چكيده به لاتين
Abstract:
Spare part usage, due to high costs, storage and such expenses, is one of the important factors to consider in any industrial factory. To lower these costs, there’s a tendency among companies to analyze and forecast their spare part demands and its use in the future. Since demand characteristics are lumpy, it is not an easy phenomenal to forecast.
In this study a forecasting method will be described to forecast spare part demands with lumpy characteristics. The proposed method is a combination of Grey GM (1,1) model and a back-propagation neural network (BPNN). Grey model is used as the main forecasting core and then the neural network is used to fine tune the model. To validate the proposed combined forecasting method, the results are compared separately with a few other forecasting methods. To compare the methods, parameters such as Mean Error, Mean Squared Errors and mean absolute percentage error have been utilized.
Final results represent some improvements in A-MAPE parameter compared to the most accurate forecasting used before.
