基于粒子群算法的汽车保有量预测方法

CAR OWNERSHIP PREDICTION BASED ON PSO

摘要:

汽车保有量关系到城市建设与规划。针对汽车保有量预测问题,提出一种基于粒子群算法的汽车保有量预测方法,建立了一种多因素汽车保有量预测模型。选取城镇人口、居民消费水平、人均地区生产总值、道路网密度、公共交通车辆运营数、公共交通客运总量、油价7个指标作为汽车保有量的主要影响因素。利用主成分分析方法确定影响因素主成分,以主成分作为自变量,汽车保有量作为因变量,建立回归分析模型。运用粒子群算法,结合主成分回归预测值对汽车保有量进行预测。以2005～2014年上海市汽车保有量数据为依据,预测出上海市2020年汽车保有量约为400万辆,并对预测结果进行了分析。

Abstract:

CarSownershipSisSrelatedStoStheScity’sSconstructionSandStheSschematization.SAimingSatStheSissueSofStheScarSownershipSprediction,SweSbringSforwardSaSmethod,SbasedSonStheSParticleSSwarmS Optimization(PSO),toSestablishSaSmulti-factorScarSownershipSpredictionSmodel.STheSsevenSfactors, includingSurbanSpopulation,consumptionSlevel,SgrossSregionalSdomesticSproduction,SroadSnetworkSdensity,SpublicStransportationScapacity,SoilSprice,SwereSchosenSasStheSmainSfactorsStoSinfluenceSthe ScarSownership.STheSprincipalScomponentsSofStheSinfluenceSfactorsSareSascertainedSbySPrincipalSCo-mponentSAnalysis(PCA).SUsingSprincipalScomponentsSasSindependentSvariableSandScarSownershipSasSdependentSvariableScanSweSbuildSupStheSregressionSmodel.STheScarSownershipSpredictionSwas ScalculatedSbyStheSParticleSSwarmSOptimizationS(PSO),ScombiningSwithSregressSpredictedSvalueSofS theSprincipalScomponents.SBasedSonStheSstatisticsSofScarSownershipSinSShanghaiSfromS1994StoS2005,theSmodelSarriveSatStheSpredictionSthatSthereSwillSbeS40millionScarsSinSShanghaiSinS2020,SandSanS analysisSisSmadeSdueStoStheSprediction.#$NLKeywords:SCarSownershipSpredictionSSPCASSRegressSanalysisSS PSO