Supply_Chain_6.ppt

Supply Chain Management--Suqian Zha * Forecasting Supply Chain Requirements * Forecasting Supply Chain Requirements (1) Nature of Forecasting Spatial versus Temporal Demand Lumpy versus Regular demand Derived versus Independent Demand Time Sales Average sales Actual Sales A random,or level ,demand pattern with no trend or seasonal elements Sales Average sales Actual Sales A random demand pattern with an increasing trend but no seasonal elements A random demand pattern with both trend and seasonal elements Sales Average sales Smoothed trend and seasonal sales Actual Sales Time Demand level Example of a Lumpy Demand Patten Useful Techniques for logisticians Exponential Smoothing For short-term forecasting. Simple, requires a minimum amount of data The most accurate among competing New forecast=a(actual demand)+(1-a)(previous forecast) a: weighting factor,between 0 and 1 For convenience,we can write this model as Ft+1=aAt+(1-a)Ft t=current time period a=exponential smoothing constant At=demand at period t Ft=forecast for period t Ft+1=forecast for period t+1 How to identify a? Self judgement. Typically range from 0.01 to 0.3 The higher the value,the greater is the weight placed on the more recent demand. The lower the value, the greater is the weight placed on forecasting demand. Example: The following quarterly data represent a demand time series for a product: We wish to forecast the demand for the third quarter of this year. Assume that a=0.2 and the previous forecast is constructed from the average for the four quarters of last year Quarter 1 2 3 4 Last year 1200 700 900 1100 This year 1400 1000 F3=? Correcting for Trend (P164) St+1=aAt+(1-a)(St+Tt) Tt+1=β (St+1 -St)+(1-β )Tt Ft+1=St+1 + Tt+1 Ft+1= trend-corrected forecast for period t St= initial forecast for period t Tt= trend for period t β=trend smoothing constant 王牌卡车运输公司必