第八章预测供应链需求

8-1预测供应链需求CR(2004)PrenticeHall,Inc.Chapter8Ihopeyou'llkeepinmindthateconomicforecastingisfarfromaperfectscience.Ifrecenthistory'sanyguide,theexpertshavesomeexplainingtodoaboutwhattheytoldushadtohappenbutneverdid.RonaldReagan,19848-2产品计划三角形ProductinthePlanningTriangleCR(2004)PrenticeHall,Inc.PLANNINGORGANIZINGCONTROLLINGTransportStrategy•Transportfundamentals•TransportdecisionsCustomerservicegoals•Theproduct•Logisticsservice•Ord.proc.&info.sys.InventoryStrategy•Forecasting•Inventorydecisions•Purchasingandsupplyschedulingdecisions•Storagefundamentals•StoragedecisionsLocationStrategy•Locationdecisions•Thenetworkplanningprocess计划组织控制TransportStrategy•Transportfundamentals•TransportdecisionsCustomerservicegoals•Theproduct•Logisticsservice•Ord.proc.&info.sys.InventoryStrategy•Forecasting•Inventorydecisions•Purchasingandsupplyschedulingdecisions•Storagefundamentals•StoragedecisionsLocationStrategy•Locationdecisions•Thenetworkplanningprocess库存战略预测客户服务目标采购和供应时间决策存储基础知识存储决策产品物流服务订单管理和信息系统库存决策运输战略运输基础知识运输决策选址战略选址决策网络规划流程8-3ForecastinginInventoryStrategyCR(2004)PrenticeHall,Inc.PLANNINGORGANIZINGCONTROLLINGTransportStrategy•Transportfundamentals•TransportdecisionsCustomerservicegoals•Theproduct•Logisticsservice•Ord.proc.&info.sys.InventoryStrategy•Forecasting•Inventorydecisions•Purchasingandsupplyschedulingdecisions•Storagefundamentals•StoragedecisionsLocationStrategy•Locationdecisions•ThenetworkplanningprocessPLANNINGORGANIZINGCONTROLLINGTransportStrategy•Transportfundamentals•TransportdecisionsCustomerservicegoals•Theproduct•Logisticsservice•Ord.proc.&info.sys.InventoryStrategy•Forecasting•Inventorydecisions•Purchasingandsupplyschedulingdecisions•Storagefundamentals•StoragedecisionsLocationStrategy•Locationdecisions•Thenetworkplanningprocess8-4供应链预测什么•Demand,salesorrequirements需求，销售或请求•Purchaseprices•购买价格•Replenishmentanddeliverytimes•补给和交货时间CR(2004)PrenticeHall,Inc.8-58.1需求预测1.需求的时间和空间特征（SpatialversusTemporalDemand）2.尖峰需求和规律性的需求（LumpyversusRegularDemand）3.派生需求和独立需求（DerivedversusIndependentDemand）8-6CR(2004)PrenticeHall,Inc.典型时间序列模式TypicalTimeSeriesPatterns:随机Random0501001502002500510152025TimeSalesActualsalesAveragesales随机性或水平发展的需求，无趋势或季节性因素8-7CR(2004)PrenticeHall,Inc.典型时间序列模式TypicalTimeSeriesPatterns:随机有趋势RandomwithTrend0501001502002500510152025TimeSalesActualsalesAveragesales随机性需求，上升趋势，无季节性因素8-8CR(2004)PrenticeHall,Inc.TypicalTimeSeriesPatterns:RandomwithTrend&Seasonal0100200300400500600700800010203040TimeSalesActualsalesTrendinsalesSmoothedtrendandseasonalsales随机性需求，有趋势，季节性因素8-9CR(2004)PrenticeHall,Inc.TypicalTimeSeriesPatterns:LumpyTimeSales尖峰需求模式8-10CR(2004)PrenticeHall,Inc.8.2预测方法•1.定性方法Qualitative调查法Surveys专家系统Expertsystemsorrule-based2.历史映射法（时间序列分析Historicalprojection）移动平均Movingaverage指数平滑Exponentialsmoothing•3.因果或联想法Causalorassociative回归分析Regressionanalysis•4.协同Collaborative8-118.3对物流管理者有用的方法8.3.1.移动平均法MovingAverageBasicformulatntiiAnMA11wherei=timeperiodt=currenttimeperiodn=lengthofmovingaverageinperiodsAi=demandinperiodiCR(2004)PrenticeHall,Inc.8-12Example3-MonthMovingAverageForecastingMonth,iDemandformonth,iTotaldemandduringpast3months3-monthmovingaverage............20120..21130360/312022110380/3126.6723140360/312024110380/3126.672513026?CR(2004)PrenticeHall,Inc.8-13加权移动平均WeightedMovingAverageperiodcurrentinforecastperiodcurrentindemandactualperiodnextforforecast0.30to0.01usuallyconstantsmoothingwhere)1(formulasmoothingexponentialonly,levelbasic,thetoreduceswhich)1(...)1()1()1(thenform,inexponentialare)(weightsIf1...1133221112211ttttttntnttttniinnFAFFAFMAAAAAAMAwwwhereAwAwAwMAaaaaaaaaaaaa8-14I.LevelonlyFt+1=aAt+(1-a)FtII.LevelandtrendSt=aAt+(1-a)(St-1+Tt-1)Tt=ß(St-St-1)+(1-ß)Tt-1Ft+1=St+TtIII.Level,trend,andseasonalitySt=a(At/It-L)+(1-a)(St-1+Tt-1)It=g(At/St)+(1-g)It-LTt=ß(St-St-1)+(1-ß)Tt-1Ft+1=(St+Tt)It-L+1whereListhetimeperiodofonefullseasonalcycle.IV.ForecasterrorMAD=|AtFNttN|1orS(AF)NFtt2t1NandSF@1.25MAD.8.3.2.指数平滑公式ExponentialSmoothingFormulasCR(2004)PrenticeHall,Inc.8-15CR(2004)PrenticeHall,Inc.ExampleExponentialSmoothingForecastingTimeseriesdata1234Lastyear12007009001100Thisyear14001000?QuarterGettingstartedAssumea=0.2.Averagefirst4quartersofdataanduseforpreviousforecast,sayFo8-16CR(2004)PrenticeHall,Inc.Example(Cont’d)Beginforecasting9754/)11009007001200(0FFirstquarterof2ndyear1000)975(8.0)1100(2.0)2.01(2.0001FAFSecondquarterof2ndyear1080)1000(8.0)1400(2.0)2.01(2.0112FAF8-17CR(2004)PrenticeHall,Inc.Example(Cont’d)Thirdquarterof2ndyear1064)1080(8.0)1000(2.0)2.01(2.0023FAFSummarizing1234Lastyear12007009001100Thisyear14001000?Fore-cast100010801064Quarter8-18CR(2004)PrenticeHall,Inc.Example(Cont’d)MeasuringforecasterrorasMAD绝对差orRMSE(std.errorofforecast)标准差nFAMADnttt1||1)(12nFASntttF8-19CR(2004)PrenticeHall,Inc.Example(Cont’d)UsingSFandassumingn=2408121080)(10001000)(140022FSNoteTocomputeareasonableaverageforSF,nshouldrangeoveratleastoneseasonalcycleinmostcases.SF=408Example(Cont’d)Rangeoftheforecast0BiasnFAn