research of dispatching method in elevator group control system based on traffic mode identify
2009InternationalConferenceonBusinessIntelligenceandFinancialEngineering
ResearchofDispatchingMethodinElevatorGroupControl
SystemBasedOnTrafficModeIdentify
JunWang
NanjingTelecommunicationInstituteNanjingTelecommunicationInstitute
Nanjing,Chinaintraweb@163.com
AirongYu
CombatSoftwareStaffRoom
PLAUniversityofScienceandTechnology
Nanjing,Chinayu_alice@163.com
XiaoyiZhang
NanjingTelecommunicationInstituteNanjingTelecommunicationInstitute
Nanjing,Chinazxy31412@163.com
Abstract—Elevatorgroupcontrolsystem(EGCSwithmulti-objective,stochasticandnonlinearcharacteristicsisacomplexoptimizationsystem.Afteranalyzingcharacteristicoftypicaltrafficmodeofelevator.ThispaperproposedanewsimulationplatformofanelevatorgroupcontrolsystemimplementedinC#usingthefuzzy-neuralnetworktechnology.Theresultofsimulationshowsthatthismethodrealizesreasonableelevatordispatchingundervariouspassengertrafficconditionsandindicatesthevalidityofthismethod.
LeiQu
NanjingTelecommunicationInstituteNanjingTelecommunicationInstitute
Nanjing,Chinaarebas@21cn.com
II.TRAFFICFLOWMODEL
Trafficflowhasregularityandrandomness,theregularityofthetrafficflowisrelatedtothepeopleinthebuilding;itsrandomnessaredifferentbecauseworkingeachthesametimeperiodthevolumeoftrafficthatisoneachfloorarerandomrequestforserviceafewpassengers,passengersandthepurposeofstartingfloorfloorsarerandom.Aworkingmodeoftransportcanbedividedintoup-peaktrafficpattern,thedown-peaktrafficpattern,Keywords-component;elevatorgroupcontrol
system;fuzzy-neural;simulation;trafficpatternrandominterlayertrafficpatternandfreetrafficpattern.
I.
INTRODUCTION
A.Up-peaktrafficpattern
Up-peaktrafficpatternsingeneralhappentoworkinWiththedevelopmentofsociety,growinglarge
high-risebuilding,elevatorhasbecomeanindispensablethemorningtime,thepassengersenteredtheelevatoronthemeansoftransport.Inlargebuildings,weoftenhavetouplinktothevariousfloorsofthebuildingwork,andinstallmanyelevatorsforpeopleinordertoimprovesecondly,thestrengthoftheuplinksmallerpeakoccurredtransportefficiencyandservicequality,soweneedtouseattheendoftheafternoonresttime.theelevatorgroupcontrolsystemfortheirreasonablejobB.Down-peaktrafficpatternscheduling,statusmonitoringandinternalmanagement.
Thishappenedattheeveningafterwork,thepeakafterElevatorgroupcontrolsystem[1]arethesamebuilding
manyelevatorsasawholetomanagethecontrolsystem,workismorestronglythanthepeakofmorning.Itexiststheanditspursuitofthegoalsarebasedondifferentflowweakerdown-peakatthebeginningofmid-dayrest.trafficconditions[2],selectareasonableschedulingC.Rrandominterlayertrafficpatternprogramtocoordinatetheliftoperation,sothatladderbaseThistrafficmodelisanormaltrafficconditions,traffictothemostappropriatewayofansweringthecalllayerdemandbetweenallfloorsofthebasicbalance.Itexistsinstationstaircasesignal.However,becauseofelevatorgroupmostofthetimeofday.controlsystemcontrolobjectivesofdiversity,aswellas
elevatorsysteminherentrandomnessandnon-linear,itisD.Thefreetrafficpatterndifficulttoestablishaprecisemathematicalmodel,simplyThistrafficpatternusuallyoccursintheeveningafterbytraditionalcontrolmethodsdifficulttoimprovethework,thenextmorningbeforeworkthistime,aswellascontrolperformanceofthesystem.therestoftheafternoontimeperiod.Atrestday,theday
willhavedifferentlevelsofthefreetrafficpatternexist.
978-0-7695-3705-4/09$25.00©2009IEEEDOI10.1109/BIFE.2009.20
46
A.Thepassengerwaitingtimewt.
NEURALNETWORKMODELOFELEVATORCONTROLSYSTEMWhenanewcallingsignalcomes,inlightofthefloorFuzzycontroltechniqueusingexpertknowledgetoFbwherethecallingsignalcomefromanditsdirectionDbobtainavarietyofcontrolrulescanbeagooddealoftheaswellasthefloorFcwheretheelevatorcurrentlylocates
anditsdirectionDc,thetimei.e.PWTthattheelevatorelevatorsystemofmulti-objective,stochasticandnonlinear.
coststoreachthedestinationfloorcanbecalculated.Thefuzzycontrolfunctionlacksforthefunctionofstudy;
itsrulescannotbemodifiedontheruntime.InordertoAssumingthatthetimetheelevatortakesfromthecurrentachieveadifferenttransportationmodeswitchingalgorithm,floortothedestinationfloorist1,thetimethattheelevator
stopsforpassengersatonefloorist2,thenumberofitisnecessarytoidentifythecurrenttransportmode,then
callingsignalstobeansweredisA,thefarthestfloorintheswitchtothecorrespondingalgorithm.ForElevatorTraffic
samedirectionisFmax,andthefarthestfloorinthereversePatternRecognitionusingFuzzyNeuralNetworks,
Neuralnetworkwithnon-linear,dynamiccharacteristicsdirectionisFmin.andastrongfunctionofstudy,applytosetupsimilartoIfDcisthesamewithDbandFbisinfrontofFc,elevatorgroupcontrolsystemforaclassofnonlineartheelevatorcanreachthecallingfloorinthesamedynamicsystems,butbecauseoftheelevatorgroupcontroldirection:systemisamulti-statesystem,inordertogettheoptimal(2wt=|Fb 