城市物流风险

KyotoUniversityGlobalCOE“HumanSecurityEngineering”IncorporatingrisksincitylogisticsProfessorEiichiTaniguchiKyotoUniversity1KyotoUniversityGlobalCOE“HumanSecurityEngineering”Outline1Conceptsofcitylogistics2Classofrisksrelatedtocitylogistics3Vehicleroutingandschedulingproblemswithtimewindows(VRPTW)4VRPTWwithuncertaintyoftraveltimes5Multi-agentsimulationwithambiguity6ConclusionKyotoUniversityGlobalCOE“HumanSecurityEngineering”Source:AnnualStatisticalReportonAutomobileTransport:InformationandResearchDepartment,PolicyBureau,MinistryofLand,InfrastructureandTransportTruckloadingefficiency45.650.047.649.347.146.242.8Rateofloadingefficiency25.035.045.055.065.075.019751977197919811983198519871989199119931995199719992001（％）TotalcommercialandprivateCommercialusePrivateuse3KyotoUniversityGlobalCOE“HumanSecurityEngineering”•Citylogisticsistheprocessfortotallyoptimisingthelogisticsandtransportactivitiesbyprivatecompanieswiththesupportofadvancedinformationsystemsinurbanareasconsideringthetrafficenvironment,itscongestion,safetyandenergysavingswithintheframeworkofamarketeconomy(Taniguchietal.2001)101ConceptsofcitylogisticsKyotoUniversityGlobalCOE“HumanSecurityEngineering”5SustainabilityLiveabilityMobilityGlobalcompetitivenessEfficiency　Environmentfriendliness　Congestionalleviation　Safety　Energyconservation　Labourforce　Security　StructureofvisionsforcitylogisticsResilienceKyotoUniversityGlobalCOE“HumanSecurityEngineering”2ClassofrisksrelatedtocitylogisticsCongestionCrashHazardousmaterialtransportFloodEarthquakeBushfireSnowfallFrequencycomplexityUncertaintyAmbiguityDay-todayRareVRPTWVRPTW-DVRPTW-PMulti-agentmodel3Vehicleroutingandschedulingproblemswithtimewindows(VRPTW)2Depotk=1,2,…K671083954VehicleRoutingandschedulingProblemswithTimeWindows(VRPTW)isdefinedastofindtheminimumcostroutesforkvehiclestoservicealltheclients.Constraints:Avehiclecannotservemoreclientsthanitscapacity.Deliveryateachclientmustbewithinsomepre-definedtimewindows.[2–3pm][4–7pm][7–8pm][10–12am][1–5pm][10–11am][10–11am][6–7pm][3–4pm]7KyotoUniversityGlobalCOE“HumanSecurityEngineering”VRPTWVariantsHardTimeWindowsDeliveryisnotpossibleoutsidethespecifiedTimeWindows(VRPHTW).SoftTimeWindowsDeliveryispossibleoutsidethespecifiedTimeWindowswithpenalties(VRPSTW).TimePenaltycost[]aibi∞∞HardTimeWindows(HTW)Inexactsolutiontechniques,waitingisallowedatnocostTime[]]aibibi’PenaltycostSimilarlyallowingwaitingwithoutcostinSoftTimeWindowsresultsin:SemiSoftTimeWindows(SSTW)KyotoUniversityGlobalCOE“HumanSecurityEngineering”ObjectivesCostofDeliveryWaitingTimeEnvironmentImpactsNOxCO2SPMDevelopanexactsolutionapproachforVRPSSTWUsingexactsolutiontechniquesforbothVRPHTWandVRPSSTW,tocomparetheirrelativecharacteristicswithrespecttoUsingsomepracticallogisticsproblemsbasedonrealroadnetworksothatabovementionedparameterscanbecalculatedonthebasisofactualtrafficcharacteristics:Traveltime,TravelSpeedAdetailedlinkbasedenvironmentalcomparisonidentifyingemissionintensitiesoneachlinkproducedbytheVRPHTWandVRPSSTWsolutionsKyotoUniversityGlobalCOE“HumanSecurityEngineering”VRPTWFormulation,},1,0{,,,,1,1,0k,1k,,1subjecttomin11),(KkAjiXKkVibSaKkAjiMXStSKkXKkChXXKXKQXqCiXXckijikiikijkijkjijkiVikiVjVjkhjkihVjkjVjkijCiiVjkijKkAjikijijKk(1)(2)(3)(4)(5)(6)(7)(8)(9)Kohletal.,1999.KyotoUniversityGlobalCOE“HumanSecurityEngineering”ExactSolutionTechnique:ColumnGenerationColumngenerationorDantzig-Wolfedecomposition,decomposestheVRPTWproblem(1–9)into:*Elementaryshortestpathproblemwithresourceconstraints(ESPPRC)(3–9).(Sub-Problem)*Setpartitioningproblem(MasterProblem).PastResearch:VRPHTWNewAlgorithm:VRPSSTWNewsubproblem:Elementaryshortestpathproblemwithresourceconstraintsandlatearrivalpenalties(ESPPRCLAP).NewLabelingAlgorithmisdeveloped.Time[]]aibibi’∞Penaltycostc’ij=cij,ifsj≤bjcij+cl(sj-bj)ifsjbjLateArrivalPenalty11KyotoUniversityGlobalCOE“HumanSecurityEngineering”ExactSolutionTechnique:ColumnGenerationSubProblemESPPRCLAPFeasibleroutesofnegativereducedcostReduceCostcij-πiMasterProblemOptimize,Prices(πi)YesIntegerSolutionNoYesEndNoBranch&BoundPpYCiYaYcpPppipPppp}1,0{,1subjecttominMasterProblemSetPartitioningLP12KyotoUniversityGlobalCOE“HumanSecurityEngineering”UpperandLowerBound29002950300030503100315032003250330033503400020406080100120Col.Gen.IterationObj.functionvalueUpperBoundLowerBoundVRPSSTWSolutionCol.Gen.stopsbutsolutionisnotintegerBranchingonNumberofVehiclesCol.Gen.stopsbutsolutionisnotinteger:BranchingonxijCol.Gen.stopswithintegerOptimumSolutionKyotoUniversityGlobalCOE“HumanSecurityEngineering”PracticalTestInstanceTestinstanceonTokyoRoadNetworkTD1_39_djkContainsasingledepotand38customers’locationsofachainofconveniencestoreDepotKyotoUniversityGlobalCOE“HumanSecurityEngineering”RoutesVRPHTWCaseVRPSSTWCaseKyotoUniversityGlobalCOE“HumanSecurityEngineering”TimePenaltycost[]aibi∞ComparisonsbetweenVRPHTWandVRPSSTWTime[]]aibibi’∞PenaltycostTotalCost0100002000030000400005000060000VRPSSTWVRPHTWCost(yen)Travel+VehiclecostPenaltycost16.3%TotalCost