demandforecasting需求预测

demandforecasting需求预测内容摘要：

Seasonal Component (季节因子 ) Repeating up amp。 down movements (重复的上下运动 ) Due to interactions of factors influencing economy (因影响经济的各种因素相互作用 ) Nonannual。 multiyear。 (多年模式的 ) Mo., Qtr., Yr. Response Cycle Components of Demand Cyclical Component (周期因子 ) Erratic, unsystematic, ‘residual’ fluctuations (无规律的、非系统的“残值”波动 ) Unexplained portion (无法解释的部分 ) Components of Demand Random Component (随机因子 )  Set of evenly spaced numerical data (一组均匀分布的数值 )  Obtained by observing response variable at regular time periods (一段时间内观察所得的数据 )  Forecast based only on past values (基于过去数据进行预测 )  Assumes that factors influencing past, present, amp。 future will continue  Example  Year: 1993 1994 1995 1996 1997  Sales: Quantitative Methods What is a Time Series? (什么是时间序列 )  Used if demand is not growing nor declining rapidly(需求模式比较平滑 )  Used often for smoothing (用于平滑 )  Remove random fluctuations  Equation where: Ft = forecast for period t (预测需求 ), n = number of periods to be averaged (具体期数 ) At1 = actual demand realized in the past period for up to n periods (真实需求 ) Time Series Analysis Simple Moving Average (简单移动平均 ) nA+...+A +A +A =F nt3t2t1ttLet’s develop 3week moving average forecasts for demand. (3周需求的移动平均预测 )  Assume you only have 3 weeks of actual demand data for the respective forecasts (只有三周的真实数据 ) Time Series Analysis Simple Moving Average (简单移动平均 ) A c t u a l F or e c a s tW e e k D e m a n d 3 W e e k1 6502 6783 7204 5Allows different weights to be assigned to past observations (对历史数据赋予相应权重 )  Older data usually less important (越久越不重要 ) Weights based on experience, trialanderror(权重靠经验 ) Equation (公式 )...... Time Series Analysis Weighted Moving Average (加权移动平均 ) F = w A + w A + w A +. . . + w At 1 t 1 2 t 2 3 t 3 n t n1=wn1=ttwt = weight given to time period “t” occurrence (weights must add to one) (权重之和为 1。 ) Determine the 3period weighted moving average forecast for period 4. (计算第4期的加权平均预测 ) Weights: t1 t2 t3 A c t u a l F or e c a s tW e e k D e m a n d 3 W e e k1 6502 6783 72045Time Series Analysis Weighted Moving Average (加权移动平均 ) Increasing n makes forecast less sensitive to changes (预测周期越长结果越不敏感 ) Do not forecast trends well (不能很好预测趋势 ) Require sufficient historical data (需要大量的历史数据 ) Time Series Analysis Disadvantages of . Methods (MA方法的不足 ) To “ramp” changes of demand (有斜面数据时 ) Demand High weight n = 2 65 55 45 35 Ramp Shift 3 2 1 T +1 2 3 4 5 6 7 8 Low weight n = 6 Time Series Analysis Responsiveness of . Methods (MA方法的敏感度 ) u Forecast lags with increasing demand, and leads with decreasing demand (上升时滞后，下降时领先 )  Premise (前提条件 )The most recent observations might have the highest predictive value (越近信息量越大 ).  Therefore, we should give more weight to the more recent time periods when forecasting (给最近的观察数据更多权重 )  Requires smoothing constant () (需要平滑系数为常数 )  Ranges from 0 to 1 (位于 0到 1之间 )  Subjectively chosen (选择有艺术 )  Involves little record keeping of past data (不需要保留很多历史数据 ) Time Series Analysis Exponential Smoothing (指数平滑法 ) Time Series Analysis Exponential Smoothing (指数平滑法 ) The equation used to pute the forecast is (公式 )... Ft = Ft1 + (At1 Ft1) = A t1) + (1 ) Ft1 where.... Ft = forecast demand (预测的需求 ) At = actual demand realized (实际的需求 )  = smoothing constant (平衡系数 ) Exponential because each increment in the past is decreased by (1  ) (对过去信息的利用以 (1  )的比率下降 ): Time Series Analysis Exponential Smoothing (指数平滑法 )  Determine exponential smoothing forecasts for periods 210 using  = (Let F1 = D1) A c t u a l F or e c a s tW e e k D e m a n d 1 650234Time Series Analysis Exponential Smoothing (指数平滑法 )  Determine exponential smoothing forecasts for periods 210 using  = (Let F1 = D1) A。