration of ergonomic factors and demand-driven cycle times. The proposed GALD framework operates in two synergistic stages. The first stage utilizes VAEs to explore a diverse solution space, generating latent representations of feasible configurations guided by life cycle cost profiles. The second stage employs GANs for iterative refinement, where a generator proposes improved assembly line configurations, and a discriminator evaluates them for cost-effectiveness and compliance with operational constraints. This two-stage generative process ensures cohesive and efficient task-equipment assignments and balanced workloads. By synthesizing generative algorithms and industrial engineering principles, GALD offers a robust and cost-efficient solution to modern assembly-line design This study provides a potential for future extensions to incorporate human-robot collaboration, aesthetics, and sustainability metrics into the line design. It provides the potential for future research to incorporate even more considerations into the assembly-line design.

Keywords: Logistic planning, Autonomous system, Facility layout
Abstract: The modern-day complex supply chains and pressures of receiving correct orders as fast as possible have driven many remarkable advances in warehouse automation. Among the latest of them is the Multi-Tote Storage and Retrieval Autonomous Mobile Robot Systems (MTSRs), which promise flexibility, scalability, and less reliance on human labor. Based on this, the presented work analyzes the performance aspects of MTSRs with regard to enhancing efficiency in dynamic warehouse settings. A mathematical model is developed to evaluate the total fulfillment time and identify critical factors influencing operational outcomes. Sensitivity analyses show how system performance may be affected by warehouse design and resource allocation. The result shows that balanced layouts with optimized shape ratios minimize travel times and that the configuration of the aisles affects both storage accessibility and operational efficiency. Results point out that great care should be taken in planning and optimizing parameters in order to take full advantage of the benefits offered by MTSRs. The study positions MTSRs as a transformative technology for logistics, enabling warehouses to handle fluctuating demand and tighter delivery windows with greater agility. However, the research also points to key areas for further exploration, such as hybrid warehousing models that integrate MTSRs with other material handling systems and strategies for improving multi-robot coordination. Addressing these challenges in future studies will help improve the implementation of MTSRs so that they can meet the evolving needs of modern supply chains.

Keywords: Design engineering, Virtual reality, Digital manufacturing
Abstract: The growing demand for innovative and customized products is intensifying the need for flexibility in modern production systems, emphasizing the relevance of rapid validation before their physical commissioning. To address this challenge, virtual commissioning has emerged as promising solution. In this context, this study proposes an immersive virtual commissioning architecture that extends traditional approaches of virtual commissioning. Specifically, it introduces the validation of human-machine interactions as an additional element of the traditional virtual commissioning process, by enabling active human involvement during simulations via Virtual Reality. A preliminary validation of the architecture is conducted in a laboratory setting, where the logical and kinematic behaviors of a production machine, along with human-machine interactions, are tested. The immersive virtual commissioning architecture contributes to improve production systems design by incorporating human-machine interactions validation prior to the physical installation, unlike current practices, which address this aspect only after the physical deployment. Future work should investigate comparative analyses against existing virtual commissioning architecture to demonstrate the benefits of active human involvement and to quantify improvements in cost and time through application in industrial contexts.

Keywords: Machine learning, Process control, Intelligent manufacturing
Abstract: Artificial Intelligence (AI) is an evolving technology with the potential to open numerous new opportunities in several fields. Predictive maintenance is one of them, where it is possible to exploit the synergy created by the combination of AI and the Internet of Things (IoT), enabling the collection and analysis of a large amount of data in real-time. This paper presents a novel AI-based methodology for failure modes identification which has been developed and validated on a real case study in automotive. The process under consideration is the quality control of a pneumatic component, carried out through a specific quality control machine. Since this machine is naturally subject to variability and susceptible to failures, the issue addressed concerns the fact that, in case of unidentified failure of the machine, the result of the quality control can lead to a false scrap. In fact, the tested component may be erroneously reported as non-conforming due to failures of the quality control machine itself, even though the variables that characterise conformity respect the control limits. The proposed methodology is built on a preliminary risk analysis dedicated to the formal definition and prioritisation of machine failures. Then, two different but complementary AI-based models are proposed, taking as input data collected by machine’s sensors and information from machine’s maintenance log. The result is a solution that allows both the prevention of the quality control machine failures in real-time and the identification of possible false scraps. Moreover, it provides significant support from an operational point of view, enhancing the power of new technologies with an innovative and strongly data-driven approach aimed at reducing inefficiencies, towards a concrete intelligent manufacturing system.

Keywords: Digital twin, Data-driven modeling, Optimization
Abstract: This paper proposes a systematic methodology for the automatic generation of Digital Twin (DT) models in manufacturing systems with the purpose to support cost-effective adaptability and scalability. The methodology integrates multiple steps to ensure the efficient creation of DT models that accurately represent manufacturing physical systems. The methodology is tested in the Industry 4.0 Lab at the School of Management of Politecnico di Milano, showcasing its assumptions of modularity and reusability and the capability to support multiple reconfigurations of manufacturing systems. Future work will focus on rapid integration of services into DT models and the achievement of fully functional DT systems.

Keywords: Intelligent manufacturing, Process planning, Knowledge engineering
Abstract: This paper investigates the role of ontologies in fostering responsiveness of manufacturing systems within dynamic environments. As the manufacturing context undergoes a transformative shift from mass production to mass customization, the need for flexibility and adaptability becomes crucial. Focused on different applications derived from a systematic literature review, the study categorizes these cases based on the application of ontologies in manufacturing and their supported functions to critical manufacturing aspects. The findings further underscore how ontologies collectively enhance responsiveness in manufacturing by facilitating adaptation to dynamic environments, fostering efficient communication, and supporting rapid configuration and reconfiguration of production systems. As manufacturing advances toward intelligent and adaptable systems, ontological integration becomes a critical element, enabling interconnected machine networks to quickly navigate changes.

Keywords: Case study of digitization or smart system, Digital manufacturing, Internet of Things (IoT)
Abstract: This study assesses the qualitative impact of deploying 286 low-cost IoT digital solutions across over 100 manufacturing SMEs between 2023 and 2025. The focus is on supporting companies to gather insights into power consumption, environmental conditions, job tracking, and more by providing fully funded pre-assembled kits which could be fitted to existing machinery and production areas. By working closely with companies to conduct structured interviews and create detailed case studies, the findings suggest that these affordable devices can improve operational visibility, costs and health and safety practices. Feedback from companies indicated that the adoption of low-cost solutions was effective in de-risking the transition into digital technologies, with 75% of company responses indicating they would be interested in investing further. However, challenges regarding the level of technical support requirements and long-term use of data were identified. This initial research demonstrates the potential of low-cost technologies in driving digital adoption strategies, offering practical guidance and examples relevant to policymakers and industrial practitioners.

Keywords: Case study of digitization or smart system, Smart system, Process control
Abstract: Wire actuators made of Shape Memory Alloys (SMA) provide an extremely good force-to-weight ratio, the possibility of flexible system integration and silent operation, making them suitable for a wide range of applications. In this paper, a concept for the use of these actuators in self-regulating (smart) air outlets for a heating and cooling system in industrial metrology rooms is presented and realized as a prototype. The activation of individual air outlets spread across the room results in more even heating or cooling of the entire metrology room particularly for digital metrology and Coordinate Measuring Machine (CMM). The core features of the SMA actuator system within the air outlet are the independent operation from the main thermostat of the heating and cooling system as well as the energy-efficient mechanical and electrical design with currentless rest positions of the air outlet. To achieve a high level of functional integration, the SMA wires required for the actuating movement are embedded directly into the structural components of the individual air outlets inside Polytetrafluoroethylene (PTFE) tubes.

Keywords: Decision making, Biology inspired system, Case study of digitization or smart system
Abstract: The fibre-cement manufacturing significantly consumes energy in steam generation for the autoclave process, which impacts both operating costs and environmental sustainability. Traditional scheduling methods for autoclaves and boilers rely on human decision-making and face difficulties due to their complex nature and dynamic constraints. This study proposes an optimisation approach that utilises Particle Swarm Optimisation (PSO) algorithm to determine the autoclave and boiler schedule. The objective is to minimise energy consumption and costs associated with steam generation. This study uses computational experiments to evaluate the effectiveness of the PSO-based method compared to the traditional manual approach. The findings show the proposed algorithms have the potential to reduce steam consumption by 8.14% to 36.47% and reduce energy costs by 6.07% to 27.34%. Additionally, this algorithm retains a comparable makespan and avoids line unbalancing issues. The outcome of optimal scheduling also increased the utilisation of biomass boilers while reducing the utilisation of natural gas boilers. However, the study focused only on the PSO algorithm and assumed several fixed operational parameters. There is potential for future research to extend more dynamic and real-time scheduling challenges and explore alternative optimisation techniques to improve the optimisation performance.

Keywords: Manufacturing execution control, Mathematical programming, Smart system
Abstract: The paper presents an analytical method for predicting the behavior of cyclic manufacturing processes executed concurrently and sharing the resources of a flexible measurement and testing cell. The approach is verified by the example of a system with two concurrent and cyclic processes sharing one machine. The possible type of the system's steady-state, its cycle time, and the waiting times of the processes are determined based on the estimates of operation times and the established rule for prioritizing access to a shared resource in conflict situations using the computational algorithm. The effect of the perturbation of the selected operation's time on the system's cycle time is studied. An example of the proposed approach is provided for determining the system's cycle time and operation times, leading to the no-wait mode of execution, potentially lowering energy consumption and meeting sustainable manufacturing requirements.

Keywords: Condition-based maintenance, Smart maintenance, Data-driven modeling
Abstract: This publication presents innovative methods for assessing tool wear, based on the use of the PELT (Pruned Exact Linear Time) algorithm for change-point detection, bandpass filtering, and signal analysis. Selected frequency bands and a novel method for analyzing change points enabled precise forecasting of tool replacement timing. The proposed approach allowed for the identification of key stages of tool wear, supporting predictive maintenance and reducing production costs as an advanced element of the manufacturing systems health management and optimization.

Keywords: Condition-based maintenance, Fault detection, Monitoring
Abstract: In many production and assembly applications, conveyor belts are used to move components and materials. Those systems are usually operated by electrical motors in combination with a gearbox. To prevent downtimes of the production lines, condition monitoring is increasingly being used to predict failures of the systems components, for example in the drive train. This research aims to develop a method for detecting pitting in worm gearboxes as an indicator of failure in such applications by carrying out a modal analysis of the worm gear components and experimental investigation of an industrial worm gearbox with undamaged and damaged worm wheel by vibration measurement using an acceleration sensor.

Keywords: Digital twin, Diagnostics, Fault detection
Abstract: In recent years, the advancement of technologies supporting Industry 4.0 has garnered significant attention, with the Digital Twin (DT) concept emerging as a prominent area of focus. Although substantial research has been devoted to exploring various methods for implementing DT, a comprehensive and standardized framework applicable across multiple domains has yet to be developed. LIVE Digital Twin is a novel methodological approach with a promising framework to address these challenges. This paper presents the employment of Simple Structural Dynamics (SSD) as an important feature of LIVE Digital Twin. The study emphasizes the development of a Digital Twin for rotary systems, specifically for prognostic and predictive maintenance purposes. A case study involving a set of rotary components is conducted, where both low-fidelity and high-fidelity digital models are constructed. The high-fidelity model is created using commercial finite element software, while the low-fidelity counterpart is generated using SSD methodology. The analysis reveals a strong correlation between the two models, highlighting SSD's superior efficiency and computational speed compared to existing programs. These advantages position SSD as a viable tool for in-situ and real-time system analysis. Furthermore, the incorporation of Machine Learning (ML) algorithms can be used in the future to enhance SSD's calibration, enabling more precise predictions and facilitating the identification of errors within rotary systems.

Keywords: Process control, Process engineering
Abstract: A closed-loop control system is proposed to optimize an acoustic device performance by maintaining resonance frequency and desired power levels, thereby maximizing acoustic energy efficiency. The control in acoustic standing waves can be used to optimize the separation and/or manipulation of particles or droplets in an aqueous medium. Experimental results demonstrate the system's ability to compensate for variations in temperature and fluid properties, ensuring consistent resonance conditions and improved separation efficiency. The findings highlight the potential of this approach to enhance energy efficiency and process accuracy in industrial applications, while paving the way for further studies to validate its scalability and applicability in dynamic flow conditions.

Keywords: Decision making, Machine learning, Data science
Abstract: Accurately estimating the hours required for maintenance, repair and overhaul (MRO) operations in the aviation sector frequently depends on the experience and personal judgment of engineers, can lead to introducing errors, increased operating costs, and time-consuming decision-making. This work presents the development of a cost-effective application to monitor and predict MRO operations in an aeronautical company. The application integrates data-driven algorithms, particularly Machine Learning (ML), with Power BI to provide a dynamic and user-friendly visualisation of historical and predicted data, improving decision-making time and facilitating operational planning. The simple linear regression model was the most effective algorithm to predict MRO operation for the case study with a R2 of 0.81, balancing simplicity and performance compared to other analysed models.

Keywords: Intelligent manufacturing, Machine learning, Digital twin
Abstract: The rise of Industry 4.0 demands intelligent, adaptive, and standardized systems to enhance the interoperability and efficiency of industrial assets. A critical aspect of modern manufacturing involves integrating software-defined mobility assets into shop floors, enabling task execution through learning and interaction for improved performance. This paper proposes a dual approach to address these needs: (1) utilizing Asset Administration Shells (AAS) as a standardization framework for the digital representation of assets, enhancing adaptability, and (2) leveraging Federated Learning (FL) as a collaborative machine learning paradigm to overcome challenges of data privacy, scalability, and resource constraints in distributed environments, thereby facilitating automation. A decentralized architecture, xFedMQTT, is introduced, combining AAS to standardize digital assets with FL to enable collaborative model training across distributed nodes. The implementation features YOLOv8, a state-of-the-art object detection framework, to evaluate the feasibility of on-device FL in real-time industrial applications. Experimental results demonstrate the effectiveness of this approach, emphasizing its potential for scalable, privacy-preserving, and efficient industrial operations.

Keywords: Machine learning, Decision making, Case study of digitization or smart system
Abstract: In the current manufacturing landscape process optimization, cost reduction, and profit maximization have become crucial. This paper aims to leverage advanced data analysis techniques such as Machine Learning (ML) to achieve these objectives, identifying issues related to operational inefficiencies, data integration challenges, and the effective application to real manufacturing problems. To do so, the paper proposes a framework to improve ML integration into manufacturing operations, highlighting the importance of customizing the strategy to suit the company needs. Through a practical implementation, this paper shows the applicability of the framework.

Keywords: Machine learning, Robotics, Intelligent agent
Abstract: This paper presents the development of a comprehensive methodology for enabling humanoid robots to perform complex human-observed activities. The proposed approach integrates motion pattern analysis, embedding space representations, and ensemble learning techniques to capture and reproduce natural human movements. The framework utilizes pattern frequency mapping, conditional probability distributions, and multi-dimensional embeddings to analyze and represent human motion sequences. The methodology incorporates a novel linear feature extraction method that maximizes separation between motion patterns and anti-patterns in the embedding space, combined with an ensemble learning architecture that ensures robust pattern recognition. Experimental results demonstrate the system's ability to effectively distinguish between natural movement sequences and irregular patterns, providing a foundation for teaching humanoid robots human-like movements while maintaining motion naturalness and physical feasibility.

Keywords: Process planning, Machine learning, Machining process
Abstract: This paper presents the concept and development of the Smart Aided Process Planning (SAPP) system, designed to address the limitations of traditional CAPP systems by leveraging machine learning techniques. The SAPP system features a modular architecture that includes a Hybrid Expert System and modules for Supervised Learning, making it an intelligent tool for CNC machining process planning. Input data are automatically downloaded from the STEP format, and proprietary solutions have been developed for automatic data loading and feature recognition. The system generates machining process plans, stores output data in an extended STEP-NC format, and utilizes predictive models to optimize process structure and parameters. The paper also discusses the creation of training sets, the application of machine learning for process generation, and the development of predictive models for various machining operations. The results indicate a promising direction for the development of CAPP systems, with potential industrial applications and benefits in efficiency, cost reduction, and sustainability.

Keywords: Additive manufacturing, Process control, Inspection
Abstract: Additive manufacturing is being increasingly focused on the production of end-use parts. Compared to the prototyping application, the production of end-use parts demands a higher level of repeatability and process quality. It has to be proven that part properties meet the requirements. This is especially demanding for properties that are evaluated by destructive test methods, like most mechanical properties. The paper shows the application of Design of Experiments DoE to analyze the interrelation between the physical and mechanical properties of additively manufactured parts made from CoCr alloy. Components with different porosity are produced by varying the production parameters. The tensile strength, the 0.2% proof stress, the Young's modulus, and the elongation at break are determined in the tensile test. In addition, the porosity is determined by measuring the density using the Archimedean method. It is shown that a strong correlation exists between the porosity and the tensile strength as well as the 0.2% proof stress for the parameter sets under validation. It can therefore be shown, that in the present application, the non-destructive evaluation of the part porosity can be applied to validate the mechanical properties of the product.

Keywords: Machine learning, Digital manufacturing, Information system
Abstract: To a large extent, the digitalization of smart manufacturing processes relies on efficiently extracting information from engineering documents. These documents often exhibit complex heterogeneous layouts, such as dense tables, small margins, and inconsistent formatting, that machines struggle to interpret. Recent developments in artificial intelligence may change this. We set out to evaluate the readiness of AI-based approaches for extracting information from engineering documents. In this work, we propose a general framework (MERI) for information extraction and evaluate it on a synthetic dataset explicitly created for parameter extraction from engineering documents. We find that converting documents into a machine-friendly intermediate format improves the performance of information extraction methods. However, further improvements are still necessary before AI systems can extract parameters reliably.

Keywords: Case study of digitization or smart system, Intelligent manufacturing, Vision systems
Abstract: The adoption of artificial intelligence applications in manufacturing is experiencing a growing trend in the last years. The impact of advanced technologies, including AI, to quality management has gained momentum as well leading to the Quality 4.0 stream of research. Despite the potential benefits, several challenges can prevent the achievement of the expected results in SMEs more than in their larger counterparts. This paper, far from aspiring to deliver a statistical analysis, pursues an explorative aim and provides empirical evidence on the adoption of AI solutions for quality control in manufacturing SMEs by discussing a sample of 7 pilot cases that participated to the Open Call of an EU-funded project. Pilots are analysed in terms of type of adopted AI, AI-human interactions approach and achieved benefits. The empirical experience gained allowed to discuss practical challenges faced by SMEs and, hence, to derive some guidelines for practitioners. Obtained results could be further refined and categorised to structure empirical knowledge on the use AI for quality in larger samples.

Keywords: Additive manufacturing
Abstract: The Laser Powder Bed Fusion (LPBF) process offers flexibility in manufacturing complex geometries but faces challenges from non-systematic sources of uncertainty that lead to Geometric and Dimensional Deviations and Surface and Material Degradation. This study investigates a range of uncertainty sources through seven experiments, classifies them, and discusses how they can be potentially avoided using generative artificial intelligence. The results highlight the variability introduced by part setup, aspect ratio, and orientation. The findings underscore the critical need for experimental validation and the potential role of Machine Learning in addressing non-systematic sources of uncertainty to improve print reliability and quality in LPBF processes.

Keywords: Human factors, Human interaction
Abstract: This paper addresses the strategic need for upskilling and reskilling human operators to meet the demands of advanced manufacturing, where automation, artificial intelligence, and digitalization are reshaping production. Beyond digital literacy, data analytics, and complex problem-solving, advanced manufacturing requires proficiency in robotics and automation, cybersecurity awareness, additive manufacturing, systems thinking, programming, project management, adaptability, and interpersonal collaboration, and this is only a partial list of skills. However, future advancements, technological and others, require a proactive strategy to harness the full potential of intelligent manufacturing systems. Keeping relevant skills and developing new important skills is a great challenge. This paper contributes a comprehensive framework for workforce development, emphasizing continuous learning and skills adaptation to ensure resilience in this disruptive technological environment. By aligning educational curricula with industry needs and promoting lifelong learning initiatives, reskilling and upskilling, stakeholders can ensure that human operators remain integral to the manufacturing ecosystem. We discuss and illustrate how successful strategic workforce development can enhance productivity and innovation. We also highlight effective cross-sector collaborations among industry, academia, and government to build sustainable talent pipelines. Future research may explore the effectiveness of specific skill training programs, reskilling initiatives, investigating personalized learning pathways, and examining AI-driven training technologies that facilitate rapid skill acquisition. Longitudinal studies could further assess the career impact of reskilling on workforce adaptability and retention, while comparative studies could help identify skill needs across sectors and regions.

Keywords: Intelligent agent, Intelligent manufacturing, Data science
Abstract: This paper presents a multimodal question-answering system designed to enhance user interaction with machine tool manuals. The system integrates text and image datasets extracted from manuals, combining Retrieval-Augmented Generation (RAG) for textual responses with image retrieval and generation using diffusion models. By leveraging ranking algorithms for image selection and generating enhanced visuals, the system provides comprehensive answers comprising text instruction, images selected from manual, and generated images. Experimental results demonstrate the system’s capability to deliver precise and visually supported responses, improving accessibility and usability for industrial users.

Keywords: Decision making, Human factors, Intelligent manufacturing
Abstract: This paper presents a co-creative methodology for the design of human-AI teaming in decision-making for dynamic environments, introducing a range of human-AI teaming design patterns, applicable to diverse domains. The methodology integrates aspects of systems design and enriches them with a typology of human-AI teaming in decision-making. It engages stakeholders in decision-making processes for the joint identification of decisions, targets, success metrics, and associated risks. This is enabled by co-creation design patterns, as part of an agile methodology that includes iterative cycles of physical and virtual collaboration, as well as synchronous and asynchronous activities between parties involved in the design, development, testing, and use of the system. The methodology is applied in a multiple case study and lessons from a manufacturing case are presented from the first phase of implementing the methodology. Keywords: AI decision making; Human – AI teaming; human-centric AI; co-creation.

Keywords: Additive manufacturing, Artificial life, Computer aided manufacturing
Abstract: W artykule przedstawiono koncepcję i implementację Inteligentny adaptacyjny system sterowania dedykowany do zadań synchronizacja w procesach produkcyjnych. System jest oparty na innowacyjnej architekturze wykorzystującej autonomiczne Moduły eAssistant (eA), które tworzą zdecentralizowaną sieć komunikacyjna umożliwiająca dynamiczną optymalizację parametry procesu w czasie rzeczywistym. Rozwiązanie integruje się zaawansowane technologie, takie jak uczenie maszynowe, sztuczna inteligencji i predykcyjnych modeli optymalizacyjnych, tworzenia Kompleksowe narzędzie wspierające poziom operacyjny procesów decyzyjnych. Weryfikacja sprawności systemu przeprowadzono w rzeczywistych warunkach produkcji, przy stanowisko obróbcze obróbcze piasty żeliwne na dwóch współpracujące tokarki CNC. System eAssistant zapewnił Precyzyjna synchronizacja toczenia zgrubnego i wykańczającego poprzez ciągłe monitorowanie i dostosowywanie parametry procesu. Uzyskane wyniki wykazały znaczna poprawa efektywności procesu produkcyjnego. Przedstawione rozwiązanie stanowi praktyczny wdrażania koncepcji Przemysłu 5.0, demonstrując Skuteczność systemów inteligentnych w optymalizacji złożonych procesy obróbki skrawaniem.

Keywords: Human factors, Smart maintenance, Human interaction
Abstract: The evolution towards smart manufacturing is leading to opportunities and challenges resulting from rapidly changing production demands, frequent changeovers, and short time to market of new products. In this context, the role of maintenance has become increasingly complex and critical. This paper focuses on the changing role in light of advanced technologies and skills that are essential for modern maintenance practices to become a resilient maintenance operator. The integration of advanced technologies is explored, including Computerized Maintenance Management Systems, Artificial Intelligence, Digital Twins, Mixed Reality, and chatbots. In addition, the importance of skills is emphasized in a broad range consisting of technical proficiency, problem-solving abilities, project management expertise, adaptability, teamwork, communication, and the digital skills to adopt advanced technologies to optimize maintenance processes and drive continuous improvement. The discussion highlights how technologies and skills not only enhance operational efficiency, but also support the development of resilient maintenance strategies capable of coping with dynamic production environments. By examining the impact of technologies and the necessary skills for maintenance professionals by multiple research queries and literature review, this paper provides a conceptual framework for the resilient maintenance operator with the identified skills paired with the key technologies, the framework focuses on the changing role in light of advanced technologies and skills that are essential to become a resilient maintenance operator within modern maintenance practices.

Keywords: Machine learning, Information system, Computational intelligence
Abstract: Automated analysis and processing of textual requirements offers great potential for more efficient quotation processes in machine engineering. The assignment of requirements to discipline-specific expert groups for an evaluation of customer requirements can be understood as multi-label classification. Traditional machine learning approaches have been studied in detail for this purpose. However, it is obvious that transformer models can also be used effectively; however, no such analysis exists yet. Therefore, this paper presents the results of a comparative study on the performance of neural networks, statistical machine learning algorithms, and transformer models in the context of industrial requirements in machine engineering. The results show the superior performance of transformer models and, in particular, their robustness to imbalances within the training dataset, as is often the case in real-world applications.

Keywords: Agent-based approach, Intelligent agent, Case study of digitization or smart system
Abstract: The calculation of reliability metrics is essential for the evaluation of complex systems, their system behaviour and economic efficiency. This paper examines the possible applications of Large Language Models (LLMs) in dynamic reliability calculations. To achieve this, an LLM system is being developed using an agent-based approach and user-defined tools within the Python-framework LangChain. The system uses an OpenAI GPT model for general language processing functionalities, while the extended, specific functionalities for reliability calculations are defined in tools and the system template. In general, the agent is able to make queries to the user and select appropriate calculation tools. Due to the nature of an LLM and the known problems of hallucination, a critical analysis of the generated answers is recommended. The developed concept and the prototype emphasize the advantages of using an LLM-agent-system to query data dynamically, and thus underline the dynamic aspect of the reliability calculation.

Keywords: Process engineering, Data science, Machining process
Abstract: This article discusses classic and today's approaches to business process improvement. The integration of modern tools, such as Process Intelligence, offers new opportunities by using data-driven insights to improve processes. In addition, a review of current applications of Generative Artificial Intelligence (Generative AI) demonstrates the potential to increase efficiency in domains ranging from human resources and e-commerce to finance and healthcare. Despite these advances, the article emphasizes the importance of a holistic approach that combines new technologies with proven methodologies. This ensures not only operational efficiency, but also strategic alignment with organizational goals.

Keywords: Robotics, Autonomous system, Machine learning
Abstract: This paper presents the development of cognitive architectures for humanoid robots operating in smart factory environments, focusing on enabling robots to perform complex manual control tasks. The research introduces novel methods for encoding robot physical states and representing process control information in cognitive space. The proposed approach combines different encoding methods for robot configurations and states with representation methods for information structures, enabling comprehensive cognitive constructs generation. Experimental evaluation demonstrates the effectiveness of the developed cognitive architecture in bridging the gap between robot physical capabilities and abstract process control requirements. The results show that the proposed solution enhances humanoid robots' ability to handle complex technological processes traditionally performed by human operators.

Keywords: Computer vision, Computational intelligence
Abstract: This paper presents an approach to combining point clouds from multiple stereo cameras. The developed method faculitate the selection of point cloud granularity, which allows for flexible adjustment of the level of detail in 3D models for various applications, with a special focus on spatial visualization. The fusion/combining process includes geometric transformations, as well as iterative and noise reduction techniques to integrate visual data from multiple cameras. The effectiveness of the proposed algorithms was validated using two depth cameras such as ZED 2 and Intel RealSense D435f (the latter with an infrared pass filter). The conducted study showed great potential for the application of such solutions in automation, robotics, industrial production management, reverse engineering, and medical imaging.