物流网络节点的动态规划选址研究

上海交通大学硕士学位论文物流网络节点的动态规划选址研究姓名：牟伦英申请学位级别：硕士专业：企业管理指导教师：黄丹2005010120021-ak0ak1a1a2a3an2002220023LOGISTICNETWORKNODESLOCATIONANALYSISUSINGDYNAMICPROGRAMMINGMETHODABSTRACTThelogisticnetworkplanningisthecoreofstrategiclogisticplanning.Logisticnetworknodeslocationmainlydecidesthestructureandscaleofthewholelogisticnetwork,anddeeplyaffectsthelogisticexpensesandservicelevelofthelogisticsystem.Logisticnodeslocationisakindoflong-termstrategicdecision.Soinpresentrapidlychangingeconomyenvironment,logisticnetworknodalpointslocationisadynamicdecisionproblem.Yet,duetocomplexityofsolvingdynamiclocationmodelwiththetimefactor,mostexistinglocationmodelsareusingstaticmethods.Basedonthereviewofformerresearchresults,thisthesisputsforwarddynamicprogramminglocationmethod.Thedynamicprogramminglocationanalysismethoddiscussedhereisanextensionofthecommonstaticlocationanalysismethods,whichprovidebothcostdataandalternativelocationsforthedynamicanalysis.Thedynamicprogrammingtechniqueservesasamathematicaltoolforfindingalogisticnodeslocation-relocationplanthatwillyieldminimumcumulativecostinagivenplanningperiod.Inordertoimprovetheaccuracyofforecasteddata,thecost-for-all-subsequent-yearsfactorisweightedbyanindexbetweenzeroandone.Usually,thelongerperiodtheforecastcovers;itwillbemoredifficulttogivingtheaccurateforecast.Hence,theweightingfactorwilldecreaseasthetimespanincreasing.20024Thepreconditionofusingthedynamicprogramminglocationmethodisthatlogisticfacilitiesshouldbeallowedtomoveduringtheplanningperiod.Whenthesavedcostbyrelocatingismuchmorethanthecostofrelocating,itshouldconsiderrelocation.What’smore,whenleasingthirdpartlogisticfacilities,itonlyneedtopayrentandnofixedinvestment,soitwillmoreeasilytorelocatewiththelowercostcomparewithself-supportcooperation.Underthiscondition,itismorenecessarytorelocatethelogisticnodesFinally,basedonthedataofthedemandforecastandshippingcostchangeforfuturefiveyearswhichprovidedbyanautomobileinteriorCo.Ltd.,thethesislocatedtransferwarehousesforeachyearwiththeminimumcost.Andthedynamicprogramminglocationmethodputforwardinthisthesisisvalidated.Fortheprospectofstrategicplanningandpermanenceoflocationdecision,thisthesisintroducedtimefactorsinthelocationmodel,andvalidatedthedynamicprogramminglocationmodel.Itisdevelopmentalforfurtherresearchofdynamiclocationanddirectivetosimilarcaseinapplication.KEYWORDS:Logisticnetworknodalpoints,Location,Static,Dynamic,Dynamicprogramming200211.1200221.22090[1][1][2]20023200241.3200252.12.1.1()20026[3][4]iiiiMinTCVRd=∑21TCViiRiidiidi()()22iiidKXXYY=---22,XY,iiXYK2221TCXYiiiiiiiiiVRXdXVRd∑=∑23iiiiiiiiiVRYdYVRd∑=∑241200272diiiiiiiiVRXXVR∑=∑25iiiiiiiVRYYVR∑=∑263252622di4di2324,XY5,XY22di645,XY123420028()[5]00111111111111minpppmqqnmnlikliklkjlkjlijlijlllikkjijqmpkkklikkilFCXCYCZFWCX============∑∑∑=++∑∑∑∑∑∑⎛⎞∑++⎜⎟∑∑⎜⎟⎝⎠11qnliklijlikjXZa==∑∑+≤l12pi12m2811qmlkjlijljkiYZb==∑∑+≥l12pj12n29272002911mnliklkjijXY==∑∑=l12pk12q210110pmlikkliXMW==∑-≤∑k12q21110plkjkjkjlYEI=∑-≥k12qj12n21210plkjkjlYMI=∑-≤k12qj12n21311pmliklikXV==∑≤∑k12q21411pmliklikXV==∑≥∑k12q21510kkW⎧⎪=⎨⎪⎩k216kjkj10kjI⎧⎪=⎨⎪⎩217,,0liklkjlijXYZ≥ll12pii12mjj12nkk12qXlikki200210YlkjjkZlijjiClikkiClkjjkClijjiWkkFkkCkkalilbljlMEkjjkIkjkjkVkkVkXlikYlkjZlij0-1WkIjkXlikYlkjkWkIkjkjWk0XlikYlkj0Ikj0Wk1Ikj011kjEkjP-P-medianApproachP-2002112.1.2H.J.HeinzCompanyNestléCompany[4]40004010-151ab200212cd2abc3ab4552-1200213Figure2-1SimulationflowchartofwarehouselocationdevelopedbyH.J.HeinzCompany2002142.1.3NP12345120021523456[6]1exchangeheuristicP-P-median2greedyheuristic3dropheuristicnmnnm4relaxheuristicP-P-medianP-P-center0-12002165approximationheuristicP-P-centerP-P-medianCavalierSherali6sequentialheuristicBaumol-wolfeKuehn-HamburgerBaumolwolfeBaumol-wolfe---KuehnHamburgermainprogrambumpandshift200217mn602.1.4[7]1220021834200219AHPAHPT.L.Saaty2070AHP[8]AHP2-2Figure2-2Evaluationindexsystemoflogisticnetworknodeslocation2002202.2FranciscoSaldanhadaGamaMariaEugeniaCaptivo/[9]EmanuelMelachrinoudisHokeyMin[10]JohnCurrent[11][12]2002212002223.1[12]JohnCurrent1997[11]///R.Bellman3.32002233.22002243.3501951R.BellmanN-N[13]3.3.1kskkkSk200225uk(sk)kskDk(sk)kskuk(sk)un(sn)kkpk,n(sk)k1p1,n(s1)skukk+1sk+131()1,kkkksTsu+=31Vk,n()1,,,,,kknknknVVsus+=L32,fk(sk)kskn(){}(),,k,1f,,,knuukknkknsVsusopt+=LL33optoptimizationminmax2002263.3.2skukDk(sk)()1,kkkksTsu+=Vk,nuk(sk)*1,np3.4nnnnABC20rnABC{ABC}cnnA-B-C-nnnn2002271nkk12n2nskskABC3skuk(sk)uk(sk)ABC4sk+1uk(sk)5kVk,n(sk,uk,sk+1,,sn+1)kskuk(sk)fk(sk)fk(sk)MinVk(sk,uk)akfk+1(sk+1)Vk(sk,uk)k0-1aka1a2a3anfn+1(sn+1)06uk(sk)p1,np1,nu1(s1)u2(s2)u3(s3)un(sn)u1(s1)-u2(s2)-u3(s3)--un(sn)1knnn1c/(1r)+nnnn1n1sn1{ABC}n{ABC}nnAnfn(A)MinVn(A,un(A))0fn(A)un(A)nBCDfn(B)un(B)fn(C)un(C)fn(D)un(D)2002282kn1n1n2c/(1r)+n1n1n1n2sn2{ABC}n1{ABC}n1n1Afn1(A)MinVn1(A,un1(A))an1fn(sn)fn1(A)un1(A)n1BCDfn1(B)un1(B)fn1(C)un1(C)fn1(D)un1(D)3fk(sk)uk(sk)4p1,nu1(s1)u2(s2)u3(s3)un(sn)u1(s1)-u2(s2)-u3(s3)--un(sn)f1(s1)n3.5-200229ak0ak1a1a2a3an2002304.1150004-12002314-1*2005200620072008200913800203002690038500480004600040800362003370032400150000001000075