简介：Animprovedenergydemandforecastingmodelisbuiltbasedontheautoregressivedistributedlag(ARDL)boundstestingapproachandanadaptivegeneticalgorithm(AGA)toobtaincredibleenergydemandforecastingresults.TheARDLboundsanalysisisfirstemployedtoselecttheappropriateinputvariablesoftheenergydemandmodel.Aftertheexistenceofacointegrationrelationshipinthemodelisconfirmed,theAGAisthenemployedtooptimizethecoefficientsofbothlinearandquadraticformswithgrossdomesticproduct,economicstructure,urbanization,andtechnologicalprogressastheinputvariables.Onthebasisofhistoricalannualdatafrom1985to2015,thesimulationresultsindicatethattheproposedmodelhasgreateraccuracyandreliabilitythanconventionaloptimizationmethods.ThepredictedresultsoftheproposedmodelalsodemonstratethatChinawilldemandapproximately4.9,5.6,and6.1billionstandardtonsofcoalequivalentin2020,2025,and2030,respectively.