Article ID: | iaor20131889 |
Volume: | 203 |
Issue: | 1 |
Start Page Number: | 255 |
End Page Number: | 277 |
Publication Date: | Mar 2013 |
Journal: | Annals of Operations Research |
Authors: | Aytac Berrin, Wu S |
Keywords: | supply & supply chains, forecasting: applications |
Most technology companies are experiencing highly volatile markets with increasingly short product life cycles due to rapid technological innovation and market competition. Current supply‐demand planning systems remain ineffective in capturing short life‐cycle nature of the products and high volatility in the markets. In this study, we propose an alternative demand‐characterization approach that models life‐cycle demand projections and incorporates advanced demand signals from leading‐indicator products through a Bayesian update. The proposed approach describes life‐cycle demand in scenarios and provides a means to reducing the variability in demand scenarios via leading‐indicator products. Computational testing on real‐world data sets from three semiconductor manufacturing companies suggests that the proposed approach is effective in capturing the life‐cycle patterns of the products and the early demand signals and is capable of reducing the uncertainty in the demand forecasts by more than 20%.