Citizen批处理for

EgyptianComputerScienceJournalVol.
40No.
2May2016ISSN-1110-2586-77-BATAlgorithmforImprovingFuzzyC-MeansClusteringforLocationAllocationofRuralKiosksinDevelopingCountriesunderE-Governance1RajanGupta,1SunilKMuttooand2SaibalKPal1DepartmentofComputerScience,FacultyofMathematicalSciences,UniversityofDelhi,Delhi,India2ScientificandAnalysisGroupLab,DRDO,MetcalfeHouse,Delhi,Indiaguptarajan2000@gmail.
comAbstractRuralKiosksareimportantinfrastructuralpillarinruralregionsforinternetandbasictechnologyfacilityallaroundtheworld.
TheyarealsoknownasTele-centersorCommonServiceCentersandaremajorlyusedbygovernmenttopromoteElectronicGovernance.
ThemajorcharacteristicofsettingupofRuralKioskistheirappropriatelocationsothatpeoplefromruralregioncanavailtheservicesatminimumtravelcostandtime.
TherearelotoftraditionalschemesusedbyresearchersinpastforlocationallocationbutthispaperproposestheusageofFuzzyC-MeansclusteringandBATalgorithmtooptimizethelocationofRuralKiosk.
Themeta-heuristicapproachhasproducedbetterresultsascomparedtonormalgraphtheoriesinpast.
Theexperimenthasbeenconductedonarandomdatasetof72villagelocationsfromIndiaandtheirclustersareformed.
ItisfoundthatusingonlyFuzzyC-MeansclusteringtoallocatethecenterandbyusingitincombinationwithBATalgorithmproducedupto25%ofefficientresults.
ThiscandrasticallyhelpthekeystakeholdersinallocationoftheseRuralKiosksatrightplacessoastomaximizetheirutility.
Keywords:BATAlgorithm,LocationAllocation,RuralKiosks,FuzzyC-Means,Clustering,E-Governance,Tele-centers,CommonServiceCenters1.
IntroductionRuralKiosksorTele-CentersorCommonServiceCenters,interchangeabletermsusedworldwide,arethecenterstoaccessinternetfacilitiesforpublicusage[1].
ItcanbeusedforaccessingE-Governmentservicesorfordifferentpurposesinvariousregions.
Theirmainaimistoconnectruralregionwiththeurbanregionandbringallthepeopleonequalplatformintermsofinfrastructureaseverybodycan'taffordtoindividualsetupindevelopingnations.
IncountrieslikeIndiaandAfrica,theseRuralKiosksbecomesevenmoreimportantasthepopulationofruralregionsisveryhighandthusmorecarehastobetakenforpeopleinthesecountries[2].
TheRuralKioskcanbeasmallcenterhavingbasicfacilitieslikeComputerSystem,Internet,Telephone,Faxandphotocopyingmachine.
Theseareusedtogetconnectedtotheworldandthusgiveabasicinfrastructuretothepeoplelivinginruralregions.
Althoughtheirmainpurposeistoprovidefacilitytothepeoplefromruralregions,buttheirlocationbecomesEgyptianComputerScienceJournalVol.
40No.
2May2016ISSN-1110-2586-78-animportantcriterionwhilesettingthemupinvariousstates,districtsandregions.
Incasethetraveltimeandtravelcostisnotappropriatefortheuserslivingintheregionofestablishment,thecitizensmayloseinterest.
ThiscandefeatthewholepurposeofsettingupoftheRuralKioskstobringintechnologyaccessrelatedequality.
Theworld-wideTele-centermovementputTele-centersontheworldmaponlyduetothefactthattheywereabletocometogetherasaglobalnetwork,andinmanycases,muchmorestronglyasnationalnetworksofTele-centers.
ThePhilippineCommunitye-CentersNetwork(PhilCeCNet),theThaitelecentre.
org,Tele-center-Europe,ColombianNationalTele-centerNetwork,ChileanNationalNetworkofTele-centers(ATACH)andtheSriLankanTele-centerFamilyareonlyafewofthemanysucheffortsaroundtheworld.
TheUnitedNationsEconomicCommissionhashelpedtoestablishregionalnetworksofTele-centersinvariousregionsoftheworld,withtheAsianone,theAsia-PacificTele-centerNetwork(APTN)currentlyhousedinThailand[3].
WithinIndiatoo,theCommonServiceCentershavebeenplannedasthekeyinfrastructuralaspectfordeliveryofeffectiveE-Governanceservices.
ThesecentersareintegralpartoftheNationalE-Governanceplanbutduesincetodaytherolloutphasehasnotcompletedandthesecentersareinefficientwherevertheyareopened.
Someofthemarenotworkingeffectivelyduetoinadequatefacilityandothersarenotworkingduetoinappropriatelocations[4].
Therefore,theaimofthisstudyistostudytheoptimizationoflocationallocationoftheseRuralKiosksorcenterssothattheirimplementationbenefitcanreachouttomaximumpeople.
Unlesstheirlocationisoptimizedwithrespecttocostandtime,itwouldbedifficultforthecenterstorunsuccessfully.
SomeschemeshavebeenstudiedbyresearchersinpastintheareaoffacilitylocationallocationbutlessworkexistsinrespecttoRuralKiosks.
Thisstudywillfirstdiscussvariousresearchesdoneinpastinthisareafollowedbyexistingschemestosolvefacilitylocationallocation.
Thentheproposedschemeanditsexperimentalresultsarementionedfollowedbydiscussionandconclusion.
2.
LiteratureReviewTherearedifferentstudiesthattookplaceintheareaoflocationallocationofRuralKiosksorTele-centers.
ThefirstofsuchstudywascarriedoutbyKolokolov&Zaozerskaya[5]in2006ontheGreekIslands.
TheseauthorsusedthediscreteoptimizationproblemapproachtoallocatethelocationtotheTele-centers.
TheyconsideredthesetoffiniteandcountableoptionsforthelocationofTele-centersandselectedthemostoptimizedlocationbasedoncostanddistance.
Similarly,Naiketal[6]alsofocusedonthefinancialaspectofthecenterintheruralregionsofIndia.
ThetotalcostforthefunctioningofCommonServiceCentersorTele-centerswasfoundanditwasconcludedthatGovernmenttoCitizenServicescannotbetheonlycriterionforthesustainabilityofthecenters.
Andifthesecentersarenotopenedatconvenientlocationforthecitizens,evenG2Caspectoffinanceswillnotbeabletoworkandfetchrevenuefromcitizen'spointofview.
Sofromthegovernment'sviewpoint,Tele-centerslocationwillplayanimportantrole.
Shourijeh&Kermanshah[7]in2012workeduponthelocationallocationofTele-centersinTehran,Iran.
ItwasamathematicsorientedmodelfortheoptimizationoflocationEgyptianComputerScienceJournalVol.
40No.
2May2016ISSN-1110-2586-79-withintheparticularregionofTehran.
Linearprogram(Bi-objective)wasdevelopedwhichprimarilyconsideredprimitivestructuredfuzzygoalprogrammingmethod.
TheregionwasdividedintovariousdistrictsforwhichcentroidswerecalculatedfortheallocationoftheTele-centers.
ThetelecommutersweregivenimportanceandtheirdistancefromtheTele-centerswasconsideredaskeyinputforsolvingtheproblem.
However,theminimumandmaximumcapacityparameterswerekeptconstantintheirsolutionwhichwasnottherealworldsituationwhilesolvingsuchproblems.
Theyproposedtoextendtheworktoamoredynamicsolution.
Kolokolov&Zaozerskaya[8]in2013presentedtheextensionoftheirworkdonein2006forbi-criteriabasedcenterlocationallocationoptimization.
Maximizationofefficiencyandminimizationofthecostwereconsideredasthetwocriterionsforlocationallocation.
SetcoverapproachwasusedandlinearintegerprogrammingwasthetechniqueusedinderivingthesolutionalongwiththeL-partitionapproach.
Thesubsetsofthepareto-optimalsetswerefoundusingthedecomposition,trade-offmethodandtheLCEalgorithm.
Asik[9]in2014workedontheoptimizationofthelocationallocationofTele-centersinruralregionsofTurkey.
HediscussedtheimportanceofsustainabilityoftheTele-centersandhowitshouldequallyservemajorityofthepersonsinaregionefficiently.
Setcoverproblemswereusedalongwithp-medianandp-centersolutions.
Theauthorappropriatelydiscussedtheadvantagesanddisadvantagesofthevarioustechniquesandalsogaveadecentbackgroundaboutdifferenttechniquesusedtosolvefacilitylocationallocation.
ApartfromTele-centerorRuralKioskslocationallocationoptimization,therearelotsofresearchersinthepastwhoworkedonotherdifferentkindsoffacilitylocations[10,11,12,13,14].
But,withadventofmoredynamicsolutionsandbetteralgorithms,theallocationofRuralKioskscanbebetteroptimizedandstudiedformoreefficiency.
Inthisregard,Kolokolov&Zaozerskaya[5]proposedtheusageofevolutionaryalgorithmsforlocationallocationoptimizationproblemsinfuture.
OtherstudiesalsoproposedtodothesameandhencethispaperfocusesupontheRural-kiosk'slocationallocationusingthemeta-heuristicalgorithm.
BATalgorithmisusedinconjunctionwiththeFuzzyC-meansclusteringwhichisaprominentclusteringtechniqueusedintheareaoflocationallocationinpastbyvariousresearchers[15,16,17].
3.
ExistingSchemeTherearevariousexistingschemesintheareaofFacilitylocationallocationbutthescopeofthispaperislimitedtothediscussionofFCMclustering[15]andBATalgorithm[18]forwhichthealgorithmsaredescribedasfollows.
BATalgorithmhasalreadybeenprovedtobeabettermeta-heuristicalgorithm[19,20,21]andithasperformedwellascomparedtootherintermittentsearchstrategies[22].
Moreoveritsperformancewithclusteringalgorithmsalsolooksencouraging[23,24],sothisparticularschemeisrestrictedfortheusageinthisstudyfromexperimentalpointofview.
EgyptianComputerScienceJournalVol.
40No.
2May2016ISSN-1110-2586-80-Algorithm1.
FuzzyCMeansClusteringInitializeclustercenter,andstoppingcriteriawhichistominimizethefollowingobjectivefunction=∑∑==‖‖Initializearandomfuzzypartitionmatrix.
CalculatetheclustercentervaluewithfollowingEquation=∑=∑=,≤≤,ForeachclusterCalculatethemembershipfunctionvalueinmatrixusingEquation=[∑(‖‖‖‖)()=],1≤≤,≤≤EndUpdatemembershipmatrixvalues.
If||MembershipCurrent–MembershipNew||≤∈(.
)ThenstopotherwisegotoclustercentercalculationstepWhere,Jmissumofsquareerrorfunctionforallthefuzzyclusterswhichisrepresentedbymembershipmatrixandrelatedsetofclustercenter;isthedegreeofmembershipfordatapointintheithcluster;istheclustercenterandmastheconstantvaluethatcontrolstheinfluenceofmembershipinperformanceindexwhichhasbeentakenas2.
Algorithm2.
BATAlgorithmObjectivefunction(),=(,….
,)Initializethebatpopulation(=1,2,…,n)andDefinepulsefrequencyInitializepulseratesriandtheloudnessWhile(t)SelectasolutionamongthebestsolutionsGeneratealocalsolutionaroundtheselectedbestsolutionendifGenerateanewsolutionbyflyingrandomlyif(<&()<())AcceptthenewsolutionsIncreaseriandreduceendifRankthebatsandfindthecurrentbestendwhilePostprocessresultsandvisualizationWhere,Eachbatisassociatedwithavelocityandalocation,atiterationt,inad-dimensionalsearchorsolutionspace.
Amongallthebats,thereexistsacurrentbestsolutionx.
istheloudnessandisthepulserateofbatatiterationt.
EgyptianComputerScienceJournalVol.
40No.
2May2016ISSN-1110-2586-81-4.
ProposedSchemeTheproposedschemecombinestheclusteringandmeta-heuristictechniquetoderivetheresult.
Figure1.
ProposedschemecombiningtheFCMClusteringandBATalgorithmIntheproposedscheme,asshowninFigure1,thecentralideaistoinitializetheFuzzypopulationwiththevillagepositions(latitudeandlongitude)andtheyareclusteredbasedonnear-nessoftheirlocationcalculatedthroughmembershipfunction.
Oncetheclustersofvillages(6villagespercluster)areformed,theircentroidbasedonlatitudeandlongitudeLatitudeandLongitudeofeverydatapointcollectedClusteringofdatapointsusingFCMclusteringPopulationInitializationforeverydatapointCentroidintermsoflatitudeandlongitudeisusedasstartinglocationofruralkioskwithintheclusterInitializerandomBATgroupalongwithfrequencyandvelocityUpdateloudnessandpulserateCalculateCostaspertheobjectivefunctionMostOptimizedsolutiongivesthebestlatitudeandlongitudepositionLocationforRuralKioskissetwithintheclusterEgyptianComputerScienceJournalVol.
40No.
2May2016ISSN-1110-2586-82-positionsarecalculated.
ThiscentroidbecomesthepointofinterestforlocationallocationofRuralKiosk.
Traditionallythecentroidshouldbetheruralkioskallocationpoint,butitmaynotbeagoodoptionifvillagepopulationvariesinthatcluster.
Sotosatisfythemajorityofpeopleinthecluster,thepopulationandcentroidareinitializedintheBATalgorithm.
ThiscentroidkeepsonchangingaspertheBATcharacteristicsandobjectivefunctioniscalculatedw.
r.
t.
travellingcostwhichisasfollows.
Oncetheminimumcostfunctionisachieved,thatlocationofRuralKioskisfinalizedforthatparticularcluster.
TravelCost=Population*TravelCostperunit*SummationofVillagedistancefromKioskposition5.
Results&DiscussionTheschemeswereimplementedinMATLAB2010aversionsoftware.
Asamplelocationofvillageswastakenfromoneofthedevelopingcountry–Indiainthisexperiment.
Since,Indiaisaggressivelymovingtowardsdigitizationandelectrificationofvillages[25],andimprovementofE-Governancehasalsoledtotheset-upofvariousRuralKiosks(CommonServiceCentersinIndianterminology),sooptimizingthelocationallocationoftheruralkiosksofsuchdatapointswouldbemorebeneficial.
Atotalof72villagepointswereconsideredfortheexperimentfrom(http://www.
mapcoordinates.
net/en)whoselatitudeandlongitudeonmapwereconsideredforthecalculations.
These72villageswereclusteredin12clustersof6villages.
ThisisaspertheIndianGovernment'snationalE-Governanceplanaccordingtowhichaclusterof6villageswillbeconsideredforsetupofoneRuralKioskinthatcluster[26].
Figure2.
ComparisonofTotalCostincurredby72villagesfromtwodifferentschemesofRuralKiosks0100020003000400050006000147101316192225283134374043464952555861646770TotalCostIncurredVillageNumberComparisonofTotalCost(Village)FCMCentroidFCM-BATEgyptianComputerScienceJournalVol.
40No.
2May2016ISSN-1110-2586-83-AsshowninFigure2,totalcostforeveryvillagewascalculatedfromboththeschemesandcomparedagainsteachother.
ForsomevillagestheFCM-BATschemehadhighercost,butforothersitwasdrasticallydownascomparedtothecostcalculatedbyFCM-Centroidscheme.
AlthoughonindividualbasistheBATalgorithminclusioncouldnotreducetheresultsforeverydatapoint,butitmadeasignificantimpactontheoverallcluster'scostasshowninFigure3.
AsshowninFigure3,thetotalcostof12clustersiscomparedagainstthetwoschemes.
ItisclearlyvisiblethatthecostcalculatedbyFCM-BATschemeislowerineachcaseascomparedtoFCM-Centroid.
Thecostreductionisjustnotmarginalbutitissignificantinalmosteverycase.
Thepercentagechangeincostforeveryclustervariesfrom4%to26%whichisveryencouraging.
Therewerealmost7clustersfoundforwhichthepercentagechangeincostwasmorethan15%.
Figure3.
ComparisonofTotalCostincurredby12villageclustersfromtwodifferentschemesofRuralKiosksBATschemehastheabilitytoovercometheweaknessofFCMclusteringi.
e.
gettingtrappedwithlocalmaxima.
Variousmeta-heuristicschemeshavebeendesignedtoovercomethelocalmaximaproblemandBATisoneofthosesuccessfulschemes.
Ithastheabilitytoproducegloballyoptimizedresultswhichcanovercomethelocalmaximaproblem.
Moreover,BATalgorithmissuperiortoothermeta-heuristicalgorithmsbecauseofitsabilitytozoomintotheareaofoptimizedsolutionandthusproducingthebestresults.
Italsohastheabilitytofindoptimizedsolutionusingswarm-intelligencebutwithvariablevaluesforcontrolparametersunlikeothersuchcategoryofalgorithms[18].
0510152025300200040006000800010000120001400016000123456789101112PercentchangeincostoftwoschemesTotalCostClusterNumberComparisonofTotalCost(Cluster)FCM-CentroidFCM-BAT%ChangeEgyptianComputerScienceJournalVol.
40No.
2May2016ISSN-1110-2586-84-BATalgorithmhasproventobeanoptimizationoptionforclusteringbydifferentauthorsinpast[19,20,21,22].
IthasbeenfoundusefulforoptimizingfuzzyclusteringschemesandevenK-meansclusteringschemes[23,24].
Butwithinthelocationallocationresearch,usageofBATalgorithmhasbeenfoundveryrareandpreviousresearchershaveonlyfocusedontheusageofmeta-heuristicschemesintheirproposedwork.
ThisstudyhasusedBATalgorithmwithFCMclusteringandfoundefficientresultsascomparedtothevariousotherlocationallocationschemesolutionsfromthepast.
6.
ConclusionRuralKiosksareanimportantinfrastructuralcomponentfordevelopingregionsaroundtheworld.
Everydevelopingcountryneedstotakeefficientmeasuresinordertogiveaccessofinternettotheruralareas.
Ifgovernmentwantstotransformtheirprocessdigitally,thereisdefinitelyaneedtosetuptheinfrastructureinsuchawaythatpeoplecanactuallyaccessit.
Clusteringisgeneralmethodwhenresourceallocationisconsideredinaparticularregionbutitmaynotproducethemostoptimumresultsasshowninthisstudy.
Itcangettrappedintheproblemoflocalmaximaandthusoveralloptimizedsolutionmaynotbeachieved.
ThisstudyproposedthecombinationofBATalgorithmandFCMclusteringschemetoproducemostoptimizedresultsintermsoftravellingcostforpeopletovisittheRuralKioskandaccesstheservice.
Itshouldbebeneficialtomajorityofpopulationofthecluster.
Theresultsshowsanimprovementofupto25%forthecostsavingandthusmorepeoplewouldbeencouragedtousetheruralkioskservices.
Futurescopeofthisworkincludestestingoftheclusteringtechniquewithotheradvancedmeta-heuristicschemesandanalyzingtheoptimizationwithintheclusters.
Also,sincethescopeofstudywaslimitedtofewdatapointsfromIndianregion,morenumberofdatapointsacrossdifferentcountriescanbestudiedw.
r.
t.
totheirlatitudeandlongitudepositiononthemap.
References[1]Toyama,K.
,&Kuriyan,R.
2007.
"ReviewofresearchonruralPCkiosks.
"MicrosoftResearchIndia.
[2]Kendall,J.
,&Singh,N.
2012.
"PerformanceofInternetKiosksinRuralIndiaGender,CasteandLocation.
"ReviewofMarketIntegration,4(1):1-43.
[3]Shadrach,B.
,&Sharma,S.
2013.
"ImpactassessmentofIndiancommonservicescentres.
"MinistryofCommunicationandInformationTechnology,GovernmentofIndia,NewDelhi.
[4]Dass,R.
,&Bhattacherjee,A.
2011.
"StatusofCommonServiceCenterPrograminIndia:Issues,ChallengesandEmergingPracticesforRollout(No.
WP2011-02-03).
"IndianInstituteofManagementAhmedabad,ResearchandPublicationDepartment.
[5]Kolokolov,A.
A.
,&Zaozerskaya,L.
A.
2006,May.
"Abicriteriaproblemofoptimalservicecenterslocation.
"InInformationControlProblemsinManufacturing12(1):425-429.
[6]Naik,G.
,Joshi,S.
,&Basavarajappa,K.
P.
2010.
"MakingE-GovernancecentersfinanciallysustainableinruralIndia:Aconceptualdesignforactionresearch.
"IIMBangaloreResearchPaper,(317).
EgyptianComputerScienceJournalVol.
40No.
2May2016ISSN-1110-2586-85-[7]Shourijeh,M.
T.
,Kermanshah,M.
,Mamdoohi,A.
R.
,Faghri,A.
,&Hamad,K.
2012.
"AMathematicalOptimizationModelforLocatingTelecenters.
"AppliedMathematics,3:251-263[8]Kolokolov,A.
A.
,&Zaozerskaya,L.
A.
2013.
"Solvingabicriteriaproblemofoptimalservicecenterslocation.
"JournalofMathematicalModellingandAlgorithmsinOperationsResearch,12(2):105-116.
[9]Asik,O.
2014.
"LocationOptimizationToDetermineTelecenterNetworkInRuralTurkey.
"Doctoraldissertation,CornellUniversity.
[10]Current,J.
,Min,H.
,&Schilling,D.
1990.
"Multiobjectiveanalysisoffacilitylocationdecisions.
"EuropeanJournalofOperationalResearch,49(3):295-307.
[11]Wen,M.
,&Iwamura,K.
2008.
"Fuzzyfacilitylocation-allocationproblemundertheHurwiczcriterion.
"Europeanjournalofoperationalresearch,184(2):627-635.
[12]Hajipour,V.
,Fattahi,P.
,Tavana,M.
,&DiCaprio,D.
2015.
"Multi-ObjectiveMulti-LayerCongestedFacilityLocation-AllocationProblemOptimizationwithPareto-basedMeta-heuristics.
"AppliedMathematicalModelling.
[13]Mestre,A.
M.
,Oliveira,M.
D.
,&Barbosa-Póvoa,A.
P.
2015.
"Location–allocationapproachesforhospitalnetworkplanningunderuncertainty.
"EuropeanJournalofOperationalResearch,240(3):791-806.
[14]Pereira,M.
A.
,Coelho,L.
C.
,Lorena,L.
A.
,&DeSouza,L.
C.
2015.
"AhybridmethodfortheProbabilisticMaximalCoveringLocation–AllocationProblem.
"Computers&OperationsResearch,57:51-59.
[15]Küükdeniz,T.
,Baray,A.
,Ecerkale,K.
,&Esnaf,.
2012.
"Integrateduseoffuzzyc-meansandconvexprogrammingforcapacitatedmulti-facilitylocationproblem.
"ExpertSystemswithApplications,39(4):4306-4314.
[16]Esnaf,.
,Küükdeniz,T.
,&Tunbilek,N.
2014.
"FuzzyC-MeansAlgorithmwithFixedClusterCentersforUncapacitatedFacilityLocationProblems:TurkishCaseStudy.
"InSupplyChainManagementUnderFuzziness(pp.
489-516).
SpringerBerlinHeidelberg.
[17]Shi,Q.
S.
,&Zheng,X.
Z.
2014,May.
"ElectricVehicleChargingStationsOptimalLocationBasedonFuzzyC-meansClustering.
"InAppliedMechanicsandMaterials,556:3972-3975.
[18]Yang,X.
S.
,&He,X.
2013.
"Batalgorithm:literaturereviewandapplications.
"InternationalJournalofBio-InspiredComputation,5(3):141-149.
[19]Shieh,C.
S.
,Horng,M.
F.
,Ngo,T.
G.
,&Dao,T.
K.
2015.
"UnequalClusteringFormationBasedonBatAlgorithmforWirelessSensorNetworks.
"InKnowledgeandSystemsEngineering(pp.
667-678).
SpringerInternationalPublishing.
[20]Cao,Y.
,Cui,Z.
,Li,F.
,Dai,C.
,&Chen,W.
2014.
"ImprovedLowEnergyAdaptiveClusteringHierarchyProtocolBasedonLocalCentroidBatAlgorithm.
"SensorLetters,12(9):1372-1377.
[21]Nanda,S.
J.
,&Panda,G.
2014.
"Asurveyonnatureinspiredmetaheuristicalgorithmsforpartitionalclustering.
"SwarmandEvolutionarycomputation,16:1-18.
[22]Yang,X.
S.
,Deb,S.
,&Fong,S.
(2014).
Batalgorithmisbetterthanintermittentsearchstrategy.
arXivpreprintarXiv:1408.
5348.
[23]Komarasamy,G.
,&Wahi,A.
2012.
"AnoptimizedK-meansclusteringtechniqueusingbatalgorithm.
"EuropeanJournalofScientificResearch,84(2):26-273.
EgyptianComputerScienceJournalVol.
40No.
2May2016ISSN-1110-2586-86-[24]Sood,M.
,&Bansal,S.
2013.
"K-MedoidsClusteringTechniqueusingBatAlgorithm.
"InternationalJournalofAppliedInformationSystems(IJAIS)ISSN,2249-0868.
[25]Asher,M.
G.
2016.
"The2016-17Budget:APositiveStepinIndia'sTransformationProcess.
"LeeKuanYewSchoolofPublicPolicyResearchPaper,(16-09).
[26]Mishra,G.
2014.
"TelecentresasaMediumforGoodGovernanceinRuralIndia.
"HandbookofResearchonDemocraticStrategiesandCitizen-CenteredE-GovernmentServices,56.