LEE- Endüstri Mühendisliği-Yüksek Lisans
Bu koleksiyon için kalıcı URI
Gözat
Son Başvurular
1 - 5 / 32
-
ÖgeExploring contact patterns among students via social network analysis: A cohort study in İstanbul, Türkiye(Graduate School, 2024-07-16)This study explores the dynamics of social interactions among middle school students using Social Network Analysis (SNA) and ultra-wide band (UWB) technology embedded in sensor cards. Conducted in a Turkish middle school, this research aimed to capture and analyze the social mixing patterns of students across three grades on separate days in March 2023. The results have significant implications for understanding disease transmission and developing effective public health interventions in school settings. SNA, which integrates sociological theories and mathematical principles from graph theory is employed in this context to understand how interactions between individuals influence social phenomena, such as cognitive processes, feelings, and behaviors. It also seeks to uncover the patterns of information, influence, and infection flow. In our study, the goal is to understand how infection spreads among students of a middle-school given the significant role children play in transmission of airborne diseases when they commute between home and school and interact with a number of peers. The data collection utilized lightweight, credit-card-sized wireless sensor devices to record close proximity interactions among students and giving us details of those contacts including who interacts with whom and for how long. These UWB-based devices are known for their high accuracy in contact tracing and were configured to detect face-to-face contacts within a 1.5-meter range. The study was set in a co-educational public middle school in Istanbul. Each student, along with their teachers and staff, was equipped with a sensor card that recorded face-to-face contacts throughout the school day. The data was collected over three separate days, each dedicated to a specific grade: 5th, 6th, and 7th grades. On each day, students were given sensor cards at the beginning of the school day, and these were collected before the last class ended. The collected data captured interactions during class times and break times, providing a comprehensive picture of the students' social networks. The context of the study is particularly significant given the recent global COVID-19 pandemic in addition to previous outbreaks that claimed a lot of lives and left detrimental impacts on public health systems and economies worldwide, thus highlighting the critical need for effective contact tracing and understanding social interactions to predict how pathogens spread and eventually prevent disease spread. Schools, being high-density environments where close contact is frequent, are prime locations for such studies. By focusing on these settings, the research aims to provide insights that can inform public health policies and interventions, such as school closings, social distancing, masking, and vaccination strategies. The contact data was analyzed to construct a network of interactions among students. Various metrics such as degree (number of contacts per student), density (ratio of actual contacts to possible contacts), clustering coefficient (degree of interconnectedness among a student's contacts), and shortest path (minimum number of intermediary nodes connecting two individuals) were calculated to understand the cohesiveness and connectedness of the network. Other data analysis methods namely calculating the distribution of contact durations and degree values were also applied to gain insight on the likelihood of transmission events, and informing disease models with the related parameters. The results obtained from this study are listed as follows: • It was found that older students tended to form more interconnected groups with stronger ties across different classroom communities. • It was determined that contact durations were short, with most interactions lasting less than a minute. • The patterns suggest that while younger students have more frequent contacts, these interactions are generally brief. The study also identified the potential for different contact durations to influence infection spread, emphasizing the need for further research on proximity distances and their effects on network dynamics. By focusing on environments with high population density and connectivity, such as schools, valuable insights can be gained about the potential effects and dynamics of virus transmission. Lastly, this thesis provides valuable insights into the social interactions of middle school students, with implications for designing effective public health interventions and improving our understanding of disease transmission in school settings. The results underscore the need for targeted strategies to manage infectious diseases, particularly in educational environments where close contact is frequent. The findings also highlight the importance of using advanced technologies like UWB sensors to gather accurate data on social interactions, which can inform more precise and effective public health measures.
-
ÖgeDöngüsel malzeme sistemi tasarımı için çok amaçlı model önerisi ve bir beyaz eşya fabrikası için uygulama(Lisansüstü Eğitim Enstitüsü, 2023)elişmekte olan dünya ile, işletmeler arası rekabetlerden ötürü, maliyet ve teknolojik üstünlük kapsamında önde olma çabası çok üst seviyelere ulaşmıştır. Bu üstünlük göstergelerini iyileştirmenin birçok potansiyel alanı olmak ile beraber lojistik faaliyetleri bunların en önemlilerinden biridir. Bu kapsamda, birçok işletme hem iç hem de dış lojistik operasyonlarında verimlilik arttırıcı yöntemlere başvurmaktadır. Bu çalışmada, bir beyaz eşya fabrikasının iç lojistik operasyonuna ait araç rotalaması üzerine bir gerçek hayat problemi ele alınmıştır. Çalışmada adı geçen fabrikada bulunan montaj hatlarının ihtiyaçlarını karşılamak için süpermarketten montaj istasyonlarına belirli periyotlarda taşıyıcılar ile malzeme dağıtımı gerçekleştirilmektedir. Bu taşıyıcılar insansız malzeme dağıtımı yapan, elektrikli ve otonom bir şekilde yönlendirilen AGV (Automated Guided Vehicles)'lerdir. Dağıtılacak malzemeler süpermarket adındaki fabrika içinde bulunan ara malzeme deposuna getirildikten sonra, burada elleçlenip, AGV'lere yüklenir. Ardından AGV'ler atanmış olan istasyonlara uğrar ve rotasını tamamlayan AGV'ler süpermarkete geri dönmektedirler. Problemde her bir AGV'nin belirli bir taşıma kapasitesi olduğu var sayılmaktadır. Çalışmada, montaj istasyonlarının ihtiyacını karşılamaya yönelik malzemelerin süpermarketten istasyonlara teslimini içeren AGV rotalama problemi üzerinde durulmuştur. Bununla birlikte her bir AGV'nin tek bir rotada ve periyotta hizmet ettiği kabul edilmiştir. Problem, literatürde NP-Zor olarak adlandırılan araç rotalama problemlerinin bir türü olan kapasite kısıtlı araç rotalama problemi olarak değerlendirilmiştir. Bu çalışma Türkiye'de faaliyet gösteren bir beyaz eşya firmasının yeni kurulacak olan fabrikası için uygulanacaktır. Kullanılan veriler yeni kurulacak olan fabrikadaki gerçek verilerdir. Bu kapsamda istasyonların ihtiyacına göre malzeme dağıtımının yapılacağı, AGV'lerin izleyecekleri rotaların ve periyotların belirlendiği, bunun yanında da ana amaçlar olarak AGV'lerin vardiyada aldıkları tur sayılarının ve istasyonların toplam bant başı stoklarının minimize edileceği çok amaçlı karma tamsayılı bir lineer matematiksel model geliştirilmiştir. Bu çok amaçlı matematiksel modelin çözümü için ise AUGMECON2 (Arttırılmış 𝜀 𝑘𝚤𝑠𝚤𝑡 yöntemi) uygulanmıştır. Metodun sonucunda problemdeki pareto optimal çözümler elde edilmiş olup, çözümlerin sonuçlarına göre yönetsel kararlar uygulanmıştır. Bu kapsamda, firma stratejileri doğrultusunda ideal bant başı stok değerleri ve vardiyadaki tur sayılarını belirlemiştir. Bunun yanında farklı amaç ve değişen parametreler ile matematiksel model değiştirilmiş ve model her bir durum için tekrar çözdürülmüştür. Modeldeki bu değişkenlerin sonuca etkisi incelenmek istenmiştir. Bunun sonucunda ise daha esnek ve verimli, uygun maliyetli üretim operasyonuna ulaşılmak hedeflenmiştir. Öncelikli olarak model 4 farklı amaç ile tekrardan çözdürülmüş olup, yeni rotalar, periyot ve rota süreleri, AGV'lerin doluluk oranları ve batarya durumları ve de bant başı stok alanlarının değişimi incelenmiştir. Bu 4 amaç sırasıyla, istasyonlardaki maksimum bant başı stokunun minimize edilmesi, AGV'lerin her periyotta aldığı rota sürenin minimize edilmesi, AGV sayısının minimize edilmesi ve AGV'lerin boşta kalma sürelerinin minimize edilmesidir. Akabinde ise modeldeki parametre değişikliklerinin modelin sonucuna etkisi incelenmek istenmiştir. Burada talep değişkenliği ve periyot sürelerindeki değişiminin, AGV kapasitesinin ve AGV sayısındaki değişiminin çıktılara etkisi incelenmiş olup, yönetsel karar destek mekanizmaları geliştirilmiştir.
-
ÖgeBaskı çoğaltma endüstrisine yönelik otonom tekliflendirme ve cihaz yönetimi stratejilerinin değerlendirilmesi: Bir karar destek sisteminin tasarımı(Lisansüstü Eğitim Enstitüsü, 2024-06-05)Tez çalışması dijital baskı endüstrisindeki stratejik karar verme süreçlerini iyileştirme, tekliflendirme sürecini otomatize etme ve sektördeki operasyonel verimlilikle birlikte pazar rekabetçiliğini artırma hedeflerine odaklanmaktadır. Çalışma, baskı endüstrisinin tarihçesine ve dijital teknolojilerin bu alana etkilerine odaklanarak geniş bir literatür taraması ile başlar. Endüstriyel baskı süreçlerinin gelişimi, dijitalleşmesi ve otomasyonun artmasıyla birlikte, firmaların karşılaştığı zorluklar ve bu zorlukların üstesinden gelmek için yapay zeka ve makine öğrenimi teknolojilerinin potansiyeli detaylandırılmaktadır. Çalışma, otonom tekliflendirme sistemlerinin geliştirilmesi ve cihaz yönetimi stratejilerinin değerlendirilmesine yönelik çözümler sunmayı amaçlamaktadır. İlk olarak metodoloji bölümünde veri toplama, veri analiz süreçleri, kullanılan makine öğrenimi modelleri, model değerlendirme ve validasyon yöntemleri, ve tekliflendirme süreci açıklanmaktadır. Çalışmada, gerçek zamanlı piyasa ve müşteri verilerini işleyebilen, dinamik tekliflendirme yapabilen ve müşteri ihtiyaçlarına göre özelleştirilmiş çözümler sunan bir karar destek sistemi geliştirilmiştir. Bu sistem, otonom olarak en uygun cihaz seçimini yapabilme ve rekabetçi fiyatlandırma stratejileri oluşturabilme kapasitesine sahiptir. Uygulama kısmında, firmadan alınan gerçek müşteri verilerinin değerlendirilmesi ve girdilerin belirlenmesi, kullanılacak makine öğrenmesi modelinin detaylandırılması ve değerlendirilmesi, geliştirilen modelin baskı endüstrisi özelinde kullanımı, kullanıcı arayüzü tasarımı, pilot uygulama ve elde edilen sonuçlar değerlendirilmektedir. Kopitek Kopyalama Sistemleri Ltd. Şti. tarafından sağlanan 8600 adet müşteri verisi, CHAID veri analizi yöntemi ile incelenmiştir. CHAID analizi, etkileşimli ağaçlar oluşturarak hangi faktörlerin müşteri tercihlerini en çok etkilediğini belirlemek için kullanılmıştır. CHAID analizinden elde edilen değişkenler, Rastgele Orman algoritması başta olmak üzere 5 farklı makine öğrenimi modeli ile test edilmiştir. Rastgele Orman, Karar Ağaçları Lojistik Regresyon, Destek Vektör Makineleri (SVM), ve K-En Yakın Komşular (KNN) algoritmaları da karşılaştırmalı olarak değerlendirilmiştir. Her bir modelin performansı, doğruluk oranı ile değerlendirilmiştir. Rastgele Orman, özellikle büyük veri setleri ve yüksek boyuttaki verilerde üstün performans gösterdiği için bu uygulama için tercih edilen makine öğrenmesi algoritması olmuştur. Rastgele Orman algoritması ile oluşturulan modelin etkinliği ve güvenilirliği, çapraz doğrulama ve karışıklık matrisleri aracılığıyla ayrıca değerlendirilmiştir. Daha sonra ise modelin sektör özelinde kullanımı adına bir kullanıcı arayüzü geliştirilmiş ve oluşturulan karar destek sisteminin gerçek zamanlı veriler ile uygulaması gerçekleştirilmiştir. Kullanıcı arayüzü, müşteri veri girişlerini kolaylaştırmak ve sistem çıktılarını anlaşılır bir biçimde sunmak üzere tasarlanmıştır. Bu arayüz, kullanıcılara dinamik tekliflendirme yapma, cihaz seçimleri sunma ve rekabetçi fiyatlandırma stratejileri oluşturma imkânı vermiştir. Sonuçlar bölümünde, tezin ana bulguları özetlenmekte ve teorik ile pratik katkılar tartışılmaktadır. Tez çalışması, baskı endüstrisindeki karar verme süreçlerini iyileştirmek için makine öğreniminin etkin bir şekilde nasıl kullanılabileceğini göstermektedir. Ayrıca, bu teknolojilerin endüstriyel uygulamalarda nasıl stratejik avantajlar sağlayabileceği üzerinde durulmaktadır. Tez aynı zamanda gelecekteki araştırmalar için önerilerde bulunarak, baskı endüstrisindeki dijitalleşme sürecinin gelişimine katkıda bulunmayı hedeflemektedir. Bu araştırma, baskı endüstrisinde faaliyet gösteren firmalara, karar verme süreçlerinde veriye dayalı, dinamik ve etkin çözümler sunma potansiyeli taşımaktadır. Geliştirilen karar destek sistemi, endüstriyel uygulamalarda yenilikçi bir yaklaşım olarak değerlendirilebilir ve baskı sektöründe dijitalleşme trendlerine uyum sağlamak, operasyonel verimliliği artırmak ve pazar rekabetçiliğini güçlendirmek için stratejik bir araç olarak ön plana çıkmaktadır. Özellikle dijital baskı teknolojilerinin sürekli gelişim gösterdiği bir dönemde, bu tür yenilikçi çözümler sektöre büyük değer katabilir ve baskı hizmetlerinin daha etkin ve verimli bir şekilde sunulmasını sağlayacaktır.
-
ÖgeTrain set complexity tunning for imbalance learning(Graduate School, 2024-05-17)Machine learning algorithms that address most classification problems are known to yield good results by assuming a balanced training set. This is because machine learning algorithms, during the training process, attempt to minimize a specific cost objective function. When undertaking this minimization process, if one class is predominant, it will tend to dominate the minimization process, essentially producing predictions biased toward the dominant class. In other words, generating predictions predominantly for the prevalent class will fulfill the minimization objective. This presents a problematic situation, as many real-world classification problems exhibit imbalanced datasets. Real-life scenarios such as fraud detection, churn prediction, and disease diagnosis often involve imbalanced datasets. To address this issue, the desired outcome in the development of classification solutions is the accurate identification of examples belonging to the minority class. In the context of these problems, the detection of instances from the minority class is typically more meaningful. Hence, the resolution of the imbalance problem in classification algorithms holds significant importance. Imbalance in real-world classification problems poses a challenge, as it hinders the proper recognition of instances from the minority class. This is particularly noteworthy in scenarios like fraud detection, churn prediction, and disease diagnosis, where the minority class represents critical instances requiring accurate identification. Therefore, addressing the imbalance problem in classification algorithms becomes crucial for the successful application of machine learning in these real-life scenarios. Various methods have been developed to address the imbalance learning problem, and these methods can manifest as manipulations at both the model level and the level of the model's learning set. Manipulating the learning set aims to create a relatively balanced dataset during the learning process. One approach involves removing some examples belonging to the majority class from the learning set to achieve this balance. However, this method may lead to information loss related to the majority class. In addition to discarding examples from the majority class, another strategy involves augmenting examples from the minority class. During this process, sampling can be performed for the minority class. A fundamental challenge in sampling from the minority class lies in the difficulty of ensuring diversity in the information represented by minority class examples, leading to the problem of incorporating repetitive information into the dataset. To address this issue, solutions have been proposed that focus on generating synthetic data for the minority class. Unlike the previously introduced solutions for the imbalance learning problem, we propose a novel approach: training set complexity ratio tuning. This proposed method diverges from traditional techniques by emphasizing the adjustment of the complexity ratio within the training set. Instead of introducing synthetic data, our approach centers around iteratively tuning the complexity of the training set. This involves a careful balance between reducing examples from the majority class and augmenting examples from the minority class to achieve a favorable training set complexity ratio. By doing so, we aim to mitigate information loss in the majority class while enhancing the diversity and quantity of minority class examples, providing a nuanced perspective in addressing the challenges posed by the imbalance learning problem. In proposed method, we will develop a solution to make classification algorithms stronger by tuning the complexity of the train set for imbalanced data sets. The complexity of the training set refers to the indicator of how many times the number of examples in the majority classs exceds the those in the majority class. We consider the complexity of the training set as a hyperparameter, crutial for effectively addressing imbalance learning problem. To achieve this, tunning the hyperparameter of the training set's complexity becomes imperative. During this process, an equal number of examples from the minority and the majority class are initially selected. Subsequently, an iterative approach is employed to incrementally increase the number of examples frım the majority class. At each iteration base model is trained on the resulting training set, and its performance on the validation set is measured and recorded. Comparing the performance metrics obtained after each iteration allows us to determined the optimal complexity of the training set, which corresponds to the iteration with highest performance. While adjusting the balance of majority class and minority class in the training set, sample selection will be made by considering the number of sample of the majority class. We aimed to enhance the representativeness of the selected samples from the majority class by employing the K-Means algorithms, instead of opting for random selection. To achieve this, we partitioned the majorit class into clusters using the K-Means algorithm. During the sample selection process, we considered the propotion of instances within each cluster, ensuring that the selected samples maintained the same ratio as the original data. This approach was designed to improve the overall representativeness of the majority class, ultimately contributing to a more robust and reliable sampling strategy. In our proposed method, in addition to tuning the complexity ratio of the training set, we compared methods for determining the ideal number of clusters for the KMeans clustering applied to the majority class. Initially, we employed the commonly used elbow curve method to identify the number of clusters for the majority class. As a second approach, for each cluster count, the train set complexity method was executed to find the optimal training set complexity ratio, and the performance of the resulting model on the validation set was recorded. This method was iteratively applied for each cluster count. The obtained results were then compared, and the number of clusters for the majority class and the corresponding training set complexity ratio were determined. Using the accuracy metric would be misleading when measuring the performance of the model built on the imbalance data set (Chowdhury et al., 2023). In the comparative analysis between our proposed model and the conventional SMOTE method, the ROC performance metric was chosen as the evaluation criterion. This metric, renowned for quantifying the discriminative prowess of a classification algorithm, proved instrumental in assessing the algorithm's capacity to effectively differentiate between distinct classes. Upon scrutinizing the results of the comparison, it is evident that the methodology we proposed has yielded more favorable outcomes when juxtaposed against the conventional SMOTE approach. The ROC performance metric, being a comprehensive indicator of the algorithm's ability to discriminate between positive and negative instances, underscores the enhanced discriminative power exhibited by our proposed model. This comparative evaluation provides empirical evidence supporting the superior efficacy of our proposed methodology over the traditional SMOTE method, underscoring its potential to enhance the overall performance and robustness of classification algorithms, particularly in the context of imbalanced datasets. In addition, to comprehensively assess the validation performance of the proposed method, we conducted a detailed examination of the confusion matrix. This analysis revealed a significant increase in the True Positive (TP) rate, which indicates that our model exhibits a superior ability to accurately predict positive classifications. This enhancement in the True Positive rate is particularly noteworthy, as it underscores the model's proficiency in identifying positive instances, thereby reducing false negatives and improving overall classification accuracy. These findings are critical as they demonstrate that the proposed method not only enhances overall model performance but also significantly improves its efficacy in specific classification tasks. The ability to accurately classify positive instances is often crucial in many real-world applications, such as medical diagnosis, fraud detection, and various safety-critical systems. Moreover, the results highlight the robustness and reliability of our model when applied to imbalanced datasets, a common challenge in many practical scenarios. Traditional methods often struggle with such datasets, leading to suboptimal performance. However, the proposed method stands out by offering a more effective and reliable solution. In summary, our newly proposed model presents a substantial advancement over other widely used techniques, providing a more effective and dependable approach to handling imbalanced data. This improvement not only demonstrates the model's potential for broader applicability but also its robustness in addressing complex and demanding classification problems. The implications of these findings are significant, suggesting that the proposed method could be widely adopted across various domains requiring precise and reliable classification performance.
-
ÖgeReducing pre-production lead time and cost through value stream mapping and the cut-to-box system: A footwear industry study(Graduate School, 2024-07-01)This master's thesis explores how Value Stream Mapping (VSM) and the innovative Cut-to-Box system can enhance efficiency and reduce costs in the pre-production stages of footwear manufacturing. Conducted at a Footwear Company in Istanbul, Turkey, the study examines pre-production challenges and optimizes processes using VSM. By identifying and eliminating waste, VSM helps reduce the overall cost of pre-production, aligning with lean accounting principles. Lean accounting focuses on supporting lean manufacturing by providing accurate, timely, and understandable financial information. This helps companies to make better decisions through continuous process improvement. To address the complexities of pre-production operations, the study implements a comprehensive value stream mapping system, optimizing existing processes. Simultaneously, the Cut-to-Box system, based on cellular manufacturing, introduces innovation to warehouse layout, effectively streamlining operations. Key findings include a reduction in lead time from 21.43 days to 16.79 days (21% improvement) and significant decreases in changeover, transfer, and waiting times. Workforce optimization reduced pre-production personnel from 27 to 21. The total monthly mileage dropped from 6 km to 0.55 km, marking a 91% reduction. The study also highlights substantial cost reductions, a key aspect of lean accounting. The average cost per unit decreased by 21%, from $ 41 to $ 32. Material costs were reduced from $ 2,400 to $ 2,123 per month, and conversion costs dropped from $ 5,800 to $ 4,364 monthly. Consequently, the value stream profit increased markedly from $ 300 to $ 2,013 per month. The non-productive rate reduced from 30% to 4%, while available capacity improved from -7% to 14%. This research underscores the importance of continuous analysis and optimization in pre-production, offering valuable insights for organizations aiming for operational excellence and profitability in the footwear industry. By integrating lean accounting metrics, this study provides a holistic view of the benefits of VSM and the Cut-to-Box system in enhancing both efficiency and profitability.