Potential inventory cost reductions using advanced time series forecasting techniques
- PDF / 626,227 Bytes
- 9 Pages / 595 x 794 pts Page_size
- 60 Downloads / 236 Views
#2001 Operational Research Society Ltd. All rights reserved. 0160-5682/01 $15.00 www.palgrave-journals.com/jors
Potential inventory cost reductions using advanced time series forecasting techniques GL Shoesmith and JP Pinder* Wake Forest University, Winston-Salem, NC, USA This paper compares demand forecasts computed using the time series forecasting techniques of vector autoregression (VAR) and Bayesian VAR (BVAR) with forecasts computed using exponential smoothing and seasonal decomposition. These forecasts for three demand data series were used to determine three inventory management policies for each time series. The inventory costs associated with each of these policies were used as a further basis for comparison of the forecasting techniques. The results show that the BVAR technique, which uses mixed estimation, is particularly useful in reducing inventory costs in cases where the limited historical data offer little useful information for forecasting. The BVAR technique was effective in improving forecast accuracy and reducing inventory costs in two of the three cases tested. In the third case, unrestricted VAR and exponential smoothing produced the lowest experimental forecast errors and computed inventory costs. Furthermore, this research illustrates that improvements in demand forecasting can provide better cost reductions than relying on stochastic inventory models to provide cost reductions. Keywords: forecasting; inventory; regression; statistics; time series
Introduction 1
Previous research (Pinder ) shows that improvements in forecasting models that identify deterministic components result in improved inventory management policies. Thus, selection of inventory management policies is based on assumptions of stationarity and finite expectation and is dependent upon the method of forecasting. This paper demonstrates potential inventory cost reductions through the application of the well-established time series forecasting techniques of vector autoregression (VAR) and Bayesian VAR (BVAR). VAR and BVAR were first introduced in 1979 and have since proven to be practical and effective economic forecasting methodologies. BVAR models, in particular, have been shown in numerous studies to be as accurate, if not more accurate, than large structural models and other time series methodologies. Perhaps the strongest evidence of the reliability of BVAR models and their contribution to economic forecasting was offered by Litterman2 and McNees,3 in which the forecasting performance of Litterman’s original national BVAR model compared favourably to that of leading institutional forecasters such as Data Resources, Inc. (DRI) and Wharton Econometric Forecasting Associates (WEFA).
*Correspondence: JP Pinder, Babcock Graduate School of Management, Wake Forest University, 7659 Reynolda Station, Winston-Salem, NC 27109, USA. E-mail: [email protected]
Although VAR and BVAR models have typically been used in cases where long histories are available for each time series, the techniques are also appropriate when data ar
Data Loading...
