Endüstri Mühendisliği Bölümü Tezleri

Permanent URI for this collectionhttps://hdl.handle.net/20.500.12416/226

Browse

Now showing 1 - 20 of 39

An application of the vehicle routing problem to a glass manufacturing firm
(2006) Seyran, İpek
This thesis presents an exact algorithm and a heuristic method for the delivery and transportation of glass plates for a glass manufacturing firm. A variant of the Capacitated Vehicle Routing Problem (CVRP) is proposed as a first attempt to solve the problem which minimizes total traveling of all the vehicles. Since the CVRP is known to be NP-hard, the solution method cannot obtain a solution to the model. Therefore an exact algorithm which is a kind of set-covering-based algorithm is proposed next. The CVRP is modeled as a set covering (SC) problem. Then column generation (CG) method is applied to the linear relaxation of the SC problem. The branch-and-price algorithm is utilized in finding an integer solution on the solution of the CG procedure. Numerical experimentations reveals that exact algorithm is slower, and fails finding a solution to larger size problems. Consequently a heuristic is developed as a generalization of petal algorithm. Initialization of this algorithm requires using some Traveling Salesman Problem (TSP) construction heuristics for finding a TSP tour, and a TSP improvement heuristic further improves the TSP tour. Then Petal Algorithm is applied to find all of the feasible petal routes to the TSP tour obtained. SP model helps the petal routes to find the best VRP routes. When the best VRP route is found, a VRP improvement heuristic attempts improving the VRP route. Finally, the number of delivery vehicles required and the vehicle routes are determined for the glass manufacturing firm
An employee transporting problem and its heuristic solutions
(2007) Önder, İlter
A typical Vehicle Routing Problem (VRP) can be described as a problem of designing the least cost routes from one depot to a set of geographically scattered points. The VRP assumes that the vehicle capacities are identical, yet in real life the vehicle capacities are different. Therefore, this thesis presents a study of VRP with heterogeneous vehicles (HVRP). A lower bound on the cost of vehicles and routing is calculated for the HVRP using some mathematical models. Various heuristics are attempted to decide which one provides better solutions on the average. The better heuristic is selected based on the deviation from the lower bound. A simple software is prepared using the best heuristic methods for the employee pickup and delivery operations of a hypothetical company
The use of resampling techniques for lifetime data analysis in industrial engineering
(2007) Aslan, Miray Hanım
This study concerns with estimating the parameters in lifetime of fragile population and the ratio of fragile population to the fragile and durable (mixed) population by using trunsored models (unification of truncated and censored models) approach. The purpose of this study is to illustrate the bootstrap resampling method used for the parameter estimation in trunsored models. The bootstrap method is especially convenient to make statistical inference when distributional assumptions are not valid. Therefore, trunsored models with bootstrapping, which follow a consistent strategy in statistical inference and data analysis, lead to more accuracy for evaluation. Like many real world cases, the thermal endurance data in material failure analysis do not follow any distribution perfectly. Furthermore, time and cost limitations prevent to observe a great number of data to analyze accurately. Thus, the trunsored model approach with bootstrapping is thought as potential to reduce the cost of destructive testing due to reduced frequency of testing, to prevent failures and to improve product reliability. The approach presented in this study may also be applied to many other real life problems
An integration of Bayesian belief networks in multi-criteria decision analysis: An application in medical decision making
(2007) Karatepe, Dilan
This thesis analyzes the Medical counseling problem regarding breast cancer disease by using Bayesian Belief Network structure to visualize the conditional dependencies between variables and their impact on each other. Analytical Hierarchy Process (AHP) method is utilized to approximate the conditional probabilities of the diagnostic Fine Needle Aspiration (FNA) test outcome. The accuracy of the results obtained from surgical biopsy are compared to the actual outcomes and evaluated in terms of consistency and error magnitude
A manufacturing system analysis combined with waste management
(2008) Elker, Sedef
This thesis presents an environmentally integrated manufacturing system analysis for companies looking for the benefits of environmental management in achieving high efficiency levels. When the relationship between environmental costs and manufacturing decisions is examined, it can be seen that the productivity of the company can be increased by using an environmentally integrated manufacturing system analysis methodology. Therefore, such a methodology is proposed and the roadmap for generating environmentally friendly and economically favorable alternative waste management solutions is elaborated. The methodology consists of data collection, operational analyses of the processes, identification of wastes and evaluation of waste reduction alternatives proposed both technically and economically. The proposed methodology is examined in a car battery manufacturing company, which generates hazardous wastes composed of lead (Pb). The main focus was on the wet-charged lead-acid battery manufacturing. It is aimed to decrease the wastes derived from the production so that the efficiency in raw materials usage is increased and the need for recycling the hazardous wastes is decreased. Following the identification of waste production points, at least one alternative is proposed for the reduction of waste. Two different alternatives are proposed for the reduction of dross formation based on the results obtained from the experiments carried out. The applicability of alternative waste management solutions is investigated from operational, technical and economical point of views and benefits and limitations are identified for each of the alternatives.
Scheduling in a three-stage dedicated hybrid flowshop with a common third-stage
(2008) Soysal, Serdar
In this study, we consider a scheduling problem of a manufacturing environment in which there are two manufacturing flow lines, where the third stage of the first line and the second stage of the second line are common. Each stage in the first flow line has a single machine whereas the second flow line contains two identical parallel machines in its first stage. Type-1 jobs are processed in the first flow line, whereas second flow line is dedicated to type-2 jobs. The last operation, of both types of jobs, must be processed on a common machine. The problem is to determine the sequence and schedule of all jobs at all stages of the two flow lines so that the makespan is minimized. We develop a mathematical model and a branch-and-bound algorithm with lower and upper bounding procedures to find optimal solution; we propose heuristic algorithms which provide good quality solutions at little computational effort when the computational effort to obtain an exact solution is prohibitive. The effectiveness of our solution approaches are demonstrated by computational analyses.
Fuzzy regression modeling of rates in a metal casting process
(2009) Kılıç, Tuna
This study has two purposes. One of them is to develop model of a metal casting process quality. This model can be used by the company to reduce the number of defects by identifying the process variables which have the most important effects on a certain defect type. The second purpose is to test and develop certain fuzzy regression approaches for the case problem. In the case study, 36 process variables are observed from the metal casting process. Initially, stepwise linear regression (LR) method is applied to the data set and 8 independent variables are selected as significant. When the process variables are examined, it is realized that instead of identifying them as crisp, expressing them with fuzzy numbers is more appropriate. In the scope of the study, all fuzzy numbers are assumed to be triangular fuzzy numbers. First the Hojati-Bector-Smimous (HBS2) method developed by Hojati et al. (2005) is generalized to multi variable modeling and then applied to the data set. In order to make a comparison between HBS2 and linear regression (LR) approach, the latter is also applied. Similarly, in order to make a comparison between HBS2 and other fuzzy methods, Fuzzy Functions (FF) method developed by Türkşen and Çelikyılmaz (2006) is used. Furthermore, Non-Parametric Improved Fuzzy Classification Functions developed by Özer (2009) is adapted to the case of fuzzy linear regression. The newly developed method called as Non-Parametric Improved Fuzzy Functions (NIFF) is applied to the same data set for a comparison with the other solutions.
An edge matching approach for two-dimensional irregular shaped cutting stock problems
(2012) Akbulut, Derya
In this thesis, a two-dimensional irregular shape cutting stock problem is considered, in which a number of irregular shaped pieces are cut out of rectangular stock sheets so that waste of stock to be minimized. This is an operational problem commonly observed in metal cutting and textile industries. In the literature, there are many algorithms proposed to find optimal or suboptimal solutions for the problem. Since the problem is NP-hard, heuristic approaches predominate among the solution methodologies. In this study, a non-linear mixed integer mathematical model formulation is developed and it is tested for small sized problems. For larger scale problems, an edge matching approach is proposed to generate cutting patterns. The approach is based on positioning of the pieces in such a way that their most fitting edges are aligned together or to the borders. In this way, the total scrap and the cutting operations are minimized. On the contrary to the most of the solution methodologies in the literature, the method enables rotating pieces by any angle during the alignment process and further more it is applicable for irregular shaped stock materials. The developed procedure is tested against the traditional cutting stock approaches using benchmark test problems reported in the literature. It is found that our procedure outperforms for a large portion of these benchmark problems.
Workforce assignment in a multi-worker multi-sided mixed-model assembly line balancing problem
(2012) Karabak, Funda
In this thesis, workforce scheduling problem in a mixed model assembly line is studied in which precedence relations, qualification of workers, walking times between tasks and worker requirements are considered. The motivation of the study comes from a real life problem at MAN Türkiye A.Ş. Mixed integer linear programming formulations for minimization of number of actively used workers and minimization of makespan value are developed. However, since the problem is NP-Hard, a heuristic approach is introduced. The developed heuristic attempts to minimize the number of actively used workers first, and then to minimize the makespan value.
Unrelated parallel machines scheduling under machine availability and eligibility constraints
(2012) Kurt, Atıl
In the literature of the parallel machines scheduling, it is generally assumed that all machines are continuously available for processing jobs and each job can be processed by any machine. However, these assumptions become unrealistic in some industrial environments. In this study, we consider the problem of scheduling n independent jobs on m unrelated parallel machines subject to machine availability and eligibility constraints, given the maximum continuous working time before the maintenance of each machine and the maintenance time. Our objective is to minimize the makespan, which is the time to complete the processing of all jobs. We consider both resumable and non-resumable jobs, and develop mathematical models and heuristic algorithm that obtain exact and near-optimal solutions, respectively, for both cases with multiple machine unavailability periods. Computational experiments are done to evaluate the performance of our solution methods in terms of both quality and time. The results show that the proposed heuristic algorithm finds near-optimal solutions in very short time
Spare parts inventory management with demand priorities
(2013) Özen, Duygu
In this thesis, we analyze a single-echelon inventory management problem of a manufacturing company which is a subsidiary company of an engineering company. The company manufactures machines and also supplies supplementary equipment and spare parts to its customers. Based on the priorities in meeting the demands, the customers of the manufacturing company are classified as high priority and low priority customers. The engineering company which is the most important customer is treated as the high priority customer whereas all other customers having the same priority are assumed to be the low priority customers. In this study, considering these two types of customers, we focus on an (R, Q) inventory model with a critical level policy where backorders and lost sales are allowed. Below the critical level the demands of low priority customers are not met to retain stock for high priority customer and therefore the unmet demands of low priority customers are lost. The demands of high priority customer are satisfied until inventory level reaches zero and under stock out any unsatisfied demand of high priority customer is backordered. Considering the possible levels of reorder point (R) and critical point (c), three different cases are defined as c=R, c>R, and c
Balancing and sequencing multi-zone mix model assembly lines
(2013) Güner, Nizamettin Doğan
Assembly based production has importance for large scale industries of developing countries. Satisfying demand and meeting production due dates of those kind of products are the prime targets of companies. One of the largest bus production facilities of Turkey, MAN Türkiye A.Ş., aims to minimize delivery costs and to meet daily production amounts by preventing delays in assembly operations. In this study, an assembly line design is suggested in order to achieve the company goals by studying single model line balancing and mixed model sequencing problems. Based on the characteristics of real life problem, three mathematical models for line balancing and sequencing are constructed. Furthermore, a heuristic algorithm is proposed for single model line balancing problem. This study is a part of a project that is partially supported by SANTEZ Program of Ministry of Science, Industry and Technology of Republic of Turkey
Freight transportation in Turkish state railways (tcdd) high-speed train system
(2013) Keskin, Merve
The aim of this thesis is to improve the freight transportation system of Turkish State Railways (TCDD) by using high speed trains (HSTs) on high speed lines as a new concept. The development of railway and HST systems in the world and in Turkey is discussed and comparison of these systems with other transportation modes is investigated. Since HST is faster than trucks and cheaper than airlines for freight transportation and increases the utilization of railway system substantially, it is very advantageous. While HSTs are used for freight transportation in some countries such as France, Germany etc., they are not used currently for freight transportation in Turkey. Therefore, in this thesis, freight is added to the existing high speed rail system which is currently used for only passengers. A HST scheduling model is constructed to observe the effects of freight when it is included in the system. The aim of the model is to complete daily travel as early as possible within a day to increase utilization of the system. Daily travel of each train consists of sequence of cities and links between these cities. In experimental study, two types of sequence of cities are used: trips using one sequence of cities and trips using two sequences of cities. In the first type, all available trains are used to perform daily travels in one sequence of cities. Then for the second type, the linkage is divided into two and available trains are allocated to these separate two linkages. Freight transportation is also analyzed in two cases which are adding separate freight trains to the system and using the same trains with passengers for freight transportation. It can be concluded that dividing the sequence of cities into two provides the completion of the train services earlier in a day and using the same trains with passengers for freight transportation provides more time savings for the system. Freight transportation in HSTs is a complex task to achieve and requires many factors to be taken into account. This thesis is a first attempt to address this complex task.
A new customer order scheduling problem on a single-machine with job setup times
(2014) Akkocaoğlu Çatmakaş, Hale
In this study, we consider a relatively new class of the customer order scheduling (COS) problem where each order consists of one or more individual jobs. All jobs in the same customer order are processed successively and delivered at the same time to the customer. Thus, the completion time of the last job processed in each customer order defines the completion time of the order. A sequence independent setup is required before the processing of each job in a customer order. However, no setup is necessary before the processing of the first job of a customer order if this first job is the same as the last job of the immediately preceding customer order. We investigate the single-machine problem for two cases in which the makespan, which is the time to complete all customer orders, is minimized in the first case while the total completion time, which is the sum of the completion time of the orders, is minimized in the second case. For some special cases of both problems, we derive the properties of the optimal solution, which can be obtained by priority rules. We show that the makespan problem is polynomially solvable. For the total completion time problem, we develop a mixed integer programming model capable of solving small-sized problem instances optimally and propose a constructive heuristic algorithm that obtains optimal and near-optimal solutions for medium and large sized problem instances. Computational experiments are done to evaluate the performance of our solution approaches in terms of both quality and time. The results show that the mixed integer linear programming model does not seem to be a useful alternative, especially for large-sized problem instances. However, the proposed heuristic algorithms find near-optimal solutions in very short time
Using containers as astorage facilities in humanitarian logistics
(2014) Şahin Arslan, Ayşenur
In humanitarian logistics, relief supplies are pre-positioned in strategic locations near disaster-prone areas. The beneficiaries are supplied from pre-positioned inventory during the initial days after the disaster; therefore, having those supplies ready to dispatch is of critical importance in disaster response. Previous studies focused on operating a permanent warehouse building for storage of relief supplies. The main objective of this thesis is to investigate how containers (e.g., 20 or 40 feet freight containers) could be used as storage facilities. Using containers as storage facilities is an appealing idea because of the ad-hoc nature of the disaster relief network. They can be shipped from unaffected locations to the disaster locations after the disaster strike. Containers can be stacked on top of each other occupying less land than warehouse buildings. Moreover, the containers that are used to store relief supplies can be used as accommodation places in the immediate aftermath of a disaster strike. In order to investigate the practicality of this idea in Turkey, a mathematical model is developed to determine the location and quantity of containers as well as the type and amount of relief supplies to store. The model is tested using earthquake risk data, estimates of population under risk, and distances between cities. To investigate how and at what cost freight containers could be used as an inventory holding mechanism instead of a transportation unit, the layout and cost comparison of two alternatives; (1) stocking in a warehouse or (2) storage in containers, are performed. Leasing and purchasing option costs of these alternatives are compared using present worth (PW) analysis. The results reveal that leasing option is not cost advantageous in the long run for both warehouse storage and container stockpiling area. Warehouse construction results with the least cost but it requires more area than a container stockpiling area and incurs more operating costs including lighting, ventilation, and maintenance as well as handling of the pallets. On the basis of the results of this research, it can be concluded that using containers for storage is cheaper than operating a warehouse building. The results of our thesis illustrate how to best use containers as storage facilities to achieve the most possible response-time benefit and also support the implementation of an ad-hoc pre-positioning strategy. The thesis is carried in conjunction with Turkish Prime Ministry Disaster and Emergency Management Presidency (AFAD) and the results of this thesis will help AFAD reach beneficiaries of a disaster in shorter time, with fewer disruptions, and in an efficient way
An intermodal humanitarian logistics model based on maritime transportation for relief item distribution in Istanbul
(2015) Özkapıcı, Dilsu Binnaz
İstanbul Türkiye'nin ekonomi başkenti ve en kalabalık şehri olup, deprem riski yüksektir. Bir deprem durumunda uluslararası ve ulusal kaynaklardan yardım malzemeleri tedarik edilecektir. Daha önceki çalışmalar, yardım malzemelerinin dağıtımı için, şehri ikiye bölen Boğaz'ı ve İstanbul'da deniz yolu ulaşımının avantajlarını göz önüne almamıştır. Bu çalışmada, İstanbul için, bir deprem durumunda insanların zararlarını azaltmak amacıyla deniz ve karayolu ulaşımını içeren ve bu yolların hasar görebilirlik olasılıklarını da göz önünde bulunduran intermodal bir yardım malzemesi dağıtım modeli geliştirilmiştir. Önerilen matematiksel model denizyolu ulaşımından ve İstanbul'un limanlarından etkin bir biçimde yararlanmakta ve Avrupa ve Anadolu yakaları arasında yardım malzemesi ulaşımına izin vermektedir. Talep noktalarına yardım malzemesi sağlamak için denizde-üs kavramı da kullanılmaktadır. Yerel ve uluslararası kaynaklardan tedarik edilen yardım malzemeleri, ana tedarik noktalarından (Haydarpaşa Limanı, Ambarlı Limanı ve Marmara Denizi'ne yerleştirilmiş bir konteyner gemisinden) doğrudan karayoluyla talep noktalarına (İstanbul'un ilçelerine) taşınmaktadır. Alternatif olarak, yardım malzemeleri denizyoluyla İstanbul'un limanlarına ve daha sonrasında karayoluyla limanlardan ilçelere taşınmaktadır. Tedarik noktalarında mevcut yardım malzemesi miktarına göre değişen çeşitli senaryolar çalıştırılmıştır ve sonuçlar üç performans ölçümü açısından incelenmiştir: (1) bir adet yardım malzemesini taşımak için gerekli ortalama süre, (2) intermodal taşıma yüzdesi ve (3) kullanılan gemi sayısı. Ayrıca, Avrupa ve Anadolu yakaları için tedarik dağılımı oranları üzerine bir analiz ve yalnızca karayolu ulaşımından yararlanan alternatif bir yardım malzemesi dağıtım modeliyle karşılaştırma yapılmıştır. Denizyolu ulaşımı ve denizde-üs kavramlarından yararlanılmasının insani yardım faaliyetlerinde esneklik sağladığı ve önerilen modelin İstanbul için etkili ve güvenilir bir afet yardımı sistemi oluşturduğu görülmüştür
Intermodal freight transportation in humanitarian logistics
(2016) Kavlak, Hasan
Intermodal freight transportation is used heavily in commercial logistics around the world. On the other hand, in humanitarian logistics, it is not considered as the primary solution. Transportation resilience during response phase is an important performance criterion for humanitarian logistics and it mostly depends on the availability of the modes used for transporting relief items. Previous studies brought multi-modal or multi-vehicle transportation to forefront. Intermodal freight transportation is not focused from a perspective of resilience in humanitarian logistics. This study aims to highlight the differences between intermodal transportation and multi-modal transportation and to present a resilient transportation system without handling of relief items. Two integer programming models are developed based on a time-space network by considering route and vehicle availabilities changing dynamically over a specified time horizon. In these models, different types of vehicles and different capacities of unit loading device and intermodal transportation unit are considered, namely, truck (highway), freight train (railway), vessel (seaway), plane (airway) and helicopter (airway). This study proposes a unit loading device for humanitarian logistics that is compatible with different transportation modes. The first model includes an integer variable representation for vehicle fleets of different transportation modes. It can be concluded that the second model includes an index representation of individual vehicles for different transportation modes. The first mathematical model with integer variable representation of vehicle fleets is more effective than the second one. Five real life scenarios are fed into these mathematical models and the results are compared. The results of the experimental study show that intermodal transportation provides better humanitarian response in terms of resilience.
Pricing and remanufacturing decisions with speculators and strategic consumers
(Çankaya Üniversitesi, 2017) Yozgat, Simge
We investigate pricing and remanufacturing decisions with speculators and strategic consumers for a single type of a product over a two-period sales horizon. A monopolist manufacturer produces a fixed quantity in the first period. Some of the sold products are returned at the end of the first period, which are collected by the manufacturer and/or speculators. Returned products are remanufactured, and then sold in the second period, along with any new products remaining from the first period. Mathematical models take into account the behavioral patterns of different types of customers to maximize the manufacturer's expected total profit. Solution to the mathematical models show that one particular customer behavior is optimal. Specifically, the manufacturer should use a fixed-pricing policy for all products –new and remanufactured alike– and set the price at the maximum level that strategic customers are willing to buy. This will force customers to wait for the second period to buy any products, and hence, will yield the maximum profit for the manufacturer. Additionally, the manufacturer is better off remanufacturing. The sensitivity analysis has shown that the profit is most sensitive to the number of strategic customers.
Regional coverage of science centers in Turkey
(Çankaya Üniversitesi, 2017) Türüdü, Özgün
In 2011, Turkish Supreme Council for Science and Technology decided to establish Science Centers with TÜBİTAK's support in all the metropolitan municipalities by 2016 and in all provinces of Turkey by 2023. Since the budgets of Science Centers are high and they are attracting visitors from other provinces, an additional decision was taken in 2016 to establish Science Centers in prioritized regional centers. The objective of this thesis is to determine the locations of Science Centers in Turkey considering the new criteria defined in the additional decision using integer and mixed integer programming models such as p-median, p-dispersion, and a multicriteria model. A hybrid model was proposed to utilize p-median and multicriteria model together. P-median model and hybrid model take into consideration population density of provinces' and distance criteria. P-dispersion model, and the multicriteria approach take into consideration only the distance criterion. Location of existing five science centers are included in all models. Models are solved for varying p (e.g. number of science centers) values of 5, 10, 13, 15, 20, 25 and 30. Four performance criteria were determined and solutions of the models are compared to each other. Cities selected using different models are compared. To the best of our knowledge, this thesis is the first study about science center location problem in the literature.
Assignment of afad warehouses to container ports in Turkey
(Çankaya Üniversitesi, 2017) Kılınç, Mehmet Can
The assignment of AFAD warehouses to container ports is a vital topic, but it has not been studied thoroughly in the literature. Turkey is a special case for using maritime transportation in humanitarian logistics, because of the geopolitical location and vulnerability of our country. The main objective of this thesis is to investigate the use of maritime transportation in humanitarian logistics to respond natural disasters effectively for Turkey via the assignment of AFAD warehouses to container ports. In this thesis, a mathematical model for assigning 25 Prime Ministry Disaster and Emergency Management Authority (AFAD in Turkish) logistics warehouses to suitable ports in Turkey is developed. The capabilities of ports to handle humanitarian logistics activities, ports' closeness to AFAD logistics warehouses, suitability of access using different transportation modes and capacities of AFAD logistics warehouses are some of the criteria that is considered in this model. Assignments are also analyzed by taking into consideration of container ports' railway connections availability. The developed approach provides an alternative solution to humanitarian operations in Turkey.

Browse

Browsing Endüstri Mühendisliği Bölümü Tezleri by Issue Date