美国大学生数学建模一等奖31552.docx
《美国大学生数学建模一等奖31552.docx》由会员分享,可在线阅读,更多相关《美国大学生数学建模一等奖31552.docx(28页珍藏版)》请在冰豆网上搜索。
美国大学生数学建模一等奖31552
Bestalltimecollegecoach
Abstract
Inordertoselectthe“bestalltimecollegecoach”inthelastcenturyfairly,Wetakeselectingthebestmalebasketballcoachasanexample,andestablishtheTOPSISsort-ComprehensiveEvaluationimprovedmodelbasedonentropyandAnalyticalHierarchyProcess.
Themodelmainlyanalyzedsuchindicatorsaswinningrate,coachingtime,thetimeofwinningthechampionship,thenumberofracesandtheabilitytoperceive.Firstly,AnalyticalHierarchyProcessandEntropyareintegrativelyutilizedtodeterminetheindexweightsoftheselectingindicatorsSecondly,Standardizedmatrixandparametermatrixarecombinedtoconstructtheweightedstandardizeddecisionmatrix.Finally,wecangetthecollegemen'sbasketballcompositescore,namelytheorderofmalebasketballcoaches,whichisshowninTable7.AdolphRuppandMarkFewarethelastcenturyandthiscentury's"bestalltimecollegecoach"respectively.Itisrealistic.Therankofcollegecoachescanbeclearlydeterminedthroughthismethods.
Next,ANOVAshowsthatthescoresoflastcentury’scoachesandthiscentury’scoacheshavesignificantdifference,whichdemonstratesthattimelinehorizonexertsinfluenceupontheevaluationandgenderfactorhasnosignificantinfluenceoncoaches’score.Theassessmentmodel,therefore,canbeappliedtobothmaleandfemalecoaches.Nevertheless,basedonthis,wehavedrawncoaches’coachingabilitydistributingdiagramunderidealsituationandnon-idealsituationaccordingtothedatawehavefound,throughwhichwegetthatiftimelinehorizonischosenreasonably,itwillnotaffecttheselectingresults.Inthisproblem,thetimelinehorizonoftheyear2000willnotinfluencetheselectingresults.
Furthermore,weputthedataofthethreetypesofsports,whichhavebeenfoundbyus,intotheaboveModel,andgetthetop5coachesofthethreesports,whichareillustratedinTable10,Table11,Table12andTable13respectively.TheseresultsarecomparedwiththeresultsontheInternet[7],soastoexaminethereasonablenessofourresults.Wechoosethesportsrandomlywhichundoubtedlyshowsthatourmodelcanbeappliedingeneralacrossbothgendersandallpossiblesports.Atthesametime,italsoshowsthepracticalityandeffectivenessofourmodel.
Finally,wehaveprepareda1-2pagearticleforSportsIllustratedthatexplainsourresultsandincludesanon-technicalexplanationofourmathematicalmodelthatsportsfanswillunderstand.
Keywords:
TOPSISImprovedModel;Entropy;AnalyticalHierarchyProcess;
ComprehensiveEvaluationModel;ANOVA
Contents
I.Introduction
Thepaperistohelp"SportsIllustrated"tofindthe“bestalltimecollegecoach”maleorfemale.
Wetacklefivemainproblems:
●Buildamathematicalmodeltochoosethebestcollegecoachorcoaches(pastorpresent)fromamongeithermaleorfemalecoachesinsuchsportsascollegehockeyorfieldhockey,football,baseballorsoftball,basketball,orsoccer,andclearlyarticulateourmetricsforassessment.
●Doesitmakeadifferencewhichtimelinehorizonthatyouuseinyouranalysis,i.e.,doescoachingin1913differfromcoachingin2013?
●Presentourmodel’stop5coachesineachof3differentsports.
●Discusshowourmodelcanbeappliedingeneralacrossbothgendersandallpossiblesports.
●InadditiontotheMCMformatandrequirements,preparea1-2pagearticleforSportsIllustratedthatexplainsourresultsandincludesanon-technicalexplanationofourmathematicalmodelthatsportsfanswillunderstand.
Totacklethefirstproblem,wesearchedtheindicatorsofTop600men’sbasketballcoachesoftheAmericancolleges.Takeselectingthebestmalebasketballcoachasanexample:
fortheexplicitfactorsthataffectassessmentstandards,wecalculateeachindicator’sweightbyusingEntropymethod;forthoseimplicitfactors,wecalculatetheweightthroughexperts’evaluation.Thedeterminationofeachindicator’sscoreshouldbegivenbyexpertsevaluationofeachindicator.Theseindicatorsarethennumericalized,andtheimportanceofeachindicatorisdeterminedthroughweightcoefficients.Thenthroughthemultiplicationofthescoresofcoaches’differentabilityindicatorwithcorrespondingweightcoefficients,wegetthecorrespondingscores,andthehighestscoreindicatesthebestchoice.
Forthesecondquestion,wefirstuseANOVAtodeterminewhethersignificantdifferenceexistsbetweenthescoresofcoachesinthelastcenturyandthiscenturyandthegenderfactorSignificancedifferenceshowsthatthetimelinehorizon,thegenderfactorhasinfluenceontheassessment,whereasinsignificantdifferenceshowsnoinfluence.Andbasedonthis,wehavedrawncoaches’coachingabilitydistributingdiagramunderidealsituationandnon-idealsituationaccordingtothedatawehavefound,whichhelpusfurtherresearchtheinfluenceoftimelinehorizonontheassessment.
Forquestion3and4,weputthedataofthethreetypesofsports,whichhavebeenfoundbyus,intotheModel,andgetthetop5coachesofthethreesports,whichareillustratedinTable10,Table11,Table12andTable13respectively.TheseresultsarecomparedwiththeresultsontheInternet,soastoexaminethereasonablenessofourresults.Wechoosethesportsrandomly,whichundoubtedlyshowsthatourmodelcanbeappliedingeneralacrossbothgendersandallpossiblesports.Atthesametime,italsoshowsthepracticalityandeffectivenessofourmodel.
Figure1.Thesourceofthebestcollegecoaches
П.TheBasicAssumption
●Expertsrecessivefactorsevaluationcriteriaevaluationisfairandequitable.
●Coaches’coachinglevelwillincreasewithincreasingage,butitwilldeclineduetomentaldeclinationandthelackofthephysicalstrength.
●Assessmentexpertsarefullyknownoncollegecoaches.
●Theevaluationcriteriaonlyconsiderthefactorsenumeratedinthispaper,withoutconsideringotherfactors.
●Theevaluationcriteriaapplyequallytomenandwomencoaches.
●Weusedthegeneraldatafromareliablewebsite,Website(seeAppendix).
Ⅲ.Nomenclature
Variable
Meaning
Indexdatanormalizationmatrix
Indexweights
Transformednormalizedmatrix
"Positiveidealsolution"
"Negativeidealsolution"
icomprehensiveevaluationindexvaluesofbeingevaluated
Indexentropy
IndexInformationutility
Fstatistic
Ⅳ.Model
4.1DataProcessing
Inordertobetterassesstheextentofoutstandingcoaches,weselectedanumberofindicatorstodeterminethecoachforthe"bestalltimecollegesportscoach".Wefoundinformationonthevariousindicatorsofdataonthesiteandgetsomereliableindicatorsdataofthesecollegecoaches.Duetothedimensionsofeachindexinconsistenciesexist,sowetransformedthedatatoeliminatetheeffectsofdimensionless.Andthroughpoorconversiongetanormalizedmatrix
isadimensionlessquantityand
.
4.2Modelanalysis
Inordertoaddresstheproblemsmentionedaboveandprovideavalid,feasibleassessmentstrategyforSportsIllustrated,wedecidetoselectsoftball,basketballandfootballbyreviewingtherelevantliterature.Coachingtime,Competitionwinningrate,Culturalqualities,Athleticability,Socialskills,Abilitytowithstand,Innovationcapacity,Abilitytoperceive,andsoon,whichareevaluationindexes.Theseevaluationindexesaredividedintodominantfactorsandrecessivefactors.SpecificfactorsofaffectingtheevaluationcriteriaareshowninFigureX.Theseindicatorswillbequantifiedanddeterminethedegreeofimportanceofeachindexbyweightcoefficient.Whenselectingcoaches,thescoresoftheindicatorsmultiplycorrespondingweightcoefficient,gettingcorrespondingscores,andthepersonwiththehighestscoreisthebestcandidate.
Multi-levelanalysismethodtodeterminetheweightismoresubjective.Itissuitabletodeterminetheweightsforhiddenfactors,whicharenotusedwidelyinbothsexesandallpossiblerequirementsforsport.Weneedtobuildamorereasonablemodeltodeterminetheweightforthedominantfactorandrecessivefactors.Finally,wedeterminethe“bestalltimecollegecoach”.
4.3Modelbuilding
Welookforthe“bestalltimecollegecoach”byestablishingamathematicalmodelinTechniqueforOrderPreferencebySimilaritytoIdealSolution.Takechoosingthebestcollegecoachorcoachesfromamongmalecoachesinsuchsportsasbasketballasanexample.Forthedominantfactor,wecalculatetheweightofeachindicatorinEntropyMethod;Forthehiddenfactors,wecalculatetheweightofeachindicatorinexpertassessmentmethod.Accordingtothesituationofthecoaches,thescoresofalllevelsshouldbedeterminedbyexperts,andtheseindicatorsshouldbequantified.Weightingcoefficientsrepresenttheimportanceofeachindicator.Thescoresoftheindicatorsmultiplycorrespondingweightcoefficienttoobtainthetotalscore,andthepersonofhighestscoreisthebestcandidate.Thismethodismoreobjective,comprehensive,accurateandwide-applicablethanthepreviousevaluationmodel.
Flowchartoflookingforthe“bestalltimecollegecoach”isshowninFigure2.
Figure2.FlowchartofModel
TOPSISModel(TechniqueforOrderPreferencebySimilaritytoanIdealSolution)wasfirstlyintroducedbyC.L.HwangandK.Yoonin1981.TOPSISModelisbasedontheproximityofalimitednumberofevaluationobjectsandidealisticgoalsandevaluatetherelativemeritsofexistingobjects.Meanwhile,TOPSISModelisanapproximationoftheidealsolutioninordermodel,themodelrequiresonlyamonotonicallyincreasing(ordecreasing)ofeachUtilityfunction.Furthermore,TOPSI