demandplanning（供应链高级计划排程课件）.ppt

《demandplanning（供应链高级计划排程课件）.ppt》由会员上传分享，免费在线阅读，更多相关内容在教育资源-天天文库。

1、Demandplanning2016-4MichaelWagnerIntroductionWhyisdemandplanningnecessary?LargebenefitswhichareachievedbySupplyChainManagementareaccreditedtothereductionofinventories,esp.tothedecrementofsafetystocksWhatisthepurposeofdemandplanning?ThepurposeofDemandPlanningistoimprovedecisionsaffectingdemandaccura

2、cyandthecalculationofbufferorsafetystockstoreachapredefinedservicelevel.Whatismainobstacles?processuncertainty(e.g.unreliableproductionprocesses,fluctuatinglead-timesetc.)demanduncertainty(differenceinplannedorestimateddemandandactualsales).CONTENTSFirstpartSecondpartThirdpartLastpartAdemandplannin

3、gframeworkStatisticalforecastingtechniquesIncorporationofjudgmentalfactorsAdditionalfeaturesAdemandplanningframeworkAdemandplanningframeworkproductdimension:product→productgroup→productfamily→productline;geographicdimension:customer→salesregion→DCregion/location;timedimension:differentbucketsize(da

4、ys→weeks→years)andhorizon.CONTENTSFirstpartSecondpartThirdpartLastpartAdemandplanningframeworkStatisticalforecastingtechniquesIncorporationofjudgmentalfactorsAdditionalfeaturesStatisticalforecastingtechniquestime-series-analysiscausalmodelsTwobasicapproachesStatisticalforecastingtechniquesStatistic

5、alforecastingtechniquestime-series-analysiscausalmodelsTwobasicapproachesStatisticalforecastingtechniques1MovingAverageandSmoothingMethods2RegressionAnalysis3ARIMA/Box-Jenkins-methodThreemostfrequentlyusedforecastingmethodsStatisticalforecastingtechniques1MovingAverageTheparameterestimatefortheleve

6、liscalculatedbyaveragingthepastndemandobservations.SmoothingMethodsTheneedtocutthetime-seriesisavoidedbytheexponentialsmoothingmethod,becauseitassignsdifferentweightstoallobserveddemanddataandincorporatesthemintotheforecastStatisticalforecastingtechniques2RegressionAnalysisWheresignificantinfluence

7、ofsomeknownfactorsispresent,itseemstobestraightforwardtousecausalmodelsintheforecastingprocess.Regressionanalysisisthestandardmethodforestimationofparametervaluesincausalmodels.Statisticalforecastingtechniq