基于神經(jīng)網(wǎng)絡(luò)的港口清淤工程預(yù)警系統(tǒng)

1、僑景忝秦⑧天畢大謦中目近代第一所^學(xué)碩士學(xué)位論文一■●0_●■●蹦|強(qiáng)■曩■●a■■■_學(xué)科專業(yè)：管理科學(xué)與工程作者姓名：謝繼勇指導(dǎo)教師：李從東教授2004年12月AbstractHarborSedimenthasbecomethekeynodusofPortManagementallovertheworld。Tofittherequirementsofsailing，eachyearamassofregulardepositclean

2、workinsearoute，havenandberthhastobedonefrequently，especiallyinthoseSiltdepositedbasedharborCurrentlythoseportcleaningcompaniesflametheirestimativecleanplansuponglancingexperienceandformerdatawithoutaccurategist，According

3、lyareasonableandeffectiveworkscheduleofequipmentmaintenanceandusingishardtoestablish，neitherthemaximumexertionoflaborforcenorthemaximumefficiencyofequipmentusageThusthecostofsedimentcleaningisgreatlyincreasedSimultaneous

4、lyalongwiththedevelopmentofcomputerscience，plentyofharborsandrelativeorganizationshasestablishedquiteconsummatedatabasetostoreabundantharborsedimentrelateddataAccordingtOthereasonabove，aimingatthelackoftheharbor—sediment

5、cleaningprojectforecast，thisthesisfindsoutdisadvantagesofforecastingtechnologyinHarborSedimentbyanalyzingthestatusquoAtthemeantime，byresearchingtheoryandcharacteristicofNeuralNetworkmethodofDataMiningandanalyzingitsnonli

6、nearinosculatingcapacity，thewriterprobedintothefeasibilityoftheapplicationofNeuralNetworkmethodofDataMiningapplyingtopredictivemeasurementoftheharbor—sedimentcleaningprojectAndhereonbasingonthemulti—layerforwardneuralnet

7、worktheoryofBPalgorithm，withthemostwidelyusedengineeringcalculationsoftwareplatformMatlab65anditsplentifulpracticalresourceneuralnetworktoolbox，moreinteractivegraphicinterfaceandvisiblesimulatingtools，analyzedtheharborse

8、dimentrelativedataandestablishedasetofeffectivesediment—cleaning—projectanticipationsystem，therebystrivedfortheintentionofsupervisingsedimentcleaningprojectpracticeFinally，putthecompletingforecastingalarmsystemintosomesu