Industrial Management Studies

Abstract The transition toward sustainable industrial systems has underscored the necessity of integrating Circular Economy (CE) principles with Industry 4.0 technologies. Despite expanding studies in both fields, an integrated and process-oriented framework that explains the systematic interaction of these approaches to enhance sustainable performance remains absent in the existing literature. This research aimed to identify existing gaps and propose an integrated conceptual framework for improving sustainable performance based on the synergy between CE and Industry 4.0. Employing a systematic literature review (SLR) following Sandelowski and Barroso's seven-step approach, a systematic search was conducted across reputable international databases for the period 2015–2026. After applying inclusion and exclusion criteria, selected articles were analyzed. A total of 407 initial codes were extracted and, following conceptual refinement, organized into 49 indicators and 33 criteria. These indicators were categorized into five sustainable performance improvement processes: procurement, design and development, production and operations, distribution, use, and maintenance. Findings revealed that the literature focuses predominantly on production and operations, while upstream and midstream value chain processes remain under-explored. The proposed framework demonstrated that Industry 4.0 technologies act as enablers for implementing the ten principles of the Circular Economy. By rearranging five organizational processes, these technologies simultaneously enhance the economic, environmental, and social dimensions of sustainable performance. The research's novelty lay in providing an integrated, process-oriented model that elucidates the mechanisms for achieving practical sustainable performance through the nexus of CE and Industry 4.0.
Introduction
The transition toward sustainable industrial systems has highlighted the necessity of integrating Circular Economy (CE) principles with Industry 4.0 technologies. Despite growing research, an integrated, process-oriented framework explaining their systematic interaction to enhance sustainable performance remains absent. This study addressed this gap by proposing a conceptual framework demonstrating how the synergy between CE and Industry 4.0 can improve sustainable performance through reorganizing five supply chain processes: procurement, design and development, production and operations, distribution, and use and maintenance. The CE literature has evolved from 3R approaches to comprehensive frameworks such as the 10R model (Potting et al., 2017). Industry 4.0 technologies (IoT, AI, big data analytics, blockchain, additive manufacturing, and cyber-physical systems) are recognized as enablers for circular practices (Rajput & Singh, 2019; Han et al., 2023). However, existing studies focused primarily on technological dimensions or conceptual relationships, with limited attention to process-oriented implementation frameworks (Rosa et al., 2020; Patyal et al., 2022; Alsaoudi et al., 2025). Most studies focused on macro-level relationships or review technologies and principles, with less attention to practical process-oriented frameworks. Many studies do not simultaneously address all three sustainability dimensions, predominantly focusing on environmental and economic aspects. A systemic approach explaining the dynamic interaction of CE principles and Industry 4.0 within supply chain processes is rarely observed. Moreover, a study that systematically extracts and classifies sustainable performance indicators within a coherent framework is not available.
Methodology
This research employed a systematic literature review (SLR) based on qualitative meta-synthesis. SLR served as the method for systematic search, screening, and selection of articles based on a specific protocol to prevent selection bias and ensure transparency and replicability (Kitchenham & Charters, 2007; Okoli & Schabram, 2010). Meta-synthesis served as the method for analyzing and synthesizing qualitative findings, enabling extraction of codes, concepts, and indicators into a new conceptual framework (Sandelowski & Barroso, 2007; Walsh & Downe, 2005). The study followed Sandelowski and Barroso's seven-step approach (2007). A systematic search was conducted across Scopus, Web of Science, and ScienceDirect, and complementary Persian databases (Civilica, Magiran, SID) for 2015–2026. After applying inclusion/exclusion criteria and CASP quality assessment, 32 articles were selected. Data analysis employed open, axial, and selective coding, extracting 407 initial codes, refined into 49 indicators and 33 criteria. Reliability was assessed using Cohen's kappa (κ = 0.847), indicating strong agreement.
Findings
Indicators were categorized into five processes: (1) procurement, (2) design and development, (3) production and operations, (4) distribution, and (5) use and maintenance. Quantitative distribution showed production and operations with the highest share (16 indicators, 32.7%), followed by use and maintenance (11, 22.4%), design and development (10, 20.4%), distribution (7, 14.3%), and procurement (5, 10.2%). This revealed a significant research imbalance, indicating upstream and midstream processes have been underexplored. Industry 4.0 technologies acted as enablers for implementing 10R principles across all five processes, simultaneously enhancing economic, environmental, and social dimensions.
Discussion and Conclusion
Sustainable performance is achieved through strategic alignment of CE and Industry 4.0 within organizational processes. The novelty lay in providing an integrated, process-oriented model that operationalizes indicators within five concrete processes and explains how technologies enable 10R implementation. This alignment between methodology and findings validated the framework as an outcome of analyzing scattered literature data, not a predetermined model. Limitations include: (1) Framework based on literature review lacking empirical testing; (2) Selected articles limited to specific databases; (3) Time frame (2015–2026) may exclude key earlier studies; (4) CASP assessment involves subjectivity; (5) Implementation requires digital infrastructure and organizational maturity. Future research should focus on: (1) empirical validation using quantitative methods across industries; (2) investigating moderating variables including organizational size, digital maturity, and institutional pressures; (3) developing social dimensions of sustainable performance; (4) applying system dynamics to analyze long-term behavior and feedback loops; (5) conducting comparative studies across industries and developed versus developing countries.

Abstract Accurate demand forecasting in online retail is a fundamental challenge in supply chain management, as demand patterns are typically influenced by complex temporal relationships, cross-product network interactions, and item-specific semantic features. To address this challenge, this study introduced a novel forecasting framework based on a Transformer architecture, capable of modeling long-lag temporal dependencies, cross-product network effects, and item semantic information. This empirical-computational research evaluated the proposed model using online retail II datasets. The proposed model was subsequently trained utilizing a multi-head attention mechanism within the Transformer framework, incorporating semantic representations extracted from product descriptions. The results demonstrated that the model achieved a Mean Absolute Error (MAE) of 2.68 and a Root Mean Square Error (RMSE) of 3.85, outperforming both the Autoregressive Integrated Moving Average model (MAE: 4.37, RMSE: 5.96) and the Gradient Boosting algorithm (MAE: 3.91, RMSE: 5.12). This improvement indicated that the framework effectively identifies complex demand patterns and cross-product relationships.
Introduction
Precise and adaptive demand forecasting is paramount in the digitally transformed retail landscape. Online platforms generate vast data, presenting opportunities and challenges. Demand is shaped by trends, seasonality, promotions, network externalities (influences between complementary/substitute products), and product characteristics. Traditional models like ARIMA capture linear temporal dependencies but often miss nonlinear relationships and external factors. While advanced machine learning models handle nonlinearity, they may not adequately integrate diverse data types (time series, network structures, textual semantics). This study proposed a unified deep learning framework, the Transformer, excelling at capturing long-range dependencies via its attention mechanism. By explicitly modeling temporal lags, product network effects, and semantic features, our approach aimed for a more comprehensive and accurate representation of demand generation in online retail, pushing forecasting boundaries.
Literature Review
Demand forecasting in online retail is vital for supply chain and inventory management. Traditional time-series models like ARIMA and Exponential Smoothing capture linear temporal dependencies but are limited by the increasing complexity of modern online retail environments, characterized by intricate temporal relationships, inter-product network effects, and rich semantic item characteristics (Tiwari, 2025; Chowdhury et al., 2025). Contemporary deep learning architectures, including RNNs (LSTMs, GRUs), model complex temporal patterns but may struggle with long-range dependencies or nuanced interactions (Cai et al., 2021).
The Transformer architecture, with its self-attention mechanism, excels at capturing long-range and context-dependent relationships, revolutionizing sequence modeling (Rahman et al., 2025). Its adoption in time-series forecasting, including retail demand prediction, is growing (Wang, 2025; Li, 2023). However, existing Transformer-based research often focuses primarily on historical sales data (temporal lags) in isolation. A significant gap remains in simultaneously modeling the multifaceted influences on demand, specifically the interplay of:

Network Effects: Product demand is interdependent; influenced by complements and substitutes (Caetano et al., 2025). Capturing these structural interdependencies is crucial for dynamic patterns, especially in large catalogs.
Temporal Lags: The impact of events or changes may not be immediate. Accurately modeling these varied time delays is vital for precise forecasting (Tarighat et al., 2025).
Product Semantics: Product characteristics and descriptions significantly influence consumer choice. Leveraging semantic information can uncover hidden relationships and improve accuracy (Rahikka & Mikkola, 2025).

While studies have explored these factors individually or in limited combinations (e.g., Cai et al., 2021; Wang, 2025), a comprehensive framework integrating temporal dependencies, network effects, and product semantics within a single Transformer architecture remains an active research area. This study bridged this gap by proposing a novel Transformer-based framework to jointly model these dimensions for a more accurate and holistic approach to dynamic retail demand forecasting.
Method
This study employed an empirical-computational approach using the Online Retail II dataset. Data preprocessing involved handling missing values, correcting errors, and removing irrelevant transactions. The dataset was split chronologically into an 80% training set and a 20% testing set.
Our core forecasting model was the Transformer architecture, utilizing its self-attention mechanism to weigh historical time steps for predicting future demand, thus capturing complex temporal patterns and long-range dependencies. Positional encodings were added to input embeddings for sequence information.
Our framework innovatively integrated three key information sources:

Temporal Lags: Explicitly accounts for lagged demand patterns. While attention captures temporal relationships, features representing past demand values at different lags can be incorporated.
Network Effects: Models product interdependencies through a product interaction graph (nodes=products, edges=relationships like co-purchase frequency). Graph-derived features (e.g., centrality) processed by techniques like Graph Convolutional Networks (GCNs) are integrated.
Product Semantics: Extracts rich semantic embeddings from product descriptions using NLP techniques (e.g., pre-trained models). These capture nuanced characteristics and are integrated alongside temporal and network features.

The model was trained end-to-end using Mean Squared Error (MSE) or Mean Absolute Error (MAE), optimized via an Adam optimizer. Hyperparameters were tuned for optimal performance. Final evaluation used MAE and RMSE on the test set.
Results
The empirical evaluation on the Online Retail II dataset showed the superiority of our proposed Transformer-based framework. On the test set, the model achieved an MAE of 2.68 and an RMSE of 3.85. This substantially outperformed baseline models: ARIMA yielded MAE 4.37, RMSE 5.96; XG-Boost achieved MAE 3.91, RMSE 5.12. These quantitative results highlighted the enhanced predictive accuracy of our integrated approach, attributed to its ability to capture the complex interplay of temporal, network, and semantic factors. Visualizations of predicted versus actual demand confirmed the model’s superior tracking of demand fluctuations. Analysis of attention weights also provided insights into influential historical periods and product relationships, offering a degree of interpretability.
Discussion
The findings underscored the advantages of a Transformer-based architecture for dynamic retail demand forecasting when augmented with explicit modeling of temporal lags, network effects, and product semantics. The accuracy improvements over ARIMA and XG-Boost highlighted the limitations of models neglecting the interconnected nature of retail demand. Our approach achieved a more comprehensive representation by jointly considering these factors. The attention mechanism effectively captured long-range temporal dependencies and contextual relationships. Explicit incorporation of network effects revealed the critical influence of product interdependencies, often overlooked in single-item forecasting. Leveraging product semantics allowed deeper understanding of demand drivers beyond historical sales. This integrated approach offered retailers enhanced capabilities for inventory management, demand planning, and strategic decision-making, potentially reducing costs and stockouts while improving customer satisfaction. Attention weight analysis offered valuable insights for model trust and adoption.
Conclusion
This research successfully developed and validated a novel Transformer-based framework for dynamic retail demand forecasting, integrating temporal lags, network effects, and product semantics. Empirical results on the Online Retail II dataset demonstrated the model’s superiority, achieving significantly lower error metrics (MAE: 2.68, RMSE: 3.85) compared to ARIMA (MAE: 4.37, RMSE: 5.96) and XG-Boost (MAE: 3.91, RMSE: 5.12). This work contributed a unified and highly accurate methodology for forecasting in complex retail environments, offering substantial practical implications for inventory management and strategic decision-making. Future research could incorporate additional external factors, explore advanced GNN architectures, and develop more sophisticated interpretability methods for real-world adoption.
Table 1. Comparison of Demand Forecasting Model Performance





Model


Mean Absolute Error (MAE)


Mean Squared Error (MSE)




ARIMA


37.4


96.5




XG-Boost


91.3


12.5




Proposed Transformer


68.2


85.3








Table 2. Sensitivity Analysis of the Proposed Model (Baseline RMSE is 1.96)





Main Group


Parameter


Baseline Value


Applied Change


New RMSE


RMSE Change (%)






Temporal Dynamics


Input Window Length


12


10
15


2.04 2.00


+4.2% +1.8%




Prediction Horizon


3


5


2.11


+7.8%




Numerical Inputs


Count


10.33


±10% Random Noise


1.98


+0.8%




Price


4.69


±10% Random Noise


1.99


+1.3%




Model Architecture


Graph Information


Present


Component Removal


2.09


+6.5%




Attention Mechanism


Present


Weighting Noise


1.97


+0.5%




Optimization


Learning Rate


0.0005


0.001 0.00025


1.99 2.01


+1.3% +2.6%

Abstract The industrial sector, due to its significant share in national electricity consumption and its strategic role in sustaining production and economic growth, holds a prominent position in energy policymaking. However, the development of renewable electrical energy in this sector requires multidimensional alignment among economic, social, environmental, and technological dimensions, which in practice faces challenges related to policies, infrastructure, and implementation mechanisms. This study aims to propose a multidimensional alignment framework for the sustainability of renewable electrical energy in the industrial sector and to analyze the interactions among influential indicators using Fuzzy Cognitive Maps (FCM). First, key indicators were identified through a systematic review of the research literature. Subsequently, the causal relationships among these indicators were determined based on the opinions of 20 experts, including energy policymakers, industrial managers from both public and private sectors, and academic scholars. Data analysis were conducted using the Fuzzy Cognitive Maps approach with the FCMapper and UCINET software tools. The results indicate that factors such as electricity price and demand, self-sufficiency in electricity generation, energy conservation in industry, installation of solar photovoltaic systems, and climate change act as the main driving forces of the system. Additionally, variables such as industrial working capital, industrial self-reliance, reduction of fossil fuel consumption, and environmental protection play bidirectional roles, while the reduction of carbon dioxide (CO₂) emissions are identified as a response variable. The proposed framework provides a basis for strategic decision-making and for developing a roadmap toward the sustainable development of renewable electrical energy in the industrial sector.

Abstract The flow shop scheduling problem with sequence-dependent setup times and operator constraints represents a complex and practically relevant production planning challenge, as it simultaneously integrates sequencing, timing, and human resource allocation decisions. Exact solution approaches become computationally intractable for medium- and large-scale instances due to the combinatorial nature of the problem. This study proposes an integrated modeling and solution framework aimed at minimizing total order tardiness while jointly incorporating key operational constraints, including technological precedence, machine non-overlapping, sequence-dependent setup times, and operator capacity limitations. A mixed-integer linear programming (MILP) model is first developed to explicitly capture sequencing, scheduling, and operator assignment decisions within a unified structure. Given the computational complexity of the model, a metaheuristic solution approach based on Particle Swarm Optimization (PSO) is designed. To accommodate the combinatorial structure of the problem, a continuous encoding mechanism combined with a constructive decoder is implemented to enforce feasibility during solution evaluation. The proposed approach is validated through an industrial case study and multiple independent runs with different random seeds. Performance is assessed using the best objective value, mean and standard deviation of results, computational time, and convergence behavior. Results demonstrate that the proposed PSO achieves high-quality and stable solutions with acceptable computational effort. Sensitivity analysis indicates that stronger exploration enhances robustness and solution quality at the cost of longer runtime, whereas exploitative settings accelerate convergence but may reduce solution quality.

Abstract In the competitive automotive parts manufacturing industry, accurately identifying and prioritizing failure modes is essential for preventing financial losses, ensuring passenger safety, and maintaining quality standards, since systematic risk management directly affects production efficiency and product reliability. This applied, descriptive-analytical study aimed to identify and prioritize failure modes in the constant-velocity (CV) joint assembly process of an Iranian automotive parts manufacturer by integrating the Fuzzy Best-Worst Method (FBWM) and Fuzzy VIKOR (FVIKOR). A panel of fifteen purposively selected experts, each with at least ten years of relevant experience, took part in a three-round Delphi survey using a five-point Likert scale to screen the candidate risk factors and failure modes; the retained risk factors were then weighted using FBWM, and the screened failure modes were ranked using FVIKOR. The findings showed that severity received the highest relative weight among the retained risk factors, and that failure modes associated with the tensile strength and dimensional and spline tolerances of the CV-joint components received the highest priority for corrective action. The pairwise comparisons of the expert panel showed a high level of consistency, and combining Delphi screening with fuzzy multi-criteria weighting and ranking improved prioritization accuracy compared with the traditional risk priority number approach. The findings provide production managers with a reliable and replicable basis for preventive decision-making and process optimization in automotive parts manufacturing.

Abstract The oil and gas industry faces challenges such as complexity of a multi-layered and global supply network, operational and geopolitical risks, cybersecurity threats in critical infrastructure, sustainability pressures, and need for real-time data-driven decision-making. The main objective of the study is to design and analyze an autonomous supply chain framework based on artificial intelligence, big data and Internet of Things in the oil and gas industry with an emphasis on sustainability and cybersecurity. The research is applied and developmental in terms of its purpose; qualitative-quantitative in terms of the nature of the data; and descriptive in terms of the data collection method, which conducted using the Fuzzy Cognitive Mapping method. The study period is summer and fall 2025 and winter 2026. The population of the study is Iranian oil and gas industry experts, which completed by 23 experts using a structured questionnaire using purposive sampling. Achieving autonomy in the oil and gas supply chain is contingent upon the synergy and application of advanced digital technologies. Within this context, operational intelligence serves as a crucial mediating factor, transforming technological capabilities into economic and environmental value. In other words, the establishment of data-driven digital infrastructure is the primary prerequisite for attaining sustainability, while resilience and trust within this ecosystem can only be realized through robust cybersecurity. The proposed model, functioning as a decision support system, empowers managers in this industry to optimally prioritize digital investments, evaluate the implications of security policies, and simulate various digital transformation scenarios prior to operationalization.

Abstract With the advent of artificial intelligence (AI), human resource management has encountered fundamental opportunities and challenges, making the enhancement and sustainability of knowledge workers’ productivity a critical concern. This study aimed to analyze the factors influencing the experience of knowledge workers in the AI era and to develop a managerial roadmap within a mission-based university context. This applied research adopted an exploratory–explanatory, cross-sectional mixed-methods design based on an inductive approach. Data were collected using researcher-developed questionnaires grounded in the ISO 34000 Human Resource Management Standard and validated through an expert panel of university specialists. Interpretive Structural Modeling (ISM) and MICMAC analysis were employed to identify and classify the relationships among the factors, while the Analytic Network Process (ANP) was used to determine their relative importance and weights. The findings revealed that intelligent leadership and organizational culture are the most influential underlying drivers, exerting the greatest impact on other dimensions of employee experience. These factors serve as the foundation for shaping and improving the work environment of knowledge employees in the AI era. Based on the results, a comprehensive managerial roadmap was developed and validated, comprising five phases: foundational preparation, managerial and cultural transformation, job experience optimization, emergence of innovation and scientific leadership, and retention of knowledge capital. The proposed roadmap provides a practical framework for effective policymaking and strategic decision-making in knowledge-based organizations, particularly higher education institutions undergoing AI-driven transformation.