International Journal of Industrial Engineering: Theory, Applications and Practice
Not a member yet
943 research outputs found
Sort by
Two-Stage Case-Based Reasoning Framework With Tree-Based Ensemble Methods for Printed Circuit Board Yield Prediction
No full textIn printed circuit board (PCB) manufacturing, yield is a critical indicator that directly impacts production costs and delivery schedules. For new products, the industry currently relies on empirical predictions by engineers based on historical yields of similar products. However, this approach suffers from limitations in accuracy and consistency due to subjective and non-systematic assessment criteria, issues that are compounded by the complex product structures and multi-stage processes used in the PCB industry. To address these challenges, this study proposes a yield prediction methodology that leverages Case-Based Reasoning (CBR) and tree-based prediction models. This methodology consists of two stages. The first stage is to predict the yield of model units before the new product is put into the first process. In this stage, we utilize SHapley Additive exPlanations (SHAP)–based variable importance to reflect the weight of each variable and determine the optimal number of clusters based on silhouette analysis to search for similar cases. Then, data from retrieved cases is analyzed with a tree-based prediction model to predict yield, which evaluates the relationship between dates and yield to detect whether a statistically significant change has occurred at a particular point in time. The second stage utilizes additional defect information from Automated Optical Inspection (AOI) reports to make more precise yield forecasts on a lot-by-lot basis. This further increases the reliability of the forecast by reflecting the quality variability that occurs during production. Based on these predictions, more accurate material input calculations can be performed, thereby improving delivery compliance rates and minimizing excess inventory
Balancing Driver Workloads Through Two-Stage Service Area Assignment in Last-Mile Delivery
No full textThis study addressed a two-stage delivery assignment problem in e-commerce with short- and mid-term planning phases. The mid-term phase groups small service areas into core areas and assigns them to drivers. These core areas are refined in the short-term phase by allocating unassigned areas based on mid-term results. This study aims to assign service areas to drivers to minimize delivery time while balancing workloads under preferences and practical constraints. It is formulated by extending traditional districting models. This study proposes two heuristic algorithms. The Relax-and-Fix heuristic relaxes decision variables to solve subproblems, then gradually fixes them to construct an initial solution. The Fix-and-Optimize heuristic fixes variable subsets and applies local search for improvement. Computational experiments using real-world data demonstrated good solution quality and computational efficiency. The heuristics also balance workloads effectively without a substantial increase in delivery time
Adaptive Kernel Learning Online Support Vector Regression for Predicting Sulfur Oxides Emissions in Coal-Based Power Plants
No full textWith the increasing global population and rapid urbanization, power plants are required to meet the growing energy demand. As a result, there has been a notable increase in environmental pollutants, particularly from coal-based plants, including carbon dioxide (CO₂), sulfur oxides (SOₓ), and nitrogen oxides (NOₓ), which pose threats to human health and the environment. Existing monitoring techniques face challenges in accurately predicting emissions due to the changing dynamics of boiler operations. Consequently, inaccuracies and insufficient control measures may occur. To address the dynamic characteristics of sensor data, a novel, accurate, and online support vector regression approach is proposed. It presents an online system for predicting SOₓ emissions from coal-fired plants, addressing dynamic changes in boiler systems. The method optimizes kernel width σ, penalty C, and error ε using the sine cosine algorithm. Based on a real-life dataset from over 400 sensors, the proposed system outperformed existing methods in predicting SOₓ emissions
Key Factors of Technology Transfer in The Moroccan Aerospace Sector: An Analysis Using Design of Experiments and SERVQUAL
No full textThe development of Morocco's aerospace sector largely depends on its ability to attract foreign companies and promote technology transfer to local actors. In this context, identifying the key factors driving this dynamic is a strategic challenge to strengthen the country's competitiveness within the global value chain. This paper addresses this issue by applying a rigorous methodology to highlight the most relevant levers for supporting the process. In fact, the Design of Experiments (DOE) methodology, well-recognized for its effectiveness across various fields, was adopted. The study was conducted in collaboration with three categories of stakeholders: Moroccan aerospace professionals, relevant government officials, and academic researchers. Together, they established a list of potential factors derived from an extensive literature review. Data evaluation relied on the SERVQUAL method, applied to three distinct responses corresponding to the three stakeholder groups. This approach allowed for more precise and reliable analysis and comparison of results. The findings provide a valuable decision-making tool for stakeholders and policymakers, enabling them to target the most influential levers to reinforce technology transfer in the Moroccan aerospace sector
Patent Citation Forecasting with Machine Learning Techniques in Supply Chain Technology Management
No full textIn today's rapidly evolving technological landscape, innovation in supply chain technologies is essential for sustaining competitive advantage. This study aims to forecast patent citations, which are useful for evaluating the quality and potential impact of patents in supply chain management. Using a dataset of 12,225 patents from lens.org, various machine learning models, including Multiple Linear Regression (MLR), Ridge, Lasso, Artificial Neural Networks (ANN), Support Vector Regression (SVR), Regression Trees (RT), and Random Forest (RF), were applied to predict forward patent citations. Model performance was assessed using RMSE and R² metrics. Among all the models, RF exhibited the highest accuracy (RMSE = 0.0821, MAE = 0.0135). These findings highlight the effectiveness of machine learning, particularly RF, in identifying high-impact patents. This approach offers valuable insights for researchers and practitioners by providing a data-driven method for assessing technological innovation and patent value in the supply chain domain
Current Status, Hot Topics, and Trends in Green Certificate Research: A Bibliometric Analysis Based on WoS Data
No full textGreen certificates are pivotal in sustainable energy governance amid global climate action. This study addresses gaps in systematic bibliometric analysis of their evolution, analyzing 459 Web of Science articles (2001–2024) via R bibliometrix across six dimensions. Four novel and valuable findings are identified. First, global scholarly interest in green certificates demonstrates sustained annual growth, with notable acceleration post-2021 when research output peaked. Second, China, the USA, and Sweden constitute the primary research contributors, collectively producing 187 documents. Third, contemporary green certificate research predominantly features solitary authorship or limited-scale collaborations. Fourth, contemporary focal points in green certificate research concentrate on policy design, system optimization for power grids, market mechanism innovation, and operational efficiency analysis of certification systems. Findings advance understanding of green certificate dynamics, offering strategic insights for optimizing policy design, market integration, and renewable energy transitions
A Hybrid AI-Driven Framework for Resilient Blood Supply Chain Optimization Under Uncertainty
No full textDesigning a resilient and sustainable blood supply chain is vital for maintaining continuous access to life-saving blood products, particularly during uncertain and disruptive conditions. This study presents a hybrid framework that combines deep-learning-based demand forecasting with a multi-objective robust optimization model. A Long Short-Term Memory network generates fuzzy forecasts under optimistic, base, and pessimistic scenarios. These forecasts are then defuzzified and used as inputs for an optimization model that aims to minimize operational cost, reduce blood shortages, and lower environmental impact through improved facility location, donor allocation, and transportation planning. The framework is tested using large-scale evolutionary optimization and validated on smaller instances with an exact mathematical solver. A real case study from Fars Province, Iran, shows improved forecasting accuracy, stronger resilience under capacity disruptions, and balanced trade-off solutions for decision-makers. The proposed approach offers a practical tool for enhancing blood supply chain performance under uncertainty
Decarbonization Through Digital Twins: Integrating Management Competencies and Industrial Energy Efficiency
No full textIndustrial sectors face mounting pressure to reduce greenhouse gas emissions while maintaining productivity, prompting growing interest in advanced digitalization strategies. Among these, digital twins (DTs) have emerged as a powerful approach for achieving significant energy efficiency gains and supporting decarbonization objectives. This review critically examines the foundations, applications, and implications of DTs in energy-intensive industries. We begin by outlining the technical principles that differentiate DTs from digital models and shadows, emphasizing their bidirectional connectivity, predictive capacity, and real-time synchronization with physical assets. We then survey a range of applications—from manufacturing processes and building energy management to grid integration and transportation systems—where DTs have delivered measurable improvements in energy optimization, maintenance scheduling, and operational flexibility. Evidence from both research and practice highlights reductions in energy consumption of 20–40% and improved integration of renewable energy sources. Beyond technical mechanisms, this review underscores the pivotal role of organizational and managerial competencies in shaping DT outcomes. Successful implementations require workforce training, leadership engagement, and effective data governance, while challenges such as high upfront costs, data quality issues, and cybersecurity risks remain barriers to widespread deployment. We also discuss controversies regarding whether DTs are overhyped or underutilized, and emphasize the need for transparent evaluation metrics and interdisciplinary collaboration. By integrating technical innovation with managerial alignment, DTs have the potential to become a cornerstone of industrial strategies for sustainability and competitiveness, accelerating the transition toward low-carbon operations
A Comprehensive Approach to Voice of Customer Extraction: A Case of Product Development for Smart Washing Machine
No full textWhile customers in the past were satisfied with products that met basic functional needs, contemporary users increasingly expect enriched experiences throughout their interactions with products and services. To address this shift, organizations are adopting customer experience frameworks aligned with human factors and ergonomics, emphasizing usability, satisfaction, and system-level performance. The Voice of the Customer (VoC) has become a critical input for capturing user requirements and informing experience-centered design. This study proposes a structured process for systematically collecting VoC related to holistic user experiences and evaluates its applicability through a case study involving a consumer appliance. Fifteen participants performed real-world tasks with a washing machine, enabling the identification of fine-grained user activities—such as IoT application control and mid-cycle laundry insertion—not documented in prior research. The process also elicited novel forms of user feedback that had previously been overlooked. As products become increasingly complex and feature-rich, accounting for diverse user interactions is essential for advancing both usability and safety. The proposed VoC collection process provides a practical contribution to human-centered product and service development, supporting the integration of ergonomic considerations for the enhancement of overall user experience
Ergonomic Logistics Optimization: Multi-Objective Ant Colony and Fuzzy Logic Approach for Fatigue Management
No full textThe rapid growth of last-mile logistics has intensified concerns regarding worker fatigue, particularly in tropical operating environments characterized by high temperature, humidity, noise, and time pressure. While most logistics optimization studies focus primarily on minimizing distance, time, and operational costs, limited attention has been given to integrating ergonomic and physiological factors into decision-making models. This gap highlights the need for optimization approaches that balance operational efficiency with worker health and safety. This study proposes an integrated ergonomic logistics optimization framework that combines Multi-Objective Ant Colony Optimization with a Takagi–Sugeno Fuzzy Inference System to jointly address routing efficiency and fatigue management. The proposed model incorporates environmental exposure indicators and physiological workload measures to estimate fatigue risk and embed it within multi-objective decision processes. The framework operates through fatigue prediction, optimization of delivery routes under ergonomic constraints, and adaptive evaluation of rest scheduling policies. The results indicate that the integrated approach produces more balanced solutions compared to conventional distance-based optimization strategies, improving system performance while mitigating excessive fatigue accumulation. The findings also reveal the limitations of static regulatory rest standards when applied to dynamic and high-stress logistics contexts. Theoretically, this study extends multi-objective metaheuristic optimization by embedding human-centered performance variables into logistics modeling. Practically, it provides a decision-support mechanism for fatigue-aware route planning and adaptive work-rest management. Overall, the research advances the development of data-driven, ergonomically informed logistics systems that promote sustainable operational performance and worker well-being