University of Isfahan Research in Production and Operations Management 2981-0329 17 1 2026 03 21 The Impact of Energy Feedback on Occupant Energy Conservation Behavior: an Agent-Based Approach The Impact of Energy Feedback on Occupant Energy Conservation Behavior: an Agent-Based Approach 1 24 29934 10.22108/pom.2025.140474.1599 FA Mohammad Zareie Department of Civil Engineering, Faculty of Engineering, Ferdowsi University of Mashhad, Mashhad, Iran Amirhossein Hosseini Sarcheshmeh Master's Degree in Construction Engineering and Management, Department of Civil Engineering, Faculty of Engineering Mojtaba Maghrebi Department of Civil Engineering, Faculty of Engineering, Ferdowsi University of Mashhad, Mashhad, Iran Journal Article 2024 12 01 Purpose: The purpose of this study is to enhance the effectiveness of energy-saving awareness events by identifying optimal participant selection methods. Given the global and national importance of reducing energy consumption, the research investigates how residents’ behavioral characteristics—such as energy consumption level, behavioral adaptability, and social influence—affect the overall impact of awareness events. The study develops an agent-based simulation model to analyze the long-term effects of different participant selection strategies on community-wide energy consumption patterns. Design/methodology/approach: This research employs an agent-based modeling (ABM) approach to simulate dynamic interactions among residents in a social network. The model integrates mathematical frameworks for opinion diffusion and behavioral change to represent how individuals’ energy consumption patterns evolve following an awareness event and through peer interactions. A synthetic community of agents is created, each defined by three attributes: the Energy Consumption Index (ECI), Susceptibility Index (SI), and Social Connections (SC). Several event participation selection methods are evaluated, including (S1) random selection, (S2) selection based on behavioral adaptability, (S3) selection based on social influence, (S4) selection based on energy consumption level, and (S5) a combined method integrating all three criteria. The model is implemented in Python and simulated across communities of varying sizes (100, 1,000, and 10,000 agents) to evaluate scalability and generalizability. Findings: Simulation results reveal that the method of participant selection significantly influences the medium-term effectiveness of energy-saving events. Specifically, selecting participants based on a combination of energy consumption level, social influence, and behavioral adaptability (S5) yields the highest energy-saving outcomes. The study also finds that the positive impact of optimized participant selection increases with community size, emphasizing the role of network effects in promoting behavioral diffusion. In addition, sensitivity analysis further shows that event success rate and participant percentage have strong effects on overall performance, while the average number of social connections has a marginal influence. The model outputs align with empirical findings from previous studies, where behavioral interventions typically achieve 5–12% energy savings, confirming the model’s validity as a predictive and decision-support tool. Research limitations/implications: Although grounded in well-established theoretical models, this study’s primary limitation is the absence of empirical validation using real-world behavioral data. Capturing accurate social interaction patterns and quantifying behavioral attributes in practice remains challenging. Future research should conduct field experiments to empirically verify the model’s outcomes, refine parameter estimation for behavioral traits, and explore cross-cultural differences in behavioral diffusion. Integrating real consumption datasets could also enhance model accuracy and policy relevance. Practical implications: The findings offer actionable insights for policymakers, utility companies, and event organizers aiming to design cost-effective energy awareness programs. By strategically selecting participants with high social influence and behavioral adaptability, the overall impact of awareness campaigns can be maximized even under budget constraints. This approach enables more efficient resource allocation, reduces campaign costs, and increases community-wide behavioral adoption rates. Additionally, the framework can guide the development of targeted reward and penalty systems that promote collective energy efficiency. Social implications: This research contributes to sustainable energy management and environmental responsibility by encouraging community-based behavioral change. The results highlight the importance of leveraging social influence to spread energy-saving habits, potentially leading to long-term cultural shifts toward energy efficiency. Implementing such optimized event strategies could improve public awareness, foster social cooperation, and enhance quality of life through reduced energy costs and environmental impact. Originality/value: This study presents a novel, integrated agent-based framework that simultaneously considers individual energy behavior, adaptability, and social influence to optimize participant selection for awareness events. Unlike previous studies that focus solely on event design or social network structure, this work uniquely bridges behavioral theory and computational modeling to propose a systematic, data-driven method for improving energy-saving program efficiency. The model can serve as a decision-support tool for designing effective social interventions in both residential and urban contexts. Purpose: The purpose of this study is to enhance the effectiveness of energy-saving awareness events by identifying optimal participant selection methods. Given the global and national importance of reducing energy consumption, the research investigates how residents’ behavioral characteristics—such as energy consumption level, behavioral adaptability, and social influence—affect the overall impact of awareness events. The study develops an agent-based simulation model to analyze the long-term effects of different participant selection strategies on community-wide energy consumption patterns. Design/methodology/approach: This research employs an agent-based modeling (ABM) approach to simulate dynamic interactions among residents in a social network. The model integrates mathematical frameworks for opinion diffusion and behavioral change to represent how individuals’ energy consumption patterns evolve following an awareness event and through peer interactions. A synthetic community of agents is created, each defined by three attributes: the Energy Consumption Index (ECI), Susceptibility Index (SI), and Social Connections (SC). Several event participation selection methods are evaluated, including (S1) random selection, (S2) selection based on behavioral adaptability, (S3) selection based on social influence, (S4) selection based on energy consumption level, and (S5) a combined method integrating all three criteria. The model is implemented in Python and simulated across communities of varying sizes (100, 1,000, and 10,000 agents) to evaluate scalability and generalizability. Findings: Simulation results reveal that the method of participant selection significantly influences the medium-term effectiveness of energy-saving events. Specifically, selecting participants based on a combination of energy consumption level, social influence, and behavioral adaptability (S5) yields the highest energy-saving outcomes. The study also finds that the positive impact of optimized participant selection increases with community size, emphasizing the role of network effects in promoting behavioral diffusion. In addition, sensitivity analysis further shows that event success rate and participant percentage have strong effects on overall performance, while the average number of social connections has a marginal influence. The model outputs align with empirical findings from previous studies, where behavioral interventions typically achieve 5–12% energy savings, confirming the model’s validity as a predictive and decision-support tool. Research limitations/implications: Although grounded in well-established theoretical models, this study’s primary limitation is the absence of empirical validation using real-world behavioral data. Capturing accurate social interaction patterns and quantifying behavioral attributes in practice remains challenging. Future research should conduct field experiments to empirically verify the model’s outcomes, refine parameter estimation for behavioral traits, and explore cross-cultural differences in behavioral diffusion. Integrating real consumption datasets could also enhance model accuracy and policy relevance. Practical implications: The findings offer actionable insights for policymakers, utility companies, and event organizers aiming to design cost-effective energy awareness programs. By strategically selecting participants with high social influence and behavioral adaptability, the overall impact of awareness campaigns can be maximized even under budget constraints. This approach enables more efficient resource allocation, reduces campaign costs, and increases community-wide behavioral adoption rates. Additionally, the framework can guide the development of targeted reward and penalty systems that promote collective energy efficiency. Social implications: This research contributes to sustainable energy management and environmental responsibility by encouraging community-based behavioral change. The results highlight the importance of leveraging social influence to spread energy-saving habits, potentially leading to long-term cultural shifts toward energy efficiency. Implementing such optimized event strategies could improve public awareness, foster social cooperation, and enhance quality of life through reduced energy costs and environmental impact. Originality/value: This study presents a novel, integrated agent-based framework that simultaneously considers individual energy behavior, adaptability, and social influence to optimize participant selection for awareness events. Unlike previous studies that focus solely on event design or social network structure, this work uniquely bridges behavioral theory and computational modeling to propose a systematic, data-driven method for improving energy-saving program efficiency. The model can serve as a decision-support tool for designing effective social interventions in both residential and urban contexts.

https://jpom.ui.ac.ir/article_29934_bff15197c8757dabad06cb88a8688fa6.pdf

University of Isfahan Research in Production and Operations Management 2981-0329 17 1 2026 03 21 Identification, localization, and Causal Modeling of Barriers to the Adoption and Implementation of Industry 4.0 in Steel Companies Identification, localization, and Causal Modeling of Barriers to the Adoption and Implementation of Industry 4.0 in Steel Companies 25 50 30081 10.22108/pom.2025.145805.1625 FA Reza Soleymani-Damaneh Department of Management, Faculty of Administrative Sciences and Economics, Vali-e-Asr University of Rafsanjan, Rafsanjan, Iran Leila Ghoroghchian Department of Management, Faculty of Administrative Sciences and Economics, Vali-e-Asr University of Rafsanjan, Rafsanjan, Iran Journal Article 2025 07 19 Purpose: The primary purpose of this study is to identify, localize, and analyze the key barriers hindering the adoption and implementation of Industry 4.0 technologies in the Iranian steel industry. Despite global momentum toward digital transformation, emerging economies face various challenges that impede the transition to smart manufacturing. Given the strategic importance of the steel industry for national development, this paper seeks to bridge the knowledge gap by developing a contextualized model of Industry 4.0 barriers tailored to Iran's socioeconomic and technological environment. The research also aims to provide managers and policymakers with actionable insights to prioritize barriers, understand their causal interactions, and design strategies to overcome the most influential impediments. Design/methodology/approach: This study adopts an applied, exploratory, and descriptive research design. The methodological framework combines the Fuzzy Delphi Method and the Fuzzy DEMATEL technique in two stages. First, an extensive systematic literature review was conducted, yielding 53 potential Industry 4.0 barriers. Through screening and expert evaluation, a refined list of 15 barriers was selected for assessment. Using Fuzzy Delphi with linguistic variables, input from 14 steel industry experts was collected to validate and localize the barriers. Barriers not meeting the threshold value were eliminated, resulting in a final approved set of 12 barriers. In the second stage, Fuzzy DEMATEL was employed to map the causal relationships among the barriers. Pairwise comparisons provided by experts were transformed into normalized matrices, defuzzified using CFCS, and analyzed to determine influence, dependence, and causal effects. Findings: The findings show that the Iranian steel sector faces a complex, interrelated set of challenges in implementing Industry 4.0. Fuzzy Delphi results indicate that three barriers—resistance to change, intellectual property issues, and sanction-related constraints—were not validated by experts. Fuzzy DEMATEL analysis identifies economic–financial constraints, technological limitations, and managerial barriers as the most influential root causes. High initial investment costs, outdated technological infrastructure, low technological maturity, and weak managerial commitment exert strong causal effects across the entire barrier network. Conversely, organizational misalignment, human resource skill gaps, lack of supply chain collaboration, and insufficient governmental support are dependent barriers heavily influenced by the root causes. The causal diagram suggests that improvements in managerial leadership, strategic planning, investment capacity, and technological readiness would produce positive effects across organizations and supply chains, thereby accelerating Industry 4.0 adoption. Research limitations/implications: This study is limited by its reliance on expert judgment within a single national context, which may constrain generalizability to other countries or industries. Additionally, as Industry 4.0 technologies evolve rapidly, some barriers may change over time or be replaced by new challenges. Future research could extend this work through longitudinal analyses as digital transformation progresses and by applying the model to other industrial sectors for comparative insights. Despite these limitations, the study offers a robust methodological framework and a localized understanding of Industry 4.0 barriers in emerging economies. Practical implications: The study provides actionable insights for steel industry managers, executives, and policymakers by identifying the most critical leverage points for successful digital transformation. The results stress that overcoming economic-financial and technological barriers—through investment in digital infrastructure, upgrading legacy systems, and strengthening R&D capabilities—is essential to enable Industry 4.0 adoption. Managerial implications include the need to reinforce strategic digital planning, develop clear roadmaps, enhance leadership commitment, and cultivate a digital culture within organizations. The findings also recommend improving collaboration across supply chain partners, expanding employee training programs, and designing targeted governmental incentives. Addressing these areas can reduce risks, boost efficiency, and support long-term transformation. Social implication: This research highlights the societal importance of accelerating digital transformation in the steel industry. Industry 4.0 can enhance resource efficiency, reduce environmental impact, improve workplace safety, and raise employees’ quality of life by minimizing hazardous manual tasks. From a broader perspective, successful Industry 4.0 implementation can strengthen national industrial competitiveness, advance sustainable development goals, and stimulate innovation ecosystems. The study’s insights can guide governmental institutions in shaping supportive regulatory frameworks, standardization policies, and investment incentives that facilitate responsible and inclusive digital transformation within the steel industry. Originality/value: This paper offers one of the first comprehensive, contextualized analyses of Industry 4.0 barriers in the Iranian steel industry using a hybrid Fuzzy Delphi–Fuzzy DEMATEL approach. Its originality lies in developing a localized set of barriers validated by domain experts and in mapping complex causal relationships among barriers using fuzzy logic. The study provides substantial value to researchers investigating digital transformation barriers and to practitioners seeking evidence-based strategies to accelerate Industry 4.0 adoption in steel manufacturing. Purpose: The primary purpose of this study is to identify, localize, and analyze the key barriers hindering the adoption and implementation of Industry 4.0 technologies in the Iranian steel industry. Despite global momentum toward digital transformation, emerging economies face various challenges that impede the transition to smart manufacturing. Given the strategic importance of the steel industry for national development, this paper seeks to bridge the knowledge gap by developing a contextualized model of Industry 4.0 barriers tailored to Iran's socioeconomic and technological environment. The research also aims to provide managers and policymakers with actionable insights to prioritize barriers, understand their causal interactions, and design strategies to overcome the most influential impediments. Design/methodology/approach: This study adopts an applied, exploratory, and descriptive research design. The methodological framework combines the Fuzzy Delphi Method and the Fuzzy DEMATEL technique in two stages. First, an extensive systematic literature review was conducted, yielding 53 potential Industry 4.0 barriers. Through screening and expert evaluation, a refined list of 15 barriers was selected for assessment. Using Fuzzy Delphi with linguistic variables, input from 14 steel industry experts was collected to validate and localize the barriers. Barriers not meeting the threshold value were eliminated, resulting in a final approved set of 12 barriers. In the second stage, Fuzzy DEMATEL was employed to map the causal relationships among the barriers. Pairwise comparisons provided by experts were transformed into normalized matrices, defuzzified using CFCS, and analyzed to determine influence, dependence, and causal effects. Findings: The findings show that the Iranian steel sector faces a complex, interrelated set of challenges in implementing Industry 4.0. Fuzzy Delphi results indicate that three barriers—resistance to change, intellectual property issues, and sanction-related constraints—were not validated by experts. Fuzzy DEMATEL analysis identifies economic–financial constraints, technological limitations, and managerial barriers as the most influential root causes. High initial investment costs, outdated technological infrastructure, low technological maturity, and weak managerial commitment exert strong causal effects across the entire barrier network. Conversely, organizational misalignment, human resource skill gaps, lack of supply chain collaboration, and insufficient governmental support are dependent barriers heavily influenced by the root causes. The causal diagram suggests that improvements in managerial leadership, strategic planning, investment capacity, and technological readiness would produce positive effects across organizations and supply chains, thereby accelerating Industry 4.0 adoption. Research limitations/implications: This study is limited by its reliance on expert judgment within a single national context, which may constrain generalizability to other countries or industries. Additionally, as Industry 4.0 technologies evolve rapidly, some barriers may change over time or be replaced by new challenges. Future research could extend this work through longitudinal analyses as digital transformation progresses and by applying the model to other industrial sectors for comparative insights. Despite these limitations, the study offers a robust methodological framework and a localized understanding of Industry 4.0 barriers in emerging economies. Practical implications: The study provides actionable insights for steel industry managers, executives, and policymakers by identifying the most critical leverage points for successful digital transformation. The results stress that overcoming economic-financial and technological barriers—through investment in digital infrastructure, upgrading legacy systems, and strengthening R&D capabilities—is essential to enable Industry 4.0 adoption. Managerial implications include the need to reinforce strategic digital planning, develop clear roadmaps, enhance leadership commitment, and cultivate a digital culture within organizations. The findings also recommend improving collaboration across supply chain partners, expanding employee training programs, and designing targeted governmental incentives. Addressing these areas can reduce risks, boost efficiency, and support long-term transformation. Social implication: This research highlights the societal importance of accelerating digital transformation in the steel industry. Industry 4.0 can enhance resource efficiency, reduce environmental impact, improve workplace safety, and raise employees’ quality of life by minimizing hazardous manual tasks. From a broader perspective, successful Industry 4.0 implementation can strengthen national industrial competitiveness, advance sustainable development goals, and stimulate innovation ecosystems. The study’s insights can guide governmental institutions in shaping supportive regulatory frameworks, standardization policies, and investment incentives that facilitate responsible and inclusive digital transformation within the steel industry. Originality/value: This paper offers one of the first comprehensive, contextualized analyses of Industry 4.0 barriers in the Iranian steel industry using a hybrid Fuzzy Delphi–Fuzzy DEMATEL approach. Its originality lies in developing a localized set of barriers validated by domain experts and in mapping complex causal relationships among barriers using fuzzy logic. The study provides substantial value to researchers investigating digital transformation barriers and to practitioners seeking evidence-based strategies to accelerate Industry 4.0 adoption in steel manufacturing.

https://jpom.ui.ac.ir/article_30081_cca71b0d6fdff79779ce772a93d6fc0f.pdf

University of Isfahan Research in Production and Operations Management 2981-0329 17 1 2026 03 21 Investigating the Impact of Wages on Human Productivity through Data Mining Techniques The Case of Yazd Province Electricity Distribution Company Investigating the Impact of Wages on Human Productivity through Data Mining Techniques The Case of Yazd Province Electricity Distribution Company 51 68 30207 10.22108/pom.2026.143121.1587 FA Mahdi Nakhaeinejad Industrial Engineering Department, Faculty of Engineering, Yazd University, Yazd, Iran Parisa Zare Industrial Engineering Department, Faculty of Engineering, Yazd University, Yazd, Iran Journal Article 2024 11 01 Purpose: This study aims to examine the relationship between wages and human labor productivity using data mining techniques. Specifically, it seeks to identify the key activities that influence wage levels and, consequently, employee productivity within electricity meter installation and testing teams. The motivation is rooted in the critical role of fair compensation as an organizational driver for improved efficiency and economic development, particularly in operational settings with diverse job tasks and non-linear performance factors. Design/methodology/approach: The research employs a quantitative, data-driven methodology based on 15 months of operational and wage data collected from 45 installation technicians at the Yazd Province Electricity Distribution Company. Feature selection was performed using statistical filter algorithms, followed by predictive modeling with the C5.0 Decision Tree and Artificial Neural Network (ANN) algorithms. Comparative analysis of model accuracy and precision validated the robustness of the results. The study follows standard data mining procedures: data preprocessing, model building, evaluation, and interpretation. Findings: The analysis revealed a significant positive relationship between wages and productivity. Key determinants of higher wages and productivity included technical inspection, debt and claims collection, switch replacement, and three-phase meter installation. Among 24 measured activity variables, the C5.0 model achieved 94.12% accuracy and 81.8% precision, while the ANN model performed slightly better with 97.05% accuracy and 90.9% precision. These findings confirm the predictive strength of neural networks while underscoring the interpretability advantage of decision tree rules. Activities identified as productivity-critical can inform performance-dependent wage structures and outsourcing strategies. Research limitations/implications: This study is subject to several limitations. It excluded personal factors such as work experience, gender, marital status, and number of children, as well as broader economic variables (e.g., inflation) and psychological dimensions like job satisfaction or organizational commitment. These elements could influence the wage–productivity relationship and warrant inclusion in future research. Methodologically, the study employed a filter-based feature selection approach; subsequent studies should explore alternative preprocessing techniques such as wrapper methods to improve classification accuracy. Although the C5.0 decision tree algorithm was used due to its suitability for the dataset, future research may compare its performance with other algorithms such as CHAID, CART, or hybrid combinations (Bayesian networks with decision trees) by evaluating model metrics like accuracy, AUC, and F-score. Furthermore, other data-mining techniques, especially support vector machines (SVM), could be applied to test the consistency and robustness of findings. Given the one-year data span, longitudinal or multi-year studies are recommended to assess productivity dynamics over time. Expansion to other sectors, such as the water and electricity industries, or comparisons across different provinces and organizational scales would enhance generalizability and practical understanding. Finally, future work may analyze productivity determinants with respect to varying levels of human-capital intensity across industrial contexts. Practical implications: The findings offer actionable insights for human resource managers and industrial operations leaders. By linking employee compensation directly to measured productivity indicators, organizations can implement performance-based wage systems and identify high-impact tasks for potential outsourcing. Improved decision support for labor planning may reduce inefficiencies and enhance electricity service delivery. Social implications: Establishing fair, data-driven wage frameworks contributes to employee motivation, equity, and organizational justice. Enhanced workforce productivity fosters sustainable economic performance, promotes job satisfaction, and strengthens social welfare by ensuring optimal use of human capital within public utility services. Originality/value: This paper is the first to apply the C5.0 decision tree algorithm alongside neural network comparison to analyze wage–productivity relationships in the electricity distribution sector. It introduces methodological diversity and clearly shows how specific technical activities affect compensation. The study contributes to productivity analysis scholarship and managerial practices in labor economics and industrial operations, and examines the feasibility of outsourcing critical activities for managerial decision-making and cost control. Purpose: This study aims to examine the relationship between wages and human labor productivity using data mining techniques. Specifically, it seeks to identify the key activities that influence wage levels and, consequently, employee productivity within electricity meter installation and testing teams. The motivation is rooted in the critical role of fair compensation as an organizational driver for improved efficiency and economic development, particularly in operational settings with diverse job tasks and non-linear performance factors. Design/methodology/approach: The research employs a quantitative, data-driven methodology based on 15 months of operational and wage data collected from 45 installation technicians at the Yazd Province Electricity Distribution Company. Feature selection was performed using statistical filter algorithms, followed by predictive modeling with the C5.0 Decision Tree and Artificial Neural Network (ANN) algorithms. Comparative analysis of model accuracy and precision validated the robustness of the results. The study follows standard data mining procedures: data preprocessing, model building, evaluation, and interpretation. Findings: The analysis revealed a significant positive relationship between wages and productivity. Key determinants of higher wages and productivity included technical inspection, debt and claims collection, switch replacement, and three-phase meter installation. Among 24 measured activity variables, the C5.0 model achieved 94.12% accuracy and 81.8% precision, while the ANN model performed slightly better with 97.05% accuracy and 90.9% precision. These findings confirm the predictive strength of neural networks while underscoring the interpretability advantage of decision tree rules. Activities identified as productivity-critical can inform performance-dependent wage structures and outsourcing strategies. Research limitations/implications: This study is subject to several limitations. It excluded personal factors such as work experience, gender, marital status, and number of children, as well as broader economic variables (e.g., inflation) and psychological dimensions like job satisfaction or organizational commitment. These elements could influence the wage–productivity relationship and warrant inclusion in future research. Methodologically, the study employed a filter-based feature selection approach; subsequent studies should explore alternative preprocessing techniques such as wrapper methods to improve classification accuracy. Although the C5.0 decision tree algorithm was used due to its suitability for the dataset, future research may compare its performance with other algorithms such as CHAID, CART, or hybrid combinations (Bayesian networks with decision trees) by evaluating model metrics like accuracy, AUC, and F-score. Furthermore, other data-mining techniques, especially support vector machines (SVM), could be applied to test the consistency and robustness of findings. Given the one-year data span, longitudinal or multi-year studies are recommended to assess productivity dynamics over time. Expansion to other sectors, such as the water and electricity industries, or comparisons across different provinces and organizational scales would enhance generalizability and practical understanding. Finally, future work may analyze productivity determinants with respect to varying levels of human-capital intensity across industrial contexts. Practical implications: The findings offer actionable insights for human resource managers and industrial operations leaders. By linking employee compensation directly to measured productivity indicators, organizations can implement performance-based wage systems and identify high-impact tasks for potential outsourcing. Improved decision support for labor planning may reduce inefficiencies and enhance electricity service delivery. Social implications: Establishing fair, data-driven wage frameworks contributes to employee motivation, equity, and organizational justice. Enhanced workforce productivity fosters sustainable economic performance, promotes job satisfaction, and strengthens social welfare by ensuring optimal use of human capital within public utility services. Originality/value: This paper is the first to apply the C5.0 decision tree algorithm alongside neural network comparison to analyze wage–productivity relationships in the electricity distribution sector. It introduces methodological diversity and clearly shows how specific technical activities affect compensation. The study contributes to productivity analysis scholarship and managerial practices in labor economics and industrial operations, and examines the feasibility of outsourcing critical activities for managerial decision-making and cost control.

https://jpom.ui.ac.ir/article_30207_c1b9db36b2d31a7a1d248ba488e18607.pdf

University of Isfahan Research in Production and Operations Management 2981-0329 17 1 2026 03 21 Optimizing the Biofuel Supply Chain with an Emphasis on Sustainability and the Mitigation of Inventory Risk Optimizing the Biofuel Supply Chain with an Emphasis on Sustainability and the Mitigation of Inventory Risk 69 92 30230 10.22108/pom.2026.146171.1631 FA Elham Mohseni Department of Management, Faculty of Administrative Science and Economics, University of Isfahan, Isfahan, Iran Dariush Zanjirani Department of Management, Faculty of Administrative Science and Economics, University of Isfahan, Isfahan, Iran Journal Article 2025 08 26 Purpose: The increasing environmental concerns linked to fossil fuel consumption—greenhouse gas emissions, climate change, and resource depletion—have heightened the global demand for sustainable, renewable energy systems. Among alternatives, biofuels have drawn considerable attention for their potential to bolster energy security, reduce environmental pollution, and make effective use of agricultural and forestry residues. However, designing and managing biofuel supply chains is inherently complex due to uncertainties in biomass availability, fluctuating market demand, transportation limitations, and operational risks. To address these challenges, this study develops a comprehensive multi-objective optimization model for the sustainable design and management of a biofuel supply chain under uncertain supply and demand. The proposed framework simultaneously minimizes total operational costs, unmet demand, carbon dioxide (CO₂) emissions, and inventory risk to enhance supply chain sustainability, resilience, and operational efficiency. A central aim of this research is to fill a gap in the literature by treating inventory risk as an independent optimization objective alongside economic and environmental sustainability indicators. Design/methodology/approach: This study proposes a multi-objective mathematical optimization framework for biofuel supply chain network design and management under uncertainty. The model integrates economic, environmental, and operational risk dimensions within a unified decision-making structure. To obtain balanced compromise solutions among conflicting objectives, the LP-metric method is employed as the primary multi-objective optimization technique. Because sustainability objectives are often contradictory, the LP-metric approach enables identification of efficient trade-offs among cost minimization, environmental protection, service-level improvement, and inventory risk reduction. Furthermore, the Fuzzy Analytic Hierarchy Process (FAHP) is used to determine the relative importance and weights of objectives based on expert judgment under uncertainty. Applying FAHP enhances the realism and flexibility of the decision-making process by incorporating ambiguity in human evaluations. In addition, sensitivity analysis and Pareto optimal solutions are examined to validate the robustness and performance of the proposed model under different supply and demand scenarios. The scenario-based analysis investigates the impacts of supply and demand fluctuations on network configuration, facility utilization, operational costs, greenhouse gas emissions, and inventory risk. Findings: The findings demonstrate that the proposed optimization framework effectively manages uncertainties while improving economic efficiency, environmental sustainability, and operational stability within the biofuel supply chain. Results indicate that integrating inventory risk into the optimization process significantly enhances the supply chain's resilience to fluctuations in supply and demand. The proposed model reduces total costs, unmet demand, and CO₂ emissions while maintaining a balanced trade-off among sustainability objectives. Scenario analysis reveals that variations in supply and demand directly affect the structure and operational performance of the supply chain network, particularly the number and utilization rates of warehouses, preprocessing centers, and active biogas facilities. Scenarios with reduced supply and demand generally lead to lower operational costs and reduced greenhouse gas emissions due to decreased production and transportation activities. However, underutilization of facilities in these scenarios may increase inventory-related risks and reduce operational efficiency. Conversely, higher supply and demand scenarios improve capacity utilization and reduce inventory risk, although they may raise transportation and operational costs. Overall, the results confirm that the proposed model provides an effective decision-support framework for identifying sustainable and resilient strategies in biofuel supply chain management under uncertainty. Research limitations/implications: Despite its contributions, this research has several limitations that suggest opportunities for future studies. The proposed model relies on simplifying assumptions that may not fully capture the complexities of real-world biofuel supply chains. In particular, the study primarily addresses economic and environmental dimensions while social and cultural sustainability indicators are considered only indirectly. Future research could incorporate broader sustainability dimensions such as employment generation, social welfare, public acceptance, and regional development impacts. Moreover, the current model treats uncertainty via scenario-based analysis; future studies may benefit from integrating more advanced uncertainty modeling approaches such as robust optimization, stochastic programming, and machine learning techniques. Additional research could also explore the application of metaheuristic algorithms to solve larger, more complex real-world supply chain problems more efficiently. Comparative analyses across different countries and renewable energy systems could further enhance the generalizability and applicability of the proposed framework. Practical implications: The proposed framework offers valuable managerial and strategic implications for policymakers, industrial managers, and energy planners engaged in renewable energy development and sustainable supply chain management. By concurrently considering economic performance, environmental impacts, and inventory risk, the model enables decision-makers to make more balanced, well-informed strategic choices regarding facility locations, storage capacity planning, biomass allocation, and transportation management. The results of this study can support the development of resilient, sustainable biofuel supply chain infrastructures, especially in countries aiming to reduce dependence on fossil fuels and enhance renewable energy utilization. The integration of FAHP and LP-metric methods also provides a practical, flexible decision-making tool capable of addressing uncertainty and conflicting objectives in real-world planning contexts. Additionally, the scenario-based analysis delivers useful insights for designing adaptive strategies under varying market and supply conditions. Social implications: The development of sustainable biofuel supply chains can generate significant social and environmental benefits by reducing greenhouse gas emissions, decreasing reliance on non-renewable fossil fuels, and promoting cleaner energy production systems. The utilization of agricultural and forestry residues as biomass feedstock can also contribute to waste reduction and improved resource efficiency. In addition, the expansion of biofuel industries may create employment opportunities in rural and agricultural regions, thereby supporting local economic development and improving social welfare. By improving supply chain resilience and sustainability, the proposed framework can assist governments and organizations in promoting long-term environmental protection and sustainable energy transitions. The findings of this study may also support national strategies aimed at achieving carbon reduction targets and strengthening energy security. Originality/value: This study contributes to the literature by proposing a comprehensive, integrated multi-objective optimization framework for biofuel supply chain design under supply and demand uncertainty. Unlike many previous studies that focused mainly on cost reduction or environmental performance, this research introduces inventory risk as an independent optimization objective within a sustainable supply chain framework. This novel perspective enhances the realism and applicability of the model for operational and strategic decision-making. Additionally, integrating FAHP for objective weighting with LP-metric optimization to generate compromise solutions represents an innovative hybrid decision-making approach capable of handling uncertainty and conflicting sustainability objectives more effectively. The scenario-based analysis provides a systematic evaluation of the simultaneous impacts of supply and demand fluctuations on economic, environmental, and operational performance. Thus, the proposed model offers both theoretical and practical contributions to advancing sustainable and resilient biofuel supply chain management. Purpose: The increasing environmental concerns linked to fossil fuel consumption—greenhouse gas emissions, climate change, and resource depletion—have heightened the global demand for sustainable, renewable energy systems. Among alternatives, biofuels have drawn considerable attention for their potential to bolster energy security, reduce environmental pollution, and make effective use of agricultural and forestry residues. However, designing and managing biofuel supply chains is inherently complex due to uncertainties in biomass availability, fluctuating market demand, transportation limitations, and operational risks. To address these challenges, this study develops a comprehensive multi-objective optimization model for the sustainable design and management of a biofuel supply chain under uncertain supply and demand. The proposed framework simultaneously minimizes total operational costs, unmet demand, carbon dioxide (CO₂) emissions, and inventory risk to enhance supply chain sustainability, resilience, and operational efficiency. A central aim of this research is to fill a gap in the literature by treating inventory risk as an independent optimization objective alongside economic and environmental sustainability indicators. Design/methodology/approach: This study proposes a multi-objective mathematical optimization framework for biofuel supply chain network design and management under uncertainty. The model integrates economic, environmental, and operational risk dimensions within a unified decision-making structure. To obtain balanced compromise solutions among conflicting objectives, the LP-metric method is employed as the primary multi-objective optimization technique. Because sustainability objectives are often contradictory, the LP-metric approach enables identification of efficient trade-offs among cost minimization, environmental protection, service-level improvement, and inventory risk reduction. Furthermore, the Fuzzy Analytic Hierarchy Process (FAHP) is used to determine the relative importance and weights of objectives based on expert judgment under uncertainty. Applying FAHP enhances the realism and flexibility of the decision-making process by incorporating ambiguity in human evaluations. In addition, sensitivity analysis and Pareto optimal solutions are examined to validate the robustness and performance of the proposed model under different supply and demand scenarios. The scenario-based analysis investigates the impacts of supply and demand fluctuations on network configuration, facility utilization, operational costs, greenhouse gas emissions, and inventory risk. Findings: The findings demonstrate that the proposed optimization framework effectively manages uncertainties while improving economic efficiency, environmental sustainability, and operational stability within the biofuel supply chain. Results indicate that integrating inventory risk into the optimization process significantly enhances the supply chain's resilience to fluctuations in supply and demand. The proposed model reduces total costs, unmet demand, and CO₂ emissions while maintaining a balanced trade-off among sustainability objectives. Scenario analysis reveals that variations in supply and demand directly affect the structure and operational performance of the supply chain network, particularly the number and utilization rates of warehouses, preprocessing centers, and active biogas facilities. Scenarios with reduced supply and demand generally lead to lower operational costs and reduced greenhouse gas emissions due to decreased production and transportation activities. However, underutilization of facilities in these scenarios may increase inventory-related risks and reduce operational efficiency. Conversely, higher supply and demand scenarios improve capacity utilization and reduce inventory risk, although they may raise transportation and operational costs. Overall, the results confirm that the proposed model provides an effective decision-support framework for identifying sustainable and resilient strategies in biofuel supply chain management under uncertainty. Research limitations/implications: Despite its contributions, this research has several limitations that suggest opportunities for future studies. The proposed model relies on simplifying assumptions that may not fully capture the complexities of real-world biofuel supply chains. In particular, the study primarily addresses economic and environmental dimensions while social and cultural sustainability indicators are considered only indirectly. Future research could incorporate broader sustainability dimensions such as employment generation, social welfare, public acceptance, and regional development impacts. Moreover, the current model treats uncertainty via scenario-based analysis; future studies may benefit from integrating more advanced uncertainty modeling approaches such as robust optimization, stochastic programming, and machine learning techniques. Additional research could also explore the application of metaheuristic algorithms to solve larger, more complex real-world supply chain problems more efficiently. Comparative analyses across different countries and renewable energy systems could further enhance the generalizability and applicability of the proposed framework. Practical implications: The proposed framework offers valuable managerial and strategic implications for policymakers, industrial managers, and energy planners engaged in renewable energy development and sustainable supply chain management. By concurrently considering economic performance, environmental impacts, and inventory risk, the model enables decision-makers to make more balanced, well-informed strategic choices regarding facility locations, storage capacity planning, biomass allocation, and transportation management. The results of this study can support the development of resilient, sustainable biofuel supply chain infrastructures, especially in countries aiming to reduce dependence on fossil fuels and enhance renewable energy utilization. The integration of FAHP and LP-metric methods also provides a practical, flexible decision-making tool capable of addressing uncertainty and conflicting objectives in real-world planning contexts. Additionally, the scenario-based analysis delivers useful insights for designing adaptive strategies under varying market and supply conditions. Social implications: The development of sustainable biofuel supply chains can generate significant social and environmental benefits by reducing greenhouse gas emissions, decreasing reliance on non-renewable fossil fuels, and promoting cleaner energy production systems. The utilization of agricultural and forestry residues as biomass feedstock can also contribute to waste reduction and improved resource efficiency. In addition, the expansion of biofuel industries may create employment opportunities in rural and agricultural regions, thereby supporting local economic development and improving social welfare. By improving supply chain resilience and sustainability, the proposed framework can assist governments and organizations in promoting long-term environmental protection and sustainable energy transitions. The findings of this study may also support national strategies aimed at achieving carbon reduction targets and strengthening energy security. Originality/value: This study contributes to the literature by proposing a comprehensive, integrated multi-objective optimization framework for biofuel supply chain design under supply and demand uncertainty. Unlike many previous studies that focused mainly on cost reduction or environmental performance, this research introduces inventory risk as an independent optimization objective within a sustainable supply chain framework. This novel perspective enhances the realism and applicability of the model for operational and strategic decision-making. Additionally, integrating FAHP for objective weighting with LP-metric optimization to generate compromise solutions represents an innovative hybrid decision-making approach capable of handling uncertainty and conflicting sustainability objectives more effectively. The scenario-based analysis provides a systematic evaluation of the simultaneous impacts of supply and demand fluctuations on economic, environmental, and operational performance. Thus, the proposed model offers both theoretical and practical contributions to advancing sustainable and resilient biofuel supply chain management.

https://jpom.ui.ac.ir/article_30230_8d9573770249ab47aa1fb627830a1113.pdf

University of Isfahan Research in Production and Operations Management 2981-0329 17 1 2026 03 21 Simulation-Based Digital Twin for Resilience Analysis of the Protein Supply chain Simulation-Based Digital Twin for Resilience Analysis of the Protein Supply chain 93 124 30263 10.22108/pom.2026.147153.1641 FA Seyed Vahid Daei Niaki Department of Industrial Engineering, Malek Ashtar University of Technology, Tehran, Iran. Abotaleb Shafaghat Department of Industrial Engineering, Malek Ashtar University of Technology, Tehran, Iran. Morteza Abbasi Department of Industrial Engineering, Malek Ashtar University of Technology, Tehran, Iran. Mir Saman Pishvaee School of Industrial Engineering, Iran University of Science and Technology, Tehran, Iran. Journal Article 2025 11 10 Purpose: Resilience has become a strategic imperative in food supply chains, particularly for protein products where perishability, tight delivery windows, and demand volatility intensify the impacts of disruptions. In such environments, even short-lived shocks can cascade through upstream sourcing, cross-docking throughput, and last‑mile distribution, resulting in significant service deterioration and extended recovery periods. This study develops and applies a simulation‑based supply chain Digital Twin to evaluate resilience performance in a protein product distribution network under both normal operations and multiple disruption scenarios. The research addresses four objectives: (i) replicating a real‑world network structure with multiple suppliers, a cross‑docking facility, and multiple customers; (ii) quantifying resilience using operational and recovery‑focused indicators; (iii) comparing the differential effects of demand, supply, cross‑dock capacity, and transportation disruptions; and (iv) identifying structural bottlenecks that hinder full post‑disruption restoration to derive actionable resilience levers. Design/methodology/approach: A scenario‑driven, simulation‑based Digital Twin prototype was implemented in anyLogistix using empirical inputs from the focal firm. Four disruption scenarios were designed to reflect salient risks in protein logistics. Performance was evaluated before, during, and after disruptions using ELT service‑level trajectories and recovery indicators. A sensitivity analysis was also conducted on demand‑shock intensity to assess system stability thresholds. Findings: The experiments reveal substantial heterogeneity in disruption consequences and recovery dynamics. Under baseline conditions, the network maintains high service performance; however, all disruptions cause measurable degradation. The demand‑surge scenario leads to rapid service erosion and backlog accumulation, and the system fails to restore pre‑shock service levels even after external demand normalizes. This “non‑recovery” behavior indicates that resilience is constrained not only by the disruption trigger but also by endogenous capacity limits in processing and distribution. The cross‑dock disruption produces similarly persistent performance losses: reduced throughput generates congestion, disrupts flow continuity, and results in incomplete recovery. These findings identify the cross‑docking node as a dominant structural bottleneck in the studied protein supply chain. By contrast, the supply disruption lowers service levels and lengthens recovery time, yet the network ultimately converges to a stable state once upstream capacity is restored, suggesting conditional resilience when sufficient time is available for replenishment. Research limitations/implications: The study is limited by its focus on a single case network and by the Digital Twin’s current implementation as an offline simulation model rather than a real‑time, synchronized system. The anyLogistix PLE environment also restricts model scalability and limits access to advanced optimization features. Future research should integrate simulation–optimization for automated policy selection, examine compound and correlated disruptions, expand performance metrics to include cost, spoilage, and emissions, and develop real‑time data integration to operationalize the Digital Twin for continuous decision support. Practical implications: The proposed Digital Twin enables managers to stress‑test resilience policies before committing to costly implementation. The observed non‑recovery phenomenon suggests that eliminating a disruption source is insufficient without establishing structural capacity buffers. Recommended interventions include increasing cross‑dock throughput flexibility, strengthening fleet redundancy, adopting multi‑sourcing for critical items, calibrating safety stock for high‑variability demand, and implementing dynamic rerouting and resource reallocation. AI‑enabled demand forecasting can further provide early warnings and trigger proactive responses, reducing emergency logistics costs and service failures. Social implications: Protein supply continuity is closely tied to food security and public welfare. Enhanced resilience reduces the likelihood of shortages and delivery delays during crises or demand surges, while timely distribution also minimizes spoilage and waste, supporting sustainability goals and strengthening public trust. Originality/value: This paper contributes a structured, scenario‑based Digital Twin methodology for resilience assessment in a real protein supply chain, integrating recovery‑focused indicators (ELT service level, recovery time, TTR) and demonstrating how cross‑docking and transportation constraints can induce non‑recovery. The framework offers a replicable research design and practical decision‑support tool for perishable food supply networks. Purpose: Resilience has become a strategic imperative in food supply chains, particularly for protein products where perishability, tight delivery windows, and demand volatility intensify the impacts of disruptions. In such environments, even short-lived shocks can cascade through upstream sourcing, cross-docking throughput, and last‑mile distribution, resulting in significant service deterioration and extended recovery periods. This study develops and applies a simulation‑based supply chain Digital Twin to evaluate resilience performance in a protein product distribution network under both normal operations and multiple disruption scenarios. The research addresses four objectives: (i) replicating a real‑world network structure with multiple suppliers, a cross‑docking facility, and multiple customers; (ii) quantifying resilience using operational and recovery‑focused indicators; (iii) comparing the differential effects of demand, supply, cross‑dock capacity, and transportation disruptions; and (iv) identifying structural bottlenecks that hinder full post‑disruption restoration to derive actionable resilience levers. Design/methodology/approach: A scenario‑driven, simulation‑based Digital Twin prototype was implemented in anyLogistix using empirical inputs from the focal firm. Four disruption scenarios were designed to reflect salient risks in protein logistics. Performance was evaluated before, during, and after disruptions using ELT service‑level trajectories and recovery indicators. A sensitivity analysis was also conducted on demand‑shock intensity to assess system stability thresholds. Findings: The experiments reveal substantial heterogeneity in disruption consequences and recovery dynamics. Under baseline conditions, the network maintains high service performance; however, all disruptions cause measurable degradation. The demand‑surge scenario leads to rapid service erosion and backlog accumulation, and the system fails to restore pre‑shock service levels even after external demand normalizes. This “non‑recovery” behavior indicates that resilience is constrained not only by the disruption trigger but also by endogenous capacity limits in processing and distribution. The cross‑dock disruption produces similarly persistent performance losses: reduced throughput generates congestion, disrupts flow continuity, and results in incomplete recovery. These findings identify the cross‑docking node as a dominant structural bottleneck in the studied protein supply chain. By contrast, the supply disruption lowers service levels and lengthens recovery time, yet the network ultimately converges to a stable state once upstream capacity is restored, suggesting conditional resilience when sufficient time is available for replenishment. Research limitations/implications: The study is limited by its focus on a single case network and by the Digital Twin’s current implementation as an offline simulation model rather than a real‑time, synchronized system. The anyLogistix PLE environment also restricts model scalability and limits access to advanced optimization features. Future research should integrate simulation–optimization for automated policy selection, examine compound and correlated disruptions, expand performance metrics to include cost, spoilage, and emissions, and develop real‑time data integration to operationalize the Digital Twin for continuous decision support. Practical implications: The proposed Digital Twin enables managers to stress‑test resilience policies before committing to costly implementation. The observed non‑recovery phenomenon suggests that eliminating a disruption source is insufficient without establishing structural capacity buffers. Recommended interventions include increasing cross‑dock throughput flexibility, strengthening fleet redundancy, adopting multi‑sourcing for critical items, calibrating safety stock for high‑variability demand, and implementing dynamic rerouting and resource reallocation. AI‑enabled demand forecasting can further provide early warnings and trigger proactive responses, reducing emergency logistics costs and service failures. Social implications: Protein supply continuity is closely tied to food security and public welfare. Enhanced resilience reduces the likelihood of shortages and delivery delays during crises or demand surges, while timely distribution also minimizes spoilage and waste, supporting sustainability goals and strengthening public trust. Originality/value: This paper contributes a structured, scenario‑based Digital Twin methodology for resilience assessment in a real protein supply chain, integrating recovery‑focused indicators (ELT service level, recovery time, TTR) and demonstrating how cross‑docking and transportation constraints can induce non‑recovery. The framework offers a replicable research design and practical decision‑support tool for perishable food supply networks.

https://jpom.ui.ac.ir/article_30263_b5b5bd80563b3a6015e0814f07626f50.pdf

University of Isfahan Research in Production and Operations Management 2981-0329 17 1 2026 03 21 A Framework for Strengthening and Sustaining Supplier Relationships in Engineering-To-Order Manufacturing Systems A Framework for Strengthening and Sustaining Supplier Relationships in Engineering-To-Order Manufacturing Systems 125 150 30279 10.22108/pom.2026.144892.1617 FA Meisam Shahbazi Department of Industrial Management and Technology, Faculty of Management and Accounting, Farabi College, University of Tehran, Tehran, Iran Mahdi Hooshmand Department of Industrial Management and Technology, Faculty of Management and Accounting, Farabi College, University of Tehran, Tehran, Iran Fahimeh Norouzzadeh Ghalhar Department of Industrial Management and Technology, Faculty of Management and Accounting, Farabi College, University of Tehran, Tehran, Iran Journal Article 2025 04 21 Purpose: Engineering-to-order supply chains play a significant role in production and technological development. Many capital goods—such as production machinery, power plant equipment, aircraft, and ships—are produced within these systems. As suppliers of production machinery, they also influence other production systems. Based on observations, one of the main challenges in ETO supply chains is the durability and reliability of supplier relationships. The items required by these systems are extremely diverse while order quantities are often very small. Consequently, there is a persistent risk that supplier relationships will be terminated prematurely due to lack of economies of scale. It is therefore essential to identify and adopt solutions to address this problem. Design/methodology/approach: After reviewing the literature, the required data were collected through interviews with experts from ETO companies, including Mapna Locomotive, Mapna Boiler, Iranian Shipbuilding Industries, and Dorna Aerospace Company. The interview results were organized and analyzed using thematic analysis. In the next phase, using the Best–Worst Method, Interpretive Structural Modeling, and MICMAC analysis, the solutions were ranked and their interrelationships examined. Findings: The results show that "postponement and process redesigning" and "guaranteeing a minimum specific purchase value" are two important strategies, respectively, from the categories of structural solutions and financial dimensions. On the other hand, diversification of the supplier portfolio has not been identified as a priority and is not recommended as an effective solution for strengthening relationships with suppliers in ETO supply chains. Based on the interrelationship analysis, “postponement and process redesigning” is at the highest level of the ISM model, which means it receives the most influence from other strategies. It is also in the dependency area in the MICMAC matrix, meaning it has a high dependency on other solutions. Conversely, "increasing ownership" is a fundamental solution for strengthening relationships with suppliers in the ETO supply chain and is located at the basic level of the ISM model. As shown in the MICMAC matrix, "guaranteeing a minimum purchase value", "increasing ownership", and "training the supplier workforce and knowledge transfer" have a high influence on other solutions and can play an effective role in maintaining the supplier base of the ETO system. Research limitations/implications: One common limitation in this type of research is the difficulty of evaluating the effectiveness of solutions in a data-driven, empirical manner. Doing so requires focusing on a specific project or product within a particular case study and assessing results through a longitudinal study. Therefore, examining the solutions identified in this study in a project-oriented company and practical context—and evaluating the outcomes over a medium-term period—is a valuable topic for future research. Although this research provides a general framework for ETO supply chains, the effectiveness and prioritization of solutions across different industries (including aircraft manufacturing, maritime industries, construction, etc.) merit separate investigation in future work. Practical implications: Postponement and process redesign reduce the degree of internal divergence in the supply chain while helping maintain economic efficiency and providing the expected external variety. In addition, guaranteeing the purchase of a minimum quantity from suppliers who provide items with high risk, critical importance, or limited demand can increase supplier confidence and make them more willing to participate in providing the required items. Social implications: The results of this research can enhance trust and interaction within the ETO supply chain as a socio‑technical system, thereby improving competitiveness and profitability while also increasing employment levels across the supply chain. Purpose: Engineering-to-order supply chains play a significant role in production and technological development. Many capital goods—such as production machinery, power plant equipment, aircraft, and ships—are produced within these systems. As suppliers of production machinery, they also influence other production systems. Based on observations, one of the main challenges in ETO supply chains is the durability and reliability of supplier relationships. The items required by these systems are extremely diverse while order quantities are often very small. Consequently, there is a persistent risk that supplier relationships will be terminated prematurely due to lack of economies of scale. It is therefore essential to identify and adopt solutions to address this problem. Design/methodology/approach: After reviewing the literature, the required data were collected through interviews with experts from ETO companies, including Mapna Locomotive, Mapna Boiler, Iranian Shipbuilding Industries, and Dorna Aerospace Company. The interview results were organized and analyzed using thematic analysis. In the next phase, using the Best–Worst Method, Interpretive Structural Modeling, and MICMAC analysis, the solutions were ranked and their interrelationships examined. Findings: The results show that "postponement and process redesigning" and "guaranteeing a minimum specific purchase value" are two important strategies, respectively, from the categories of structural solutions and financial dimensions. On the other hand, diversification of the supplier portfolio has not been identified as a priority and is not recommended as an effective solution for strengthening relationships with suppliers in ETO supply chains. Based on the interrelationship analysis, “postponement and process redesigning” is at the highest level of the ISM model, which means it receives the most influence from other strategies. It is also in the dependency area in the MICMAC matrix, meaning it has a high dependency on other solutions. Conversely, "increasing ownership" is a fundamental solution for strengthening relationships with suppliers in the ETO supply chain and is located at the basic level of the ISM model. As shown in the MICMAC matrix, "guaranteeing a minimum purchase value", "increasing ownership", and "training the supplier workforce and knowledge transfer" have a high influence on other solutions and can play an effective role in maintaining the supplier base of the ETO system. Research limitations/implications: One common limitation in this type of research is the difficulty of evaluating the effectiveness of solutions in a data-driven, empirical manner. Doing so requires focusing on a specific project or product within a particular case study and assessing results through a longitudinal study. Therefore, examining the solutions identified in this study in a project-oriented company and practical context—and evaluating the outcomes over a medium-term period—is a valuable topic for future research. Although this research provides a general framework for ETO supply chains, the effectiveness and prioritization of solutions across different industries (including aircraft manufacturing, maritime industries, construction, etc.) merit separate investigation in future work. Practical implications: Postponement and process redesign reduce the degree of internal divergence in the supply chain while helping maintain economic efficiency and providing the expected external variety. In addition, guaranteeing the purchase of a minimum quantity from suppliers who provide items with high risk, critical importance, or limited demand can increase supplier confidence and make them more willing to participate in providing the required items. Social implications: The results of this research can enhance trust and interaction within the ETO supply chain as a socio‑technical system, thereby improving competitiveness and profitability while also increasing employment levels across the supply chain.

https://jpom.ui.ac.ir/article_30279_770b0973efc1b7257e2551dbb580b215.pdf