This study examines Industry 4.0-based technologies, focusing on the barriers to their implementation in European small- and medium-sized enterprises (SMEs). The purpose of this research was to determine the most significant obstacles that prevent SMEs from implementing smart manufacturing, as well as to identify the most important components of such an operationalization and to evaluate whether only large businesses have access to technological opportunities given the financial complexities of such an adoption. The study is premised on the notion that, in the setting of cyber-physical production systems, the gap between massive corporations and SMEs may result in significant disadvantages for the latter, leading to their market exclusion by the former. The research aim was achieved by secondary data analysis, where previously gathered data were assessed and analyzed. The need to investigate this topic originates from the fact that SMEs require more research than large corporations, which are typically the focus of mainstream debates. The findings validated Industry 4.0′s critical role in smart process planning provided by deep learning and virtual simulation algorithms, especially for industrial production. The research also discussed the connection options for SMEs as a means of enhancing business efficiency through machine intelligence and autonomous robotic technologies. The interaction between Industry 4.0 and the economic management of organizations is viewed in this study as a possible source of significant added value.
Vision Language Models (VLMs) have shown strong performance on multimodal reasoning tasks, yet most evaluations focus on short videos and assume unconstrained computational resources. In industrial settings such as pharmaceutical content understanding, practitioners must process long-form videos under strict GPU, latency, and cost constraints, where many existing approaches fail to scale. In this work, we present an industrial GenAI framework that processes over 200,000 PDFs, 25,326 videos across eight formats (e.g., MP4, M4V, etc.), and 888 multilingual audio files in more than 20 languages. Our study makes three contributions: (i) an industrial large-scale architecture for multimodal reasoning in pharmaceutical domains; (ii) empirical analysis of over 40 VLMs on two leading benchmarks (Video-MME and MMBench) and proprietary dataset of 25,326 videos across 14 disease areas; and (iii) four findings relevant to long-form video reasoning: the role of multimodality, attention mechanism trade-offs, temporal reasoning limits, and challenges of video splitting under GPU constraints. Results show 3-8 times efficiency gains with SDPA attention on commodity GPUs, multimodality improving up to 8/12 task domains (especially length-dependent tasks), and clear bottlenecks in temporal alignment and keyframe detection across open- and closed-source VLMs. Rather than proposing a new "A+B" model, this paper characterizes practical limits, trade-offs, and failure patterns of current VLMs under realistic deployment constraints, and provide actionable guidance for both researchers and practitioners designing scalable multimodal systems for long-form video understanding in industrial domains.
Modern business environments demand continuous reconfiguration of cross-functional processes, yet most enterprise systems remain organized around siloed departments, rigid workflows, and hard-coded automation. Meanwhile, large language models (LLMs) demonstrate strong capabilities in interpreting natural language and synthesizing unstructured information, but they lack deterministic, auditable execution of complex business logic. We introduce Autonomous Business System (AUTOBUS), a system that integrates LLM-based AI agents, predicate-logic programming, and business-semantics-centric enterprise data into a unified neuro-symbolic architecture for executing end-to-end business initiatives. AUTOBUS models a business initiative as a network of interrelated tasks with explicit pre- and post-conditions, required data, evaluation rules, and API-level actions. Enterprise data is organized as a knowledge graph, whose entities, relationships, and constraints are translated into logic facts and foundational rules that ground reasoning and ensure semantic consistency. Core AI agents synthesize task instructions, enterprise semantics, and available tools into task-specific logic programs, which are executed by a logic engine that enforces constraints, coordinates auxiliary tools, and produces deterministic outcomes. Humans specify task instructions, define and maintain business semantics and policies, curate tools, and supervise high-impact or ambiguous decisions, ensuring accountability and adaptability. We detail the AUTOBUS architecture, the structure of AI-generated logic programs, and the human-AI collaboration model and present a case study that demonstrates accelerated time to market in a data-rich organization. A reference implementation of the case study is available at https://github.com/cecilpang/autobus-paper.
The integration of Large Language Models (LLMs) into Industrial IoT (IIoT) systems enables 30-50% faster fault diagnosis and 25% reduction in unplanned downtime through predictive maintenance and quality control. However, deploying LLMs on resource-constrained edge devices (e.g., <4 GB PLCs) faces challenges in real-time processing (<10 ms latency) and compliance with industrial privacy standards (IEC 62443/GDPR). Federated Learning (FL) emerges as a critical enabler, allowing distributed training across sensors, robots and PLCs without raw data sharing. This paper presents the first comprehensive survey and taxonomy for FL+LLM in manufacturing, validated through case studies across automotive, electronics and pharmaceutical production. We systematically analyze: (1) compressed architectures (e.g., TinyBERT achieving 4ms inference), (2) EMI-resistant protocols for factory floors (tolerating 25% packet loss), and (3) privacy-accuracy tradeoffs (e.g., homomorphic encryption adding <15% latency overhead). Key unresolved challenges include sub-5ms inference on legacy PLCs and cross-factory generalization under non-IID data. The work provides concrete design guidelines for implementing FL+LLM systems that meet Industry 4.0 requirements for security, reliability, and real-time performance.
Alberto Pliego Marugán, Jesús M. Pinar-Pérez, Fausto Pedro García Márquez
Efficient maintenance has always been essential for the successful application of engineering systems. However, the challenges to be overcome in the implementation of Industry 4.0 necessitate new paradigms of maintenance optimization. Machine learning techniques are becoming increasingly used in engineering and maintenance, with reinforcement learning being one of the most promising. In this paper, we propose a gamma degradation process together with a novel maintenance model in which repairs are increasingly imperfect, i.e., the beneficial effect of system repairs decreases as more repairs are performed, reflecting the degradational behavior of real-world systems. To generate maintenance policies for this system, we developed a reinforcement-learning-based agent using a Double Deep Q-Network architecture. This agent presents two important advantages: it works without a predefined preventive threshold, and it can operate in a continuous degradation state space. Our agent learns to behave in different scenarios, showing great flexibility. In addition, we performed an analysis of how changes in the main parameters of the environment affect the maintenance policy proposed by the agent. The proposed approach is demonstrated to be appropriate and to significatively improve long-run cost as compared with other common maintenance strategies.
Modern enterprise computing systems integrate numerous subsystems to resolve a common task by yielding emergent behavior. A widespread approach is using services implemented with Web technologies like REST or OpenAPI, which offer an interaction mechanism and service documentation standard, respectively. Each service represents a specific business functionality, allowing encapsulation and easier maintenance. Despite the reduced maintenance costs on an individual service level, increased integration complexity arises. Consequently, automated service composition approaches have arisen to mitigate this issue. Nevertheless, these approaches have not achieved high acceptance in practice due to their reliance on complex formal modeling. Within this Ph.D. thesis, we analyze the application of Large Language Models (LLMs) to automatically integrate the services based on a natural language input. The result is a reusable service composition, e.g., as program code. While not always generating entirely correct results, the result can still be helpful by providing integration engineers with a close approximation of a suitable solution, which requires little effort to become operational. Our research involves (i) introducing a software architecture for automated service composition using LLMs, (ii) analyzing Retrieval Augmented Generation (RAG) for service discovery, (iii) proposing a novel natural language query-based benchmark for service discovery, and (iv) extending the benchmark to complete service composition scenarios. We have presented our software architecture as Compositio Prompto, the analysis of RAG for service discovery, and submitted a proposal for the service discovery benchmark. Open topics are primarily the extension of the service discovery benchmark to service composition scenarios and the improvements of the service composition generation, e.g., using fine-tuning or LLM agents.
The scarcity of water and energy in South Africa impacts multiple domestic and economic sectors, such as agriculture, industry, energy production, and ecosystems. This leads to diminished output, environmental degradation, and loss of biodiversity. The challenges are closely associated with the ongoing overexploitation of natural resources, attributable to the transformative effects of industrialisation and urbanisation. Current trends in energy and resource consumption, agricultural systems, and urban expansion are largely unsustainable and necessitate strategic interventions to mitigate environmental damage and promote long-term resilience. In the absence of oversight, these trends could exacerbate detrimental climate change patterns and limit economic growth. This study investigates the ongoing challenges related to the manufacturing sector's impact on water and energy resource depletion, while also assessing the potential cost savings from adopting cleaner production technologies. This study provides a significant contribution by presenting empirical evidence from the iLembe District, utilising a quantitative research design. A survey was performed across multiple factories in the iLembe Municipal District of South Africa. Of the 262 businesses surveyed, 184 out of 191 responses were considered reliable for data analysis. The study objectives were evaluated through descriptive and correlation analyses to assess the impact of digital technology on water, energy, and operational cost savings in manufacturing organisations within the district. The findings suggest that manufacturing organisations in the iLembe Municipal District demonstrated a lack of commitment to adopting digital technology as a waste management strategy to decrease electricity and water consumption. Respondents concurred that their organisations have achieved cost reductions by implementing cleaner production processes. The research demonstrated that digital technology has a beneficial effect on water and energy conservation, cost reduction, and the improvement of organisational performance. Management must understand the beneficial effects of digital technology to guarantee sustainable organisational performance..
The pharmaceutical industry is undergoing a significant digital transformation to improve efficiency, productivity, and regulatory compliance. One critical aspect of this transformation is ensuring data integrity, which is essential for maintaining the quality and safety of pharmaceutical products. This article explores the reasons behind the digital transformation in the pharmaceutical industry, focusing on the need for better data integrity. It examines the challenges pharmaceutical companies face in achieving data integrity, such as legacy systems, data silos, resistance to change, cybersecurity risks, and the validation of computerized systems. The article also discusses the benefits of regulatory compliance, including enhanced product quality and patient safety, reduced risk of non-compliance, improved operational efficiency, and increased brand reputation and customer trust. The author proposes strategies for successful digital transformation and data integrity, such as developing a comprehensive roadmap, investing in modern IT infrastructure, implementing data governance and quality management systems, fostering a culture of continuous improvement, and collaborating with technology partners and industry consortia. The article concludes by discussing future trends and opportunities, such as the adoption of blockchain technology, the integration of Internet of Things devices, the use of big data analytics for predictive quality control, and the need for collaboration with regulators to develop industry-wide data integrity standards. This comprehensive review provides valuable insights for pharmaceutical professionals, researchers, and regulators seeking to navigate the complex landscape of digital transformation and data integrity in the pursuit of improved patient outcomes and business success.
J. Pereira, Ana Julia Dal Forno, L. M. Kipper
et al.
Purpose: This study aimed to analyze the occupational health and safety data in the textile industry to guide the implementation of the ISO 45001 standard. Theoretical Framework: The global estimates of the International Labor Organization show that the world economy loses about 4% of GDP annually to occupational diseases and accidents, which, in addition to human losses, result in a loss of productivity due to unsafe or unhealthy environments. Motivated by the transition from the OHSAS 18000 standard to the ISO 45001, it is necessary to understand the scenario of industries and the impact that accidents cause. Design/Methodology/Approach: From the collection of data from the state of Santa Catarina in the southern region of Brazil, a diagnosis is presented that may serve as a starting point for improvement actions regarding worker health and safety and as a benchmark for other companies in other sectors. The methodology began with analyzing the state of the art in occupational health and safety management, accident concepts, and the history of this theme worldwide. Findings: The results showed that the main accidents that occurred in the textile factories of Santa Catarina from 2012 to 2022 were with machinery and equipment, followed by accidents with chemical agents, transport vehicles, and biological agents. As for the most affected body parts, these were the fingers, followed by feet, hands (except wrists and fingers), and eyes. Another research question was to identify the sectors of the textile industry that had the most accidents in the period, which were the spinning, weaving, and textile processing sectors. Also, there were two thousand days lost in 2021 alone and, cumulatively, 45,900 days lost in this interim. Research, Practical & Social Implications: The absence of studies of this type for the textile industry and also a starting point for improvement actions regarding worker health and safety and as a benchmark for other companies in other sectors. The adoption of an Occupational Health and Safety Management System with the application of the ISO 45001 standard is a preventive and necessary measure to reduce the rates of accidents and diseases raised in the research carried out. Originality/Value: The relevance of the topic is demonstrated by the large number of accidents registered in Brazil and worldwide and, at the same time, the absence of studies of this type for the textile industry. The clipping of this sector helps to understand the data regarding the most affected body parts, the number of registered work accidents, expenses and waste for companies, their causes, and the sector in which the most accidents occur, thus guiding managers to the implementation of an effective occupational health and safety management system.
The fourth Industrial Revolution (4th IR) i.e., Industry 4.0 features the deployment of extensive automation in the form of smart factories and smart machines facilitating improved information and data flows across the value chains. It facilitates the integration of intelligent digital technologies into the production processes for enhanced productivity, efficiency, flexibility, quality, transparency, intelligent decision making, and customization. Consequently, research has been embarked on studying the main elements of the 4th IR i.e., technology pillars (Autonomous Robots/Advanced Robotics, Augmented Reality/Virtual Reality, Smart Sensors, Big Data Analytics/Machine Learning, Cloud Computing/Cognitive Computing/Edge Computing, Additive Manufacturing etc.), role of people within the highly automated environments, level of mass personalization/adaptability, sustainability performance, and the economic prospects. In this pursuit, a few studies also concentrated on listing common challenges associated with Industry 4.0 deployment within Manufacturing Systems. Nonetheless, a dearth of research is observed with regards to listing, comprehensive analysis, and prioritization of barriers towards the implementation of the 4th IR within small to medium-sized enterprises. Hence, this study aims at identifying and prioritizing an exhaustive set of barriers for successful deployment of Industry 4.0 within Manufacturing Systems through Fuzzy DEMANTEL Analysis towards making the manufacturing processes more intelligent, efficient, and precise.
Digital transformation (DX) has recently become a pressing issue for many companies as the latest digital technologies, such as artificial intelligence and the Internet of Things, can be easily utilized. However, devising new business models is not easy for compa-nies, though they can improve their operations through digital technologies. Thus, business model design support methods are needed by people who lack digital tech-nology expertise. In contrast, large language models (LLMs) represented by ChatGPT and natural language processing utilizing LLMs have been developed revolutionarily. A business model design support system that utilizes these technologies has great potential. However, research on this area is scant. Accordingly, this study proposes an LLM-based method for comparing and analyzing similar companies from different business do-mains as a first step toward business model design support utilizing LLMs. This method can support idea generation in digital business model design.
We present TinyLLaVA Factory, an open-source modular codebase for small-scale large multimodal models (LMMs) with a focus on simplicity of code implementations, extensibility of new features, and reproducibility of training results. Following the design philosophy of the factory pattern in software engineering, TinyLLaVA Factory modularizes the entire system into interchangeable components, with each component integrating a suite of cutting-edge models and methods, meanwhile leaving room for extensions to more features. In addition to allowing users to customize their own LMMs, TinyLLaVA Factory provides popular training recipes to let users pretrain and finetune their models with less coding effort. Empirical experiments validate the effectiveness of our codebase. The goal of TinyLLaVA Factory is to assist researchers and practitioners in exploring the wide landscape of designing and training small-scale LMMs with affordable computational resources.
Henry Beuster, Kevin Tebbe, Thomas Doebbert
et al.
In the past few years, there has been a growing significance of interactions between human workers and automated systems throughout the factory floor. Wherever static or mobile robots, such as automated guided vehicles, operate autonomously, a protected environment for personnel and machines must be provided by, e.g., safe, deterministic and low-latency technologies. Another trend in this area is the increased use of wireless communication, offering a high flexibility, modularity, and reduced installation and maintenance efforts. This work presents a testbed implementation that integrates a wireless framework, employing IO-Link Wireless (IOLW) and a private 5G cellular network, to orchestrate a complete example process from sensors and actuators up into the edge, represented by a programmable logic controller (PLC). Latency assessments identify the systems cycle time as well as opportunities for improvement. A worst-case estimation shows the attainable safety function response time for practical applications in the context of functional safety.
Tiago M. Fernandez-Carames, Oscar Blanco-Novoa, Ivan Froiz-Miguez
et al.
In this paper we present the design and evaluation of a UAV-based system aimed at automating inventory tasks and keeping the traceability of industrial items attached to Radio-Frequency IDentification (RFID) tags. To confront current shortcomings, such a system is developed under a versatile, modular and scalable architecture aimed to reinforce cyber security and decentralization while fostering external audits and big data analytics. Therefore, the system uses a blockchain and a distributed ledger to store certain inventory data collected by UAVs, validate them, ensure their trustworthiness and make them available to the interested parties. In order to show the performance of the proposed system, different tests were performed in a real industrial warehouse, concluding that the system is able to obtain the inventory data really fast in comparison to traditional manual tasks, while being also able to estimate the position of the items when hovering over them thanks to their tag's signal strength. In addition, the performance of the proposed blockchain-based architecture was evaluated in different scenarios.
With the advancement of digital transformation in the manufacturing industry, the industrial internet has become a key technology for improving production efficiency and product quality, and achieving intelligent manufacturing. This paper takes Foxconn Technology Group as an example to explore how it has utilized the industrial internet platform BEACON for digital transformation, as well as the practical experience and effectiveness in this process. Foxconn has integrated cloud computing, the Internet of Things, big data, mobile internet, and smart factory technologies by constructing the BEACON platform, achieving automation, intelligence, and networking in production, and enhancing management efficiency and supply chain response speed. In addition, Foxconn also focuses on talent cultivation and corporate culture construction, developing a large number of talents with innovative thinking and digital skills through internal training and educational platforms. The research conclusion of this paper points out that the industrial internet is an important driving force for the digital transformation of the manufacturing industry, and proposes policy recommendations such as strengthening infrastructure construction, promoting in-depth integration of industry, academia, research, and application, and formulating and improving relevant policies and regulations to promote high-quality development and competitiveness enhancement of the manufacturing industry.
Matti Yli-Ojanperä, S. Sierla, N. Papakonstantinou
et al.
Abstract Industry 4.0 architecture has been studied in a large number of publications in the fields of Industrial Internet of Things, Cyber Physical Production Systems, Enterprise Architectures, Enterprise Integration and Cloud Manufacturing. A large number of architectures have been proposed, but none of them has been adopted by a large number of research groups. Two major Industry 4.0 reference architectures have been developed by industry-driven initiatives, namely the German Industry 4.0 and the US-led Industrial Internet Consortium. These are the Reference Architecture Model Industry 4.0 and Industrial Internet Reference Architecture, which are being standardized by the International Electrotechnical Commission and the Object Management Group, respectively. The first research goal of this article is to survey the literature on Industry 4.0 architectures in a factory context and assess awareness and compatibility with Reference Architecture Model Industry 4.0 and Industrial Internet Reference Architecture. The second research goal is to adapt a previously proposed advanced manufacturing concept to Reference Architecture Model Industry 4.0. With respect to the first research goal, it was discovered that only a minority of researchers were aware of the said reference architectures and that in general authors offered no discussion about the compatibility of their proposals with any internationally standardized reference architecture for Industry 4.0. With respect to the second research goal, it was discovered that Reference Architecture Model Industry 4.0 was mature with respect to communication and information sharing in the scope of the connected world, that further standardization enabling interoperability of different vendors’ technology is still under development and that technology standardization enabling executable business processes between networked enterprises was lacking.
Purpose Second-order system-dynamics engineering (SOSDE) involves constructing and running enterprise manufacturing simulation models with new proposals for operational processes, byproducts, supply chain and/or downstream marketing designs. This paper aims to describe sustainability the principal lessons from enacting SOSDE research for achieving goals in large manufacturing firms. Design/methodology/approach This study is a case research commentary in the agricultural industry that contributes abductively derives six principal lessons from SOSDE research on introducing sustainability-focused manufacturing and product innovations. Operational processes in large-scale poultry processing plants in an emerging market represent the specific industry and firm domain of this case study. Alternative SOSDE simulation models of decisions, materials flow and outcomes with versus without operational innovations were constructed following one-to-one interviews with experienced farm managers and entrepreneurs. Findings The principles demonstrate how large farms in a developing nation (i.e. Bangladesh) go about adopting radically innovative manufacturing, supply chain and marketing operations to improve traditional operations. This study confirms and expands on the general observation that SOSDE can help achieve sustainability and environmental, social and governance goals, contribute new value outcomes by converting unused production wastes into valuable byproducts and introduce design efficiencies in production, supply chain and marketing processes. SOSDE complements, while being a revolutionary departure from, “six sigma management programs” that focus on achieving exceptional and near mistake-free manufacturing operations. Both represent distinct philosophies and sets of actions that sometimes can conflict with one another. Embracing both successfully in the same enterprise is a goal that may appear unreachable, seemingly impossible to achieve and yet represents a manufacturing/marketing epitome that is observable in exceptional enterprises. Research limitations/implications This paper may generate controversy as well as advance interest in applying SOSDE in introductions of improved manufacturing, supply chain and marketing operations aiming to accomplish radical improvements in sustainability goals. Practical implications This commentary describes how using SOSDE and running alternative production simulations with versus without including superior, radically new, process innovations enable the firm to find and eliminate glitches in system changes and reduce the fear associating with breakdowns and financial losses due to inadequate knowledge of operating new industrial procedures and outcomes. Social implications Introductions of superior radically new innovations in industrial manufacturing and marketing via SOSDE frequently include manufacturing firms embracing new environment sustainability objectives and additional marketable byproducts from the firm's main productions lines. This commentary offers details on how this process is enacted in poultry manufacturing in an economically emerging nation. Originality/value Running simulations in SOSDE research offers a low-cost, fast and in-depth method to test “what-if” impacts of enhanced and radical innovations into product/service manufacturing operations – benefits supporting the recommendation to apply systems dynamics in business and industrial marketing.