Optimization of material requirements planning in automotive glass production

The study includes the design and application of MRP for the materials supplied from outside by a company which operating as a glass manufacturer in the automotive supply industry in Turkey. With this study, MINLP optimization model has been created to calculate the order quantity of materials and stock quantities to be held. With the optimization model; it is aimed minimizing of the total cost consists order cost, holding cost and non-holding cost. ABC analysis was performed to 78 materials in the product tree of the products that were requested in 2019 for the selection of materials used in the model. As a result of ABC analysis, 7 windshield materials in A class have been determined to be used in a mathematical model. The model which was created by considering the firm unit costs and constraints, was expressed mathematically in a closed format in the LINGO program. The solution report of the model was examined and its economic interpretation was made. As a result of the solution of the model, the total cost for 7 materials was obtained as 1.434.046 TL. As an alternative to the installed model, FOQ, EOQ, LFL, FPQ, POQ, MEOQ and WW order size determination methods were used, order sizes and total costs of the materials were calculated. In addition to the study, for the materials used in the mathematical model, the order information given by the firm's MRP planning team in 2019 was taken from the company and the total costs of the orders placed according to the firm unit costs were calculated. As a result; the model, methods and the total costs of the orders given intuitively were compared. LFL, MEOQ and POQ total costs, which are among the order determination methods, yielded 30% lower than the total model cost. However, it was determined that these 3 models do not have stocks and it was concluded that they are not applicable in the automotive industry. When the total costs resulting of the model solution and the WW method were analyzed, it was obtained that the total cost of the model solution was lower for 4 materials with an average of 6% difference, the total cost of the WW method was lower for 3 materials with an average of 2% difference. When total costs of the model solution and FOQ, EOQ and FPQ total costs, which are among the order determination methods, were compared the model solution, on average, at 64%, 22% and 46%, respectively, calculated lower costs. When the total cost of the model solution and the orders given intuitively are compared, it was obtained that the total cost of the model solution is 75% less than the total cost of the orders given intuitively.