资源描述
2002 Georgia TechInternet-basedData Envelopment Analysisfor Warehousing 2002 Georgia TechOutlineThe problemThe current solutionThe needA new solutionHow it worksInternet deployment and results to dateFuture directions 2002 Georgia TechPerformance AssessmentHow well are you performing?Do you have opportunities to improve?2002 Georgia TechWarehouse OperationsReceiving Function(inbound)UnloadInspectPut AwayStorageFunctionShipping Function(order fulfillment)LoadPackOrder PickStorageFunctionStorageFunction 2002 Georgia TechSingle Factor Productivity MetricsProductivity=outputinput 2002 Georgia TechTraditional Performance MetricsFill rateInventory turnsLines/hourOrders/hour$/line$/order 2002 Georgia TechWork ok whenRequirements are not changingTechnology is not changingCompetition is not changingIts very hard to interpret a single factor productivity metric when the environment is subject to rapid change in products,customer requirements,technology,or competition.2002 Georgia TechBut in a dynamic world Cant compare over timeCant compare across locationsCant compare to other companiesAt least not without a lot of additional explanatory data and information!2002 Georgia TechBenchmarkingRelative performance levelBest(effective)practices 2002 Georgia TechINPUTSOUTPUTSSystem-oriented Performance MeasureThe NeedResourcesServicesActivities 2002 Georgia TechTotal Factor Productivity?Cant solve the pricing problem 2002 Georgia TechINPUTSOUTPUTSONEPERFORMANCE INDEXData Envelopment AnalysisResourcesServicesActivities 2002 Georgia TechSystem-based assessment methodResources:capital,labor,overheadActivities:inbound,order fulfillmentServices:lines/qty shipped,fill rate,etc 2002 Georgia TechCompare to other warehousesAll other warehousesAll other warehouses in your industryAll other warehouses in your companyYour warehouse in the past 2002 Georgia TechProduction Function Theoryfor one input,one outputResource/InputProduction/Output 2002 Georgia TechSystem Efficiency ConceptResource/InputProduction/OutputOBASystem efficiency of warehouse B is the ratioOAOB 2002 Georgia TechDEA Model:Charnes,Cooper,and RhodesConstant Returns to Scale 2002 Georgia TechData Envelopment AnalysisAllows us to consider multiple“inputs”Allows us to consider multiple“outputs”Determines the reference point on the production function by constructing a hypothetical“best practices”warehouse using real warehouse dataBest possible*not average*from data 2002 Georgia TechDEA Performance ScoreContribution to Profit 2002 Georgia TechInput/Output Specification(the Frazelle/Hackman model)EfficiencyWarehouseLines ShippedStorage FunctionAccumulationTotal StaffingEquipment“Replacement”CostWarehouse area 2002 Georgia TechHtmldocumentsSolverDatabaseAt your siteGT ServerWeb-based ToolOver the internet 2002 Georgia TechOver 150 qualified users 2002 Georgia TechResults to Date 2002 Georgia TechExperienceExisting databaseMore than 150 warehousesNot segmented by industry(yet)No“descriptive”data to use for segmentingCan segment based on inputs and outputs 2002 Georgia Tech 2002 Georgia Tech 2002 Georgia Tech 2002 Georgia Tech 2002 Georgia Tech 2002 Georgia Tech 2002 Georgia Tech 2002 Georgia Tech 2002 Georgia Tech 2002 Georgia Tech 2002 Georgia Tech 2002 Georgia TechOutput Segmentationbroken case:49full case:32pallet:13mix:65total:159 2002 Georgia TechInput-oriented,all 159 together 2002 Georgia TechBroken Case,Input EfficiencyCompared Within(49/49)2002 Georgia TechPick Rate for Broken Case PickingAve=17SD=27 2002 Georgia TechFull case,Input Efficiency Compared Within(32/32)2002 Georgia TechPick Rate for Case PickingAve=14SD=27.7 2002 Georgia TechPallet,Input Efficiency Compared within(13/13)2002 Georgia TechPick Rate for Pallet PickingLines/Labor hour(pallet)02468100.010.020.030.040.050.0100.0150.0200.0Morelines/labor hourFrequencyAve=25SD=27.7 2002 Georgia TechMixed,Input Efficiency Compared Within(65/65)2002 Georgia TechPick Rate for Mixed PickingLines/Labor Hour(mix)01020304050600.0020.0040.0060.0080.00100.00lines/labor hourFrequencyAve=10.6SD=23 2002 Georgia TechAggregate Pick Rate for All 159 2002 Georgia TechWhere do we go from here?2002 Georgia TechMany Opportunities to Improve the Benchmarking ToolEnhance the basic input/output modelEnhance the ability to benchmark for technology,practice,&requirements 2002 Georgia TechSome Suggested MetricsInputsSpaceCapitalLaborInventory#of skusturnsOutputsInboundreceipt mixreceipt variabilityreturnstime to availabilityFulfillmentpick volumepick variabilitypick accuracyfill rate but Sorta 2002 Georgia Tech“Marker”Analysis 2002 Georgia TechPerformance“Marker”AttributeDEA Performance Score 2002 Georgia TechPerformance“Marker”PracticeDEA Performance Score 2002 Georgia TechResultsBigger is not always better,at least with regard to equipment and labor.There is,however,some evidence that more warehouse space leads to better system efficiency.Labor hours was not found to be a significant factor,by itself,in predicting system efficiency.However,the interaction of labor with investment was found to be significant in the sense that labor hours mitigates the effect of investment(in other words,though high investment warehouses tended to be less efficient than low investment warehouses,the differences becomes less prominent the higher the labor hours).2002 Georgia TechMore ResultsThe interaction of investment and area was found to be significant.This means that high investment warehouses are even less efficient if they are also large.No matter how we segment the data,a very large proportion of warehouses are operating at or below 50%system efficiency.While this may reflect seasonal fluctuations in customer orders,it still represents a very significant opportunity for improvement.2002 Georgia TechMore ResultsThe opportunity for improvement seems largest for the segment of warehouses doing predominantly full case picking.In that segment,a smaller proportion of the warehouses are efficient than in any other segment,and a larger proportion are operating below 50%efficiency.2002 Georgia TechOther Data RequirementsType(wholesale,retail,manufacturing)Industry(pharma,auto,electronics,)Order volumeSku spreadOrder size distributionPallet/case/each distributionPlanning horizonDegree of automationPush vs PullPicking strategy Sorta 2002 Georgia TechStill More Data RequirementsPracticesWMS?Compliant shipping?Space utilization?Velocity based slottingetcDiagnosticscost/carton shippedvalue added servicesinventory accuracyetc 2002 Georgia TechQuestions?
展开阅读全文