دانشگاه اصفهانپژوهش در مدیریت تولید و عملیات2981-032917220260622Monitoring Simple Multivariate linear Profiles in Phase II using MHWMA Chartsپایش پروفایل های خطی سادۀ چندمتغیره در فاز 2 با نمودارهای MHWMA1243039410.22108/pom.2026.146557.1637FAزهرهقاسمیدانشآموخته دکتری، دانشکده مهندسی صنایع و سیستمها، دانشگاه صنعتی اصفهان، اصفهان، ایراناحمداحمدی یزدیاستادیار گروه مهندسی صنایع، دانشکده فنی و مهندسی، دانشگاه یزد، یزد، ایرانJournal Article20250911<strong>Purpose: </strong>This paper aims to develop and propose three new multivariate control charts based on the MHWMA (Multivariate Homogeneously Weighted Moving Average) approach for monitoring multivariate simple linear profiles in Phase II of statistical process control. The primary motivations for this research are the limitations of existing methods in effectively detecting sudden shifts in profile parameters and the need for more sensitive monitoring schemes that accommodate the multivariate nature of regression relationships between dependent response variables and a single explanatory variable.<br /><strong>Design/methodology/approach: </strong>To achieve the research objectives, three control methods—MHWMA, MHWMA/χ², and MHWMA-2/MMECD—are developed, each employing a distinct strategy for analyzing profile coefficients and covariance structures. The proposed charts use a uniformly weighted moving average to enhance detection capability. Performance is evaluated using the CVRL (Cumulative Variance of the Run Length) criterion via an extensive simulation study. Additionally, a sensitivity analysis assesses the robustness and efficiency of the proposed methods with respect to their design parameters. Finally, two real-world case studies validate the applicability and operational effectiveness of the proposed control charts.<br /><strong>Findings: </strong>The simulation results indicate that the MHWMA/χ² monitoring method generally outperforms the other two proposed methods in detecting deviations in multivariate simple linear profiles across most scenarios. The sensitivity analysis demonstrates the stability and robustness of the proposed charts under various parameter settings. Furthermore, the two case studies confirm the effectiveness and practical applicability of the proposed control charts in real operational conditions.<br /><strong>Research limitations/implications:</strong> This research is limited to Phase II monitoring of multivariate simple linear profiles. The proposed methods assume a known in-control profile model and normally distributed errors. Future research directions include extending the proposed approaches to nonlinear profiles, multivariate profiles with more than one explanatory variable, and Phase I analysis. Additionally, investigating the performance of the proposed charts under non-normal error distributions and auto-correlated data would be valuable.<br /><strong>Practical implications:</strong> The proposed control charts provide quality engineers and process managers with effective tools for real‑time monitoring of complex industrial processes in which product quality is characterized by multivariate linear profile relationships. Implementing these charts can lead to earlier detection of process disturbances, reduced nonconforming output, lower rework and scrap costs, and improved overall process efficiency. The ease of interpretation and the demonstrated superior performance of the MHWMA/χ² method make it a practical choice for industrial applications.<br /><strong>Social implications:</strong> By enabling more effective quality monitoring in complex industrial processes, this research contributes to improved product reliability and safety, directly benefiting consumers and society at large. Enhanced process control can reduce material waste and energy consumption, supporting environmental sustainability goals. Furthermore, the methodologies developed here can inform industry standards and regulatory policies related to statistical quality control practices, ultimately contributing to a higher quality of life through more consistent and reliable manufactured products.<br /><strong>Originality/value: </strong>This paper introduces three novel control charts (MHWMA, MHWMA/χ², and MHWMA‑2/MMECD) for monitoring multivariate simple linear profiles, extending the application of the MHWMA framework to a previously unaddressed area. The proposed MHWMA/χ² method, in particular, demonstrates superior detection performance compared to existing alternatives. The comprehensive evaluation using the CVRL criterion, sensitivity analysis, and real case studies provides strong empirical support for the proposed methods. This research offers significant value to academic researchers in statistical process control as well as industrial practitioners seeking more sensitive tools for profile monitoring.در بسیاری از فرآیندهای صنعتی پیچیده، کیفیت محصول با استفاده از مدل پروفایل خطی سادۀ چندمتغیره توصیف میشود که شامل روابط رگرسیونی میان متغیرهای پاسخ وابسته و یک متغیر توضیحی است. در این پژوهش، سه روش کنترلی جدید مبتنی بر نمودار<strong> </strong>MHWMA<strong> </strong>برای پایش پروفایلهای خطی سادۀ چندمتغیره در فاز دوم کنترل فرآیند پیشنهاد شده است. این نمودارها با بهرهگیری از ساختار میانگین متحرک وزنی یکنواخت، توانایی بالایی در شناسایی تغییرات ناگهانی در پارامترهای مدل پروفایل دارند. سه روش پیشنهادی با عناوین <strong> </strong>MHWMA،<strong> </strong>MHWMA<strong>/χ²</strong>و MHWMA-2/MMECD بهگونهای طراحی شدهاند که هریک با رویکردی خاص، ضرایب مدل و ساختار پراکندگی پروفایلها را تحلیل میکنند<strong>.</strong> بهمنظور ارزیابی عملکرد این روشها، شاخص CVRL در یک مطالعۀ شبیهسازی به کار رفت و نتایج به دست آمده نشان داد روش پایش در بیشتر موارد، نسبتبه دیگر روشها عملکرد بهتری در تشخیص انحرافات دارد. علاوه بر این، بهمنظور ارزیابی پایداری و کارایی روشهای پیشنهادی، تحلیل حساسیت بر پارامترهای روشهای کنترلی انجام شد و نتایج آن، پایداری عملکرد نمودارهای پیشنهادی را نشان داد. درنهایت، برای ارزیابی کاربردپذیری روشهای پیشنهادی در محیط واقعی، دو مطالعۀ موردی بررسی و اثربخشی نمودارهای کنترلی پیشنهادی در شرایط عملیاتی نیز تأیید شد.https://jpom.ui.ac.ir/article_30394_ae2ef3da9ac28e0600c973e2a16e53e3.pdfدانشگاه اصفهانپژوهش در مدیریت تولید و عملیات2981-032917220260622Facility Location for Organ Transplant Networks with Backup Transplant Centersمکانیابی تسهیلات در شبکۀ پیوند عضو با در نظر گرفتن مراکز پشتیبان25513057510.22108/pom.2026.147333.1644FAالهامشریفیدانشکده مهندسی صنایع و سیستمها، دانشگاه صنعتی اصفهان، اصفهان، ایرانعلیشاهندهدانشکده مهندسی صنایع و سیستمها، دانشگاه صنعتی اصفهان، اصفهان، ایرانمهدیایرانپوردانشکده مهندسی صنایع و سیستمها، دانشگاه صنعتی اصفهان، اصفهان، ایرانJournal Article20251202<strong>Purpose: </strong>Organ transplantation is one of the most critical healthcare services for patients with end‑stage organ failure, yet a substantial gap persists between organ supply and demand. Limited transport time and short organ viability often result in long waiting lists and high patient mortality. Designing an efficient and resilient organ transplant network is therefore essential to improve organ utilization and patient outcomes. This study develops a multi‑objective facility location–allocation model for organ transplant networks. The model simultaneously determines optimal locations for organ procurement and transplant centers, assigns hospitals and demand regions, and manages organ flows while accounting for potential disruptions at transplant centers. Unlike previous studies, this work incorporates backup transplant centers to enhance network resilience and ensure continuity of service when primary centers fail due to technical issues, capacity overload, or emergencies.<br /><strong>Design/Methodology/Approach:</strong> A bi‑objective mixed‑integer linear programming model is formulated to design an efficient organ transplant network. The model minimizes total cost and maximizes weighted organ flow while considering travel‑time and ischemia constraints. The network includes hospitals, organ procurement centers, transplant centers, transportation modes, and demand regions over a multi‑period horizon. Each primary transplant center is assigned a backup center to improve reliability during disruptions. Because the problem becomes computationally challenging for medium‑ and large‑scale instances, a multi‑objective simulated annealing (MOSA) algorithm is developed to obtain high‑quality solutions efficiently. The model is evaluated using test instances and compared with the augmented ε‑constraint method, demonstrating superior computational performance while maintaining solution quality.<br /><strong>Findings:</strong> The results indicate that the proposed location–allocation model effectively balances total network cost and organ flow while improving system resilience. The MOSA algorithm generates high‑quality Pareto solutions in significantly shorter computational time, particularly for medium‑ and large‑scale problems. Although the augmented ε‑constraint method performs slightly better for small instances, the proposed heuristic becomes more efficient as problem size increases. Sensitivity analysis shows that higher disruption probability increases total cost and reduces organ flow, underscoring the importance of backup transplant centers. Additionally, organ demand increases cost only until supply becomes the limiting factor. Overall, the findings confirm that incorporating resilience considerations substantially improves organ transplant network design.<br /><strong>Research limitations/implications: </strong>The study assumes deterministic demand, supply, and transportation conditions, along with a predefined disruption probability. However, real organ transplant systems involve uncertainties such as weather conditions, traffic congestion, donor availability, and unexpected operational failures. Future research may extend the model by incorporating more practical factors, including probability of successful transplantation at each center, patient waiting times, and the use of air transportation. Further studies may also examine capacity constraints, strategies to reduce organ wastage, and joint location of transportation agencies with other facilities to better assess their impact on time and cost.<br /><strong>Practical implications:</strong> The proposed decision‑support framework assists healthcare managers in identifying optimal locations for organ procurement and transplant centers and making more effective allocation decisions. By incorporating backup transplant centers, it supports reliable planning and ensures service continuity during disruptions. The model helps reduce transportation time, improve transplant success rates, and minimize organ wastage, leading to more efficient resource use and saving more lives. From a managerial perspective, it reduces operational costs and enhances overall system performance. Moreover, the MOSA algorithm enables efficient solution of large‑scale planning problems, making the framework practical for real‑world decision‑making.<br /><strong>Social implications:</strong> Improving organ transplant network design yields significant social benefits. By reducing organ losses caused by transport delays and increasing reliability during disruptions, the proposed approach can lower patient mortality, shorten waiting lists, and improve access to transplant services. Better utilization of donated organs also strengthens public trust in organ donation systems and supports fairer access to healthcare. Overall, it enhances patients’ quality of life and contributes to a more resilient healthcare system.<br /><strong>Originality/value:</strong> This study offers several original contributions. First, it incorporates backup transplant centers into a strategic facility location–allocation model to improve resilience against disruptions. Second, it simultaneously maximizes weighted organ flow while considering travel‑time and ischemia constraints. Third, it integrates hospitals as the first operational level and models the transfer of brain‑dead donors from hospitals to procurement centers, an aspect rarely addressed in prior research. Finally, the development of a multi‑objective simulated annealing algorithm provides a practical solution approach for large‑scale organ transplant network design problems.پیوند عضو از شاخههای اصلی حوزۀ سلامت و از روشهای درمانی پیچیده در علم پزشکی است که در برخی موارد، ازجمله بیماریهای حاد کبدی، قلبی و ریوی، تنها درمان ممکن است. بااینحال، تعادلنداشتن بین عرضه و تقاضا، به لیستهای انتظار طولانی و مرگومیر بیماران پیش از دریافت عضو منجر میشود. این مشکل، توجه به بهینهسازی فرآیند اهدای عضو و استفادۀ کارآمد از آن را ضروری میکند. با توجه به اینکه تعداد مراکز پیوند در سطح یک کشور محدود است و کل تقاضا باید<strong> </strong>از سوی این مراکز پوشش داده شود، در صورت از دسترس خارج شدن هرکدام از مراکز در دورههای زمانی، خسارات جبرانناپذیری به سیستم پیوند عضو وارد و موجب از دست رفتن اعضای اهدایی و پایمالشدن حق بیماران داوطلب میشود. در این پژوهش، با لحاظ احتمال از دسترس خارج شدن مراکز پیوند، علاوه بر تعیین تعداد و مکان بهینۀ تسهیلات شبکۀ پیوند، برای هرکدام از مراکز پیوندی که استقرار مییابند، یکی از مراکز پیوند پشتیبان در نظر گرفته میشود تا در صورت لزوم، تقاضای پیوند در مرکز پیوند پشتیبان انجام شود و خسارات به حداقل برسد. در مدل برنامهریزی چندهدفۀ ارائهشده، علاوه بر حداقلسازی هزینههای استقرار و جریان بین مراکز، حداکثرسازی جریان عبوری بین مراکز نیز در نظر گرفته میشود تا کیفیت عمل پیوند انجامشده افزایش و احتمال از دست رفتن عضو، کاهش یابد. بهمنظور حل مسائل در ابعاد بزرگ، یک الگوریتم ابتکاری مبتنی بر شبیهسازی تبرید چندهدفه ارائهشده است. نتایج بر دادههای وزارت بهداشت نشان میدهد الگوریتم پیشنهادی در مقایسه با روش محدودیت اپسیلون تعمیمیافته، در حل مسائل متوسط و بزرگ، عملکرد بهتری دارد.https://jpom.ui.ac.ir/article_30575_72a6a1ede69eb4db85a75dcc2ba11117.pdfدانشگاه اصفهانپژوهش در مدیریت تولید و عملیات2981-032917220260622Designing a Deployment Model for Artificial Intelligence in Supply and Production Processes: a Case Study of a Telecommunications Equipment Manufacturerطراحی مدل استقرار هوش مصنوعی در فرآیندهای تأمین و تولید: مطالعۀ موردی یک شرکت تولیدکنندۀ محصولات مخابراتی53723064410.22108/pom.2026.149216.1666FAامیرسرآبادانیگروه مدیریت و مهندسی صنایع، دانشگاه صنعتی مالک اشتر، تهران، ایرانحمیدرضاطهوریگروه مدیریت و مهندسی صنایع، دانشگاه صنعتی مالک اشتر، تهران، ایرانمحمدرضازاهدیگروه مدیریت و مهندسی صنایع، دانشگاه صنعتی مالک اشتر، تهران، ایرانافشینعلی پورگروه مدیریت و مهندسی صنایع، دانشگاه صنعتی مالک اشتر، تهران، ایرانJournal Article20260609<strong>Purpose:</strong> Telecommunications equipment manufacturing is among the most complex production sectors, marked by rapid technological change, shortened product cycles, global supply chains and rising customer expectations. Even minor disruptions can quickly translate into lost market share. While artificial intelligence offers considerable promise for improving demand forecasting, inventory management, supplier evaluation, maintenance and quality inspection, many organizations struggle with successful implementation. The literature review reveals that primary obstacles are predominantly organizational and cultural rather than technical. Prior research falls into three categories: macro‑level reviews lacking operational depth, quantitative studies testing predetermined hypotheses without contextual consideration and descriptive barrier inventories failing to capture dynamic interactions. These gaps are especially pronounced in Iran’s telecommunications industry, where firms additionally contend with sanctions, cultural norms and resource constraints. This study aims to design and empirically validate a context‑sensitive model for AI deployment tailored to supply and production processes in an Iranian telecommunications equipment manufacturer.<br /><strong>Design/methodology/approach:</strong> This research adopted a sequential exploratory mixed‑methods design. The qualitative phase employed systematic grounded theory (Strauss and Corbin) with its paradigm model. Data were collected from 26 in‑depth semi‑structured interviews with executives, managers and IT specialists. Purposive sampling with snowball techniques was used until theoretical saturation. Analysis proceeded through open, axial and selective coding. The quantitative phase used variance‑based structural equation modeling (PLS‑SEM) through SmartPLS. A 17‑item Likert‑scale questionnaire, derived from qualitative findings, was completed by 100 managers and experts, exceeding the minimum 40 cases suggested by the ten‑times rule. The measurement model was evaluated for reliability, convergent validity and discriminant validity. Structural model testing used bootstrapping with 5,000 resamples.<br /><strong>Findings:</strong> Qualitative analysis yielded categories mapped onto the paradigm model. Causal conditions captured internal and external pressures: rising costs, compressing margins, competition, demand fluctuations and the need for faster decisions. The core phenomenon emerged as behavioral intention to deploy AI, representing commitment through management advocacy, strategies, budgets and project teams. Contextual conditions encompassed culture, structure, IT infrastructure, leadership and executive backing. Intervening conditions were constraining: talent shortages, budget limitations, sanctions, employee resistance, fragmented AI understanding and cybersecurity concerns. Strategies included pilots, staff training, recruitment, outsourcing and partnerships. Consequences covered lower costs, improved quality, reduced downtime, agility and resilience. Quantitative analysis confirmed reliability and validity. Path analysis revealed that causal conditions significantly shape strategies (β=0.485, t=6.762), and intervening conditions strongly influence strategies (β=0.470, t=7.327). Strategies predicted consequences (β=0.448, t=7.069), which exerted the strongest effect on behavioral intention (β=0.746, t=17.090). Intervening conditions directly influenced behavioral intention (β=0.160, t=2.728). Contextual conditions showed no significant direct path to strategies (β=-0.198, t=1.749), suggesting they function primarily as foundational enablers and prerequisites rather than direct drivers of strategic choices. The moderating effect of intervening conditions on the causal‑conditions‑to‑strategies relationship was significant (β=-0.165, t=2.304), indicating that pronounced obstacles weaken the influence of causal pressures on strategy selection.<br /><strong>Research limitations/implications:</strong> The single case study limits generalizability. Future comparative studies across multiple sectors such as automotive, pharmaceuticals or petrochemicals could test broader applicability. Cross‑sectional data restrict observation over time; longitudinal designs would be valuable for tracking dynamic feedback loops. Environmental moderators including government policies, sanctions intensity and national digital maturity were not separately examined. Despite limitations, the mixed‑methods design and empirical support provide a solid platform for theoretical refinement and practical application.<br /><strong>Practical implications:</strong> Managers should adopt a pilot‑oriented approach based on the robust path from strategies to consequences and the stronger link from consequences to behavioral intention. Instead of large, expensive projects, managers might select well‑defined operational areas with clear AI potential, design pilots with explicit goals and metrics, build teams combining technical and operational expertise and document results transparently. Successful pilots build trust and reinforce intention to continue. Managers must actively address intervening conditions through upskilling investments, dedicated budgets, compelling business cases and structured change management. Framing AI as augmenting rather than replacing human capability reduces resistance. Creating systematic feedback loops that measure and communicate tangible outcomes—cost reductions, quality improvements—generates positive reinforcement cycles sustaining long‑term transformation.<br /><strong>Social implications:</strong> AI contributes to sustainable production through reduced waste and optimized resource use. Enhanced quality control and predictive maintenance mean fewer defective products and less downtime. The model emphasizes upskilling over automation, alleviating job‑displacement fears. In Iran’s context, this locally grounded approach builds internal capabilities, strengthening national self‑sufficiency. Transparent communication fosters public trust. Benefits include preserving skilled jobs, improving product quality and supporting responsible environmental practices, aligning with corporate social responsibility and informing policy.<br /><strong>Originality/value</strong>: This research offers several novel contributions. First, it employs a sequential exploratory mixed‑methods design combining qualitative depth with quantitative validation, uncommon in developing‑country manufacturing. Second, it produces a dynamic paradigm model capturing causal interplay, revealing two specific findings: the direct effect of intervening conditions on behavioral intention and the moderating role of intervening conditions in the causal‑conditions‑to‑strategies relationship. Third, the focus on telecommunications equipment manufacturing fills a gap where no prior localized AI deployment model exists. Fourth, the study undertakes thorough localization for Iran’s business environment, explicitly incorporating sanctions, limited technology access, organizational culture and scarce AI talent. The result is an empirically grounded theory offering researchers a framework for testing and practitioners a roadmap for navigating AI adoption complexities.هدف این پژوهش، طراحی مدلی بافتمحور برای استقرار هوش مصنوعی در فرآیندهای تأمین و تولید یک شرکت محصولات مخابراتی در ایران است. مطالعۀ حاضر با رویکرد ترکیبی متوالی اکتشافی، ابتدا با روش نظریۀ دادهبنیاد نظاممند (اشتراوس و کوربین<sup><strong>[i]</strong></sup>)، مسئله را بهطور عمیق تحلیل میکند. دادههای کیفی از مصاحبۀ نیمهساختاریافتۀ عمیق با ۲۶ نفر از مدیران و کارشناسان خبره گردآوری شد. مقولههای اصلی در قالب مدل پارادایمی شامل شرایط علّی، زمینهای، مداخلهگر، راهبردها، پیامدها و قصد رفتاری برای استقرار هوش مصنوعی (پدیدهمحوری)، سازماندهی و روابط بین مؤلفهها شناسایی و مدل اولیه ترسیم شد؛ سپس با روش مدلسازی معادلات ساختاری (حداقل مربعات جزئی)، روابط مدل آزمون شد؛ به این صورت که پرسشنامهای ۱۷ سؤالی با طیف لیکرت از سوی ۱۰۰ نفر تکمیل و با SmartPLS تحلیل شد. یافتهها پایایی و روایی مدل تأیید شد و نشان داد شرایط علّی و مداخلهگر بر راهبردها و راهبردها بر پیامدها و پیامدها بر قصد رفتاری برای استقرار هوش مصنوعی (پدیدهمحوری) تأثیر معنادار دارند. اثر تعدیلگری شرایط مداخلهگر نیز معنادار است. این پژوهش مدل بومی و راهکارهای عملیاتی مشخصی را ارائه میکند.<br /> <br />[i] Strauss & Corbinhttps://jpom.ui.ac.ir/article_30644_187ad4c9e1aea52923fe4eb54bb25e19.pdf