Optimal radio frequency identification deployment in a supply chain network.docx

Optimal radio frequency identification deployment in a supply chain network.docx

《Optimal radio frequency identification deployment in a supply chain network.docx》由会员分享，可在线阅读，更多相关《Optimal radio frequency identification deployment in a supply chain network.docx（48页珍藏版）》请在冰豆网上搜索。

Optimalradiofrequencyidentificationdeploymentinasupplychainnetwork

Optimalradiofrequencyidentificationdeploymentinasupplychainnetwork

SeokcheolChanga,

DiegoKlabjana,

andThomasVossenb,

aIndustrialEngineeringandManagementSciences,NorthwesternUniversity,Evanston,IL,USA

bLeedsSchoolofBusiness,UniversityofColorado,Boulder,CO80309-0419,USA

Received7July2009; 

accepted8January2010. 

Availableonline29January2010.

Abstract

Inrecentyears,theuseofradiofrequencyidentification（RFID）withinsupplychainmanagementhasreceivedconsiderableattention.Asignificantupfrontinvestmenttogetherwithuncertainreturnoninvestment,however,remainimpedimentstodeploymentsofRFIDsystems.ThispaperproposesanovelapproachtoanalyzethepotentialbenefitsofRFIDsystems.WesuggestdeploymentstrategiesthatdetermineanoptimallocationofRFIDwithinasupplychainnetwork.Theresultingmodelscanbeusedtoevaluatethevalueofinformationwithrespecttoitemlosses.Inaddition,wediscussmethodsforsolvingtheresultinglarge-scaleoptimizationproblems,andanalyzehowvarioussystemcharacteristicsimpactoptimaldeploymentstrategiesaswellastheoverallRFIDbenefits.

Keywords:

RFID;Supplychainmanagement;Facilitylocation

ArticleOutline

1.Introduction

2.Literaturereview

3.RFIDplacementinserialnetworks

3.1.Leadtimeexpressions

3.1.1.Norecovery

3.1.2.Fullrecovery

3.1.2.1.Constantsearchtime

3.1.2.2.Proportionalsearchtime

3.1.3.Partialrecovery

3.1.3.1.Constantsearchtime

3.1.3.2.Proportionalsearchtime

3.2.Optimalplacement

4.RFIDplacementingeneralnetworks

5.Computationalstudy

5.1.Experimentalsetup

5.2.Impactoflossprobability

5.3.Impactofrecoveryprobability

5.4.Impactofinstallationcost

5.5.OverlapprobabilityofDCsinthepath

6.Summaryandconclusions

Acknowledgements

AppendixA.Nomenclature

References

1.Introduction

Overthelastfewyears,radiofrequencyidentification（RFID）hasemergedasanimportantnewtechnologytotrackthemovementofgoodsinasupplychain.AnRFIDsystemconsistsofthreeprincipalcomponents.Thefirstcomponentisatagwithanelectronicproductcode（EPC）toidentifypallets,cases,orindividualproductitems.Inaddition,anRFIDsystemalsorequiresreadersthatcaninitializeandreceiveinformationfromthetags.Thecommunicationbetweenthetwoisbyradiowaves.Thefinalcomponentisaninformationsystemwithadatawarehouse,whichisusedtostoreandmanagethedatacapturedfromthereaders.TooperateanRFIDsystem,readerssendradio-wavesignalsovercertainpre-specifiedfrequencies.Thesesignalsarereceivedbytagsinthevicinityofthereaders.Thetagsthentransmitstoreddata（suchastheEPCnumber）backtothereader.Thereader,inturn,decodesandtransfersthisinformationtothedatawarehousetogetherwiththetimeandlocationstamp.

Asaresult,RFIDhasseveralimportantadvantagesovertraditionalbarcoding:

itemscanbereadfromadistancewithoutopticallineofsight,multipletagscanbereadsimultaneously,anditemspecificdatacanbewrittenonatag（Galeetal.,2005）.TheRFIDtechnologyyieldsseveralimportantbenefitswithinsupplychains.Amongothers,automateddatatrackingwithouthumaninterventionoffersfasterprocessing,moreaccurateinventoryrecords,andadvancedshippingnotices.Thus,productshrinkage,transactionerrors,chargebacks,misplacedproducts,andincorrectproductidentificationcanallbereduced（[KangandGershwin,2005]and[Özelkanetal.,2006]）.Moreover,unprecedentedvisibilityoftheproductflows（seeDelenetal.,2007,forexamplesofthedetailedlevelofinformationthatRFIDcanprovide）providesconsumerswiththerightamountofproductattherighttimewhilereducingsafetystockandtheriskofstockouts.

Assuch,theuseofRFIDhasenormouspotentialtoincreasesupplychainefficiency.WhilecompaniesintheU.S.invested$1trilliononupgradingandimprovingsupplychainsinyear2000alone（Galeetal.,2005）inefficienciesremainhigh.AccordingtotheNationalRetailSecuritySurvey,forexample,theproductshrinkagecostforU.S.retailerswasover$31billionin2001;asimilarnumber,$27.5billion,hasalsobeenreportedasthecostofshrinkageinWesternEurope（Bamfield,2004）.Inadditiontothisdirectfinancialimpact,however,thislackofvisibilityalsohasbroaderimpactonsupplychainperformance,leadingtoexcessinventoryand/orlostsales.DeHoratiusandRaman（2008）andKangandGershwin（2005）presentempiricalstudies,whichshowthatinventoryrecordinaccuraciesaresubstantialandwidespread.Asaresult,companiesmayfailtoorderwhenneededorplaceordersthatareunnecessary.Thus,itisperhapsnotsurprisingthatinvestmentsinRFIDtechnologyhavebeengrowingrapidly,spurred—inpart—bymandatesfromindustrygiantssuchasWal-MartandtheU.S.DepartmentofDefense.AccordingtothemarketresearchgroupIDTechEx（2008）,theglobalRFIDmarketwillreach$5.3billionby2008,withafivefoldincreaseexpectedby2018.

Nevertheless,theintroductionofRFIDsystemsposesanumberofimportantissues.Impedimentstodeployment,forexample,ariseduetotheabsenceofinternationalstandards,securityandprivacyconcerns,andlowerreadaccuracyinenvironmentswithliquidsormetals.Inparticular,thehighcostsofimplementationoftenrestricttheadoptionofRFIDsystemsinsupplychainmanagement,especiallyinlightoftheuncertainreturnoninvestment.Onlyabout30%ofWal-Mart'stop100suppliers,forone,haveintroducedfull-scaleRFIDdeploymentsand95%ofthemdidnotexpectreturnoninvestmentwithintwoyears（[Galeetal.,2005]and[ChuangandShaw,2005]）.

Asaresult,thejustificationofRFIDinvestmentsremainsamajorchallenge.Giventhatitisdifficulttomeasuresupplychainbenefitsthatresultfromvisibilityimprovements,practitionersoftenresorttoroughestimatesandsuperficialanalysis（seeLeeandÖzer,2007,foranexcellentdiscussion）.Generallyspeaking,assessmentsofRFID'sreturnoninvestmentproceedbycomparinga“no-visibility”scenario,correspondingtothecurrentsituation,witha“full-visibility”scenariothatisbasedonacompletedeploymentofRFID.These,however,canbeviewedasextremeswithinabroadspectrumof“partial-visibility”optionsavailabletoacompanywhendecidingupontheextentandscopeoftheirRFIDinvestments.Themainobjectiveofourwork,therefore,istoanalyzehowdifferentRFIDdeploymentlevelscanbenefitthesupplychain.Indoingso,ourgoalistoprovidenewinsightsonhowvisibilityimpactssupplychainperformance,whichmayproveusefulwhencompaniescontemplateRFIDinvestments.

Toaddresstheseissues,weevaluatetheplacementofRFIDdeployments,i.e.,thelocationsofRFIDdeploymentsinthesupplychainnetwork.Adeploymentpresentsasignificantcost,whichistypicallyestimatedatuptoseveralhundredthousanddollarsatasinglelocation（Simchi-Levietal.,2007）.WhilethecostofareaderitselfisinthethousandsofU.S.dollars,themajorityofthisamountresultsfromsoftwareandintegrationneeds.DecisionmakershavetoconsiderdeploymentstrategiesthatbalancethebenefitsofRFIDwiththeirdeploymentcostsatanaggregatenetworklevel.WhilerecentresearchconsiderstheperformanceimprovementsthatmayresultfromRFIDinstallationsatasinglelocation（[Özelkanetal.,2006]and[Sounderpandianetal.,2006]）,webelievethatconsiderationsofthesemacro-leveltrade-offshavenotbeenaddressedbefore.

WestudyapproachestoevaluatesuchaggregatebenefitsofRFIDinasupplychain.Specifically,weproposemodelstodeterminetheoptimallocationsofRFIDsystemsinthenetworkbytradingofftheirpotentialbenefitswithRFIDinstallationcosts.Toevaluatethepotentialbenefitsofaninstallationstrategy,weconsideritsimpactontheeffectiveleadtimesinthesystem,whichtakeintoaccountthetimespenttorecoverand/orreshipitemsthatarelost.Here,theeffectiveleadtimesrepresentnotonlythetimeneededtoreceiveadirectorderfromthesupplier,buttheorder-to-deliverytimethatincludesallofthestepsbetweenhittingthereorderpointandplacingareplenishmentorderaswellasthetimetoreceivegoods.Leadtimesareanimportantfactorindeterminingsupplychainperformance,giventheirimpactonpipelineinventorylevels,safetystock,transportationcosts,etc.WebelievethatthecostbenefitsresultingfromRFID-inducedleadtimereductionsprovideanattractiveandnovelwayofmeasuringthebenefitsofincreasedvisibility.

Assuch,ourstudyconcentratesonthemarginalbenefitsthatarisefromanearlydetectionofsupplyandshipmentdisruptions.Withinthisframework,weconsideravarietyofmodelalternativestoaddressdifferentsystemcharacteristicsandrecoverystrategies,eachresultinginacomplexnon-linearlocationproblem.Tosolvetheseproblems,weproposeanovelsolutionmethodologythatappliesdynamicprogrammingtechniqueswithinabranch-and-priceframework,togetherwithaheuristictransformationtechniquetoobtainintegralsolutions.Thecontributionsofourpaperarethereforetwofold.Tothebestofourknowledge,ourworkisthefirstonetoevaluatetheaggregatesupplychainbenefitsofanRFIDdeploymentatthesystemlevel.Inaddition,wealsoproposenovelmethodologiestosolvethemodelsthatarisefromourapproach.

OurmodelsandsolutionmethodologiesdonotexplicitlyuseanyparticularfeaturesofRFID.Anytechnologythatimprovesvisibilitythroughanabilitytotrackgoodscanbedeployedbasedonthemodelsprovided.R