抄録
Developing a reliable forecasting method enables to optimize the overall planning of production for remanufacturer. Owing to the uncertainties in timing of product returns and demands, remanufacturers perceive difficulties in forecasting and are still exploring superior forecasting methods. Prior research proposed two forecasting methods for remanufacturing, based on the Holt-winters model and the Weibull distribution model. In this study, a pattern based demand forecasting method was proposed. In this method, the defects of the Weibull distribution model were analyzed and modified by applying constraints and categorization rules. To verify the effectiveness, actual time series data of sales for remanufactured alternators of an independent remanufacturer were used. The results provided average errors between actual value and forecasted value were 35.3% for the Holt-winters model, 41.7% for the Weibull distribution, and 24.9% for the proposed model.
