TR Dizin İndeksli Yayınlar Koleksiyonu / TR Dizin Indexed Publications Collection
Permanent URI for this collectionhttps://hdl.handle.net/20.500.14365/4
Browse
7 results
Search Results
Article Kısmi Gözlemlenebilir Çok Bileşenli Sistemler için Bakım Politikalarının Pekiştirmeli Derin Öğrenme Yöntemleri ile Belirlenmesi(Pamukkale Univ, 2025) Karabağ, OktayBu çalışmada, kısmi gözlemlenebilir çok bileşenli sistemler için bakım/onarım kararları incelenmiştir. Bu tip sistemler genellikle servis sağlayıcının uzakta olduğu koşullarda işletilmekte ve bileşenlerin aşınma seviyeleri genellikle sensörler yardımı ile tam olarak izlenememektedir. Rüzgâr türbinleri, bu tarz sistemlere birebir uyan bir örnek oluşturmaktadır. İlgili sistemlerde, servis sağlayıcı ne zaman bakım/onarım yapacağına, bakım kararı ile birlikte hangi parçaları bakım noktasına sevk edeceğine ve bakım noktasındaki incelemesinin ardından hangi sistem bileşenlerinin değiştirilmesi gerektiğine karar vermektedir. Çalışmamızda, bahsi geçen bu komplike karar problemi kısmi gözlemlenebilir Markov karar süreci olarak modellenmiş ve ilgili nümerik çözümler aktör kritik pekiştirmeli öğrenme yöntemi kullanılarak elde edilmiştir. Yaptığımız nümerik çalışmalar, pekiştirmeli öğrenme algoritması ile elde edilen çözümlerin pratikte ve literatürde yaygın olarak kullanılan sezgisel bakım/onarım politikalarına kıyasla daha iyi sonuçlar verdiğini göstermiştir. Bazı durumlarda, bu çözümlerin ortalamada %10-%15 düzeyinde bir iyileştirme sağladığı gözlemlenmiştir. Ayrıca, düzeltici bakım maliyeti, acil sipariş maliyeti ve fazla yedek parçayı geri döndürme maliyeti arttıkça, pekiştirmeli öğrenme algoritması ile elde edilen çözümlerin diğer sezgisel politikalara kıyasla daha fazla avantaj sağladığı da belirlenmiştir.Article Citation - WoS: 2Günlük Planlanmış Bakım ve Onarım Servis Planlaması için Zaman Pencereli Bir Çok Depolu Araç Rotalama Problemi(Pamukkale Univ, 2023) Toru, Elif; Yılmaz, GorkemKocaeli/Dilovası bölgesinde üretim yapan bir kompresör üreticisi, Marmara bölgesi ve çevresinin ertesi günkü bakım ve onarım talebini karşılamak için günlük olarak araç rotalama ve personel planlaması yapmaktadır. Servis planlamasından önce servis türleri ve süreleri müşteri ile kararlaştırılır. Servis personelinin başlangıç ve bitiş noktaları ikametgah adresleri olmak üzere, belirli bir planlama günü için araçlar ve ilgili operatörleri bilinmektedir. Tüm talepler, müşteriler tarafından verilen zaman pencerelerinde karşılanır. Problem, Zaman Pencereli Çok Depolu Araç Rotalama problemi (MDVRPTW) olarak ele alınmaktadır ve bir karma tamsayılı doğrusal programlama modeli geliştirilmiştir. Matematiksel model çözümü şirketin problemini çözmeye yeterlidir. Büyük örnekleri çözmek için, kısa sürede yeterli bir çözüm sağlayan bir kümeleme algoritması geliştirilmiştir.Article Certainty Factor Model in Paraphrase Detection(Pamukkale Univ, 2021) Metin, Senem Kumova; Karaoglan, Bahar; Kisla, Tarik; Soleymanzadeh, KatiraIn this paper, we address the problem of uncertainty management in identification of paraphrase sentence pairs. Paraphrase sentences are simply sets/pairs of sentences that express the same facts and/or opinions using different words or order of words. We propose the use of certainty factor (CF) model in paraphrase detection. A set of succeeding paraphrase detection features (generic and distance based features) is built by filtering and this set is used as evidences in CF model. The CF model is evaluated by F1 and accuracy measures on Microsoft Research Paraphrase corpus. The results are compared to the well-known Bayesian reasoning. The experimental results showed that CF model is an alternating paraphrase detection method to Bayes model.Article Citation - WoS: 1Image Quality Assessment Based on Manifold Distortion(Pamukkale Univ, 2021) Turkan, MehmetAn image quality metric is proposed by introducing a new framework for full reference image quality assessment from the perspective of image patch manifolds. Assuming that most natural scenes are sampled from low dimensional manifolds or submanifolds, perceived image degradations in structural variations can be quantitatively evaluated on the surfaces of highly nonlinear image manifolds. Manifold distortion image quality index first characterizes intrinsic geometric properties of the locally linear manifold structures of spatially local patch spaces, and then measures the deviation from the original smooth manifold structure to calculate the distortion index. Experimental results demonstrate a strong promise with a comparison to both subjective evaluation and state-of-the-art objective quality assessment methods.Article Estimating the Difficulty of Tartarus Instances(Pamukkale Univ, 2021) Oguz, KayaTartarus is a commonly used benchmark problem for genetic programming. However, it has never been fully explored for its difficulty tuning property. Using the data from a previous study in which we have executed millions of Tartarus instances, we contribute to the literature with an equation to estimate their difficulty. Our approach uses four metrics that are embedded into the equation. These metrics are related to the number of clusters and clusters sizes, the distances of boxes to the edges of the board grid, the number of boxes around the agent, and the minimum number of actions for the agent to reach the largest cluster. The coefficients of these metrics have been fit to the data using the general linear model and a mean residual error of similar to 0.1 has been achieved. This is the first study that can estimate the difficulty of a Tartarus board without modifying the problem in any way.Article Design and Analysis of Bi-Objective Flexible Job Shop Scheduling Problem: a Case Study in Construction Equipment Manufacturing Industry(Pamukkale Univ, 2020) Resat, Hamdi GirayThis paper presents a design and development of mixed-integer linear optimization model for scheduling of flexible job-shop production problem under capacity constraints by using exact solution algorithm. Modelling approach is designed in order to introduce data analysis in real situations, minimize production time in production lines, reduce total production costs, and reveal important features of mathematical programming problem in detail. The main purpose of this study is to obtain faster and efficient Pareto solution sets for bi-objective problem by using epsilon-constraint method. Generated Pareto frontier using real life data is shared with decision makers. The GAMS programming language is used during the solution phase of a mixed-integer linear optimization model for bi-objective problem and production efficiency of the company is increased around 16.6% in terms of production cost.Article System Identification Work on 199+325 Steel Railroad Bridge and Development of Its Calibrated Finite Element Model(Pamukkale Univ, 2018) Ozcelik, Ozgur; Girgin, Ozgur; Amaddeo, CarmenRailroad bridges maintained and operated by the State Raid Road Agency (TCDD) constitute the main passage ways and junction points of the railroad network of the country. Most of these bridges have been under service for more than 100 years. These bridges are exposed to larger service loads as compared to the highway bridges, and are open to external actions leading to changes in their dynamic parameters. Due to these reasons, the railroad bridges must routinely be checked and serviced. The routine checks done by TCDD are based on visual inspection, and highly subjective and dependent on the technician's experience. This increases the chance of making mistakes and missing hidden structural damages. Vibration-based structural health monitoring offers a more objective framework which has the potential to reduce operator dependent nature of the routine checks. This study presents modal parameter estimation studies by in-situ experiments and a developed reference numerical model of the 199+325 steel railway bridge located in Usak. The dynamic response of the bridge was measured in four different test setups and in two different temperature states, and under ambient vibration conditions. Modal parameters of the bridge are estimated using two different output-only system identification methods, namely, Enhanced Frequency Domain Decomposition and Data-driven Stochastic Subspace Identification methods. The identification results obtained under different temperature conditions are compared in assessing the effects of temperature change in identification results. Three dimensional finite element model of the bridge is created using FEDEASLab software. Trial-and-error type model updating study is conducted. Therefore a reference numerical model of the bridge representing its current condition is obtained. This model will be facilitated in the future for damage identification purpose using the sensitivity based finite element modeling updating method.