供水庫(kù)群優(yōu)化調(diào)度的計(jì)算智能方法及應(yīng)用研究.pdf

1、河海大學(xué)博士學(xué)位論文供水庫(kù)群優(yōu)化調(diào)度的計(jì)算智能方法及應(yīng)用研究姓名：王德智申請(qǐng)學(xué)位級(jí)別：博士專業(yè)：水文學(xué)及水資源指導(dǎo)教師：董增川20070601AbstractOptimaldispatchoffeedingreservoirgroupplaysaveryimportantroleinwaterresourcesdevelopmentutilizationandmanagementTheapplicabilityandvalidityof

2、availableoptimizationtechniquessometimesmaybearen’tveryexcellentinsolvingtheproblemofoptimaldispatchoffeedingreservoirgroupItneedstobeimprovedandperfectedAimingattheresearchhotspotsonComputationalIntelligence’Sapplicatio

3、nincomplexwaterresourcessystembasedonoptimaldispatchoffeedingreservoirgroup，inordertoconductsomeresearchesonforecastingandoptimization，thispapertakesoptimaldispatchproblemoffeedingreservoirgroupinBao’anDistrict，ShenzhenC

4、ityasanexampleonthewholeseveralresearcheshavebeenconductedasfclllows：FirstlytheproblemofrunoffforecastingisresearchedforwatersupplyreservoirgroupbasedonPhaseSpaceReconstructionTheoryAftertheconsistencyofrunof!fsequenceis

5、analyzed，thestatisticmethodofBDSisappliedtoproveitsnonlinearityandthelargestLyapunovexponentiscomputedwhichmanifoststhatthereischaoticcharacteristicsintherunoffsequenceofreservoirgroupSingle—dimensionalandmultidimensiona

6、lrunoffforecastmodelsarebuiltandanalyzedbasedonStateSpaceReconstructionTheoryArtificialNeuralNetworkandGeneticAlgorithmTheirperformancesinpracticearecomparedandanalyzedSecondlyahybridintelligentalgorithmcalledGAAisdesign

7、ed，whichisbasedonGeneticAlgorithmandAntAlgorithmandhasthecharacteristicofrandomsearchanddeterministicresearchPenaltyfactorbasedonsimulatedannealingideaiSintroducedtodealwiththeoptimalproblemwithconstrainteffectivelyItiSa

8、relativelyuniversalframeforoptimalproblemsThisalgorithmiscomparedwiththeunimprovedalgorithmthroughnumericalexperimentFurthermore，inpracticeitembodiesitsrelativelyfeasibilityapplicabilityandversatilityMeanwhileinordertoov

9、ercometheproblemof“DimensionalityDisaster”causedbymanywatersupplydistricts，fuzzyoverallevaluationmodelisusedtoaggregatemanywatersupplydistrictsintooneaggregatedistrictbyassigningarelativewei曲tforeachdistrictAccordinglyad

10、ecompositionalgorithmisdesignedConsideringthefunctionofwaterconservationandwaterreleaseofreservoirs，thenotionoffeasibleintervalofiointoperationforfeedingreservoirgroupisputforwardandtheaccordingalgorithmisdesigned，whichi

11、sprovedfeasibleinpracticeThirdlyEvolutionAlgorithm，whichbelongstoComputationalIntelligence，isadaptedtogeneratenoninferiorsolutionformultiobjectivereservoirgroupoptimaloperationandthenon—inferiorsolutionsareevaluatedbyfuz

12、zytheoryAftermultiobjlectiveevolutionalgorithmtheoryisreadupanditscriticaltechnologyisstudiedindetail，aevolutionaryalgorithmbasedonParetostrengthisdesignedandappliedtosolvearealcaseAfterthenon—inferiorsolutionsaregenerat