多重博弈的Dijkstra算法快递行业航空网络应用研究.docx

多重博弈的Dijkstra算法快递行业航空网络应用研究.docx

《多重博弈的Dijkstra算法快递行业航空网络应用研究.docx》由会员分享，可在线阅读，更多相关《多重博弈的Dijkstra算法快递行业航空网络应用研究.docx（14页珍藏版）》请在冰豆网上搜索。

多重博弈的Dijkstra算法快递行业航空网络应用研究

Dijkstraalgorithmofmultiplegametheoryinexpressprofessionaviationnetworkapplicationresearch

JunminLv1XiangchaoLiu1HaoLi1

（1TianFuCollegeofSouthwesternUniversityofFinanceandEconomics,MianYang,Sichuan,621000,China）

ABSTRACT：

Withexpressindustry’sdevelopmentandcustomer’sexperiencerequirementincreased,airtransportationhasbecomeanindispensableparttoguaranteethecustomerserviceandcustomerexperience.Mostofthetraditionalresearcharebasedontheangleofaviationline‘seconomicfactors,theselectionofhub,andthemainresearchtoolsare,RobustOptimizationAlgorithm,SimulatedAnnealingAlgorithm,AntColonyAlgorithm,tosingledimensionthesemethodsarevaluable.Buttherealityisdimensional.Atpresent,severaltraditionalstudiescombinedtheGametheoryandtheShortestPathAlgorithmintheapplicationresearchofexpressindustryaviationnetwork.Thispaper,fromtheviewofexpressindustry,basedontheinternalrelationshipbetweenthetimeandcost,andcomeupwiththeMultipleGameDijkstraAlgorithmtoplantheaviationnetworkandconsiderabouttime,cost,anddistancedimensional,inordertoachievethegoalofhighefficiencyandlowcost.Thealgorithmiseasyforusedincomputerprogramming,anditprovesthatitsdominanceandfeasibility.

Keywords:

ExpressdeliveryAviationnetworkMultiplegameDijkstraalgorithm

1Introduction

IntheincreasinglyfierceExpressmarketcompetition,theairtransportation’sdegreeofadvantageefficiencyhasbecomeakeyfactorwhencomparedwiththecompetitionsanditisbecomeavitalfactorwhichcouldinfluenceexpressspeed.What’smore,it’salsoakeyfactorincustomerexperience.Domesticstudiesusuallyemphasizeparticularlyontheairlinesiteselectionandairrouteplanning,buttheydidn’tconsidertheairlinenetworklocationandairrouteplanningissuesinExpressindustry.Aviationnetworkplanningissimilartolandtransportationtosomeextent,buttheyhavemuchdifferentinessence.Thefactorsandparametersconsidereddomesticstudiesinlandtransportationrouteplanningstudyalmostcannotbeappliedinairtransportation,becausethefactorswhichneedtoconsiderinairtransportationaredifferentfromlandtransportation,sothelandtransportationlayoutmodelinthedomesticstudycouldn’tbeappliedinmodernairlinetransportation.

Inthedomesticandforeignrelevantresearch,TaoJiang,JinFuZhu（2006）[1]usedtheRobustOptimizationAlgorithmintheairlineshublocationanalysis;XiangyuanShu,MingYang,YanpingWang（2010）[2]usedtheSimulatedAnnealingAlgorithminoptimizingallocationofaviationprojectresources;HongZhou,JianxinOu,ZhendaoLi（2008）[3]studiedairfreightcenterlogisticssystembyusingsimulationmodel;FuQingDai,RuiWang（2007）[4]combinedthesinglehub-airportlocationandflightroutenetworklayoutcomprehensiveoptimizationbyusingIterativeOptimizationAlgorithm,andvalidateit;GuiJieYu,Yubingpen,YanChangChu（2006）[5]studiedtheapplicationofcomplexnetworktheoryintheairlinenetwork;MingguoBo,JinFuZhu（2006）[6]usingThreePhaseMethodinthedesignoftheaviationnetwork;JunChaoWang（2010）[7]researchedthecomplexityoftheChinaaviationnetwork;HanyiYang（2010）[8]studiedtheapplicationofnetworkstructureinChina'scivilaviationnetworkbyusingSingleAxistheoryHannula,M,Huttunen,K,Koskelo,J,Laitinen,T,Leino,T（2008）[9]comparedthedifferenceofestimatebetweenartificialneuralnetworkandMultipleLinearRegressionModelintheairlinenetwork.

Fromthepreviousstudies,itcanbeeducedthatmostresearchesismainlystudiedthecivilaviationoraviationhublocationproblem,buttheydidn’tconsiderthespecificcharacteristicofExpressaviationnetwork.Atpresent,themethodsofcivilaviationtransportationplanningaremature,andtheroutingplanningbetweencivilaviationtransportationandExpressaviationaredifferent,theycan’tbeuseduniversally.ChinaExpressindustrydevelopsrapidlyatthemoment,theurgentproblemistheantinomybetweenefficiencyandcost,howevernotraditionalstudyconsideredExpressaviationnetwork.

Inallusiontothissituation,thispapercombinedtheGameTheorywiththeShortestPathAlgorithmDijkstraalgorithm,basedontheMultipleGameDijkstraAlgorithmtoplanaviationnetwork,andoptimizeaviationnetworkpathtomaketheExpressaviationnetworkoflowcostandhighefficiency.Thealgorithmissimpletocomputerprogramming,andithasstrongapplicability.Intheend,wetakeanExpresscompanyaviationnetworkasanexample,itprovesthatitsdominanceandfeasibility.

2Modelillustration

2.1.Basicassumptionsandsymbols

Inordertodescribethemodelclearly,nowintroducethedefinitionofeachsymbolanditsmeaningasfollows:

N:

Representsthenumberofairterminals,namelyanalysisobjecthasNaviationhubscurrently,needtotransportthegoodstoNnumbersofairterminals.

Xij:

RepresentsthedemandamountsofeachairaviationhubtootheraviationhubiandjrepresentstheNumbersofeachaviationhub.

Tij;Representstheflighttimeofeachaviationhubtootheraviationhubs,thesubscriptiandjrepresentstheNumbersofeachaviationhub.

Assumethattheshortertheaircraftflighttimeis,thelessfuelconsumptionneeded,andthecostisless,andflightdistanceandflighttimeisproportional.Accordingtothereality,theamountofdemandfromeachaviationhubcomeandbackisdifferent,thebackandforthdemandbetweeneverytwohubofdemandisdifferent.

2.2.Algorithmmodelanalysis

Thealgorithmmodelregarddemandasitsfirstvariables,everyoptimalroutewillgothroughtheroutethatmeetthemaximumdemand,soweadopttheShortestPathAlgorithm,Dijkstraalgorithmtocalculatetheoptimalvalue.Beforethecalculation,weneedtoflipthedataofdemandasDijkstraisamethodtocalculatetheshortestpathalgorithm.

Dijkstraalgorithmisusedineveryairline,whenitgothroughthelineaddapointstotheline,thesameroutegoonceisonemorepoints.Thismakestherouteandtherelativedifferencebetweenroute,finallytheconclusionshowsthedifferencebetweeneachroute,thenobtainthefinalresults.

Inthisalgorithmmodelusetheflighttimebetweenaviationhubasitssecondvariables,betweeneachaviationhubhavedifferentflighttime,theshorterflighttimethelowerofthecost.CompareallaviationhubofflighttimewiththeresultDijkstraalgorithmthecometothefinalgameagain.Afterthegame,choosethemultiplesupplydemandandrelativelylowcost,thenthealgorithmisend.

3Algorithmdesign

1.Willtheaviationhubforbusinessdemand（inandoutintosum）frombigtosmallsorting,establishprocessinganumberline.Regarda0asthestartingpointofanumberline.SelectthelargestbusinessdemandfordataXijastheendofanumberline.Andwillendonehalfofdata（1/2Xij）asthemiddleofanumberlineanumberline.

2.Willmorethananumberline1/2Xijamongthedigitalrowamongtherighttoanumberline,amongthenumberoflessthananumberlinewasanumberlineontheleft.Useanumberlinebetween1/2XijtotherightoftheXijdemandrespectivelybetweenminusanumberline,getanumberX’ij.Useanumberlineamongthedigitalminus,as1/2Xij-X’ij,settoHij,thisdigitalinevitableamonglessthananumberline.Willthisnumberareamongthelefttoturnanumberline;Useanumberlinebetweenminusamongtheleftsideofthebusinessmodelrespectivelydemand（Xij）,getX’ij,useanumberlinewiththeNumbers.Itmeans1/2Xij+X’ij,settoHij.Thisdigitalinevitableamongmorethananumberline1/2Xij,willthisnumbertoaligntoflipanumberline.Amongtheright,asshowninfigure1show:

Figure1Firstdealwithanumberline

Willdealwiththedemandofanumberlinedataandturndataapart,asshowninfigure2shows:

Figure2Finaldisposalofanumberline

Throughthedataprocessing,willbebigdemandforsmallerNumbers,conversionofconvenientoperationafter.

3.Accordingtostep2dataprocessingresultsDijkstraalgorithm,usefortheshortestroute.NotofigureG=（V,E）,thelengthofthesideE[i]thateachforw[i],findthevertexV0totheotherseachpointoftheshortestpath.

Dijkstraalgorithmisdescribedbelow:

Aboutidentification,Suchasnodeidentificationfor[20,4]:

Thefirstfiguresaidfromthebeginningnodetothenodenumberfromtheseconddistance,saidthenodenodetostartonthepathofpleaseanodeintheNumbers.

Step1:

givenode1identification[0,S].0marknode1tothedistanceto0is0,Ssaidnodeisthestartingpoint.

Step2:

visitfromnodetonodeofthedirectlyto1,andgivesatemporarylogo.

Step3:

marknodedeterminefromtemporarywithminimumdistancenode,andmarkthenodeisthepermanentmarks.Ifallnodesareallpermanentidentifier,turntostep5.

Step4:

thenewpermanentmarker,thenewlogomarkingbegantoexaminethepermanentlogocannotbedirectlytopermanentmarkingnode.

Iftheinvestigationfortemporarymarknode,nodethenewlogoofpermanentmarkingnodeandthenewlogodistancevalueofpermanentmarkthatpointdirectlytothedistancefromthenodevalueaddingtogether.Ifitsandlessthanthedistancebetweentemporarilylogopoint,istodeterminetheminimumdistancevalueistheclosesttothedistancevalue.

Ifthenodeistheinvestigationofthenode,notmarkthenewlogoofpermanentmarkingnodewithnewmarkdistancevalueofpermanentmarkthatpointdirectlytothedistancefromthenodevalueaddingtogether.Andasacontingentlogotothepoint.Returntostep3.

Step5:

Thepermanentlogoisdete