Development of demand forecasting model for automotive electric component remanufacturing

Abstract

Developing a reliable forecasting process is a crucial step for optimization of the overall planning process of product remanufacturing. This study examined the effectiveness of demand forecasting in remanufacturing by time series analysis (Holt-Winters model), product lifetime model (Weibull distribution), and incorporation of the two methods. To verify the effectiveness, the actual data of the time series of the sales of remanufactured alternators of an independent remanufacturer was used. For the forecasting over a year, the results provided average errors of 35.3% for Holt-Winters model, 42.2% for Weibull distribution, and 29.3% for the incorporated model. The results indicate the forecasting accuracy can improve by appropriately incorporating different methods. The results, implications, and future steps are discussed.