İki aşamalı soğuk tedarik zincirinde nsga-II algoritma uygulaması = Nsga-İİ application in a two echelon cold supply Chain

Abstract

İki aşamalı soğuk tedarik zinciri; dondurulmuş gıda, taze/yeşil ürün veya ilaç gibi dikkat gerektiren ürünlerin taşınması, depolanması ve dağıtılması aşamalarını içeren bir sistemdir. Üzerinde çalışılan iki aşamalı soğuk tedarik zincirinde üç temel eleman mevcuttur. Bu elemanlar; bir tedarikçi, bir depo ve birden fazla müşteridir. İki aşamalı soğuk tedarik zincirlerinde birçok ürünün tedarik planlaması yapılabilmektedir. Çalışmada kalitesi aşamalı olarak düşen taze/yeşil bir ürünün taşınacağı kabul edilmiştir. İki aşamalı soğuk tedarik zincirinde; maliyet en küçüklemesi, ürün kalitesi optimizasyonu, ürün taşıma koşulları optimizasyonu, ürün depolama koşulları optimizasyonu, vb. gibi birden fazla amaç fonksiyonu belirlenebilmektedir. Birden fazla amaç fonksiyonun optimize edilebilmesi için kurulmuş olan matematiksel modelde maliyetleri ve ürün yaşlarını minimize eden iki ayrı amaç fonksiyonu oluşturulmuştur. Literatürde özellikle geleneksel çözüm yöntemleri birden fazla amaç fonksiyonu içeren matematiksel modellerin çözümünde iyi bir performans göstermemektedir. Bu nedenle iki aşamalı soğuk tedarik zinciri için kurulan matematiksel modelin çözümü için; Genetik Algoritma geliştirilerek oluşturulan Baskın Olmayan Sıralama Genetik Algoritması II (NSGA-II) kullanılmıştır. Belirli talep altında, 7 günlük periyot için yapılacak olan planlama NSGA-II ile optimum planlamaya ulaşılmaya çalışılmıştır. Yapılan çalışma sonucunda elde edilen sonuçlar; NSGA-II'nin iki aşamalı soğuk tedarik zincirindeki rotalama performansının literatürde yer alan geleneksel çözümlere göre yüksek olduğu ve önemli ölçüde optimum sonuca ulaşılmasını sağladığını göstermiştir.

With the increasing consumption and technological advancements, changes occur in the structure of supply chains. As a result of these changes, various problems arise at different points in supply chain management. To solve these problems, Supply Chain Management has been examined in more detail, and new solution methods have been developed. This study aims to work on a two-stage cold supply chain with a supplier and a warehouse, under a specific demand, selling a product in the market. The fresh/green product obtained from the supplier will be stored in a single warehouse before being distributed to customers. It is assumed that the quality of the product transported through the supply chain gradually decreases. Therefore, it is intended that the storage time of fresh/green products in the interim warehouse be kept short. To minimize the expected cost and achieve the best solution, a mathematical model has been proposed. The Non-dominated Sorting Genetic Algorithm II (NSGA-II), one of the metaheuristic algorithms, was used to solve the mathematical model. The results obtained from the solution are shown in the study, and recommendations are made based on the obtained results. The Cold Supply Chain is a type of supply chain that ensures the preservation of products that need to be kept within a specific temperature range from production to consumer delivery. The Cold Supply Chain involves maintaining and sustaining the appropriate temperature conditions throughout the process of production, storage, distribution, and consumption of perishable products such as frozen food, fresh/green produce, pharmaceuticals, and other perishables. The cold supply chain is critically important for ensuring food safety, extending product shelf life, and reducing costs. In this supply chain, products need to be transported under specific temperature and other conditions. A two-stage cold supply chain can be defined as a supply chain model that involves the management and transportation of products through two distinct stages. In this type of supply chain, elements such as suppliers, cold storage warehouses, distribution/transportation vehicles, and customers are involved. The two-stage cold supply chain is a system that includes the transportation, storage, and distribution of products requiring special attention, such as frozen food, fresh/green produce, or pharmaceuticals. The examined two-stage cold supply chain consists of three main elements: a supplier, a warehouse, and multiple customers. In such a system, supply planning can be conducted for many products. In this study, it is assumed that a fresh/green product with gradually decreasing quality will be transported. In a two-stage cold supply chain, various objective functions can be determined, such as cost minimization, product quality optimization, optimization of product transportation conditions, and optimization of product storage conditions. To optimize multiple objectives, two separate objective functions that minimize costs and product age are included in the mathematical model. Traditional solution methods in the literature generally perform inadequately in solving mathematical models with multiple objective functions. Therefore, the Non-Dominated Sorting Genetic Algorithm II (NSGA-II), developed from the Genetic Algorithm, is used to solve the mathematical model for the two-stage cold supply chain. Planning for a 7-day period under certain demand is attempted to be optimized using NSGA-II. In the two-stage cold supply chain, the supplier is considered singular as it collects green products from producers. The fresh/green product obtained from the supplier will be stored in a single warehouse before distribution to the customers. Customers are located within a circular area with a radius of 25 km centered around the warehouse. The distance of each customer to the warehouse is within this circle. The demands of all customers are determined to follow a normal distribution. The interval between two consecutive delivery periods to each customer will be 2 periods. At the same time, multiple vehicles cannot deliver to a customer in the same period. It is assumed that the quality of the product transported within the supply chain gradually decreases. To determine the quality criterion of each product, the age of each product will be present. Products are considered to be zero years old when delivered to the warehouse. After a product is delivered to the warehouse, its age increases by 1 for each period. If the product age in the warehouse exceeds 33% of the expiration date, the related products will not be delivered to the customers and will be removed from the warehouse inventory. The main objectives of the proposed mathematical model are to: - Minimize routing and inventory costs, - Minimize the number of vehicle uses, - Minimize the number of spoiled fresh/green products. A mathematical model is one of the engineering methods used for solving problems. It also serves as a basis for other solution methods by representing the problem. The mathematical model developed in this study is based on the optimization model established in the 2019 study "A Two-echelon Inventory Routing Problem for Perishable Products" by S.U.K Rohmer et al. A new objective function has been added to the referenced mathematical model. Constraints are crucial for reflecting the current conditions and assumptions in the model to achieve optimal results with the defined objective functions. A total of 18 constraints have been determined for this study. Under these constraints, the solutions to the objective functions and the optimal transportation and storage plan in the supply chain will be achieved. NSGA-II is a metaheuristic algorithm derived from the Genetic Algorithm and is commonly used to solve complex problems. While the Genetic Algorithm is widely used, NSGA-II has been made more effective with additions to address certain shortcomings. The NSGA-II algorithm was introduced to the literature by Deb et al. in their 2002 study as a multi-objective metaheuristic algorithm. It emerged as a result of the enhancement of the NSGA algorithm, which Srinivas and Deb developed in 1995. The basic structure of the NSGA-II algorithm is based on the Genetic Algorithm (GA). Steps for dominance ranking and crowding distance calculation have been included in the fundamental steps of the Genetic Algorithm. This study examines the effectiveness of NSGA-II in solving the two-stage cold supply chain problem, demonstrating its superior performance in cost and routing optimization compared to traditional methods. NSGA-II effectively minimizes both overall costs and costs associated with spoiled products. The study results show that NSGA-II significantly outperforms traditional solutions in minimizing costs and optimizing routing in the two-stage cold supply chain, providing substantially optimal results. To prevent fresh/green products that have reached a certain age from being delivered to customers, a new objective function and constraints were added to the mathematical model. This addition, solved using NSGA-II, successfully minimized the cost of spoiled products and eliminated waste. Products that have reached a certain age can be repurposed in various ways: - Discounted sales - Donations - Production of processed goods: For example, older fruits can be used to make jam or juice. - Production of compost and animal feed - Energy production: Fresh/green products that have aged or spoiled can be used in biogas plants to generate energy, thus contributing to energy production while repurposing the products. Future studies can expand on this model by exploring how aged products can be utilized in different fields, adding a new level to the model. This enhancement could help plan how aged or spoiled products contribute to the system. The model could be reevaluated using NSGA-II or other algorithms, or by integrating artificial intelligence for optimal solutions. The dynamic nature of the algorithm developed in this thesis allows it to be applied to various two-stage supply chain problems in future academic studies. The problem and solution model presented in this thesis can also serve as a multi-objective optimization problem in academic research. The prepared study consists of five main sections. Section 1 includes the subject, aim, and importance of the study. Section 2 covers applications of the two-stage cold supply chain model, methods, the Non-dominated Sorting Genetic Algorithm II (NSGA-II), and literature examples related to the supply chain. Section 3 provides details about the two-stage cold supply chain model used in the study and the NSGA-II applied. Section 4 presents the application and results of the NSGA-II model. In the final section, Section 5, conclusions and recommendations regarding the study's results are summarized, and future studies are discussed.