美赛常用模型65123主成分分析.docx

美赛常用模型65123主成分分析.docx

《美赛常用模型65123主成分分析.docx》由会员分享，可在线阅读，更多相关《美赛常用模型65123主成分分析.docx（49页珍藏版）》请在冰豆网上搜索。

美赛常用模型65123主成分分析

T2

TeamControlNumber

F2

65123

Forofficeuseonly

Forofficeuseonly

T1

F1

T3

ProblemChosen

F3

T4

F4

E

SmartGrowthTheoriesinCityDesign

Summary

Withthedevelopmentofglobalurbanization,urbanplanninghasbeenahotspotofmostconcern.Sincethetraditionalurbansprawlplanhasbeengraduallyunabletomeetthedevelopmentneedsofthecity,smartgrowthisproposedin1990’s.Inthispaper,aseriesofmethodsaredevelopedtohelptoimplementsmartgrowththeoriesintocityde-sign.

Intask1,basedonthegatheredabundantdataaboutsmartgrowthofcity,25indi-catorsareselectedbyprincipalcomponentanalysis.Thenweconstructanewthree-levelindicatorsystem,wherethefirstlevelhas1indicator,thesecondlevelhas5indicatorsandthethirdlevelhas25indicatorstoevaluatethesuccessdegreeofthesmartgrowth.Moreover,weusetheentropyweightmethod（EWM）togetweightvectoroftheindica-torsinthethirdlevelandthegroupdecisionmakingmethod（GDM）togettheweightvectorofindicatorsinthesecond-class.Meanwhile,thestandardofthesuccessofthesmartgrowthisobtainedthroughK-meansalgorithm.Ametricsystemwhichincludetheintensityofland,socialequitydegree,trafficquality,coordinationdegreeandeconomicdevelopmentlevel,iseventuallybuilttomeasurethesuccessofsmartgrowthofacity.

Intask2,PittsburghandNingguoareselectedtoresearchtheircurrentgrowthplan.ThroughthemetricproposedinTask1,thestrengthsandweaknessesofthecurrentgrowthplanofselectedcitiesismeasured.

Intask3,oursmartgrowthplanisdevelopedbasedontheindexvalueofthecities.Thenthesupportvectormachine（SVM）andweightedmovingaveragemethod（WMAM）areadoptedtopredictthechangeofindicators.Thecombinedpredictionmodelissubse-quentlyadoptedtominimizethepredictionerror.Eventually,bycomparingtheindicatorbetweencurrentcityplanandoursmartgrowthplan,wecanseethesuccessofoursmartgrowthplan.

Intask4,thesingleeffectoftheindividualinitiativesisevaluatedbyourmetrictorankthepotential.TheresultshowsthatthemostpotentialtothesuccessdegreeofthesmartgrowthistheinitiativeoftheeconomicdevelopmentinNingguo,whileinPitts-burghistheinitiativeofimprovinglandintensiverate.

Intask5,sincethepopulationstructurebetweenthetwocitesisdifferent,twodif-ferentpopulationgrowthmodels（PGM）areadoptedtopredictthechangeofthepopula-tion.ThegrowthratecalculatedbyPGMdemonstratesthatourplancanadjustadaptivelytosatisfiedtheincreasingneedofpopulation.

Finally,weanalyzethestrengthsandweaknessesofthemethodsweproposedinthispaper.TheresearchcanalsosatisfytheneedoftheICMandmakereasonablegrowththeoriesintocitydesign.

Contents

1Introduction3

1.1Background3

1.2ProblemStatementandAnalysis3

2.AssumptionandSymbolExplanation4

2.1Assumption4

2.2SymbolExplanation4

3.Task14

3.1DataPre-processing4

3.2PrimaryIndicatorSystem5

3.3WeightingModelsofEvaluationIndicators7

3.4ComprehensiveEvaluationIndex10

3.5MetricofSmartGrowth12

4Task213

4.1GrowthPlanofSelectedCities13

4.2CityIndexEvaluation14

4.3AnalysisofCurrentGrowthPlan14

5Task314

5.1OurSmartGrowthPlan（SGP）14

5.2ThePredictionModelsforEvaluation15

5.3TheEvaluationofOurSmartGrowthPlan16

6Task417

6.1RankingtheIndividualInitiatives17

6.2ComparingtheInitiativesofTwoCities18

7Task518

7.1TheAdaptablePlanForPittsburgh19

7.2TheAdaptablePlanForNingguo19

8.SensitivityAnalysis20

9.StrengthsandWeaknesses21

9.1Strengths21

9.2Weaknesses21

References21

Appendix22

Team#65123Page3of23

1Introduction

1.1Background

Inrecentyears,urbanizationhasbecomeaninevitabletrend.Ithasbeenpredictedthatby2050,about66percentoftheworld'spopulationwillbeaddedtothecity.[1]Consequently,formulatingsuitableurbanplanshavebecomeaconcernofgovernments.Atanearliertime,somecitieschose‘urbansprawl’strategy,namely,justlikestandingpancake.Urbansprawlisaspecialformofsuburbanization.Itincludesthemarginalexpansionoftheexistingurbanizationareaswithverylowpopulationdensity,occupy-inglandthathasneverbeendeveloped.Gradually,manyproblemswereexposedwhenurbansprawlwascarriedout,suchasenvironmentalpollution,thelandinexcesscon-sumption,High-costinfrastructureconstruction,etc.

Inviewofthoseproblems,theconceptofsmartgrowthwasproposedbyGeoffAllderson.Tenprinciplesforsmartgrowthareproposedsubsequentlyin[2].SGisameanstocurbcontinuedurbansprawlandreducethelossoffarmlandsurroundingur-bancenters.Itwillmakethetownorcityapproachamoreeconomicallyprosperous,sociallyequitable,andenvironmentallysustainableplace.

1.2ProblemStatementandAnalysis

Sincesmartgrowthisessentialtourbanconstruction,itissignificanttoproposeanefficientmethodtogiveplanstomeettheprincipleforsmartgrowthinthisproblem.Therefore,indexsystemisneededtoevaluatethesuccessofsmartgrowthofacityfirstly.Twospecificcitiesareselectedtoevaluatetheeffectoftheirgrowthplansthroughindexsystem.Thenwehavetoproposefeasibleplanstohelpthegrowthofourselectedcities.Tomeasurethesuccessofourplans,theindexsystemisneededinasecondtime.Moreover,thepotentialoftheindividualinitiativesshouldberankedinthisproblem.Eventually,weneedtoexplaininwhatwaysourplansupportsthislevelofgrowthifthepopulationwillincreasebyanadditional50%by2050.

Throughtheaboveanalysis,theflowchartofthispaperisshowninfig.1asfel-lows.

Fig.1Theflowchartinthispaper

Team#65123Page4of23

2.AssumptionandSymbolExplanation

2.1Assumption

lThegovernmentpolicywillnotchangeintheshortterm.

lThedatasourceisactualandreliable.

lNeglecttheexplosivechangeswhenforecastingoverthefewdecades.

lScorematrixgivenbyexpertsisnotaffectedbysubjectivefactors.

2.2SymbolExplanation

Symbol

SDSG

IOL

SED

EDL

TQ

CD

J（ck）wi

Definition

thesuccessdegreeofsmartgrowthofacity

Theintensityofland

Thesocialequitydegree

Theeconomicdevelopmentlevel

Thetrafficquality

Thecoordinationdegree

thesumofsquaresofthedistancesfromthesortcenterTheweightofthei-thsecondaryindicator

3.Task1

3.1DataPre-processing

3.1.1DataCollection

Collectingsufficientdataisthebasisofdevelopingacompleteindexsystem.Wesearchedthedatabaseandfound88indicatorsoftwocitiesfirstly.ThedataofPitts-burghinPaisfromcitydata1andthedataofNingguowhichislocatedinAnhui,Chi-naarefromNationalBureauofStatisticsofthePeople'sRepublicofChina2.Thenwesearchedthedataofthetencitieswhichisusedtosetstandardsbyclustering.

3.1.2DataFilling

Theavailabilityofdataisanessentialissue.Nomeasures,regardlessofitsvalue,canprovideefficientassessmentsifbasedonunreliableoruntruthfuldata.Consequent-ly,itisessentialtoensurethecontinuityandauthenticityoftheresearchdata.Never-theless,somedataismissingbecausenotalldataisprovidedinthewebsite.

Toamelioratethissituation,fourmethodsareproposedtocompletethedata,whichareasfollows,

lIftheindicatorvaluesaresmooth,previousdatacanbeadoptedtoreplaceit.

lIfthedatabeforeandaftercanbeobtained,theaveragevaluecanbetakenasthemissing.

lIftwogroupsaresimilar,thenthemissingdatainonegroupcanbereplacedbythevalueofthesamelocationintheothergroup.

1http:

//www.city-

2

Team#65123Page5of23

lTheinterpolationmethodisusedindatafitting.

3.2PrimaryIndicatorSystem

3.2.1SelectprimaryindicatorsbyPCA

88initialindicatorsareselectedfirstlytorepresentthesmartgrowthofcities.Be-causethereareagreatdealofindicatorsrelatedtosmartgrowth,themethodofPCA（PrincipalComponentAnalysis）isadoptedtoreducethenumberofprimaryindicators.Theselectionoftheinitial88indicatorsforsmartgrowthisbasedon3E’sprinciplesand10principles.

3E’sprinciplesreferstoEconomicallyprosperous（EP）,sociallyEquitable（SEQ）,andEnvironmentallysustainable（ESU）.Namely,keepeconomysustainedandsteady,guaranteetherighttofairandensureenvironmentalprotectionifthegovernmentadoptsmartgrowthstrategy.

Sincethesmartgrowthwaspresentedin1990’s,thetenprinciplesforsmartgrowth（SGP）havebeenproposed[2].TheSGPisalsoadoptedasreferenceprinciplesinourindexselection.TheSGPismainlyasfollows:

（1）Mixlanduses.

（2）Takeadvantageofcompactbuildingdesign.

（3）Createarangeofhousingopportunitiesandchoices.

（4）Createwalkableneighborhoods.

（5）Fosterdistinctive,attractivecommunitieswithastrongsenseofplace.

（6）Keepopenspace,farmland,naturalbeauty,andcriticalenvironmentalareas.

（7）Strengthenanddirectdevelopmenttowardsexistingcommunities.

（8）Provideavarietyoftransportationchoices.

（9）Makedevelopmentdecisionspredictable,fair,andcosteffective.

（10）Encouragecommunityandstakeholdercollaborationindevelopmentdeci-

sions.

Fromtheaforementionedselections,theinitialselectionoftheprimaryindicatorsisselected.ThenthemethodofPCAisadoptedtoreducethenumberoftheindicator.Eventually,25indicatorsareobtainedasourprimaryindicators.PrimaryindicatorsareshownasinTab.1.

Tab.1Frameworkofsustainabi