Estimation of processing time using machine learning and real factory data for optimization of parallel machine scheduling problem

Traditionally, mathematical optimization methods have been applied in manufacturing industries where production scheduling is one of the most important problems and being actively researched. Extant studies assume that processing times are known or follow a simple distribution. However, actual time factory often unknown likely follows complex Therefore, this study, we consider estimating using machine-learning model. Although there use machine learning for itself, it should be noted purpose study to estimate an time. Using models, can distribution while further improving schedule computed importance variable. Based on above, propose system models when data. The advantages proposed its versatility applicability real-world unknown. method was evaluated process information with each sample provided by research partner companies. Light gradient-boosted (LightGBM) algorithm Ridge performed best MAPE RMSE. parallel estimated our resulted average reduction approximately 30% makespan. On other hands, results probabilistic sampling which Kernel Density Estimation, Gamma distribution, Normal Distribution shown poorer performance than ML approaches. In addition, used deduce variables affect estimation times, demonstrated example feature from experimental