The single machine schedule problem with raw material based and sequence dependent setup time in plastic injection machines: An automotive industry application

It is possible to produce in different sizes and features with the molds attached to the injection machines where plastic production is made. In addition to its difficult and demanding production, it is also difficult to plan the products to be produced because of connecting different molds to a machine and producing them. The problem addressed in this study is the only machine scheduling problem in an automotive company with a manual sequence dependent setup time. The production plans specified in the problem are for sequencing and scheduling the productions in plastic injection machines and are based on experience. For this reason, calculating the sequence-dependent setup time caused by raw materials, sorting and scheduling is not done in a way that will give the best results. In the study, a model that aims to minimize the total delay and will give the best result has been developed and this model has been investigated on GAMS software. The study consists of two stages. In the first stage, using the data required for the plastic injection production plan, the number of clusters was determined by the elbow method in the Python program and the K-means clustering algorithm was established. In the second stage, important constraints were determined using cluster information and a mathematical model was established to give the best schedule. In this problem, which describes the real-life situation, the fact that the products have too many colors and types of raw materials and the calculation of the setup time according to these are the constraints that make the model different. Setup time has been calculated very comprehensively. This model was solved using GAMS optimization software. The results obtained were compared and interpreted with the solutions made manually in the company.