Page 26 - 2023年第54卷第10期
P. 26

[22] 蔡杰.基于 MLP - ARX模型盾构自动掘进技术研究与应用 [J].隧道建设,2022,42(10):1797 - 1803.
                [23] ISABONAJ,IMOIZEAL,OJOS,etal.Developmentofamultilayerperceptronneuralnetworkforoptimalpredic
                       tivemodelinginurbanmicrocellularradioenvironments[J].AppliedSciences - Basel,2022,12(11):5713.
                [24] ZHANGY,CHENGP,MALIKOP,etal.Anartificialneuralnetworkbasedadaptivepower - system stabilizer
                       [J].IEEETransactionsonEnergyConversion,1993,8(1):71 - 77.
                [25] 王雨虹,王淑月,王志中,等.基于改进蝗虫算法优化长短时记忆神经网络的多参数瓦斯浓度预测模型研
                       究[J].传感技术学报,2121,34(9):1196 - 1230.
                [26] LUOJ,CHENH L,ZHANGQ,etal.Animprovedgrasshopperoptimizationalgorithmwithapplicationtofinan
                       cialstressprediction[J].AppliedMathematicalModelling,2018,64:654 - 668.



                       IGOA - MLPdynamicpredictionmodelforsimulationparametersofhighcore
                                rockfilldam constructionundertransferlearningframework

                                                                           1
                                                                                           1
                                                              1
                                    1
                                                     2
                              LFei,ZHONGDenghua,YUJia,ZHANGJun,ZHANGYunuo
                      (1.StateKeyLaboratoryofHydraulicEngineeringSimulationandSafety,TianjinUniversity,Tianjin 300072,China;
                         2.CollegeofWaterResourcesandCivilEngineering,ChinaAgriculturalUniversity,Beijing 100083,China)
                  Abstract:Forconstructionsimulationofhighcorerockfilldam,theparametersarethekeytoensuringitsaccura
                  cy.However,existingparameterpredictionmethodsusedhistoricaldataandignorethedifferencesbetweenthe
                  constructionprocessesofdifferentlayers ,andthereisofteninsufficientormissingdataatthebeginningofanew
                  layer.Inaddition ,theparametersareaffectedbymanyfactorssuchasmeteorologicalconditionsandoperating
                  stateofthemachine.Tosolvetheaboveproblems ,thispapertakesadvantageofthetransferlearning’ scapability
                  ofmodelingwithsmallsamplesthroughknowledgetransferandconsidersthequantitativeinfluenceofvariousfac
                  tors.Animprovedmulti - layerperceptrondynamicpredictionmodel (IGOA - MLP)isproposedforconstruction
                  simulationparametersofhighcorerockfilldamundertheframeworkoftransferlearning.Firstly ,theIGOA - MLP
                  predictionmodelisestablishedthatconsideringtheinfluenceofmultiplefactors.Thegrasshopperoptimizationalgo
                  rithmisimproved (IGOA)bynonlinearreductionfactorandCauchy - Gaussianhybridmutationmode,andtheeffi
                  cientglobaloptimalsearchcapabilityofIGOAisutilizedtooptimizethehyperparametersofmulti - layerperceptron
                  (MLP).Secondly,thetransferlearningstrategyisintroducedtorealizetheknowledgetransferbetweenthehistori
                  calandnewconditionsandsolvetheproblemofinsufficientormissingdatainthenewconditions.Thetrainingset
                  isdividedintosourcedomainandtargetdomain ,andanadaptivelayerisaddedtothehiddenlayerofMLPtore
                  presentthedifferencebetweensourcedomaindataandtargetdomaindata.Acasestudyshowsthatcomparedwith
                  othermachinelearningmethodssuchasMLPmodelandIGOA - MLPmodelwithouttransferlearning,themeanab
                  solutepercentageerror (MAPE)oftheproposedmethodisreducedby54.68% and40.57%,respectively.Itis
                  provedthattheproposedmodelcanpredicttheparametersofconstructionsimulationmoreaccuratelyandprovidea
                  reliabledatabasisforsimulation.
                  Keywords:transferlearning;highcorerockfilldam;constructionsimulation;multi - layerperceptronoptimized
                  byimprovedgrasshopperoptimizationalgorithm;parameterprediction


                                                                                    (责任编辑:李福田)
















                —  1 1 2 —
                     6
   21   22   23   24   25   26   27   28   29   30   31