团购网站相关外文翻译学士学位论文.docx

团购网站相关外文翻译学士学位论文.docx

《团购网站相关外文翻译学士学位论文.docx》由会员分享，可在线阅读，更多相关《团购网站相关外文翻译学士学位论文.docx（15页珍藏版）》请在冰豆网上搜索。

团购网站相关外文翻译学士学位论文

外文翻译

原文1

MatchingModelsforPreference-sensitiveGroupPurchasing

Matchingbuyersandsellersisoneofthemostfundamentalproblemsineconomicsandmarketdesign.Aninterestingvariantofthematchingproblemariseswhenself-interestedbuyerscometogetherinordertoinducesellerstoofferquantityorvolumediscounts,asiscommoninbuyingconsortia,andmorerecentlyintheconsumergroupcouponingspace（e.g.,Groupon）.Weconsiderageneralmodelofthisprobleminwhichagrouporbuyingconsortiumisfacedwithvolumediscountoffersfrommultiplevendors,butgroupmembershavedistinctpreferencesfordifferentvendorofferings.Unlikesomerecentformulationsofmatchinggamesthatinvolvequantitydiscounts,thecombinationofvaryingpreferencesanddiscountscanrenderthecoreofthematchinggameempty,inboththetransferableandnontransferableutilitysense.Thus,insteadofcoalitionalstability,weproposeseveralformsofNashstabilityundervariousepistemicandtransfer/paymentassumptions.Weinvestigatethecomputationofbuyer-welfaremaximizingmatchingsandshowtheexistenceoftransfers（subsidizedprices）ofaparticularlydesirableformthatsupportstablematchings.Wealsostudyanontransferableutilitymodel,showingthatstablematchingsexist;andwedevelopavariantoftheprobleminwhichbuyersprovideasimplepreferenceorderingover“deals”ratherthanspecificvaluations—amodelthatisespeciallyattractiveintheconsumerspace—whichalsoadmitsstablematchings.Computationalexperimentsdemonstratetheefficacyandvalueofourapproach.

CategoriesandSubjectDescriptors:

I.2.11[DistributedArtificialIntelligence]:

MultiagentSystems;J.4

[ComputerApplications]:

SocialandBehavioralSciences—Economics

GeneralTerms:

Algorithms,Economics,Theory

AdditionalKeyWordsandPhrases:

stablematching,preferences,demandaggregation,grouppurchasing,volumediscounts,dailydeals,cooperativegames.

1.INTRODUCTION

Matchingbuyersandsellersisoneofthemostfundamentalproblemsineconomicsanddeal”providerslikeGrouponandLivingSocial（andservicesthataggregatesuchdeals）haspropelledgroupdiscountsintothepublicconsciousness.Groupbuyinganddemandaggregationhasbeenstudiedfromseveralperspectives,andmanymodelshavebeenproposedfortheiranalysis.However,weconsideravitalingredientofgroupbuyingthathasreceivedinsufficientattentionintheliterature,namely,thefactthatbuyersoftenhavedistinctpreferencesfortheofferingsofdifferentvendors.Mostmatchingmodelswithvolumediscountsassumethatvendorofferingsareindistinguishabletobuyers,whichsignificantlylimitstheirapplicability.Forinstance,supposetwobuyersXandYare（jointly）comparingtheoffersoftwovendorsorsomeitem:

Aoffersapriceof10foroneunit,butadiscountedpriceof8ifbothbuyfromhim;andBoffersasinglepriceof9perunit.IfAandBareindistinguishable,XandYshouldcooperateandbuyfromA.ButsupposeXprefersB（withvaluation11.5）toA（valuation10）.Inthiscase,XwouldprefertostickwithBunlessYofferssomepaymenttoswitchvendors（YwouldgladlysharesomeofhergeneratedsurpluswithXforthispurpose）.Withouttheabilitytoexpresspreferencesovervendors,“groupbuying”wouldnotemergeeveninthistrivialexample.marketdesign.Awidevarietyofmodelsandmechanismshavebeendevelopedthatreflectdifferentassumptionsaboutthedemands,valuations/preferences,andknowledgeofthemarketparticipantsandtheirabilitytocooperate.Eachleadstoitsowncomputationalchallengeswhendevelopingalgorithmsforcomputingstable（core）matchings,Nashequilibria,clearingpricesorothersolutionconcepts.Inthispaper,weaddresstheproblemofcooperativegroupbuying,inwhichagroupofbuyerscoordinatetheirpurchasestorealizevolumediscounts,mitigatedemandrisk,orreduceinventorycosts.Groupbuyinghaslongbeenusedforcorporateprocurement,viaindustry-specificbuyingconsortiaorbroadlybasedgrouppurchasingorganizations（GPOs）[ChenandRoma2010].TheadventoftheInternet,inparticular,hashelpedbusinesseswithnoprioraffiliationmoreeasilyaggregatetheirdemand[AnandandAron2003].Consumer-orientedgrouppurchasinghasalsobeengreatlyfacilitatedbytheweb;andtherecentpopularityofvolume-basedcouponingand“dailydeal”providerslikeGrouponandLivingSocial（andservicesthataggregatesuchdeals）haspropelledgroupdiscountsintothepublicconsciousness.Groupbuyinganddemandaggregationhasbeenstudiedfromseveralperspectives,andmanymodelshavebeenproposedfortheiranalysis.However,weconsideravitalingredientofgroupbuyingthathasreceivedinsufficientattentionintheliterature,namely,thefactthatbuyersoftenhavedistinctpreferencesfortheofferingsofdifferentvendors.Mostmatchingmodelswithvolumediscountsassumethatvendorofferingsareindistinguishabletobuyers,whichsignificantlylimitstheirapplicability.Forinstance,supposetwobuyersXandYare（jointly）comparingtheoffersoftwovendorsforsomeitem:

Aoffersapriceof10foroneunit,butadiscountedpriceof8ifbothbuyfromhim;andBoffersasinglepriceof9perunit.IfAandBareindistinguishable,XandYshouldcooperateandbuyfromA.ButsupposeXprefersB（withvaluation11.5）toA（valuation10）.Inthiscase,XwouldprefertostickwithBunlessYofferssomepaymenttoswitchvendors（YwouldgladlysharesomeofhergeneratedsurpluswithXforthispurpose）.Withouttheabilitytoexpresspreferencesovervendors,“groupbuying”wouldnotemergeeveninthistrivialexample.Whilematchingbecomesmuchmoresubtleinsuchmodels,assigningbuyerstovendorsinawaythattriggersvolumediscounts,whileremainingsensitivetobuyerpreferences,offersflexibilityandefficiencygainsthatgreatlyenhancetheappealofgroupbuying.ConsideragroupofbusinessesorbuyersworkingwithaGPOtoprocuresupplieswithinaspecificproductcategory（e.g.,manufacturingmaterials,packaging,transportation,payrollservices,etc.）.TheGPOisabletonegotiatevolumediscounts

fromahandfulofsuppliersorvendors,possiblywithmultiplediscountthresholds.Buyersgenerallyhavedifferentvaluationsfortheofferingsofdifferentvendors（e.g.,buyersmayhaveslightlydifferentmanufacturingspecifications;ormaypreferthecontract,paymentordeliverytermsofcertainvendors）.Asuitablematchingofbuyerstovendorsmusttradeoffthesepreferenceswiththetriggereddiscountprices.Thesameissuesariseinconsumerdomains.Supposeadailydealaggregatorcreatesa“marketplace”forsomeproductcategory,say,spas.Multiplespasofferdealsthatonlytriggerifacertainquantityissold.Buyersarefacedwithadilemma:

theymaywantonlyoneitem,butareuncertainaboutwhichdealwilltrigger.Iftheyonlyoffertobuy（i.e.,conditionallypurchase）theirmostpreferredspa,theymaynotgetanydealiftheirpreferreddealdoesnottrigger.Butiftheyofferonmultiplespastohedgethatrisk,theyruntheoppositeriskofobtainingmoreitemsthantheywant.Amatchingmodelthatallowsconsumerstospecifypreferencesforitemsrelativetotheirdiscountedpricesprovidesflexibilitythatbenefitsbothconsumersandretailers.Ourmodel.Inbroadstrokes,ourmodelassumesasetofvendorsofferingproducts（e.g.,withinaspecificproductcategory）.InteractingwithsomeGPOorinformalbuyinggroup,vendorsoffer（possiblymultiple）volumediscountsthattriggerifthegroupcollectivelybuysinacertainquantity.Weassumetheseareproposedornegotiatedin

advance,andtakethemtobefixed,postedprices.Foreaseofexposition,weassumebuyershaveunitdemand,hencetreatitemsaspartialsubstitutes.Eachbuyerhasvaluationsforeachitemandquasilinearutility.Sincevendorpricesarefixed,ouraimistofindanallocationofitemstobuyersthatmaximizessocialwelfare（i.e.,sumofbuyers’utilities）giventhediscountsthattrigger,whileensuringstability,orbuyer“satisfaction”withtheresultingallocationatthetriggeredprices.Weconsidertwomainvariantsofthisproblem.Inthetransferableutility（TU）model,thegainsduetodemandaggregationcanbetransferredbetweenbuyerstoensurecooperation.Inthenon-transferableutility（NTU）model,eachbuyerpaysthe（triggered）priceofherallocateditem.Bothmodelshavearoletoplayinspecificbusinessandconsumerapplications.Wealsoconsidervariousformsofknowledgeandrecourseonthepartofthebuyer（e.g.,whethertheyknowonlywhichdiscountstriggered,orhaveknowledgeoftheentireallocationanddiscountschedule）.Ourresults.Sincevendorpricesarefixedgivensomedemandedquantity,themodelinducesacoalitionalgameamongthebuyers,whichwerefertoasadiscountmatchinggame.Vendordiscountsintroducesignificantexternalitiesinthecorrespondingmatchingproblem:

thisleadstotheemptinessofcoreofsuchgamesincertaininstances,bothintheTUandtheNTUsense.Asaconsequence,weconsiderunilateraldeviationsfromthematching,andfocusontheweakernotionofNashstabilityunderseveraldifferentepistemicassumptions.Wefocusfirst（andprimarily）onTUgames.Weestablishthatstablematchings（underallepistemicassumptions）notonlyexist,butthattheymaximizesocialwelfare.Moreover,theycanberealizedusingtransfersonlybetweenbuyersthatarematchedtothesamevendor.Wethenconsidercomputationofsocialwelfaremaximizingmatchings:

weshowthatthecorrespondingdecisionproblemisNP-complete,butthat,givena（fixed）setofdiscountprices,computinganoptimalallocationcanbedoneinpolynomialtime.Asaresult,am