Distributon planning of a closed-loop supply chain with product recovery

Increasing population, climate changes, and rapid consuption of natural resources are the factors that led organizations to reduce sources and to decrease costs. Today's highly competitive environment, increasing advancements in technology, growing customers environmental consciousness and legislations force organizations to produce environmentally friendly products, and to concentrate on waste management and recovery approaches of products, components, and materials. All these factors require to examine reverse logistic activities, to plan reverse flows of products, components, and materials and to manage effectively and efficiently forward and reverse logistic activities. One of the important topics in reverse logistics is recovery approaches. There are different approaches which vary depending on the organization operate and even on the product, component and material. Therefore, choosing the appropriate recovery approach is both complex and important topic for the organizations. The researches have highlighted that consideration of all these issues and planning of forward and reverse flows of products help organizations for decreasing costs, reducing resource consumptions while increasing productivity and proficiency, as well as protecting environment. Therefore, effective management of both forward and reverse flows allows organizations to gain competitive advantage in the market. For all these reasons, closed-loop supply chain and reverse logistics have been one of the critical and popular topics drawing attentions of researchers and practitioners for more than two decades. In the literature some of the studies concentrate only on the aspect of reverse flows while some other integrate both the forward and the reverse flows. In this dissertation, closed-loop supply chain integating both the forward and the reverse flows of the products is examined. By conducting a comprehensive literature review, papers with regard to closed-loop supply chain and reverse logistics are analyzed and classified depending on the problem they handle. The literature review points out that the studies mainly focus on some of the topics such as network design, inventory control, life-cycle analysis, and routing problems. However, this dissertation concentrates on production and distribution planning problem of a closed-loop supply chain for recoverable products by considering forward and reverse flows of products. The review also highlights that there are few studies made on reuse which is one of the product recovery options. However, a large amount of studies are conducted on particularly recycling and remanufacturing recovery options. In addition, there are some studies integrating couple of product recovery options such as recycling and repairing, recycling and remanufacturing while there are a few number of studies conducted for reusable products and/or items. The review also indicates that that there are only some studies using fuzzy logic methodology for recoverable products. However, researchers emphasize that the fuzzy logic and fuzzy modelling provide efficient solutions for handling uncertainties, vaguness and lack of information in the real life applications. For this reasons, fuzzy closed-loop supply chain models are reviewed in the literature by considering uncertainties and lack of information in the closed-loop supply chain and reverse logistics topics. In the dissertation, the fundamental concepts on fuzzy sets, extension principle, fuzzy numbers and fuzzy arithmetic operations are defined. Then, fuzzy mathematical modelling is analyzed in two different ways in which the previous deals with the fuzziness in the objectives and in the constraints while the last considers the fuzziness in the parameters (technological coefficients and right hand side values in the constraints). By considering all these information stated above, in this dissertation, production and distribution planning problem for a closed-loop supply chain is analyzed under fuzzy environment. In this pursuit, a various mathematical models are developed for bringing solutions to production and distribution planning problems of different organizations. For each model, graphical illustrations marking forward and reverse flows of the products are given in the dissertation to make the model more understandable. Also, indices, variables and parameters are defined and matematical models are formulated. The developed models are solved for numerical examples to check if the models are running. The results obtained from the solutions are presented using graphics. The results show that all the developed models run. This study systematically analyzing the closed loop supply chain production and distribution problem from a broad perspective illustrates the application to the real life industrial problem. In the application part, the developed models are revised for the real-life industrial problem. All the necessary parameters are collected from the factory operating in Istanbul. The factory prefered to change proportionally the real data. The model considering multiple reusable products is composed of a single producer, multiple distributors and multiple regional warehouses as well as a large number of customers that demand the reusable products from the factory. The model is conducted for the city of Istanbul and constructed for a single time period. A varity of objectives are evaluated in the model. One of the objectives having an economical perspective is to minimized total costs of a closed-loop supply chain. The second objective having enviromental perspective is to minimize carbon dioxide emissions occuring during forward and reverse flows of the products. This objective considers all the forward and the reverse transportation activities between a factory and distributors, factory and regional warehouses, distributors and customers in the model. This is one of the important objectives since carbon dioxide is playing essential role as green house gas. Additionally, carbon dioxide emission is also one of the popular and important topics that researchers and practitionars work on. The last objective deals with minimizing total number of products that can not be collected back from the customers. The objective is essential especially for the organizations that reuse products or components. The reduction in lost products help organization to reduce consumption of sources, to protect environment and to reduce costs such as procurement costs, and transportation costs. Thus, this objective is also essential for the organizations. On the other hand, in real life applications, there are uncertainties, vaguness and lack of information with regard to closed-loop supply chain and reverse logistics. One of the uncertainties are with regard to return rate of the products. The other uncertainty is on quality of the returned products or items. There are also uncertainties related to timing of returning items and even demands of the reusable products. To deal with the uncertainties, in this dissertation the developed multiobjective mixed-integer linear programming model is solved by using the concepts of fuzzy sets and fuzzy modelling. In the application of the proposed model, the uncertainties associated to the problem are determined by reviewing the literature and by consulting to the factory management. The uncertainties are evaluated in two different aspects. Firstly, all the mentioned objectives and demand constraint are taken as fuzzy. Demand constraint is specifically chosen in the analysis because it affects many important decisions that the managers should make on such as inventory, production and logistics. For the solution of the first model, weighted additive approach is used for converting the fuzzy multiobjective model into crisp linear mathematical model having a single objective. This approach is prefered for the solution of the problem since the approach is easy to understand and easy to conduct. In addition, it is also easy to apply and it gives effective solutions. In the approach, weights obtained from fuzzy Analytic Hierarchy Analysis are assigned to the fuzzy objectives and fuzzy constraint. Secondly, some of the important model parameters such as product return rate and product demand, are also evaluated under fuzzy environment. In this approach, there is no weight attained to the objectives and constraints. The model aims to reach a single satisfaction level for all the uncertainties and maximize this level. For solving both of the single objective crisp mathematical models, an optimization software called GAMS (general algebraic modelling system) 24.04 with CPLEX solver is used. Both of the models are solved in couple of seconds and optimal solutions are obtained for both of them. The solutions of both models given in the application section demonstrate different results. Even though demands of the customers are all satisfied, there are some differences in the results related to the distributors that the factory have to send the products, the amount of products should be sent to distributors by the factory, and the distributors should serve to the customers. In addition, the values calculated for each objective also changed. For instance, according to the first model total cost, total CO2 emissions and total lost products are calculated as 20.316.460 TL, 39,6 tonne and 668.540 products, respectively. On the other hand, total costs, total CO2 emissions and total lost products are obtained as 22.552.820 TL, 47,5 tonne and 675.580 products, respectively. In this context, the verification and validation analysis are conducted. First, distributors are selected randomly and the capacity of them are assigned as "0". Then, the distances between factory and distributors are increased around 10.000 km. Thus, the model is tested if it gives rational results. The results of these analysis show that the model does not choose the distributor having the capacity of zero and choose the distributor which is far away from the factory. Sensitivity analysis is also performed to check if there is a change in the solutions when the parameters are changed. In this analysis, the weights of the fuzzy objectives and fuzzy constraints are systematically changed. Starting from the first objective, the weights are changed in the range of [0,10;1.0] with an incremental increase of 0,10 while the rest of the weights are changed proportionally. The results are illustrated by figures. Lastly, the discussions, conclusions and suggestions for the further studies are represented.