Belirsizlik altında heterojen filo ve zaman pencereli rotalama problemi: Hızlı tüketim sektöründe bir uygulama
Heterogeneous vehicle routing with time windows under uncertainty: Implementation in fast moving goods industry
- Tez No: 510576
- Danışmanlar: PROF. DR. FERHAN ÇEBİ
- Tez Türü: Yüksek Lisans
- Konular: Endüstri ve Endüstri Mühendisliği, Industrial and Industrial Engineering
- Anahtar Kelimeler: Belirtilmemiş.
- Yıl: 2018
- Dil: Türkçe
- Üniversite: İstanbul Teknik Üniversitesi
- Enstitü: Fen Bilimleri Enstitüsü
- Ana Bilim Dalı: İşletme Mühendisliği Ana Bilim Dalı
- Bilim Dalı: İşletme Mühendisliği Bilim Dalı
- Sayfa Sayısı: 95
Özet
Günümüz artan rekabet koşullarında firmalar sürekli gelişerek değişen koşullara uyum sağlamaya ve pazardaki paylarını arttırmaya çalışmaktadırlar. Pazarda bir çok alternatifin bulunduğu günümüzde, firmalar rekabet güçlerini ellerinde bulundurabilmek için yüksek kalite ve çeşitlilikte olan ürünlerini düşük maliyet ve yüksek servis hızı ile müşterilene ulaştırmak zorundadırlar.Ancak müşteri memnuniyetini arttırmak için hızlı ve esnek yapılar kurulurken, karlılıkların da kontrol altında tutulabilmesi önemlidir.Bu sebeple firmalar sürekli olarak maliyetlerini düşürme hedefindedirler.Etkin bir tedarik zinciri yönetimi, maliyetleri dengelemek için etkili bir araçtır. Müşterilere gereken ürünleri istenilen miktar ve zamanda ulaştırırken bir yandan da tedarikçilerle etkin bir biçimde çalışarak ham maddeleri istenilen zamanda ve düşük maliyetlerle tedarik etmek gerekmektedir.Lojistik maliyetleri işletmelerin toplam maliyetlerinin %6-20'sine karşılık gelmektedir. Bu sebeple lojistik yönetiminin en önemli faaliyetlerinden biri olan dağıtım alanında yapılan iyileştirmeler firmaların karlılıklarına önemli katkılar sağlamaktadır. Araç rotalama problemleri dağıtım yönetiminin merkezini oluşturmaktadır.Araç rotalama problemlerinde talebin yapısı, araçların çeşitliliği, taşınacak malzemelerin tipi, dağıtım/toplama noktalarının lokasyonları, kapasite ve zaman gibi girdilere göre araçların müşterilere/tedarikçilere atamasını gerçekleştirerek, süreçlerdeki verimliliklerin arttırılıp, lojistik maliyetlerinin düşürülmesini amaçlanmaktadır. Yapılan bu çalışma, bir firmanın ihtiyaçlarından yola çıkılarak üretilmiştir. İşletmenin lojistik maliyetlerini düşürebilmek için ham madde tedarik ağı planlaması yapılmıştır.İncelenen işletmenin depolama, zaman gibi kısıtları ve maliyet düşürme amacı gözönüne alınarak problem bir araç rotalama problemine dönüştürülmüştür.Problem yapısının daha iyi anlaşılabilmesi için öncelikle araç rotalama problemlerinin türleri ve litaratürde kullanılmakta olan çözüm yöntemlerinden bahsedilmiştir.Tezin daha sonraki kısımlarında ise, problemin amaçlarında bazı belirsizlikler olması nedeni ile bulanık kümeler ve bulanık programlama üzerinde durulmuş, tez kapsamında geliştirilen çok amaçlı karma tam sayılı doğrusal programlama modelinin çözümünde bulanık küme ve bulanık modellemeden yararlanılmıştır.Bulanık çok amaçlı modelin geleneksel tek amaçlı doğrusal programlama modeline dönüştürülmesinde arttırılmış max-min modeli kullanılmıştır. Modelin çözümünde GAMS 24.1 solver'ından yararlanılmış ve firma yetkililerinin belirlediği başarılma derecelerine göre en iyi çözümler elde edilmeye çalışılmıştır.Son olarak elde edilen sonuçlara göre firmanın kazanımları tartışılmış ve gelecek çalışmalar için önerilerde bulunulmuştur.
Özet (Çeviri)
In today's increasing competitive conditions, firms try to adapt to changing conditions and increase their market share with constantly evolving. Nowadays, there are many alternatives in the market because of that companies must deliver high quality and diversity products to customers with low cost and high service speed in order to maintain their competitive power. However, while increasing customer satisfaction with having a quick and flexible structure, also profitability need to be kept under control. Therefore, firms need to constantly aim to reduce their costs. An efficient supply chain management is an effective tool for balancing costs. While delivering the required products to customers in the desired quantity and time, it is necessary to work with suppliers efficiently to supply raw materials at the desired time with low costs. Supply chain does not contains only manufacturers, suppliers and distributors. It is also contain warehouse management, carriers and customers. In recent years, firms realized that they can not gain competitive advantage in the market with only established their marketing strategies.They should design effective supply chain network to sustain their market share. Logistics costs correspond to between six and twenty percent of the total cost of companies. Distribution is one of the most important activity of logistic management. For this reason, improvements in distribution area provide significant contributions to the profitability of companies. Vehicle routing problems constitute the center of distribution management. According to inputs such as the structure of demand, diversity of vehicles, the type of materials, locations of distribution and collection points, capacity and time, the assignment of vehicles to customers or suppliers is done to increase the efficiency of processes and reduce logistics costs. In general, the main objectives of vehicle routing problems are as follows: Minimizing the total route cost, minimizing the costs of vehicles (fixed costs or variable costs), minimizing the total route time, minimizing the number of vehicles, minimizing routes times, if there are customers who cannot meet the their demand on time or in full, minimizing the total penalty cost for these customers. In addition, some studies in the literature have been used often workload balancing as a second goal of vehicle routing problems. The first study in vehicle routing problems was done by Dantzing and Ramser in 1959. This study aims to provide an optimal routing plan with using homogeneous vehicle fleets. Five years after this study, Clarke and Wright extended this study as optimisation problem and created today's vehicle routing problem. However, with the increase of the complex structures of today's firms, presented models in 1960s are insufficient to respond to company needs. For this reason, for solving the real life problems heuristic and meta- heuristics approaches have been developed. In many methods developed for vehicle routing problems, the problem parameters are are considered to be known, precise, and unchanging data. But in real life,especially in the transportation industry there are many uncertainties, and these uncertainties have made the problems very complicated. Using the exact solution methods for such problems and trying to achieve the best solution make the problems more complicated by making the solution impossible. Therefore, considering today's competitive conditions, instead of wasting unnecessary money and time for the best solution, searching for the best solution for the best will bring more viable solutions. Heuristic and Meteheuristic methods aim to find the closest solution to the best to minimize the time spent solving the problem. On the other hand, in real life problems there are uncertainties due to lack of information. For the solutions of this kind of vague, probability theory and statistics have been used for many years. Both methods use a logic that relies on the truth in suggestions, meaning that a proposition is either true or false. However, expressions may not always be clear. They can be partially true or partially wrong and this cause to fuzzy environments. If information is not mathematically precise and clear, solution of the classical methods does not provide results. For this reason Zadeh was developed fuzzy logic and fuzzy set theory. According to Zadeh, if we have more information about the system, complexity in the problems is reduced. On the other hand, as the complexity increases, definite and meaningful expressions can not be established and uncertainty occurs. Zadeh celled this situation as fuzzy. The developed model under this study is provided by the needs of an industrial organization which is operating in the fast moving consumer goods industry. Due to there is no enough storage areas in the plant, the company has rented another warehouse. But due to the high cost of renting a warehouse and material movements between the main depot and the rental warehouse, significant increase in production expenses was realized. For this reason, efforts have been initiated to reduce transportation and storage costs by planning and logistics departments. In this study, it was determined that, the 53% of raw materials which hold in the rental warehouse are belong to only eight local suppliers. For this reason, to reduce shipping and storage costs, vehicle routing plan was designed for raw material collection network. With this new design, assigned trucks collect the goods and delivery the goods to directly to plant. The trucks should leave from the plant and should return back to the factory by collecting raw materials from suppliers. Purchased products should be delivered directly to the main warehouse in the factory without being stocked in the rental warehouse. The first goal to be achieved with the solution of the problem is minimizing the total route cost. In addition, completion material collection operations in the minimum time is important for the company. The manufacturer is working with a carrier company for raw material procurement. This company has different types of trucks. For this reason, total processing times have been added to the model as a second goal and the problem has been transformed to multi-objective mixed integer linear programming model with heterogeneous fleet. Due to each supplier has defined time zones for material loading, time windows were added the models.In addition that firm has two different satisfaction degrees for goals of models, and this situation cause the uncertainty. To eliminate uncertainties, multi-objective mixed-integer linear programming model is solved by fuzzy modelling. For the solution of the model, augmented max-min model was used to transform fuzzy objective model to the single objective function model. Augmented model was preferred due to give more balanced solutions. GAMS 24.1 solver was used for the solution of the model. According to the results obtained, products from eight suppliers can be taken with a total of four vehicles. Total cost of routes was calculated as 3026 TL and all vehicles return to factory until 14 o'clock so the factory can be used same trucks for internal movements again. Because the products were directly placed in the production site without being stored in the rental warehouse, thirty percent profit was achieved in total cost due to the reduction in storage and internal transports. In this master thesis, the data of a real problem was used and the solution was calculated according to this data. But some assumptions were used for simplify the problem. As an example, the changes in demand are not considered as part of the thesis. In real life applications there will be cases where demand is uncertain. Therefore, it is suggested that the model can be expanded for cases where the demand is uncertain. In addition that, If the number of suppliers and vehicles in the model increases, the solution of the model will take a long with exact solution methods. For this reason heuristic and meta-heuristics approaches can be used in future studies.
Benzer Tezler
- Talep belirsizliği altında kapasite kısıtlı yer seçimi ve araç rotalama problemi için hibrit sezgisel bir çözüm önerisi
A hybrid heuristic solution proposal for capacitated location routing problem under demand uncertainty
ENGİN PEKEL
Doktora
Türkçe
2018
Endüstri ve Endüstri MühendisliğiYıldız Teknik ÜniversitesiEndüstri Mühendisliği Ana Bilim Dalı
DOÇ. DR. SELİN SONER KARA
- Contextual multi-armed bandits with structured payoffs
Yapılandırılmış getirili bağlamsal çok kollu haydutlar
MUHAMMAD ANJUM QURESHI
Doktora
İngilizce
2020
Elektrik ve Elektronik Mühendisliğiİhsan Doğramacı Bilkent ÜniversitesiElektrik-Elektronik Mühendisliği Ana Bilim Dalı
DR. ÖĞR. ÜYESİ CEM TEKİN
- Heterogeneity and strategic sophistication in multi-agent reinforcement learning
Çoklu etmen pekiştirmeli öğrenmede heterojenlik ve stratejik karmaşıklık
YÜKSEL ARSLANTAŞ
Yüksek Lisans
İngilizce
2024
Elektrik ve Elektronik Mühendisliğiİhsan Doğramacı Bilkent ÜniversitesiElektrik ve Elektronik Mühendisliği Ana Bilim Dalı
DR. ÖĞR. ÜYESİ MUHAMMED ÖMER SAYIN
- Deterministic and stochastic models for practical scheduling problems
Uygulamalı çizelgeleme problemleri için deterministik ve stokastik modeller
ELVİN ÇOBAN GÖKTÜRK
Doktora
İngilizce
2012
Endüstri ve Endüstri MühendisliğiCarnegie Mellon Universityİşletme Yönetimi Ana Bilim Dalı
PROF. DR. ALAN SCHELLER-WOLF
- Expectation heterogeneity and wealth inequality
Beklentilerin heterojenliği ve servet eşitsizliği
FURKAN SARIKAYA
Yüksek Lisans
İngilizce
2019
EkonomiSabancı ÜniversitesiEkonomi Ana Bilim Dalı
DR. ÖĞR. ÜYESİ REMZİ KAYGUSUZ