18年美赛O奖论文四

ForofficeuseonlyTeamControlNumberForofficeuseonly80560T1F1T2F2T3ProblemChosenF3T4CF42018MCM/ICMSummarySheetSummaryANewKeynesianApproachtoOptimizingEnergyCompactInourpaper,weconstructanEROIevaluationsystemforthefourstatesusingdatascienceandsucceedindeterminingoptimalgoalsfortheinterstateenergycompact.First,weoperateonthedata.Dataarescreenedaccordingtotheirintegrityandusefulnessoftheinfomation.ThenweselectandmergedifferentvariablesusingCoin-tegrationandMultipleDimensionalScaling(MDS)basedontheindependenceandrep-resentativenessoftheattributes.Forthereservedvariablesandstatisticsbyyear,weuseMeanSubtitutiontoconductdataimputation.Then,weclassifytheprocesseddatabasebyusage,sourcesandsectors.Classificationontheenergysourcesiseventuallymadeaccordingtothecorrespondingenvironmentalimpact.Second,weconstructaEROIevaluationsystem,whichisanimprovementofRe-turnonInvestment(ROI).Weclassifyvariouskindsofenergiesinto10distinctgroups.Allvariablesofpricesareadjustedinordertooffsettheinfluencebyinflationandgeo-graphicaldifferences.Afterthat,wefindthattheexternalcostisrelatedtotheintensityofpollution,soitisusedtomeasuretheinfluenceonenvironment.Also,wetakesectorinfluenceandelectricenergylossintoconsideration.OurdatashowsthatCaliforniahasthebestprofileforuseofcleanerenergysince1974.Third,ourpredictingmodelsfeaturebothMathematicalandEconomicmodels.SincethedatagivenarenotstableinTimeSeries,wedonottakeARMAorARCHmodelintoconsideration.Alinearmodelisinitiallyadoptedtoregressthedata,butitturnsouttohavelimitedaccuracyandfailstofitshort-termfluctuationsorlong-termtrends.Asaresult,weadoptadynamicNewKeynesianIS-LMmodelandincludeforward-lookingexpectationsinthemodel.Wecanthereforepredictfutureenergyconsumptionandstructurewithbetteraccuracy.What’smore,tosimulatepolicyeffects,demandshocksandsupplyshocksareaddedtotheenhancedmodel,sothatweareabletoprovidegovernorswithquantitativepredictionofpolicies.Finally,sensitivityanalysisisaddedtotestandverifyourmodels.Thesatisfyingresultsallowustoputmodelsintorealsituationsandtosolverealproblems.Wedeter-minetherenewableenergyusagetargetsthatin2025,Californiamayreach42%ofclean,renewableenergytothetotalconsumption.Otherstatescanreach35%.Andin2050Allstatesmayreachdifferentfrom38%to51%.Fourstates’governmentshouldsubsidizecleanandrenewableenergyandimposepollutiontaxonothers.Otherkindsofdirectinvestmentandlong-termpolicycanalsobeusedtomeettheenergygoals.Keywords:NewKeynesian;IS-LMModel;LinearRegression;TimeSeries;MDSTeam#80560Page1of24Contents1Introduction11.1ProblemBackground...............................11.2OverviewofOurWork..............................11.3Assumptions....................................22DataProcessing22.1DataScreening...................................22.2DataImputation..................................32.3DataClassification................................33EnergyProfile43.1OverviewProfileoftheFourStates.......................43.2Characterizetheenergyprofile.........................44EROIEvaluationSystem54.1EROIDefinition..................................54.2TheRevisedEROIEvaluationSystem.....................64.3ResultsofEROIEvaluationSystem.......................75PredictiveModeling95.1LinearRegressionModel.............................95.2DynamicNewKeynesianIS-LMModel....................105.3EnhancedNKIS-LMModelwithDemandandSupplyShock.......125.4ClimateChangeCompactinAction......................166SensitivityAnalysis177StrengthandWeakness177.1Weakness......................................188Conclusions189Memo21Appendices22Team#80560Page1of241Introduction1.1ProblemBackgroundEnergyproductionlaysasolidfoundationforthedevelopmentofthewholenationandservesastheessentialimpetusforthefunctionoftheentiresociety.Theutilizationofcleanerandgreenenergyisagrowingtrendworldwideforpurposeofthesustainabledevelopment.Therearemassesofcleanenergyorientedcontractsbeingsignedallovertheworld,whileveryfewofthosearecarriedoutduelargelyparttotheirunrealisticandfar-fetchedgoals.Therefore,settinggoalsreasonableenoughcontributesalottotheoptimalreconstructionoftheenergystructureregardingvariouscountriesorstates.Thepastyearshavewitnessedunprecedentedboominthedevelopmentofbigdata.Datasciencehaspenetratedintoeveryaspectsofourlife,andplaysasignificantroleinstatistics,marketintelligence,businessanalysisandsoon.Moreover,datasciencecanbeappliedtoofferafeasiblesolutionandsetarealisticgoalandthusfacilitatesthedecision-makingprocess.Thereforewegivefullplaytothedatascienceinaddressingtheoptimalissue.NowtherearefourstatesalongtheUSborderwithMexico,California(CA),Arizona(AZ),NewMexico(NM),andTexas(TX)thatwishtoformarealisticandpracticalnewenergycompactfocusedonincreasedusageofcleaner,renewableenergysources.With50yearsofdatain605variablesoneachofthesefourstates’energyproductionandcon-sumptioncollected,wecanperformdataanalysisandmodelingtofigureoutaseriesofreasonablegoalsfortheinterstateenergycompact.1.2OverviewofOurWorkFirst,wefindafewkeypointsinthisquestion:Createanenergyprofileforthefourstatesrespectively.Characterizetheenergypr