供应链外包的DEA和风险价值模型2

SupplyChainOutsourcinginEnterpriseRiskManagement:ADEAVaRModel–DeshengDashWu•UniversityofToronto•ReykjavikUniversity–RiskLab–[e-mail]ominanceandStochasticDEAforVendorEvaluation.IntJofProductionResearch.2007(1).Nov,2008Callforpaper•Computers&OperationsResearch(SCI/EI,Impactfactor1.147)www.elsevier.com/authored_subject_sections/S03/S03_cfp/CAOR_CfP_DashWuRiskManagement.pdf–“ORinRiskManagement”–Duedate:March31,2009–Guesteditor:DeshengDashWu,DavidL.OlsonandJohnBirge–Approximatedateforfinalsubmissionofacceptedmanuscripts:Nov,2009Outline•Introduction•EnterpriseRiskManagement(ERM全面风险管理)•SupplyChainOutsourcing,VendorEvaluation–Contribution:•ERMstepsinSupplyChainOutsourcingRisk•Dataenvelopmentanalysis(DEA)+ValueatRisk(VaR):Intuition•ConclusionsandFutureResearch(银行链,金融危机)ReviewofRiskManagementTools风险管理工具介绍•RiskManagementtools–mean-varianceframeworkofportfoliotheoryi.e.,selectionanddiversification(Markowitz1952)–CapitalAssetPricingModel(Sharpe1964;Lintner1965;Mossin1966)–ArbitragePricingTheory(Ross,1976)–Optionpricingtheory(Black1972;Black1973)–ValueatRisk(VaR),RiskMetrics(Jorion1997)•Prob{1dayLoss≤VaR}=1-α•Min{VaR︱P(VaR)≥α}–EnterpriseRiskManagement•Professionalorganization,Consultant,Ratingagency,Academics•31%adoptedERMinCanadianrisk&insurance[Kleffner2003]•WhyERM?ToyotaReviewofRiskManagementTools(cont.)VariousRisks:$MeasurementDefinitionofERM•Systematic,integratedapproach–Manageallrisksfacingorganization•External–Economic(market-price,demandchange)–Financial(insurance,currencyexchange)–Political/Legal–Technological•Internal–Humanerror–Fraud–Systemsfailure–DisruptedproductionStochasticORModelsforRiskManagement(Beneda[2005],Dash&Kajiji[2005]))•Multiplecriteriaanalysis–Subjective•Simulation–Probabilistic;Canbesubjective(systemdynamics)•Dataenvelopmentanalysis(DEA)–Optimization•Objective,subjective,probabilisticERMResearchandStepsStep1:Determinethecorporation’sobjectives•Step2:Identifytheriskfactors,exposures•Step3:Quantifythefactors,exposures–Assesstheimpact•Step4:Examinealternativeriskmanagementtools•Step5:Selectappropriateriskmanagementapproach•Step6:ImplementandmonitorprogramMorethan80frameworks:problem-oriented,descriptive,frameworksSpecificERM:SupplyChainOutsourcingRiskSupplierManufacturerRetailerEndcustomerWarehouseASupplyChainModelSupplyChainVendorSelection•SupplyChainVendorSelectiongoodsinputbads(risk,uncertainty?)(risk,uncertainty?)•Efficiency=output/inputSupplierPerformanceDataEnvelopmentAnalysis(DEA)-Deterministic{Charnes,Cooper,Rhodes}•nVendors(DMUs)tobeevaluated.•mdifferentinputsXij,sdifferentoutputsYrj.kTkTXVYUMax1..jTjTXVYUtsnj,,2,10,0TTVUThedeterministicDEAmodeljTjTmiijisrrjrjXVYUxvyuE11nj,,2,1DEAefficiencyforDMUj:DeterministicDEA(cont.),TTmax..10,0TkkTkTTwvYstXvYXvCCRMultiplierformDEAVaR-StochasticmodelmaxE()s.t.Pr()1,=1,2N0TkvTjjjTvyvyjvβj:aspirationlevel;αj:riskcriterion;0≤αj,βj≤1Intuition:1)Atwhatconfidencelevel,itisefficienttoselectthe?thVendor?2)Atwhatconfidencelevel,itisenoughtoreducethe?thcostinordertomakethe?thVendorefficient?(1)[1,,]kNStochasticDEAAssumingmultivariatenormaldistribution:1maxE()s.t.(1)0TkvTjjjjTvyvyVarv(2)Equivalentlinearprogramming:1maxs.t.((1)),1,,,0TkvTjjjjTvyvbysjNvs(3)MetricsinVendorSelection{Olson&Wu}CriteriaNumberofstudiesusingPrice/cost12Acceptance/quality12On-timeresponse/logistics12R&Dintechnology/innovation/design7Productionfacilities/assets6Flexibility/agility6Service4Management&organization2DataSetMoskowitz,Tang&Lam,2000,DecisionSciences31,327-360•9vendorsVj–Mean–Standarddeviation–Normallydistributed•12CriteriaeachwithweightWi1.Qualitypersonnel2.Qualityprocedure3.Concernforquality4.Companyhistory5.Price-quality6.Actualprice7.Financialability8.Technicalperformance9.Deliveryhistory10.Technicalassistance11.Productioncapability12.ManufacturingequipmentSampledatademonstrationCriteriaV1V2V3V4V5V6V7V8V91Qualitypersonnel85(5.2)82(4.2)90(3.1)78(12.8)95(1.5)75(2.9)90(1.7)70(12.2)75(2.8)2Qualityprocedure80(3.3)88(4.2)85(5.1)90(4.2)75(5.6)82(2.2)82(4.2)90(33)78(3.8)3Deliveryhistory80(4.7)83(5.5)70(5.5)75(14.3)85(5.8)85(1.9)75(5.9)90(2.4)90(1.1)4Companyhistory90(5.5)88(4.5)75(7.0)85(5.6)70(5.6)80(4.1)80(4.6)85(4.5)82(3.7)SimulatedweightsandParameterSensitivity•Equalweights–Usefultoidentifydominatedsolutions•{V20.03,V40.08,V60.36,V80.53}•Ordinalweights–Reflectdecisionmakerpreference•Moreusefultomakedecision:selectnondominatedsolutions•Usedcentroidweights{Olson&Dorai}–{V20.71,V40.22,V60.07,V80}•Adjustedriskcriterion0≤αj≤1•AdjustedRHSswithβjDEAefficiencyscores:equalweight%V1V2V3V4V5V6V7V8V9AverageV195.4094.3393.5894.6275.1195.3395.1694.3289.7291.95V293.5695.6094.6395.0279.3793.9394.5392.1590.0292.09V394.9885.1795.3792.2794.8388.5592.9694.7192.2592.34V489.6190.2895.9398.1189.2393.8894.3297.4594.8893.74V585.8683.0191.0495.6398.1083.6488.8091.0886.1689.26V692.6992.8792.8492.1192.3293.4793.2988.4692.8692.32V79