IN4PL 2024 Abstracts

Area 1 - Industry of the Future

Full Papers

Paper Nr:	13
Title:	Multi-Agent Deep Q-Network with Layer-Based Communication Channel for Autonomous Internal Logistics Vehicle Scheduling in Smart Manufacturing
Authors:	Mohammad Feizabadi, Arman Hosseini and Zakaria Yahouni
Abstract:	. In smart manufacturing, scheduling autonomous internal logistic vehicles is crucial for optimizing operational efficiency. This paper proposes a multiagent deep Q-network (MADQN) with a layer-based communication channel (LBCC) to address this challenge. The main goals are to minimize total job tardiness, reduce the number of tardy jobs, and lower vehicle energy consumption. The method is evaluated against nine well-known scheduling heuristics, demonstrating its effectiveness in handling dynamic job shop behaviors like job arrivals and workstation unavailabilities. The approach also proves scalable, maintaining performance across different layouts and larger problem instances, highlighting the robustness and adaptability of MADQN with LBCC in smart manufacturing.

Paper Nr:	17
Title:	Advanced Process Monitoring and OEE Metrics: Leveraging AASs for Efficiency
Authors:	Aaron Zielstorff, Dirk Schöttke, Fiona Buettner, Thomas Kämpfe and Stephan Schäfer
Abstract:	The article shows a way of collecting data to prepare the Overall Equipment Effectiveness (OEE) using the asset administration shell (AAS) in a case study. A systematic approach and the use of established methods are essential for this. The established KPI metric for evaluating equipment effectiveness has proven itself and is used across all industries for quantitative productivity measurement. The digitalization of systems generates a large amount of data from the production environment, which forms the basis for modern analysis methods. These enable, among other things, the detection of anomalies in the plant environment and an increase in machine effectiveness. With the increasing establishment of the AAS, there is a need for pragmatic integration of data that supports the process of asset evaluation and optimization. In this context, the use of the “Time Series Data” submodel offers an effective solution. It defines a uniform standard for the integration and semantic description of time series data. The effective integration of time series data into the AAS environment, which enables the comprehensive use of generated process data, is explained using an example.

Paper Nr:	20
Title:	Machine Learning Tool for Yield Maximization in Cream Cheese Production
Authors:	Loic Parrenin, Ambre Dupuis, Christophe Danjou and Bruno Agard
Abstract:	Artificial intelligence tools and data collection on the shop floor are enhancing flexibility and productivity in industry, addressing labor shortages and skills attrition by leveraging the tacit knowledge of workers. This study focuses on the cream cheese production sector, where operator expertise is essential for controlling the ultrafiltration concentration factor, a critical parameter affecting product moisture content. To ensure continuous and flexible production despite workforce challenges, a machine-learning tool was developed using the CRISPDM approach to maximize cream cheese yield on a Canadian production line. A decision tree algorithm applied to real production and quality data yielded promising results, with an RMSE of 0.061 and an R² of 0.91 when predicting the ultrafiltration concentration factor used by an experienced operator to maximize yield while complying with quality standards. The implementation saw positive operator acceptance due to comprehensive training and an inclusive approach. This research marks a pioneering effort to harness tacit knowledge in the dairy industry for machine parameter control, highlighting data acquisition and quality as key areas for further investigation to enhance tool performance and adaptability.

Paper Nr:	24
Title:	Multiscale Clustering to Improve Anomaly Detection in Nuclear Equipments
Authors:	Amaratou Mahamadou Saley, Thierry Moyaux, Aicha Sekhari, Guillaume Bouleux, Vincent Cheutet and Jean-Baptiste Danielou
Abstract:	In the nuclear industry, the relentless pursuit of operational excellence and reliability is paramount, especially considering the critical importance of safety and efficiency. This sector, a key component of global energy supply, is constantly driven to innovate and optimize its equipments. Early fault detection plays a central role in this endeavor, as it not only helps to prevent potential incidents, but also maximizes production. Traditional anomaly detection methods, primarily based on clustering algorithms, often overlook minor yet warning faults. Addressing this, our paper introduces a multiscale clustering approach, advancing beyond classical methods. This methods achieves a satisfactory classification of anomalies and culminates in a silhouette score of 90%. Additionally, it facilitates the computation of a preventive maintenance indicator.

Paper Nr:	40
Title:	A Framework for Resilient Integration of Industry 4.0 Components into Production Systems
Authors:	Héctor Hostos, Virginie Goepp and Patrick Sondi
Abstract:	Industry 4.0 has made possible the integration of new technologies, also called 4.0 components, into production systems in order to enhance their productivity and performance in general. However, this integration has brought new types of risks that have to be mitigated to avoid production systems to stop operating due to a failure of the 4.0 components. In this context, resilience, defined as the capability and ability of an element to return to a stable state after a disruption, represents a promising domain to explore. For industrial systems and Industry 4.0, current works dealing with resilience focus on specific types of disruptions and technologies without providing a general framework. This article proposes a general framework to integrate 4.0 components into production systems in a resilient manner. The framework is inspired by the concepts of resilience in socio-ecological systems in an attempt to provide production systems the ability to overcome unexpected disturbances and persist in time. This includes the adaptation of the concepts of state variables, stability landscape and the aspects of resilience (resistance, latitude and precariousness). It has a service-oriented architecture that connects the 4.0 components to all the services that they enable in the production system. This allows to define several system states according to the service levels enabled by the 4.0 components. The potential use of the framework is illustrated on the Internet of Things (IoT) as a specific 4.0 component.

Paper Nr:	45
Title:	A Comprehensive Framework Integrating ML, Automation Pyramid, and KPIs for Industry 5.0
Authors:	Pedro Ponce, Brian Anthony, Russel Bradley, Wenhao Xu, Juana Isabel Méndez and Arturo Molina
Abstract:	The manufacturing industry continually seeks advanced technologies to enhance performance per evolving customer requirements. Machine learning (ML) emerg-es as a pivotal assistive technology essential for strategic integration with Key Performance Indicators (KPIs). Traditionally, KPIs monitor and measure indus-trial system performance. This paper proposes a framework leveraging KPIs to integrate ML across the automation pyramid in Industry 5.0. The framework ena-bles early detection of malfunctions and areas for improvement, preventing productivity loss. Validated across various industries, the framework demon-strates enhanced operational efficiency, sustainability, and human-centric benefits. Information and Communication Technologies advancements facilitate real-time data collection and analysis, aligning with ISO 22400 standards for manufactur-ing operations management. ML techniques generate actionable insights crucial for sustainable development in industries such as automotive, which require ho-listic goal assessments. Industry 4.0 marked a significant shift towards automa-tion and data exchange, leveraging IoT, cloud computing, and big data analytics. Industry 5.0 emphasizes human-machine collaboration, customization, and sus-tainability, evolving KPIs to include worker satisfaction, customization capabili-ties, and social and environmental impact metrics. This evolution spans various sectors: manufacturing, pharmaceuticals, retail, e-commerce, high-energy-use in-dustries, and consumer goods. ML minimizes downtime, enhances product quali-ty, optimizes supply chains, and improves worker safety by analyzing data from wearables and sensors. Integrating ML with KPIs in Industry 4.0 and 5.0 enables industries to be more efficient, adaptive, and responsive to market and environ-mental changes, improving decision-making, operational efficiency, and align-ment with business and sustainability goals.

Paper Nr:	49
Title:	Unleashing the Potential of Agility, Resilience and Business Continuity: A Systematic Literature Review
Authors:	Kunruthai Meechang, Margherita Pero and Khaled Medini
Abstract:	Challenges in the recent business environment have forced manufacturers to adapt production processes while sustaining operational continuity in response to changes in consumer behaviors and the working environment post-pandemic. Even an agile company struggled to survive during COVID-19 due to restriction measures causing supply chain disruption. Thus, there is a need to extend capabilities that allow companies to be fast and flexible despite a crisis. This paper aims to investigate the potential of three manufacturing paradigms, namely agility, business continuity, and resilience, for adverse situations. A systematic literature review has been conducted to analyze existing studies and build a holistic under-standing of how they are complemented manufacturing. The results indicate the relevance of the paradigms that strive to survive through changes from different perspectives: agility focuses on quick adaptation, business continuity keeps products or services delivered, and resilience recovers operations from disruptions. Integration of these paradigms is a promising approach to surviving through crisis, adapting manufacturing, and growing competitiveness for future industries. In addition, the enablers for paradigms are identified, including process, strategy, technology, and people. The knowledge gathered from this study enables future researchers to explore an explicit integration from both conceptual and operational points of view.

Paper Nr:	82
Title:	Integrating Digitalization and Sustainability: An Innovative Approach to Assess Digitainability in Manufacturing
Authors:	Pietro Andrea Miciaccia, Fabrizia Devito, Concetta Semeraro and Michele Dassisti
Abstract:	Manufacturing faces significant sustainability challenges due to the intensive use of natural resources. Recently, scholars and practitioners have developed numerous tools to assess the environmental, economic, and social impacts of manufacturing, grounded in the Triple Bottom Line (TBL). Despite these collective efforts, the application of these tools in real-world manufacturing scenarios remains limited in terms of scalability (with few focusing on the manufacturing process or system), applicability (many are tailored to specific manufacturing sectors), and the comparability of the results obtained. Simultaneously, the adoption of the Best Available Techniques (BAT) concept has introduced a new benchmark for sustainability performance against which to compare in terms of sustainability per-formance. In Industry 4.0 and 5.0, rethinking sustainability assessment as a tool to guide management in the digitalization process becomes necessary. Therefore, the concept of “digitainability”, i.e. the convergence of sustainability and digitalization, will drive new manufacturing contexts. This paper presents an innovative approach by which to assess sustainability performance and integrate to it the paradigms imposed by digitalization in the vision of the factory of the future. The DigiSustain Manufacturing Assessment Model (DSMAM) offers a comprehensive framework for evaluating sustainability and provides insights for redesigning and reengineering manufacturing systems, identifying areas for digital innovation.

Paper Nr:	85
Title:	Training Operator in VR: A Scalable Solution for the Creation of VR Training Scenes
Authors:	Geoffrey Melzani, Tony Quach, Henrik Söderlund, Dan Li, Puranjay Mugur and Björn Johansson
Abstract:	In the automotive industry, just like in other complex product-based industries, fierce competition is ongoing, and companies must cope with the challenges. In such industries, operator training is a cornerstone, as many of the most important product manufacturing processes rely on human-performed assembly. Thus, with the factory environments that are getting increasingly complex, and efficien-cy-dedicated, the importance of training is enhanced. With the generalization of Virtual Reality (VR) technologies, VR operator training emerges more rapidly. However, the many advantages of VR training, such as safety, cost, and flexibil-ity, have not yet been fully realized in the industry as large-scale implementations still have not been reached. Creating VR training scenes is still time-consuming, and therefore expensive, hindering large-scale implementation and adaptation. This paper first recapitulates the data needed to populate such VR training scenes, then it exemplifies the automatic generation of VR training scenes through an in-dustry use case, enabling large-scale automated implementation. Finally, it high-lights the challenges related to data availability and handling and that companies can encounter on the way to the large-scale implementation of VR training.

Short Papers

Paper Nr:	16
Title:	An Out-of-Sample Clustering Ensemble Method for Defect Detection and Classification in Metal Additive Manufacturing
Authors:	Sylvain Chabanet, Adil Han Orta and Mathias Kersemans
Abstract:	Unsupervised learning methods, and in particular clustering algorithms, have found many applications in manufacturing, ranging from customer segmentation to quality monitoring. It has, however, been demonstrated that no clustering algorithm can be suitable for all applications and data structures. Ensembles of clustering algorithms have emerged as a partial answer to this limitation, aiming at increasing the robustness of clustering algorithms by aggregating partitions discovered by many models. This robustness, however, comes at the cost of increased computational requirements to generate and aggregate partitions. The ability to quickly predict a cluster for new, out-of-sample data points without having to recompute the whole clustering algorithm from scratch is, therefore, a desirable property for many real-world applications. Such out-of-sample methods, however, are not straightforward in the context of clustering ensemble, and few models include one. As a step toward filling this gap, this article proposes a novel out-of-sample method for clustering ensemble algorithms following the median consensus framework. An application of this method is proposed for the detection and classification of defects in metal parts produced by additive manufacturing processes. The proposed method is compared with state-of-the-art algorithms on both artificial and experimental datasets, demonstrating its high performance and robustness.

Paper Nr:	21
Title:	Comparative Evaluation of Irregular Shape Strip-Packing Algorithms
Authors:	Niccolò Giovenali, Giulia Bruno and Paolo Chiabert
Abstract:	The nesting problem of 2D shapes, which has impactful application in the cutting and packing fields, has been studied for many years. Previous papers are mainly focused on proposing new algorithms and prove their efficiency in terms of pack-ing density or computation time. However, the results are reported only on few datasets and the comparison is done only with respect to few competing algo-rithms. The aim of the paper is to analyse and compare the results obtained by strip-packing algorithms published in the last 20 years. The results show that the effectiveness of the algorithms varies widely across different datasets, and there is a lack of comprehensive benchmarking that considers both the quality of solu-tion and the computational time required to achieve it. Furthermore, since no algo-rithm clearly outperforms all the others, further methods to address the nesting problem with reinforcement learning and neural networks could be investigated to improve the generalization ability on the nesting problem.

Paper Nr:	31
Title:	Building Realistic Environment from Computer Vision Approach Applied to Manufacturing Simulation in the Digital Twin Context
Authors:	Marcelo Rudek, Ana P. R. Valle and Ricardo Bertolin
Abstract:	Digital twins to manufacturing simulation require a complete virtual representation of the environment for immersive human interactions. From computer vision techniques, it is possible to model the respective real scenery based on a set of digital images. This paper addresses a strategy to create virtual layouts through 3D reconstruction from photogrammetry. The main context is to help humans to be trained using synthetic information, preparing their actions to perform the respective real operations of productive processes as aligned with industry 5.0 requirements. Here we present a selection of the emerging technologies categorized to perform the shop-floor layout digitalization as a first stage of a digital twin creation of a real industrial scenario as a starting point to the industrial metaverse architecture. Also, we present an experimental digitalization operation of a complex scenery and the respective discussion about its use in manufacturing simulation.

Paper Nr:	33
Title:	A Multi-Objective Genetic Algorithm Approach for Multi-Component Products Recovery and Remanufacturing Planning
Authors:	Latifa Belhocine, Mohammed Dahane and Mohammed Yagouni Mohammed Yagouni
Abstract:	The remanufacturing process has gained recognition primarily for its effectiveness in addressing environmental concerns related to End-Of-Life (EOL) and End-Of-Use (EOU) products. Consequently, a growing number of companies specialise in remanufacturing various product types. This practice not only prolongs product lifespan but also reduces manufacturing costs. This paper examines the challenges encompassing all stages of the remanufacturing process: product recovery, transportation, and remanufacturing operations for customers with similar product types over a finite horizon. The problem involves planning the recovery of used products for remanufacturing and grade enhancement. The main decisions include selecting customers for product recovery and replacement, optimising transportation for used product retrieval, and making decisions for the post-remanufacturing grade. The objective is to minimise both economic and environmental costs. To address this, we propose an NSGA-II (Non-dominated Sorting Genetic Algorithm) based multi-objective solution approach to tackle this problem.

Paper Nr:	36
Title:	An Innovative Fault Detection Robotic Tool for Overhead Cranes in Industries: Magnetic Wheel Modelling and Experimental Validation
Authors:	Arun Kumar Yadav and Janusz Szpytko
Abstract:	In this paper, we present the innovative robotic tool design and modelling, validation of a new family of climbing robots that are capable of adhering to vertical surfaces through permanent magnetic wheels. The robotic system is composed of two modules, a sensing module, and magnetic wheel module which are arranged in a sandwich configuration, with the surface to climb inter-posed between them. To perform the inspection task, few magnetic wheeled climbing robots have been proposed for many industrial applications of ferro-magnetic structures. To achieve a reliable system, good payload abilities, and minimize the power consumption of robot, the design of magnetic circuit and calculation of adhesion forces of magnetic wheel are drive factor to achieve all these. In this paper a four permanent wheeled robot is proposed to climb on the bridges and girders of the overhead cranes for health monitoring. An improved design of magnetic wheel is presented in this paper. Finite element method is performed on the wheel to get the magnetic flux distribution and calculate the attractive forces between the wheels and inspection area. By the simulation distribution of magnetic flux lines were compared on the plane inspection area and a curved concave and convex surface. By keeping the main objective of recognizing the fault into the ferromagnetic structures a wireless robotic system is proposed for the over-head cranes.

Paper Nr:	38
Title:	Enhancing Human – Robot Collaboration in the Industry 5.0 Framework with Physics-Informed Neural Networks: Application to Collision Detection
Authors:	Francesco G. Ciampi, Thierno M. L. Diallo, Faïda Mhenni, Stanislao Patalano and Jean-Yves Choley
Abstract:	In the context of rapidly evolving technology and increased attention to social and environmental dimensions, the industrial sector is transitioning from Industry 4.0 to Industry 5.0. This new paradigm emphasizes human centrality in highly auto-mated environments, necessitating the exploration of collaboration mechanisms between humans and robots. This study investigates the application of Physics-Informed Neural Networks (PINNs) to enhance Human-Robot Collaboration (HRC). PINNs integrates the traditional data-driven approach based on machine learning models with a prior physical knowledge of the system, providing a valu-able solution when data are scarce or physical models are too complex or incom-plete. The study first focuses on the main aspects of interest in the field of HRC through a state-of-the-art analysis, evaluating the application of Physics-Informed Neural Networks (PINNs) in HRC and robotics. It then delves into the definition of PINNs and their implementation. Finally, as a proof of concept, the model is applied to a case study concerning collision detection in a 6 DoF robotic arm. This is achieved by predicting the joint currents and comparing them with the measured values to identify the contribution due to external forces such as collisions. The results demonstrate that the PINNs model outperforms a traditional neural network, achieving an average error below 10%. Additionally, the collision detection application shows an f1_score of 0.80, indicating strong performance.

Paper Nr:	39
Title:	AI-Driven Smart Air Conditioning System for a Sustainable and Energy-Efficient Industrial Future
Authors:	Cherifa Nakkach and Yvan Picaud
Abstract:	This paper examines short-term predictions of industrial air-conditioning loads using the data from smart meters. Using IoT, data can be collected from sensors that measure temperature, humidity, and other relevant parameters. This enables real-time monitoring of environmental conditions. We present a comprehensive architecture for an energy-efficient and sustainable solar air conditioning system for an efficient industrial future. To optimize operational efficiency and reduce energy costs, this system must predict energy consumption. Various machine learning models were explored and tested, including Convolutional Neural Networks - Long Short-Term Memory (CNN-LSTM), Recurrent Neural Networks (RNN), Gated Recurrent Units (GRU) and TimeGPT to identify the most effective approach for energy prediction. RMSE and MAPE metrics showed that the TimeGPT model outperformed all other models in terms of accuracy and reliability in forecasting energy consumption. TimeGPT’s results provide evidence that industrial environments can improve their energy efficiency and sustainability by using the TimeGPT model. In addition to better energy management, this approach reduces costs and reduces carbon footprints. Energy consumption can is therefore predicted to optimize air conditioner operation, plan preventive maintenance, detect potential problems, and take corrective action to reduce the unnecessary consumption of energy.

Paper Nr:	43
Title:	A Reinforcement Learning Algorithm for Dynamic Job Shop Scheduling
Authors:	Laura Alcamo, Giulia Bruno and Niccolò Giovenali
Abstract:	The job shop scheduling problem, a notable NP-hard problem, requires scheduling jobs with multiple operations on specific machines in a predetermined order. A strong assumption is that all the information of the manufacturing environment is known in advance and there is no modification during the scheduling process. However, the real-world environment is significantly affected by uncertainties. The dynamic job shop scheduling is a variant of the job shop scheduling problem in which the scheduling environment is subject to changes over time including variations in job arrival times, processing times, machine breakdowns, resource availability and job priority. To address this issue, this paper presents a single-agent reinforcement learning algorithm, which implements a proximal policy optimization that uses masking to reduce the search space and improve efficiency. The algorithm was tested in both deterministic and dynamic environments and compared to traditional scheduling methods. The results demonstrate that the proposed approach is comparable to traditional methods in deterministic cases and outperforms them in dynamic environments. These findings emphasize the potential of reinforcement learning in addressing and optimizing complex scheduling challenges.

Paper Nr:	46
Title:	A Novel Pipeline for Data Management and Analysis that Integrates Data Lakehouse Architecture into the Aeronautics Industry
Authors:	Nelson Freitas, Diogo Vaqueira, Andre Dionisio Rocha, José Barata, Fábio Serrano, Luís Silva and Manuel Madeira
Abstract:	Currently, the majority of processes or ssystems aim to benefit from the data produced by their own or other relevant systems, with the objective of increasing efficiency. This is especially true in the field of industrial systems, where a multitude of devices attempt to publish their metrics and data into the system, often resulting in characteristics that can be classified as big data. However, companies often struggle with the correct and useful utilization of this harvested data. Therefore, this paper focuses on a use case of a data pipeline system with a data lakehouse in an airplane parts factory. The developed architecture shows that with some adjustments to the classic data lakehouse architecture, it is possible to achieve higher parallelism in order to simultaneously store data in the data lake and data warehouse. Additionally, a visualization tool was developed to highlight how metric calculation and outlier detection can be automated or facilitated with the utilization of data, as opposed to manual labor.

Paper Nr:	50
Title:	Accelerating Industry 4.0 and 5.0: The Potential of Generative Artificial Intelligence
Authors:	Pedro Antonio Boareto, Anderson Luis Szejka, Eduardo Freitas Rocha Loures, Fernando Deschamps and Eduardo Alves Portela Santos
Abstract:	Industry 4.0 (I4.0) integrates technologies like the Internet of Things (IoT), Artificial Intelligence (AI), and robotics to create interconnected, intelligent, and autonomous production environments. This transformation drives innovation and competitiveness but poses challenges, including system integrations, investments, and workforce upskilling. This work explores the potential of Generative Artifi-cial Intelligence (GAI) as an accelerator for I4.0 adoption considering Industry 5.0 (I5.0) requirements, using I4.0 reference architectures and the four key dimensions: Smart Manufacturing (SM), Smart Working (SW), Smart Products and Services (SPS), and Smart Supply Chain (SSC) to guide the analysis. The literature indicates that GAI has been applied in various domains, but there is a gap in comprehensive research to the industrial context. GAI's potential contributions to SM and SW include generating insights, optimizing operations through Digital Twins (DT), predictive maintenance, and enhancing human-machine collaboration, aligning with the I5.0 concept on personalization and human-centric technology solutions. In SPS and SSC, GAI aids in product development, mass customization, production simulation, and inventory control, providing real-time support and improving supply chain efficiency. The paper concludes that GAI holds significant promise for enhancing I4.0 and I5.0. But further research is needed to learn its impact at each organizational level, develop best practices, and address data quality and integration challenges. Future work will involve a systematic literature review to deepen insights into the integration of GAI with I4.0 and I5.0, with a particular focus on the role of DT and their potential to create more connected and cognitive industrial solutions.

Paper Nr:	53
Title:	Optimizing Internal Logistics using Automated Guided Vehicles: An Evaluation of Heuristic Approaches
Authors:	José Oliveira, Bernardo Fernandes and Marcelo Henriques
Abstract:	Future factories, underpinned by Industry 4.0 and Industry 5.0 frameworks, in-creasingly depend on Automated Guided Vehicles (AGVs) and Intelligent Au-tonomous Vehicles to enhance internal logistics efficiency. These vehicles are in-tegral to achieving greater efficiency, safety, and adaptability in modern manufac-turing environments. AGVs are transformative technologies that significantly en-hance internal logistics operations in modern manufacturing. While they offer numerous advantages, addressing the associated challenges is essential for their successful integration and long-term viability. This paper investigates the operational efficiencies of AGVs within industrial set-tings, focusing particularly on the scheduling and sequencing of AGV move-ments. Traditionally, the First In, First Out (FIFO) method has been employed to decide the next job for the AGV. However, through computational experiments, we analyse various heuristic methods to enhance the throughput and reduce idle times or 'empty' movements of AGVs, thereby optimising internal logistics and supporting robust production processes. The FIFO approach, while straightforward, often leads to suboptimal scheduling outcomes, especially in complex and dynamic manufacturing environments where job priorities frequently shift. It is increasingly evident that FIFO is inadequate for the nuanced demands of modern AGV operations, which benefit significantly from the adaptability and efficiency provided by heuristic solutions. This study suggests that advanced heuristics, particularly those designed for solving TSP or VRP, are better suited to these tasks. These methods not only offer more flexible and efficient route planning but also significantly improve the overall utility of AGVs in high-demand scenarios.

Paper Nr:	55
Title:	On Capability Needs for AI Utilization in Innovation Networks: Critical Literature Review
Authors:	Pasi Hänninen and Jyri Vilko
Abstract:	The potential of AI has been a growing trend in research in recent years. From a business perspective, the use of AI places new demands on skills and resources. Innovation networks are one option to tap into resources that companies them-selves do not have. Although there has been research on combining AI and inno-vation, there is not enough research on the capabilities needed for an innovation network to use AI. This study aims to present a snapshot of the research on the topic and to show how the capabilities of the network for using AI are formed. The findings of the study are that research on innovation network capabilities for using AI does not match the requirements of using AI, and it is argued that inno-vation network capability needs remain the same in terms of the terms presented, but that the use of AI places entirely new requirements on them.

Paper Nr:	56
Title:	Towards the Automation of the Product Definition Process for Design-to-Order Manufacturing Systems
Authors:	Maude Beauchemin, Jonathan Gaudreault, Xavier Zwingmann and Nadia Lehoux
Abstract:	In the last decade, the focus of automation in the manufacturing industry has been on the physical and logical aspects of the production system, supposing the product definition is well known. The steps required to create this product definition (in terms of design, engineering, and industrialization) are still mostly achieved by humans, although making use of the computer-aided tools that compose the CAx chain. This is efficient when the same product is afterward produced thousands of times, making the extensive human labor profitable. However, for design-to-order manufacturing systems, such a methodology suffers from long lead times and exorbitant costs directly transferred to the client. Industrial visits and discussions with design-to-order manufacturers followed by a detailed modelization of their current manufacturing processes led us to target the need for automation of the product definition process. We propose a new approach focusing on product families defined by data and algorithms which encapsulate product family semantics. This approach allows clients to generate their own personalized product definitions from this family in a few clicks. Our approach is currently implemented in SMEs from different industries, ranging from the construction industry to the medical implant sector, as well as our own research factory producing personalized micro-trailers.

Paper Nr:	65
Title:	A Framework for Ontology-Based Engineering Systems: Advances and Open Questions about Knowledge Capture and Use in Aerospace Manufacturing
Authors:	Fernando Mas, Rebeca Arista, Murillo Skrzek, Manuel Oliva, Domingo Morales-Palma and Anderson Luis Szejka
Abstract:	Modern Ontology-Based Engineering (OBE) systems, built over the founda-tions of Knowledge-Based Engineering (KBE), leverage ontology-based tech-nologies and the semantic web to capture and formalise knowledge holistical-ly. This comprehensive knowledge repository can subsequently feed design and decision support tools with robust generative features. However, the inte-gration and reuse of knowledge poses challenges for which it is necessary to develop appropriate methods and tools capable of handling diverse types of knowledge from both human experience and databases. The objective of this paper is to discuss the current limitations of OBE sys-tems to address knowledge integration and reuse, based on the research and experiences of the authors in the aerospace industry and manufacturing. This aims to provide a set of Research Questions (RQ) and the author's position to them, to serve as future research work and topics of discussion for the scien-tific community.

Paper Nr:	68
Title:	Boosting Governance-Centric Digital Product Passports Through Traceability in Footwear Industry
Authors:	Hugo Moço, Cristovão Sousa, Ricardo Ferreira, Pedro Pinto, Carla Pereira and Rui Diogo
Abstract:	Since supply chains have become complex and tracking a product's journey, from raw materials to the end of it´s life has become more difficult. Consumers are demanding greater transparency about the materials origins and environmental impact of the products they buy. These new requirements, togeher with European Commission Green Deal strategy, lead to the concept of digital product passport (DPP). DPP could be seen as an instrument to boost circularity, however the DPP architecture and governance model still undefined and unclear. Data Governance in the context of the DPP acts as the backbone for ensuring accurate and reliable data within these "passports" or data models, leading to flawless traceability. This article approaches the DPPs and it´s governance challenges, explaining how they function as digital repositories for a product's life cycle information and the concept of Data Governance. By understanding how these two concepts work together, we will explore a short use case within the footwear industry to show how DPP governance architecture might work in a distributed environment.

Paper Nr:	83
Title:	Towards a Human-Centric Industry 5.0: Exploring Team Roles to Improve Human-Machine Collaboration
Authors:	Domenico Monopoli, Mariateresa Caggiano, Concetta Semeraro and Michele Dassisti
Abstract:	Industry 5.0 aims to enhance the strengths of the current Industry 4.0 paradigm and broaden its scope by emphasizing the human operators within the working environment through harmonious interactions with machines and other human collaborators. One of the three pillars of Industry 5.0 is the redefined role that human operators would assume within rapidly evolving production systems due to significant technological improvements descending from the I4.0 paradigm. This new role is becoming a major organizational and technological challenge to ensure maximum productivity while preserving the health and well-being of the operator. In this context, the primary objective of this study was to investigate the impact of various human-machine team configurations on process performance and worker well-being. To achieve this, a collaborative workstation was designed and implemented to simulate different configurations of collaborative scenarios, each involving distinct roles for humans and machines. Preliminary findings indicate that key factors such as experience and attitude toward technology necessitate different collaborative setups. Specifically, experienced operators prefer supervisory roles within the team, with the machine primarily used for final inspections. In contrast, less experienced operators tend to rely more on the machine’s decisions, adopting a more apprentice-like role.

Paper Nr:	84
Title:	An Ontology Framework for Human-Robot Interoperability in Dynamic Construction Environments
Authors:	Pantelis Karapanagiotis, Felix Koester and Christos Emmanouilidis
Abstract:	Humans and robots increasingly work collaboratively in work environments. Their synergy varies between simple co-existence, wherein they perform independently in shared spaces, to deep collaboration, wherein the joint outcome of their co-working cannot be attributed to either one of the two types of actors alone. Relevant concepts are that of human agency, when robots operate under human oversight, with humans assuming control in appropriate circumstances, and sliding autonomy, where control slides back and forth between humans and robots, depending on the situational context. This study introduces an ontology framework for human-robot interoperability (HRI) in dynamic construction environments. The ontology serves as a context model, facilitating task allocation and collaboration between humans and robots under a Construction Sliding Work Sharing (CSWS) scheme. While previous ontology designs focused either on aspects of construction processes or human-robot collaboration, the ontological formalization of the intersection of HRI and dynamic construction environments is unexplored. The CSWS ontology integrates human-centricity, dynamic task distribution, and human-robot interoperability concepts in construction settings. The ontology development process combines concepts distilled from qualitative content analysis of a construction case study with relevant concepts from previous literature and ontologies. The ontology is consistent with the Sliding Work Sharing (SWS) concept, an extension of sliding autonomy for dynamic task allocation across human and robot actors. The capacity of the CSWS ontology to support dynamic task allocation and human-robot collaboration is exemplified via use cases querying for agent safety, process errors, and human-robot collaborative activities.

Paper Nr:	14
Title:	Digital Twin Data Broker with Assisted Mapping into a Knowledge Base
Authors:	Thomas Schmeyer, Kai Krämer, Anna-Lena Peh, Boris Brandherm, Margarita Chikobava and Gian-Lucca Kiefer
Abstract:	The frequent usage of digital twins to communicate between physical objects is resulting in more complex cyber-physical systems. To simplify the individual components’ integration and to optimize their usage, a data broker is being developed. Therefore, digital twins need to be semantically organized in an ontology that provides the advantage of reasoning methods. An assisted workflow is being developed to automatically enter subgraphs into an ontology. As a digital twin representation, the Asset Administration Shell format is used to have an international standard technology. Based on this, a new domain-specific language is developed, allowing experts to configure the generation process. This process maps the digital twin’s information into a graph representation of the ontology. The preconfigured generation process enables the user to efficiently register new digital twins without having expert knowledge of the underlying ontology. Additionally, a Large Language Model vector embedding and text reasoning support is implemented analysing the digital twin to create entity suggestions. The presented data broker is an automation tool for bridging the gap between semantic descriptions and digital twin formats in order to unite the advantages of both representations.

Paper Nr:	27
Title:	A Distributed Framework for Cooperative Scheduling of Production, Transportation, and Maintenance Using Multi-Agent Systems
Authors:	Yassir Haoudi, Agnès Letouzey and Xavier Desforges
Abstract:	Manufacturing systems often face conflicting demands between production, maintenance, and transport services, as the activities of each can disrupt the op-erations of the others. To address these conflicts and ensure a harmonized work-flow, this paper introduces an advanced multi-agent system with three distinct SCEP (Supervisor, Customers, Environment and Producers) models for produc-tion, maintenance, and transport. The objective is to facilitate distributed joint scheduling of these three critical areas, with considerations about the health state of the machines. This approach allows the integrated scheduling of production, transport, and predictive maintenance activities based on real-time health assess-ments of the machines. By doing so, the system aims at resolving operational conflicts among the services, ensuring that all stakeholders of these three domains collaborate effectively towards a common goal.

Paper Nr:	35
Title:	Reinforcement Learning for Optimizing Routing in Production Supply of Matrix Production Systems
Authors:	Florian Ried, Simon Niederdränk and Johannes Fottner
Abstract:	Matrix production systems offer the flexibility to meet an increasingly individual-ized and volatile customer demand. However, production supply processes with-in these systems have rarely been investigated in detail despite playing an integral role in their performance. Contributing to closing this research gap, this work uti-lizes reinforcement learning for routing in the production supply of matrix pro-duction systems. In particular, it focuses on dispatching orders to the vehicles and scheduling the orders within a route. Various constraints are considered to simu-late a realistic setting, including order time windows, vehicle battery limitations, and a vehicle capacity allowing to transport multiple items at once. A reinforce-ment learning framework is conceptualized and implemented, assigning orders to vehicles based on various route construction heuristics. Its observation space contains abstract information about current orders of the matrix production sup-ply environment and specific data on the vehicles for the reinforcement learning agent to select both a vehicle and a heuristic. The action and observation spaces are complemented by a multi-criteria reward function, prompting the agent to learn not to violate any constraints of the environment while simultaneously choosing actions that lead to the most cost-effective routes after route optimiza-tion. The reinforcement learning route constructor approach is trained and de-ployed on a discrete-event simulation of a matrix production system, which is connected to the reinforcement learning framework via a socket interface. The ap-proach has proven to be successful by outperforming two non-reinforcement learning heuristics for route construction.

Paper Nr:	37
Title:	Current Trends and Future Challenges to Put Circular Manufacturing in Practice
Authors:	Yasamin Eslami, Chiara Franciosi, Maroua Nouiri, Adriana Giret, Elisa Negri and Pascale Marangé
Abstract:	Addressing Sustainable Development issues in the industrial sector is a hot topic in the existing literature. Different production and consumption strategies have been introduced as drivers for sustainable development in manufacturing such as Circular Economy (CE). It covers a prominent position in boosting sustainable development. The adoption of the CE concept to manufacturing leads to Circular Manufacturing (CM). Different strategies and approaches have been investigated toward the transition to a Circular Manufacturing System (CMS). Although abundant studies have been focused on the adoption of CM strategies in manu-facturing systems, the literature still lacks an overview of the challenges raised through this implementation in practice. To address this gap, the present study centres around looking through various definitions of CM, different methodolo-gies and frameworks to adopt CM strategies and finally the challenges faced in putting these strategies into practice.

Paper Nr:	77
Title:	Contribution of Fuzzy-Possibility Approach to Assessing the Complexity Level of IT-Systems Design
Authors:	Igor Kimyaev, Alexander Spesivtsev, Vasily Spesivtsev and Nikolay Shilov
Abstract:	The problem of assessing the level of the design work complexity in the IT in-dustry is of vital importance. This is due to the fact that design in this industry in-volves the consolidation of a wide variety of knowledge sectors: programming, network technologies, engineering infrastructure, etc. From such diversity fol-lows the absence of generally accepted methods for quantitative assessment of the complexity and volume of design work based on the most general requirements for the designed IT system. In this research, a method for quantitative indicative assessment of the complexity of one of the main phases of IT system design is proposed: architectural planning. Using a seven-factor structure of fuzzy linguis-tic variables, a fuzzy possibility model was developed based on expert knowledge and experience. This model allows for quantitative assessment and forecasting of the complexity of tasks in the architecture of information technology systems.

Area 2 - Logistics

Full Papers

Paper Nr:	30
Title:	The Role of Actor’s Creative Self-Efficacy in AI Enabled Value Creation
Authors:	Tuomas Hongisto and Jyri Vilko
Abstract:	Researching creativity remains crucial despite the growing use of AI, as human creativity drives innovation, problem-solving, and value creation in ways that AI alone cannot replicate. This study examines the impact of Solution-Focused (SF) coaching on individuals' creative self-efficacy (CSE) within AI-enabled environments. Core findings indicate that SF coaching significantly enhances CSE by increasing participants' awareness of their creative potential and cognitive abilities related to creativity. The study reveals that a clear coaching structure, personalized tools, and the quality of the coach-client relationship and interaction are essential for maximizing the benefits of SF coaching. These findings underscore the importance of fostering human creativity alongside AI to achieve dynamic and human-centric value creation in modern organizations.

Paper Nr:	48
Title:	Multi-Agent Path Planning for Logistics Cargo Environment Using LSTM Based Reinforcement Learning
Authors:	Gun Rae Cho, Sungho Park, Eui-Jung Jung, Hyunseok Shin, So Eun Son and Yong Choi
Abstract:	In the operation of logistics cargo, temporal efficiency in handling logistics is a critical issue. One solution to enhance this efficiency is the deployment of multiple autonomous ground vehicles (AGVs). This paper proposes a reinforcement learning approach based on Long Short-Term Memory (LSTM) for multi-agent path planning in the logistics cargo environment. When AGVs are treated as moving obstacles to each other, the application of LSTM allows for path planning that aptly addresses changes over time within the environment. Additionally, to solve the well-known problem of the sparse reward in pathfinding, we propose a reinforcement learning architecture for multi-agent path planning that uses the path planning results of a single agent, guided by Q-learning, as the guide path. Furthermore, we have established state variables independent of the number of agents by setting interest window-based state variables, and introduced revisit rewards, effectively resolving the issue of local minima caused by repetitive movement-avoidance behavior between agents. Simulation results, emulating a real-world logistics warehouse environment, demonstrate the proposed technique's capability for effective multi-agent path planning in such settings.

Paper Nr:	52
Title:	Evaluation of Intermodal Transport Routes: Environment and Biodiversity Perspectives
Authors:	Mladen Krstic, Snežana Tadić, Leonardo Agnusdei, Pier Paolo Miglietta and Donatella Porrini
Abstract:	Intermodal transportation integrates multiple modes of transport, enhancing logis-tics efficiency and reducing environmental footprints. It alleviates road conges-tion, reduces greenhouse gas emissions, and supports sustainable economic growth. One of the main issues in intermodal transport planning is the selection of adequate routes, especially in terms of their impact on the environment. Strate-gic route planning can mitigate transportation impacts on biodiversity and ecosys-tems, preventing habitat fragmentation and wildlife disruption. Thus, incorporat-ing ecological considerations in transport planning is vital for sustainability. This paper aims to define a framework for evaluating and ranking intermodal transport routes based on their environmental and biodiversity impacts, promoting resilient and environmentally responsible transportation systems. For solving the defined problem this paper proposes a geographic information system (GIS)-based multi-criteria decision-making (MCDM) framework. It employs a Comprehensive dis-tance-based Ranking (COBRA) MCDM method that uses GIS-based input data to evaluate alternative intermodal transport routes and to rank them. The proposed framework is tested on a real-life case study, i.e., on evaluating and comparing in-termodal transport routes between North European ports and North Italy inter-modal terminals, performing on two Trans-European Transport Network (TEN-T) corridors. The results indicate that the Scandinavian-Mediterranean corridor between the port of Hamburg and the intermodal terminal Verona Quadrante Eu-ropa is more favorable regarding environmental impact and biodiversity protec-tion. The main contributions of the study include the development of a framework for route evaluation and the establishment of criteria that can be used for evaluat-ing and ranking routes.

Paper Nr:	57
Title:	Stockyard Planning and Optimization Using Intelligent Search
Authors:	Adele Pißarek, Petra Hofstedt and Sven Löffler
Abstract:	Stockyard planning and optimization in ports is an important logistic task. It describes the calculation of efficient task sequences for unloading of bulk cargo from import vessels, the transportation, the storage and blending of material and for loading export ships. At this, the optimization goal is to minimize the lay times of the import and export ships by optimized material operation sequences. The paper presents a new approach on stockyard planning by intelligent search methods (greedy, beam, and Monte-Carlo Tree Search) and using domain-specific expert knowledge. The approach is evaluated through a series of tests on the model of a real-world port configuration.

Paper Nr:	59
Title:	Forecasting Multivariate Time Series with Trend and Seasonality: A Random Forest Approach
Authors:	Zahira Marzak, Rajaa Benabbou, Salma Mouatassim and Jamal Benhra
Abstract:	Random Forest algorithm, a powerful machine learning technique recognized for its flexibility and high predictive accuracy has been used in different applications across various domains. This paper presents a comparative analysis of the per-formance of Random Forest in forecasting multivariate time series data exhibiting trend and seasonality, a task that poses a significant challenge since the traditional methods often used to forecast time series can fail to accurately model these com-plex characteristics, leading to poor forecasting performance and unreliable pre-diction. In this study, different optimization techniques are used to enhance the perfor-mance of the RF models, such as feature engineering, using various features (temporal, lagged, rolling window, and interaction features), and parameter tuning (Manual, Random Search, and Grid Search). The models are evaluated using three accuracy metrics: Root Mean Square Error (RMSE), Mean Absolute Error (MAE), and Mean Absolute Percentage Error (MAPE) with lower values indi-cating better performances, the results are then compared to traditional forecasting methods. The proposed forecasting framework demonstrated the capability of the RF model to capture trends and seasonality in multivariate time series and provide competitive results.

Paper Nr:	67
Title:	Leveraging Large Language Models for Supply Chain Management Optimization: A Case Study
Authors:	Sumaya Abdul Rahman, Sanjay Chawla, Mohammed Yaqot and Brenno Menezes
Abstract:	This research examines the transformative potential of Large Language Models (LLMs) and Generative AI (GAI) in supply chain management (SCM) and operations research (OR). By leveraging advanced Natural Language Capabilities (NLP) capabilities in models such as OpenAI’s GPT-4o, we explore how these technologies can support modeling tasks and streamline complex supply chain problems for optimization steps. Our study specifically focuses on translating mathematical formulations into executable code and interpreting solver outputs. Our work shows that LLMs complement and augment traditional solvers by speeding up the end-to-end process of problem formulation, solution, and interpretation, while also enhancing overall efficiency and reducing the need for manual adjustments. This work systematically identifies the strengths and limitations of LLMs in SCM applications, highlighting their ability to enhance efficiency, accuracy, and decision-making. We conducted a proof-of-concept demonstration using GPT-4o to prepare the model for solving and interpretation of three increasingly complex transportation problems. The results demonstrate that LLMs not only excel at providing error-free code but also exhibit enhanced capabilities in reasoning and interpreting complex outputs from optimization solvers. The proposed framework provides a practical guide for practitioners and researchers in SCM and OR, demonstrating how LLMs can automate and improve optimization tasks.

Short Papers

Paper Nr:	18
Title:	Obsolescence Forecasting of Intel Processors: A Transformers Approach
Authors:	Manelle Nouar, Amel Souif, Bertrand Decocq and Marc Zolghadri
Abstract:	This study investigates the issue of electronic product obsolescence, focusing on Intel processors within Orange’s data centers. As digitization advances, managing stock and addressing supply chain disruptions become critical, particularly when dealing with the obsolescence of key components like servers. A proactive approach to obsolescence management is explored, emphasizing the importance of predicting the life cycle of server components to ensure operational continuity. The research highlights the use of advanced predictive techniques, specifically transformers, to forecast the life cycles of electronic devices. The study aims to provide insights into mitigating the risks associated with obsolescence, thereby aiding in the formulation of effective long-term strategies for maintaining and managing essential electronic components.

Paper Nr:	22
Title:	An Inventory Management Support Tool Through Indirect Q-value Estimation: A Combined Optimization and Forecasting Approach
Authors:	Amanda Rodrigues Delfiol, Camélia Dadouchi, Bruno Agard and Philippe St-Aubin
Abstract:	Effective inventory management is crucial in manufacturing and wholesale businesses to reduce operation costs and meet service level guarantees. Due to the continuous increase in product catalogues and highly volatile demand, inventory management complexity continues to grow. This paper introduces a decision support tool designed to aid in inventory management through an indirect Q-value estimator technique. The proposed tool employs simulation, optimization and forecasting techniques to enable purchase actions evaluation for large horizons. By integrating both simulation and optimization into a supervised learning algorithm, the tool provides an easy to interpret cost estimation that can directly be used to make informed procurement decisions. A case study in the textile industry demonstrates its use and its performance in a single-echelon supply chain setting. This research presents a comprehensive step by step framework to support the creation of a decision support tool that can offer valuable aid for decision-making processes across different supply management contexts.

Paper Nr:	25
Title:	Clustering Analysis for Forecasting Medicine Consumption
Authors:	Douglas Mateus Machado, Zakaria Yahouni, Gülgün Alpan and Denis Koala
Abstract:	This study investigates machine learning for forecasting medicine consumption in hospitals to optimize resource allocation and logistics. We use two approaches: a unified approach that combines data from multiple hospitals and a separated approach that forecasts for individual hospitals. We explored both K-means clustering and manual pair clustering based on consumption trends. While K-means clustering did not yield improvements, manual clusturing identified specific pairs of medicines with significantly enhanced forecast accuracy (e.g., Medicine 15 at Hospital 1: MAPE decreased from 19.70% to 3.30%). However, the unified approach did not consistently benefit all hospitals (e.g., Medicine 9). This underscores the need to balance accuracy gains in some hospitals against potential losses in others. Overall, manual clustering within the separated approach shows promise. Future work should explore advanced automated clustering techniques like Dynamic Time Warping (DTW) and leverage larger datasets for further validation.

Paper Nr:	32
Title:	Towards a Process-Based Industry 5.0 Maturity Model: A Feasibility Study in Supply Chain
Authors:	Sara Himmiche, Jean-Luc Maire, José-Fernando Jimenez, Magali Pralus and Laurent Tabourot
Abstract:	A focus on sustainability, resilience, and human-centricity is necessary to imple-ment the global vision of Industry 5.0. However, assessing the readiness of these pillars within the business processes of an organization remains a significant challenge. This paper addresses this gap by providing a comprehensive and pro-cess-based maturity model in Industry 5.0 context. The model adopts a top-down approach assessing the maturity of business processes at different layers and fo-cusing on the formalization of practices and their correlation with maturity levels. The applicability of the model is demonstrated through its instantiation to the supply chain process category, with a focus on the schedule production process. Furthermore, the paper explores the integration of AI-based techniques to auto-mate the creation of maturity grids and streamline the assessment process. Over-all, the results of this paper highlight the prescriptive characteristic of the pro-posed maturity model, providing a robust framework to guide organizations to-wards a step-by-step transition through the Industry 5.0 vision.

Paper Nr:	44
Title:	A Blockchain-Powered Framework for Traceable and Secure Pharmaceutical Delivery with Crowdsourced Logistics
Authors:	Kadim Lahcen Nadime, Jamal Benhra, Rajaa Benabbou and Salma Mouatassim
Abstract:	Current pharmaceutical last-mile delivery systems often rely on outdated manual processes and paper-based documentation, leading to significant inefficiencies and security vulnerabilities. This paper proposes a comprehensive framework that leverages smart contracts, blockchain technology, and crowdsourcing to enhance the traceability and security of pharmaceutical deliveries. By utilizing blockchain, the framework automates order processing, tracking, and delivery, ensuring a transparent and tamper-proof record of each transaction. Additionally, the framework incorporates a flexible, crowd-sourced model for selecting carriers through a competitive bidding mechanism, improving efficiency and responsiveness. Our research includes a site visit to a drug distribution company to analyze the existing delivery process and identify critical issues. Based on these insights, we developed a solution that integrates digital and physical tracking measures, such as reusable drug packaging with sensors. The solution was tested locally in a simulated blockchain environment to validate its functionality. The paper emphasizes the need for tailored studies to address international regulatory differences and ethical implications in implementing a blockchain-based pharmaceutical delivery system. This approach not only simplifies the delivery process but also ensures enhanced visibility, reliability, and accountability across the supply chain.

Paper Nr:	51
Title:	Last-Mile Delivery Optimization Using Mixed Electric Vehicles, UAVs and Full Truck Delivery Based on Artificial Intelligence Algorithms
Authors:	Bouhia Hanaa, Jamal Benhra and Wissal Ed-dehbi
Abstract:	This paper presents the Electric Vehicle Routing Problem with Drones (EVRPD), a novel strategy aimed at enhancing last-mile delivery through the combined use of Electric Vehicles (EVs) and drones. The primary objective is to reduce energy usage and CO2 emissions while maintaining effective delivery operations. The proposed system employs a diverse fleet where EVs act as mobile bases, and drones handle the final leg of parcel delivery. Both EVs and drones can deliver packages, with drones being utilized for smaller, time-sensitive shipments. The system emphasizes optimal payload distribution to boost energy efficiency and drone endurance. Drones are resupplied on the EVs' rooftops via IoT technology. To address the routing challenge, the solution leverages bio-inspired algorithms, including Ant Colony Optimization (ACO) for EVs and Particle Swarm Optimization (PSO) for drones. This methodology enhances energy efficiency and drone autonomy, ensuring effective parcel delivery. By incorporating these advanced techniques, EVRPD presents a promising approach for a more sustainable and effective urban logistics network.

Paper Nr:	54
Title:	Layout Optimization Strategy Based on Three-Stage Cutting Pattern
Authors:	Jiaojiao Li, Zeqiang Hou, Diaoyin Tan, Xin Liu and Jiawu Peng
Abstract:	The popularity of “personalized customization” has put forward higher requirements for industrial production, and how to achieve mass customization production has become one of the main problems plaguing manufacturing enterprises. This paper focuses on the layout optimization problem of square parts, and adopts a three-stage 2D rectangular cutting method, i.e., following the cutting idea of “sheet-strip-stack-item”. To solve this problem, we develop a mixed-integer linear programming model with the optimization objective of using as few square sheets as possible while satisfying the order demand and related constraints. Afterwards, we adopt the CPLEX solver to invoke the branch-and-bound algorithm package to solve the model. The experimental results show that the method can accomplish the scheduling of all the orders in the given four datasets and obtain a satisfactory solution within 30 minutes. The utilization rate of sheets in all datasets reaches more than 85%, proving that the method can utilize raw materials more efficiently, and achieve good economic benefits.

Paper Nr:	58
Title:	Enabling Decentralized Collaboration Among Transporters through an Optimizing Trading Network for Transport Orders
Authors:	Renaud De Landtsheer, Aline Goeminne and Quentin Meurisse
Abstract:	Nowadays, at least 15-20% of the moving trucks in the European Union are completely or nearly empty. We suggest that this is mainly due to a large quantity of transport companies being small; despite optimizing their routes, they will have to perform empty travels. Some approaches have been proposed to enable transport companies to consolidate their route through collaboration. These approaches either rely on global auction systems or require companies to share their information in a centralized optimizing system. We propose a novel approach where optimization software owned by different transport companies would trade transport orders through a standardized communication protocol to constitute a large and decentralized optimization system. We propose to build such a protocol and system based on the Very Large-Scale Neighborhood search (VLSN) optimization algorithm. It would maintain the confidentiality of private information and let the companies keep the control of the consolidation that the system will be able to explore. We conclude by listing research challenges related to this approach.

Paper Nr:	60
Title:	Real-Time Bearing Health Monitoring Using Diffusion-Based Spectral Analysis: A Self-Adaptive Approach to Predictive Maintenance
Authors:	Guillaume Prevost, Jérôme Boutet, Esteban Cabanillas and Cornel Ioana
Abstract:	This paper introduces a novel approach to real-time health monitoring of rolling bearings using a diffusion-based model. We propose a method that leverages Denoising Diffusion Implicit Models (DDIM), exploiting their intrinsic denoising capability, to process frequency spectra of vibration signals and construct a health index without prior system information. Our approach employs a continuous-time DDIM to model the complex distribution of frequencies in bearing vibration signals, enabling the detection of subtle spectral changes indicative of incipient faults. The method operates in real-time, adapting to specific bearing characteristics and learning normality over time. The health index (HI) is constructed using the reconstruction error between input features and their denoised versions and standardized using Welford's algorithm for online mean and variance estimation. We validate our method on the XJTU-SY bearing dataset. Results demonstrate the efficacy of our diffusion-based indicator, particularly in complex degradation scenarios. The method's ability to capture complex distributions and early degradation signs without prior system information presents a promising tool for predictive maintenance.

Paper Nr:	63
Title:	The Impact of Logistics Performance on the Convergence of Economic Growth in the Countries of the World
Authors:	Pablo Coto-Millán and Miguel Angel Pesquera
Abstract:	The article highlights the importance of beta convergence and generative artificial intelligence in facilitating the transition to a more generative paradigm. A panel data of 120 countries between 2007 and 2022 is used to estimate the effect of the Logistics Performance Index (LPI) on economic growth, along with other variables such as human capital, physical capital and employment. The results show that the most significant LPI indices are Customs, Infrastructure and competitive international shipping prices, and that logistics has facilitated convergence in the growth of less developed countries. It is concluded that conditional beta convergence has allowed for an inverse relationship between the growth trend and the income level of each starting country, indicating that conditional convergence exists and that logistics has played a key role in facilitating economic convergence between less developed and developed countries. Overall, the results of the study provide valuable information for policy makers in formulating appropriate strategies to promote global economic growth. In summary, beta convergence and technologies such as generative artificial intelligence are important tools to facilitate the transition to a more generative paradigm in logistics performance in countries around the world.

Paper Nr:	70
Title:	An Improved Hybrid Recommendation Algorithm for Vehicle-Cargo Matching in the Logistics
Authors:	Yiting Wang and Lifen Wei
Abstract:	This paper aims to optimize the vehicle-cargo matching of the logistics platform through the hybrid recommendation algorithm to achieve the effective matching between vehicle and cargo sources. This paper constructs a hybrid recommendation model taken into account the content-based recommendation algorithm and the item-based collaborative filtering algorithm. The content-based recommendation algorithm has advantages in matching the inherent attributes of vehicle and cargo sources, whereas the item-based collaborative filtering can deal with the problem of interest drift, and the advantages of these two algorithms are integrated in the hybrid model in this paper. The matching factors are pick out from the data of current logistics platforms, and the feedback competition method is applied for the calculation of the weights, and then the weight similarity is worked out for the recommendation. Multiple aspects are considered in the hybrid model, so as to satisfy the users by recommendation results, thereby achieving vehicle-cargo matching. This paper focuses on the matching of vehicles and cargoes in multiple dimensions, where the personalized and diversified comprehensive recommendation results are acquired to meet the needs of car owners and cargo owners.

Paper Nr:	75
Title:	E-Waste Collection Under Recycling Hub Demand and Partial Information: The Benchmark Solution in a Pilot Case
Authors:	Alessia Ciccarelli, Marta Flamini, Maurizio Naldi and Elpidio Romano
Abstract:	E-waste collection is a case of reverse logistics in a circular economy approach where e-waste is recycled, refurbished, or properly disposed of. In some cases, the recycling process is driven by companies that request specific types and amounts of e-waste to be collected and included back in their manufacturing cycle. The collection process should be optimized to properly match demand and offer at the minimum cost. We study a realistic collection scenario where a recycling hub demands a certain amount of e-waste items of different types to be collected among a set of stations whose inventory availability is unknown. This scenario represents a benchmark for what we can obtain when more information is available. We propose a routing algorithm that assigns the requested items to vehicles and plans their routes with the goal of maximizing the hub's demand fulfilment while minimizing operating costs, all within the constraints of vehicle capacity. Through a simulation conducted on the pilot case in the city of Rome, we show that the demand satisfaction level can exceed 75\% in more than 75\% of the simulation instances, and fuel costs represent the dominant component.

Paper Nr:	76
Title:	Improved Package Orientation Estimation Using RFID Signal Analysis with Machine Learning
Authors:	Joaquin Dillen, N. Simões, João M. Faria, Luis Vilas Boas, Inês Caetano, Luis Cardoso, João Borges and António H. J. Moreira
Abstract:	Radio-frequency identification (RFID) technology has become increasingly popular in various applications due to its low cost, energy efficiency, and compact size. Its versatility extends into diverse fields, prompting further research beyond its initial uses. This study explores the application of commercial RFID antennas for estimating package orientation, focusing on Phase and RSSI values. By employing machine learning algorithms, we examined sixteen different configurations and orientations of the package. Our findings indicate that the optimal RFID tag location for pose estimation can achieve an 87.7% success rate, demonstrating that the best setup involves using one antenna and one tag. This research introduces a novel framework for posture identification, showcasing RFID technology as a vital tool for both traceability and monitoring.

Paper Nr:	81
Title:	Optimization of Resource Allocation and Distribution in Industrial Supply Chains and Logistics Networks: A Hybrid Approach based on Genetic Algorithm and Game-Theoretic Analysis
Authors:	Chandadevi Giri and Nitin Harale
Abstract:	Efficient resource allocation and distribution are essential for optimizing industrial supply chains and logistics networks. This paper presents a novel hybrid approach that combines genetic algorithms (GAs) with game-theoretic analysis to tackle the complex, multi-objective challenges inherent in these systems. The proposed framework focuses on minimizing transportation costs and carbon emissions while optimizing routing from source locations through intermediate facilities to final demand points. The genetic algorithm effectively navigates the search space to identify near-optimal solutions, while the game-theoretic analysis models strategic interactions among stakeholders, promoting stable and equitable resource allocation. The proposed approach exhibits advantages as it enhances cost efficiency and reduces emissions. The results provide the practical implications for improving the efficiency and sustainability of logistics operations across various industries. This study contributes to the advancement of resilient and adaptive supply chains by offering a comprehensive optimization framework that leverages the strengths of both genetic algorithms and game theory.

Paper Nr:	87
Title:	A Novel Model for Multi-Robot Task Assignment in Smart Warehouses
Authors:	Meryem Bamoumen, Riheb Sioud and Nadia Hamani
Abstract:	Efficient task assignment to multiple robots becomes a core challenge for smart warehouse operations. This work falls within the context of automation and task management of modern intelligent logistics. A new mathematical model for solving the open path multi-depot asymmetric traveling salesman problem is developed in this paper, aiming at finding an optimized task allocation scheme for the robots in an intelligent warehouse. Our model aims at determining optimal task sets for each robot while minimizing operational costs and in order to enhance productivity without requiring each robot to return to its starting depot. By using CPLEX, we solved the small and medium sized scenarios and demonstrated superior performance compared with other models in computation time and solution quality.

Paper Nr:	26
Title:	A Decision Support System of the Configuration of a Supermarket in a Components Company for the Automotive Industry
Authors:	Telma Pereira, José Oliveira and António Vieira
Abstract:	The present study took place in the industrial environment of a company that is part of the automotive industry. The characteristics of this type of industry require company to constantly optimise the configuration of its warehouses. The aim of this paper is to present an algorithm that allows the company to (re)configure the location of items in the Central Shelf Supply System (CSSS), which is the main supplier of parts for production lines. The algorithm was developed in Visual Basic .NET and operates sequentially, considering the characteristics of each item, providing a solution for the distribution of items in the CSSS, considering the balance of picking workload and ergonomic conditions of the picking opera-tion. To evaluate the performance of the proposed solution for items location in CSSS, a simulation tool was developed using SIMIO. This tool demonstrated the capability to simulate different real-world scenarios in a virtual environment effi-ciently, quantifying important performance indicators. Two scenarios were con-sidered when modelling the configuration of the CSSS system. Scenario 1 was modelled considering the locations of the items in the CSSS, while Scenario 2 was modelled taking into account the locations of the items proposed by the de-veloped algorithm. With this simulation tool, it is now possible to easily assess over time whether the current item location solution can be maintained or needs adjustment. This developed simulation solution also enables the creation and test-ing of alternative layouts.