Algorithm for an efficient material requirements planning of consumable supplies : results of an empirical analysis of german supply data

الملخص EN

The present paper aims at describing a standardized algorithm, with which the consumption of tie-in products can be forecasted with different quantitative forecasting techniques.

Among the techniques are AR-, MA-, ARMA-, ARIMA- and structural regression models.

By means of algorithmic procedures, an optimal forecasting model of performance criteria will be evaluated, which will be used for requirements planning.

For all determined forecasting horizons, the ARIMA-model proved to be the best model followed by simple Moving Average models.

The importance of autoregressive procedures decreases with the length of the forecasting horizon.

Structural approaches seldom prove to be the best forecasting models, even if the importance of these increase according to the length of the forecasting period.

The algorithm allows for a considerable part of tie-in products a good forecasting quality.

Considering elaborated premises the algorithmic procedure simplifies the forecast of the outflow of goods for a considerable part of consumable supplies.