毕业论文(设计) 外文翻译 外文原文: Evaluation of forecasting error measurements and techniques for intermittent demand Peter Wallstroma, Anders Segerstedt 1. Introduction The most adopted technique for short-term forecasting is probably the single exponential smoothing (SES) from Brown(1959, 1962). The initial advantages were low requirement of memory storage and the putational effort appropriate for yesterdays puter capacity. Today SES is still a widespread method in puter systems and forecasting programs. However, the ability for SES to forecast slowmoving items or an intermittent demand, when for an item the forecasting time periods often have zero demand, has been questioned. Croston (1972) presented a method that separates the forecasts in two parts; in time between withdrawals or demand and demand size. The forecasts are updated only when there is a demand. The usefulness of Croston’s method is verified by . by Willemain et al (1994). Leve ´n and Segerstedt (2004) suggested a modification of the Croston method where a demand rate is directly calculated when a demand has happened. Syntetos and Boylan (2005) mended an adjustment of the Croston method due to a systematic error notified by Syntetos and Boylan(2001). It is necessary to examine and evaluate the performances of the forecasts. Silver et al (1998) point out; ‘‘no single measure is universally best’’. Still, evaluations
