Browsing by Author "Toy, Ayhan Özgür"

Now showing 1 - 17 of 17

Citation - WoS: 12
Citation - Scopus: 11
An efficient procedure for optimal maintenance intervention in partially observable multi-component systems
(Elsevier Ltd, 2024) Oktay Karabağ; Önder Bulut; Ayhan Özgür Toy; Mehmet Murat Fadiloglu; Toy, Ayhan Özgür; Karabağ, Oktay; Bulut, Önder; Fadıloğlu, Mehmet Murat
With rapid advances in technology many systems are becoming more complex including ever-increasing numbers of components that are prone to failure. In most cases it may not be feasible from a technical or economic standpoint to dedicate a sensor for each individual component to gauge its wear and tear. To make sure that these systems that may require large capitals are economically maintained one should provide maintenance in a way that responds to captured sensor observations. This gives rise to condition-based maintenance in partially observable multi-component systems. In this study we propose a novel methodology to manage maintenance interventions as well as spare part quantity decisions for such systems. Our methodology is based on reducing the state space of the multi-component system and optimizing the resulting reduced-state Markov decision process via a linear programming approach. This methodology is highly scalable and capable of solving large problems that cannot be approached with the previously existing solution procedures. © 2024 Elsevier B.V. All rights reserved.
Citation - Scopus: 1
An Implementation of Flexible Job Shop Scheduling Problem in a Metal Processing Company
(Springer Science and Business Media Deutschland GmbH, 2021) Ali Ozan Özkul; Gamze Küçük; Ceren Çelik; Nevzatcem Öztuna; Mert Demirkan; Ezgi Çağlar Nizam; Yıldırımşan Büyükmertoğlu; Damla Yüksel; Ayhan Özgür Toy; Nizam, Ezgi Çağlar; Demirkan, Mert; Küçük, Gamze; Öztuna, Nevzatcem; Özkul, Ali Ozan; Toy, Ayhan Özgür; Çelik, Ceren; N.M. Durakbasa , M.G. Gençyılmaz
In this study we consider the flexible job-shop scheduling problem at a local metal sheet processing company. We aim to develop a model and an algorithm to generate a weekly production plan for the company. The objective is to minimize the makespan while meeting the demands of products for a given planning horizon. First we provide an LP formulation of this problem. The computational complexity of the problem is NP-hard hence the input data prohibits obtaining the optimal solution in a reasonable time. Therefore we implement a metaheuristic and several rule-based heuristics. These are Genetic Algorithm Giffler and Thompson’s Algorithm and three other Rule-Based Heuristic Algorithms that we developed. We first test our model and heuristics over a set of sample instances then we solve for the real data. Our experimental study indicates that one of the rule-based heuristics we developed outperforms others in most of the instances. © 2020 Elsevier B.V. All rights reserved.
Citation - Scopus: 2
Analysis of inoculation strategies during COVID-19 pandemic with an agent-based simulation approach
(Elsevier Ltd, 2025) Oray Kulaç; Ayhan Özgür Toy; Kamil Erkan Kabak; Toy, Ayhan Özgür; Kabak, Kamil Erkan; Kulaç, Oray
Background: The severity of recent Coronavirus (COVID-19) pandemics has revealed the importance of development of inoculation strategies in case of limited vaccine availability. Authorities have implemented inoculation strategies based on perceived risk factors such as age and existence of other chronic health conditions for survivability from the disease. However various other factors can be considered for identifying the preferred inoculation strategies depending on the vaccine availability and disease spread levels. This study explores the effectiveness of inoculating different groups of population in case of various vaccine availabilities and disease spread levels by means of some performance metrics namely: Attack Rate (AR) Death Rate (DR) and Hospitalization Rate (HR). Method: In this study we have implemented a highly detailed Agent-Based Simulation (ABS) model that extends classical SEIR Model by including five more additional states: Asymptomatic (A) Quarantine (Q) Hospitalized (H) Dead (D) and Immune (M) which can be used as a decision support tool to prioritize the groups of the population inoculated. The approach employs the modelling of daily mobility of individuals their interactions and transmission of virus among individuals. The population is heterogeneously clustered according to age family size work status transportation and leisure preferences with 17 different groups in order to find the most appropriate one to inoculate. Three different Disease Spread Levels (DSL) (low mid high) are experimented with four different Vaccine Available Percentages (VAP) (25% 50% 75% and 85%) with a total of 84 scenarios. Results: As the benchmark under the No Vaccine case Attack Rate Hospitalization Rate and Death Rate goes as high as 99.53% 16.96% and 1.38% respectively. Corresponding highest performance metrics (rates) are 72.33% 15.95% and 1.35% for VAP = 25%, 50.25% 9.55% and 0.94% for VAP = 50%, 24.53% 2.62% and 0.25% for VAP = 75%, and 11.51% 0.002% and 0.08% for VAP = 85%. The results of our study shows that the common practice of inoculation based on the age of individual does not yield the best outcome in terms of performance metrics across all DSL and VAP values. The groups containing workers and students that represent highly interactive individuals i.e. Group (9 10) Group (9 11 10‾) and Group (9 10 11 12‾) emerge as a commonly recommended choice for inoculation in the majority of cases. As expected we observe that the higher is the VAP levels the more is the number of alternative inoculation groups. Conclusions: Findings of this study present that: (i) inoculation considerably decreases the number of infected individuals the number of deaths and the number of hospitalized individuals due to the disease (ii) the best inoculation group/groups with respect to performance metrics varies depending on the vaccine availability percentages and disease spread levels (iii) simultaneous implementation of both inoculation and precautions like lock-down social distances and quarantines yields a stronger impact on disease spread and its consequences. © 2025 Elsevier B.V. All rights reserved.
Capacity Expansion Through Bottleneck Analysis for an OEM Company
(Springer Science and Business Media Deutschland GmbH, 2024) Arda Ali Bayraktar; Gülşah Coşkunseda; Ege Salman; Ayhan Özgür Toy; Nazlı Karatas Aygün; Coşkunseda, Gülşah; Salman, Ege; Toy, Ayhan Özgür; Aygün, Nazlı Karataş; Bayraktar, Arda Ali; N.M. Durakbasa , M.G. Gençyılmaz
In collaboration with Totomak A.Ş this study addresses a critical challenge in heavy machining: identifying and resolving bottlenecks in a recently established crankshaft production line. Our approach to the problem includes a series of steps and advanced computer simulations that meticulously uncover problems and formulate solutions to ensure that production seamlessly matches original plans. In addition the research examines various scenarios subjecting each scenario to a comprehensive cost analysis. These efforts contribute significantly to the overall goal of increasing operational efficiency in the heavy machining sector. Beyond the immediate benefits the study provides a practical and evidence-based method that can address the complex details of manufacturing systems. This has the potential to improve efficiency and strengthen competitiveness in this important industry providing a clear path to greater productivity and success. © 2024 Elsevier B.V. All rights reserved.
Citation - Scopus: 2
Course Scheduling Problem and Real-Life Implementation
(Springer Science and Business Media Deutschland GmbH, 2023) Ali Berk Behrenk; Simge Güçlükol Ergin; Ayhan Özgür Toy; Güçlükol Ergin, Simge; Toy, Ayhan Özgür; Behrenk, Ali Berk; N.M. Durakbasa , M.G. Gençyılmaz
Course scheduling and classroom assignment problem is a common problem for all educational fields. It is an NP Hard problem. Especially universities should handle this problem while preparing course timetabling for each level due to elective courses and students taking upper/lower-level courses. There is a vast literature on this problem both for modelling and solution approaches. Although the essence of the problems is similar for the most each problem has some unique restriction and/or parameters. We study the timetabling of the courses of Industrial Engineering program at Yaşar University. We develop a mathematical model to maximize the minimum lecturer satisfaction. Due to the computational complexity of the problem we proposed a heuristic solution method namely the Genetic Algorithm. Gene structure we use ensures the feasibility of many constraints of the mathematical model. Computational results of optimal and heuristic method are compared. As a real-life implementation a university in Turkey Industrial Engineering Department data is used and results were reported. © 2024 Elsevier B.V. All rights reserved.
Design of job assignment and routing policies in service logistics
(PAMUKKALE UNIV, 2019) Zehra Duzgit; Ayhan Ozgur Toy; Simge Coban; Zeynep Alibasoglu; Ozlem Tok Ozkeskin; Mert Karakaya; Yucel Bayrak; Tok Ozkeskin, Ozlem; Çoban, Simge; Özkeskin, Özlem Tok; Alibaşoğlu, Zeynep; Bayrak, Yücel; Toy, Ayhan Özgür; Karakaya, Mert; Düzgit, Zehra
In this study we consider to improve efficiency of an after-sales technical service in home appliances industry. The efficiency measure is the total time spent in a day to serve customer requests. Hence the objective is to minimize total working hours spent in a day in the aftersales services. We first analyze the system to identify causes of delays in job completion. Upon findings of our analysis we choose to focus on job assignment and job sequencing to improve efficiency. We propose a mixed integer programming model for the assignment of technicians to jobs and sequencing of jobs for each technician to minimize total time spent in a day. Through this model we solve the problem with expected job durations. We present a numerical study to illustrate the proposed solution procedure.
Elektrikli araçların sürdürülebilirlik açısından benimsenmesinin analizi
(2025) Şenocak, Gökcem Mısra; Öztürkoğlu, Yücel; Toy, Ayhan Özgür
Ulaşım küresel kalkınmanın en temel unsurlarından biri olup,insanları birbirine bağlayan ve yerel ile küresel pazarlar arasında köprü kuran bir yapıya sahiptir. Tüm bu olumlu özelliklerine rağmen, çevreye en fazla zarar veren unsurlardan biri olarak da öne çıkmaktadır. Bu nedenle, ulaşımda sürdürülebilir yaklaşımların benimsenmesi büyük önem taşımaktadır. Elektrikli araçların yaygınlaşması, sürdürülebilirlik açısından kritik bir rol oynamaktadır. Literatür taramasında elektrikli araçların yaygın olarak benimsenmesini etkileyen çeşitli ekonomik, çevresel ve sosyal faktörler tespit edilmiştir. Bu çalışma, elektrikli araçların benimsenmesini sürdürülebilirlik perspektifinden değerlendirmek için Best Worst Method kullanarak temel kriterleri belirlemektedir. Ekonomik faktörlerin, yüksek başlangıç maliyetleri ve batarya maliyetlerinin, elektrikli araçların benimsenmesinde en önemli engeller olduğunu göstermektedir. Çevresel faktörler arasında ise batarya üretiminin ekolojik etkileri ve elektrikli araçların yaşam döngüsü boyunca ortaya çıkan çevresel etkiler önemli zorluklar oluşturmaktadır. Bunun yanı sıra, elektrikli araçlar hakkında bilgi eksikliği, düşük model çeşitliliği gibi sosyal faktörler de elektrikli araçlara geçişi engellemektedir. Çalışma, elektrikli araçların benimsenmesini hızlandırmak için ekonomik engellerin azaltılması, sürdürülebilir üretim uygulamalarının geliştirilmesi ve tüketici farkındalığını arttırmaya yönelik stratejilerin uygulanması gerektiğini vurgulamaktadır.
Multi-stage Inventory Management and Transportation Planning
(Springer Science and Business Media Deutschland GmbH, 2023) Canan Özsümbül; Mert İşbilen; Deniz Simge Oriyaşın; Beste Bandioğlu; İsmail Egemen Akpınar; Mücahit Taha Kaya; Remzi Sertan Altınkaya; Simge Güçlükol Ergin; Ayhan Özgür Toy; İşbilen, Mert; Oriyaşın, Deniz Simge; Bandioğlu, Beste; Kaya, Mücahit Taha; Özsümbül, Canan; Akpınar, İsmail Egemen; Toy, Ayhan Özgür; N.M. Durakbasa , M.G. Gençyılmaz
Effective inventory management enables us to meet customer demand by ensuring the delivery of the right products that customers need on time. Likewise the correct design of the supply chain contributes to the effectiveness of stock management. Companies increase customer satisfaction by making effective stock planning. A petroleum company which located in Izmir also aims at accurate and effective stock management. The company decides to determine the product variety and product quantity that should be in the stocks of the micro warehouses to be established in 5 different regions. In this study our objectives are (i) to identify which product should be stored at each micro depot based on the demand forecasting model we develop (ii) to allocate storage capacities to the products (iii) to determine ordering instances and (iv) to conduct an optimal aggregate transportation plan with the given fleet size of different vehicles. The most demanded products for each region were determined by performing ABC analysis with 2 criteria. Since the products selected according to the ABC analysis will be transported to the micro warehouses by 3 different types of vehicles a transportation planning model has been constructed. In order to decide on the ordering policy of the specified products in the micro warehouses 2 different dynamic simulation models were created on a unit basis and on a pallet basis by using the Arena Simulation program with the objective of satisfying a certain service level. © 2024 Elsevier B.V. All rights reserved.
Ortak teslim tarihli ve ağırlıklı toplam erkenlik, gecikme, birim erkenlik ve sabit atama maliyetli paralel makine çizelgeleme
(2025) Yeginoğlu, Berkin; Toy, Ayhan Özgür; Bulut, Önder
This thesis addresses the parallel machine scheduling problem with a common due date, aiming to minimize total weighted earliness, tardiness, unit earliness and fixed assignment costs. The objective function introduces a novel cost structure by extending traditional earliness and tardiness penalties with two additional components: (i) unit earliness costs, and (ii) fixed machine assignment costs. These extensions are motivated by practical considerations in production systems, such as inventory holding costs, quality preservation, and work-in-process (WIP) stocking costs. A mixed-integer linear programming (MILP) formulation is proposed to optimally solve small-sized instances and serve as a benchmark for evaluating heuristic methods. Due to the NP-hard nature of the problem, exact methods become computationally infeasible for larger instances. To address this, a problem-specific Genetic Algorithm (GA) is developed, incorporating a tailored solution representation, one-point crossover and cusromized mutation operators, and the revised V-shaped scheduling property to guide sequencing decisions. A comprehensive experimental study is implemented to assess the GA's performance across various problem sizes and parameter settings. GA parameters are tuned through preliminary testing to ensure a problem adaptive convergence behavior. Extensive computational experiments are conducted using both benchmark problems from the literature and newly generated datasets. Results indicate that the proposed GA consistently achieves near-optimal solutions in significantly less time than the MILP model, especially in larger instances. The algorithm demonstrates strong performance in terms of solution quality and computational efficiency making it a promising approach for solving complex scheduling problems in practical manufacturing environments.
Paralel makine çizelgeleme: Konfeksiyon endüstrisinde bir uygulama
(2023) Çini, Gülce; Toy, Ayhan Özgür; Bulut, Önder
Bu çalışmada, giyim endüstrisine ait bir firmanın dikim aşaması incelenmiş olup, çizelgeleme problemi terminolojisi kullanılarak altı farklı tam sayılı çizelgeleme problemi formülasyonu sunulmuştur. Makinelerin aynı hız ve özelliklere sahip olması nedeniyle, sistemin özdeş paralel makine çizelgeleme problemine uygun olduğu tespit edilmiştir. İlk model, basit bir maksimum tamamlanma süresi (makespan) minimizasyonu olarak verilmiştir. Ardından, sıraya bağlı kurulum süreleri, işe hazır olma tarihleri, makine uygunluğu ve iş bölme gibi kısıtlamalar eklenerek problem genişletilmiştir. Probleme, işlerin teslim tarihleri de eklenerek amaç fonksiyonu toplam erken ve gecikmelerin minimizasyonu olarak belirlenmiştir. Geliştirilen son model bir genetik algoritma ile birleştirilmiştir. Ancak, gerçek hayattaki problemlerde, operatör öğrenme eğrisi, yorgunluğu ve makine bakım gereksinimleri gibi faktörler nedeniyle bir işin işlem süresi değişkenlik gösterebilmektedir. Aynı şekilde kurulum süreleri de insan ve teknik faktörlere bağlı olarak değişebilmektedir. Bu faktörler nedeniyle problem bulanık işlem süreleri ve bulanık kurulum süreleri ile problemin bir uzantısı olarak verilmiştir. Bu yaklaşım, bir rastgele arama algoritması, bir Monte Carlo simulasyonu, aracılığıyla karşılaştırılmıştır. Geliştirilen modeller OPL Cplex ile doğrulandıktan sonra çeşitli veriler yardımı ile optimum çözümün sonuçları ve algoritmaların çözüm sonuçları karşılaştırılmıştır ve önerilen metodolojilerin çözüm kaliteleri sunulmuştur.
Parallel Machine Scheduling with Fuzzy Processing Times and Sequence Dependent Setup Times: An Application in a Textile Company
(SPRINGER INTERNATIONAL PUBLISHING AG, 2022) Gulce Cini; Ayhan Özgür Toy; Önder Bulut; Toy, Ayhan Özgür; Çini, Gülce; Bulut, Önder; C Kahraman; AC Tolga; SC Onar; S Cebi; B Oztaysi; IU Sari
The textile company we consider receives orders from customers for different types of products and with different quantities. The company has several production lines as a machine and an order for a product type as a job to comply with the scheduling terminology. We assume that not all machines are suitable for processing all jobs i.e. jobs can only be processed on their eligible machines. When there is more than one eligible machine those machines are identical in terms of their setups and processing times. We assume that all jobs and machines are ready to be processed at time zero. We also assume that the times required to set up the machine for the next job depend on the job completed and the job in order hence setup times are sequence-dependent. The processing time of a job is not deterministic due to various factors such as operator learning curve and fatigue and machine maintenance requirements. Likewise setup times also vary depending on human and technical factors. We choose to model these uncertainties by assuming fuzzy processing times and fuzzy setup times. Our objective is to assign and schedule jobs to minimize the makespan the completion time of the last job. Our solution approach relies on the comparison of solutions through a randomized search algorithm a Monte Carlo simulation with an improvement routine. We conduct a numerical study and present the solution quality of the proposed methodology.
Citation - WoS: 2
Citation - Scopus: 4
PERFORMANCE COMPARISON OF META-HEURISTICS FOR THE MULTIBLOCK WAREHOUSE ORDER PICKING PROBLEM
(UNIV CINCINNATI INDUSTRIAL ENGINEERING, 2021) Zehra Duzgit; Ayhan Ozgur Toy; Ahmet Can Saner; Toy, Ayhan Özgür; Saner, Ahmet Can; Düzgit, Zehra
This study focuses on streamlining the order-picking process in a warehouse. We consider determining the picking sequence of items in a pick-list to minimize the total traveled distance in a multiblock warehouse where a low-level picker-to-parts manual picking system is employed. We assume that the items are stored randomly in the warehouse. First we construct a distance matrix of the shortest path between any pair of items. Next using the distance matrix we implement two meta-heuristics-the tabu search algorithm and the iterated greedy algorithm-to determine the picking sequence with the minimum total traveled distance. Through a numerical study the performances of the meta-heuristic algorithms are compared with those of popular rule-based heuristics (S-shape largest gap and Combined+) and the bestknown solutions. We conducted the numerical study in two stages. In the first stage we considered a two-block rectangular warehouse and in the second stage we considered a three-block rectangular warehouse. The performance of the heuristics was calculated based on the optimal solution when available or the best calculated bound when the optimal solution is not available. We observed that the iterated greedy algorithm significantly outperforms the other heuristics for both stages.
Production Planning at a Chocolate Company: A Two-Phase Approach by Aggregation
(Springer Science and Business Media Deutschland GmbH, 2020) Zehra Düzgit; Ayşe Beyza Kuzuoğlu; Hazal Kalelioğlu; Hazal Kolay; Merve Başak Güler; Senem Akyol; Ayhan Özgür Toy; Kalelioğlu, Hazal; Kolay, Hazal; Güler, Merve Başak; Kuzuoğlu, Ayşe Beyza; Akyol, Senem; Toy, Ayhan Özgür; Düzgit, Zehra; M.N. Osman Zahid , R. Abd. Aziz , A.R. Yusoff , N. Mat Yahya , F. Abdul Aziz , M. Yazid Abu , N.M. Durakbasa , M.G. Gençyilmaz
There are some factors to be considered while developing production planning policy of chocolate and chocolate-based products. One of these factors is that chocolate is a perishable food therefore it has limited shelf life. Another factor is that, its demand is not always easy to forecast. Holidays mothers’ day teachers’ day valentines’ day and new year are examples of the peak periods. For these special days demand generally follows a seasonal demand pattern where seasonality may also contain trend for some specific product types. Moreover there are two religious holidays (Ramadan Feast and Feast of Sacrifice) in Turkey whose dates shift each year. This phenomenon makes forecasts challenging. Underestimating demand causes loss of customer goodwill lost customers and market share whereas overestimating demand causes excess inventory to keep in stock and risk of fat blooming. Accurate forecasting is critical since it provides a fundamental input for the production plan. The study is conducted in a chocolate company in Turkey. The company does not implement a systematic planning method for chocolate production instead the planning is based on past experiences with respect to experts’ opinions. The objective of this study is to determine the optimal production and ending inventory levels for the period of 2018 so as to minimize total production and inventory holding cost subject to production inventory and capacity related restrictions. A two-phase optimization method is adopted as a solution method. Firstly a mathematical model is developed for the monthly aggregate production planning on product group basis. In order to solve the problem monthly demand for product groups are forecasted based on the sales data of the previous two years. Secondly a mathematical model for weekly disaggregate production planning is developed for each end item using the outputs of the aggregate planning as input. The disaggregate production plan gives the weekly planned production and inventory levels for end items for year 2018 with minimum deviation from the aggregate plan. Although the proposed solution model is implemented for year 2018 it can be used for the coming years by updating some parameters. © 2022 Elsevier B.V. All rights reserved.
Quantitative Methods for Agri-Food Supply Chain Resilience: A Systematic Literature Review Using Text Mining
(Elsevier, 2025) Çali, Sedef; Toy, Ayhan Özgür; Ekren, Banu Yetkin
Agri-food Supply Chains (AFSCs) face increasing disruptions from natural disasters, pandemics, and economic crises, necessitating robust quantitative analysis for resilience. This study conducts a Systematic Literature Review (SLR) using text mining and Latent Dirichlet Allocation (LDA) to identify six key research themes, including risk management, pandemic effects, simulation-based resilience, climate change, market price volatility, and optimization models. Findings reveal that multi-criteria decision-making, simulation, optimization, and machine learning are widely used, yet gaps remain in Artificial Intelligence (AI)-driven risk prediction, real-time data integration, and adaptive decision-making This review offers insights for researchers and practitioners, emphasizing the need for AI, digital twins, and blockchain to enhance AFSC resilience. Copyright (C) 2020 The Authors. Published by Elsevier B.V. This is an open access article under the CC BY-NC-ND license (http://creativecommons.org/licenses/by-nc-nd/4.0/)
Rüzgar türbini verimliliğinin analizi için bulanık çıkarım sistemi ve uyarlamalı ağ tabanlı bulanık çıkarım sisteminin uygulanması
(2023) Özcan, Gülcan İncu; Toy, Ayhan Özgür; Ulutagay, Gözde
Bu çalışma rüzgar türbinlerinin performansını değerlendirmek için bir bulanık çıkarım sistemi (FIS) ve uyarlanabilir bir nöro-bulanık çıkarım sistemi (ANFIS) kullanılmasını önermektedir. Rüzgar hızı, kanat uzunluğu ve diğer türbinlere olan mesafe gibi rüzgar türbini performansını etkileyen temel faktörler, veri analizi yoluyla belirlenir. Bu faktörlerin anlaşılması, türbin işletimi ve tasarımına ilişkin bilinçli karar vermeye olanak tanıyarak genel verimliliğin ve maliyet etkinliğinin iyileştirilmesine katkıda bulunur. Yazarlar, rüzgar türbinlerinin birden çok girdi ve çıktıya sahip karmaşık sistemler olduğunu ve bunun da geleneksel analiz yöntemlerini doğru sonuçlar vermede yetersiz kıldığını ileri sürüyorlar. FIS ve ANFIS, rüzgar türbini verilerindeki belirsizliği modellemek ve türbin verimliliğinin daha doğru bir şekilde değerlendirilmesini sağlamak için tasarlanmıştır. Bu çalışma, yaklaşımın etkinliğini Türkiye'deki bir rüzgar santralinden gerçek dünya verilerini kullanarak göstermektedir. Çalışma sonucunda FIS'in rüzgar türbini verimliliğini analiz etmek için umut verici bir araç olduğu ve rüzgar türbini sahlarının performansını iyileştirme potansiyeline sahip olduğu sonucuna varılıyor. Bu konuda çalışmamızdaki temel motivasyonumuz, müşterinin iki türbini daha fazla enerji üreteceği için değiştirme talebidir. Veri eksikliğinden dolayı bu durumu matematiksel olarak ifade edemesek de bu kararda kullanılacak değerli bilgiler paylaşılmıştır. Keywords: rüzgar türbini, bulanık çıkarım sistemi, uyarlamalı ağ tabanlı bulanık çıkarım sistemi, yenilenebilir enerji, verimlilik
Sıraya bağımlı kurulum süreleri ile tek tip paralel makine çizelgelemesi üzerine bir çalışma
(2022) Yıldız, Beste; Toy, Ayhan Özgür; Kandiller, Levent
Scheduling problems are essential for decision-making in many academic disciplines, including operations management, computer science, and information systems. Since many scheduling problems are NP-hard in the strong sense, there is only limited research on exact algorithms and their efficiency when implemented on parallel computing architectures. This master's thesis considers the uniform parallel machine scheduling problem with sequence-dependent setup times to minimize the maximum completion time (makespan). We present an IP formulation, which clearly describes our problem and can be used to obtain optimal solutions for small-sized problems. As our problem is NP-hard, we propose a randomized heuristic with an improvement subroutine. The performance of the proposed heuristic through a computational study was tested with 320 instances. We created these instances using the full factorial design of experiment (DOE) with five different factors. Our computational study indicates that the proposed mathematical model takes 22.88 minutes on average, and the heuristic algorithm achieves these results only in 0.062 minutes. The average solutions obtained with the heuristic have an approximately 4% Gap value for average CPLEX solutions. Also, the contribution of the improvement subroutine step to the overall performance of the heuristic is 73.34%. Keywords: parallel machine scheduling, sequence-dependent setup time, full factorial design, randomized heuristic, uniform machines, total completion times
Citation - WoS: 1
Citation - Scopus: 1
Uniform Parallel Machine Scheduling with Sequence Dependent Setup Times: A Randomized Heuristic
(SPRINGER-VERLAG SINGAPORE PTE LTD, 2022) Beste Yildiz; Levent Kandiller; Ayhan Ozgur Toy; Yıldız, Beste; Kandiller, Levent; Toy, Ayhan Özgür; NM Durakbasa; MG Gencyilmaz
We consider the uniform parallel machine scheduling problem with sequence-dependent setup times to minimize the total completion times. This problem is known to be NP-hard. We propose a simple randomized heuristic with an improvement subroutine. We analyze the performance of the proposed heuristic through a computational study. Our computational study indicates that the heuristic performs well in terms of optimality gap and solution time