城市需水量的预测研究

《城市需水量的预测研究》由会员上传分享，免费在线阅读，更多相关内容在学术论文-天天文库。

1、安徽理工大学硕士学位论文城市需水量的预测研究姓名：何忠华申请学位级别：硕士专业：计算机应用技术指导教师：孟祥瑞20090601安徽理工大学硕士学位论文AbstractAsthepopulationofcityincreasesandtheeconomydevelopsrapidly,waterdemandbecomesmoreandmorelarge．TheexistingwaterresourceandfacilitiesCan’tmeetthefuturedemand．Thecontradictionbetweensupplyanddemandisbecom

2、ingmoreandmoreremarkable．Inordertochangethepresentserioussituation,fmdingnewwellspringandsavingwatermustbemadetoinstructthefarfuturewatersupplyinginfrastructureconstruction．However,alltheseprogramsarebasedoncitywaterdemandforecast．Thus，waterdemandforecastcarriesanimportantsensetoward

3、sthedevelopmentofwaterfacilitiesincity．TheurbanwaterdemandforecastingCanbedividedintofourparts：industrialwaterdemand、agriculturewaterdemand、lifewaterdemandandmunicipalwaterdemand．Structureanalysisaboutwaterdemandisconducted,basedonresearchingthehistoryandeconomyofacity．Andrelatedfact

4、orsforwaterdemandarestudieddeeply．UnderthehelpofSPSSsoftware，correlationanalysisoffactorsismadeandsixfactorsarechoseasthecriticalfactorsofcitywaterdemandforecast．BPnetwork(namely,Back-Propagationnetwork)isakindofartificialneuralnetworks．ItCanapproachanyconsecutivefunction,justdesigne

5、d、析Ⅱlonehiddenlayer．Thisnonlinearcharactermakesitgainawideapplicationroominforecastingfield．However,thetraditionalBPalgorithmsholdsmanydrawbacks，suchas：beingpronetofallintoalocalminimBm、trainingwithaslowspeed、beingpooringeneralizing、beinghardtodecidethenumberofhiddenlayerunits．Thispa

6、permakesanimprovementforBPonthebasisoftheanalysisofdefectroot．Matlabsofrwareisanexcellentdevelopenvironmentwhichsupportssystem’sbuildingfromconceptiondesign、buildingmodel、simulationtorealization、析mrealtime．Itisalsocharacterized硒manymerits，suchas：programmingrapidly,powerfuldrawingperf

7、ormance，beingeasytovisualize、expressoperation(esp．whenoperating诵Ⅱlanmatrixoranarray)．Inaddition,powerfulBPnetworktoolboxandexpressinterfacebuildingmade孤advancedimpulseforittowardsinformationsystemdevelop．Basingonallconsiderationsabove，thispaperbuildsanurbanwaterdemandforecastingmodel

8、诚mBPnetwork、