第1題ImportantbranchesofintelligentcontroldonotincludeAexpertcontrolBneuralnetworkcontrolCfuzzycontrolDadaptivecontrol第2題Inthefollowingfeatures,theonethatdoesnotbelongtointelligentcontrolisAoptimizationabilityBadaptabilityChighrealtimeDlearningfunction第3題CharacteristicsofintelligentcontrolresearchobjectsdonotincludeAuncertainmodelBhighnon-linearmodelCcomplexcontrolrequirementsDfastcomputation第4題Thebiggestadvantageofintelligentcontrolisthatitdoesnotrequireanaccuratemathematicalmodel.第5題FuJingsunproposedthedualismofintelligentcontrol,includingartificialintelligenceandautomaticcontrol.第6題Thecontroleffectofintelligentcontrolinanysituationisbetterthanthatofclassiccontroltheory.第7題Intelligentcontrolismainlyusedtosolvecomplexsystemproblemswhichareverydifficulttobesolvedbyclassicalmethods.第8題Thelearningofintelligentcontroltheoryisaimedtosolvetheproblemthatsignalsaredifficulttobedetectedinautomaticcontrol.第9題Thetaskofintelligentcontrolistodesignacontrollerwiththefunctionsoflearning,abstracting,reasoning,decision-makingandsoon.第10題InJanuary1987,thefirstInternationalConferenceonIntelligentControlwasheldintheUnitedStates,indicatingtheformationofintelligentcontrolfield.第1題FeaturesofexpertcontroldonotincludeAflexibilityBadaptabilityCrobustnessDnetworkstructure第2題KeytechniquesofexpertcontroldonotincludeArepresentingmethodsofknowledgeBidentifyingandobtainingquantitativecontrolsignalsfromsensorsCconvertingqualitativecontrolsignalsintoquantitativecontrolsignalsonlyDmathematicalmodellingofcontrolledobject第3題InthefollowingdescriptionaboutexpertPIDcontrol,theonethatisincorrectisAcontrolrulesbasedonthedeviationandthedeviationchangerateBaccordingtotheactualcontroltaskanderrorcurve,theexpertcontrolrulesandparameterscanbeadjusted.Cgenerally,incrementalPIDisusedastheoutputvalueofthecontroller.Dthefunctionofdifferentiationistoeliminatethestaticerrorofthesystem.第4題Thedirectexpertcontrollerisusedtoreplacetheconventionalone,directlycontrollingtheproductionprocessorthecontrolledobject.第5題Expertsystemisactuallyacomputerprogram,whichcontainstheknowledgeandexperienceofexpertsinacertainfieldandtheabilitytosolvespecializedproblems.第6題Expertcontrolisdividedintodirectexpertcontrollerandindirectexpertcontroller.ExpertPIDcontrollerbelongstotheindirecttype.第7題Thekeypartsofexpertsystemstructureareknowledgebaseandinferenceengine.第8題Commonknowledgerepresentationmethodsare:productionrule,framework,semanticnetwork,andprocess.第9題Theforwardreasoningininferenceengineistogettheoriginaldataandknownconditionsfromtheresults.第10題Therulebaseofexpertcontrolisgenerallyrepresentedbytheproductionrule,IFcontrollingthesituationandTHENoperatingtheconclusion.TestQuestionsforChapter3第1題ThedeterminationmethodofmembershipfunctionisAfuzzystatisticalmethodBsupervisorexperiencemethodCeuralnetworkmethodDalloftheabove第2題Settwofuzzysetsdefinedinthesamedomain,A=[0.20.4;0.60.9],B=[0.80.7;0.50.3],thentheresultofsynthesizingAtoBisA[0.40.3;0.60.6]B[0.40.6;0.20.9]C[0.80.5;0.30.6]D[0.70.5;0.30.6]第3題GiventherelationRof“ifitiscold,wearmoreclothes”,then“whattowearwhenitisverycold”.ThisreasoningprocessbelongstoAforwardreasoningBbackwardreasoningCtwo-wayreasoningDnoneoftheabove第4題Theelementsofafuzzysetbelongtothissettosomeextent,whiletheelementsofanordinarysethaveamembershipof0or1.第5題TheUnion,intersectionandotheroperationsofelementscanbeperformedbetweentwofuzzysetsdefinedindifferentfields.第6題Theelementsofafuzzyrelationshipmatrixcanbegreaterthan1.第7題SettwofuzzysetsAandBdefinedinthesamedomain.TheresultofsynthesizingAtoBisequaltothatofsynthesizingBtoA.第8題Thefuzzysetisexpressedas“A=u1/x1+u2/x2+u3/x3+......”Whereurepresentsmembershipdegree,Xrepresentsdomainelement,“+”and“÷”meanthesameastheydoinmathematics.第9題GiventhatthefuzzysetArepresents“abigapple”,thenthefuzzysetBis“averybigapple”withamoodoperator,andthemembershipdegreeoftheelementsinBisequaltothemembershipdegreeoftheelementsinAplus(1/2)poweroperation.TestQuestionsforChapter4第1題ThefuzzycontrollerincludesAfuzzyinterfaceBknowledgebaseCreasoningandde-fuzzifyinginterfaceDalloftheabove第2題ThecommonmethodsofdefuzzificationincludeAthemaximummembershipmethodBcentroidmethodCweightedmeanmethodDalloftheabove第3題Inthefollowingdescriptionoffuzzycontrol,theonethatisfalseisAtheinputofafuzzycontrolisgenerallytheerrorandthechangerateoftheerror.Bfirstly,thedeviationbetweenthesensorandthesetvalueshouldbetransformedintothediscretedomainbythequantizationfactor.Ctheresultoffuzzyreasoningisobtainedbytheprincipleofmaximizingmembershipdegree,andthenconvertedintotheactualvaluethroughthescalingfactor.Dthemembershipdegreeoffuzzycontrolcanbegivencasually.第4題Thepurposeofde-fuzzingistotransformthefuzzysetstothefixedquantitativesignals.第5題Thefuzzycontrollercanbedesignedwithoutfuzzyinterface.第6題Thefuzzyinterfacegenerallyusesthemaximummembershipprincipletorealizethemappingfromtheactualvaluetothefuzzyset.第7題Intheactualfuzzycontroltask,thefuzzyinferencecandesignafuzzyruletableintheofflinestate,andrealizethefuzzyinferencebyaquerytable.第8題Fuzzycontrolhascertainuncertainty,soitcannotachieveanaccuratecontrol.第9題Whenmakingthefuzzyquerytable,itneedstobesetaccordingtotheactualexperience.第10題Whenthereareseveralknownrelationsinfuzzyreasoning,theresultofeachsubrelationshipcalculationisthen“intersected”togetthefinalreasoningrelationshipofR.TestQuestionsforChapter5第1題ThethresholdvalueinthemathematicalmodelofartificialneuronissimulatingtheactualAroleofneuronsinhibitingsignalsBbondingstrengthamongneuronsCroleofneuronsinsignaltransmissionDroleofneuronscollectingsignals第2題ThefalsedescriptionofBPneuralnetworkisAnegativegradientdescentalgorithmbasedonerrorbackpropagationisusedtoadjustnetworkweightvalueBBPnetworkisaforwardnetworkstructureCthereisnoconnectionbetweenlayersofneuralnetworkDthereisapartialconnectionbetweentheneuronsintheformerlayerandtheneuronsinthelatterlayer.第3題ThefalsedescriptionofneuralnetworkisAthedestructionofasingleneuronintheneuralnetworkwillnotaffecttheresult,soithasabetterfault-tolerantmechanismBtheneuralnetworkhastheabilityofparallelcomputingCthethresholdparametersofneuronscannotbeomittedinthecomputationofneuralnetworkDtheNegativegradientdescentalgorithmisusedinerrorbackpropagation第4題Themathematicalmodelofartificialneuronisestablishedaccordingtotheworkingcharacteristicsoftheactualneuron.第5題Whentherearemanyextremumpointsintheerrorobjectivefunction,thenegativegradientdescentalgorithmusedinBPneuralnetworkcanstillfindtheglobaloptimalsolution.第6題Learningaredividedintoonlinelearningandbatchlearning.第7題Oneneuronisnotsmart,butthousandsofneuronsconnectedtoeachotherinalimitedspacewillproduceintelligence,thisisoneofthephenomenaofswarmintelligence.第8題Thelearningprocessofneuralnetworkistheprocessofweightadjustmentinthenetwork.第9題Therearealotofparameterstobeoptimizedinthedeepnetworkmodel,sobigdatasamplesareneededfortraining.第10題Thegeneralizationabilityofneuralnetworkreferstothefittingandapproximationeffectofsampledata.TestQuestionsforChapter6第1題TheonewhichbelongstoneuralnetworkcontrolstructuresisAneuralnetworksupervisorycontrolBneuralnetworkdirectinversecontrolCneuralnetworkself-correctingcontrolDalloftheabove第2題Theneuralnetworkcontrolgenerallyincludesthesystemidentificationofcontrolledobjectandthedesignofneuralnetworkcontroller.第3題Theneuralnetworkdirectinversecontrolisconnectedinparallelwiththecontrolledobject,sothattheoveralltransferfunctionis1.第4題WhenneuralnetworkandPIDcontrollersarecombined,theoutputisproportionalintegralanddifferentialcoefficients.第5題Whenusedforsystemidentificationofcontrolledobjects,neuralnetworkneedstocollectalargenumberofinputandoutputdatatoformtrainingsamples.第6題Inessence,neuralnetworkisamappingfunctionbetweeninputvariablesandoutputvariables.第7題Theso-calledblackboxmodelmeansthatthemathematicalrelationshipbetweeninputandoutputvariablesisdetermined.第8題Theavailabilityofneuralnetworkdirectinversecontroldependsontheaccuracyoftheinversemodeltoagreatextent.第9題Thenumberofnodesintheinputlayerofneuralnetworkisrelatedtothenumberoftrainingsamples.第10題Asacontroller,neuralnetworkcaneffectivelycontroluncertainsystemorunknownsystem,sothatthecontrolsystemcanachievetherequireddynamicandstaticcharacteristic.TestQuestionsforChapter7第1題BasicoperationsofgeneticalgorithmdonotincludeAreplicationBcrossoverCmutationDinhibition第2題ThecomponentsofgeneticalgorithmareAchromosomecodingmethodBindividualfitnessevaluationCgeneticoperatorsandoperationparametersDalloftheabove第3題Thebiggestadvantageofgeneticalgorithmisthatitcanquicklyfindtheoptimalsolutionoftheobjectivefunctioninthesolutionspace.第4題Ifthefitnessfunctionhastwovariables,thebinarycodesofthetwovariablescanbeconcatenatedtoformachromosome.第5題Theadaptabilityofthechromosomeafterreplication,crossoverandmutationishigherthanthatofthechromosomeintheoriginalpopulation.第6題Themutationoperationenablesthechromosomesequencetogeneratenewchromosomecodes.第7題Intheprocessofreplication,weuseroulettetoscreenoutthechromosomesinvolvedincrossoverandmutation.Obviously,Individualswithhighfitnessaremorelikelytobeselected.第8題Becausemultiplechromosomesareinvolvedingeneticoperationintheiterativeprocess,ithasthecharacteristicsofparallelcomputationandissuitablefortheoptimizationoflarge-scalecomplexproblems.第9題Whentwochromosomescross,theyexchangeallthechromosomesafteracertainbit,thentheybelongtothetypeofsinglepointcrossing.第10題Theresultofgeneticoperationneedstobeconvertedintothevalueofactualvariablethroughdecoding.TestQuestionsforChapter8第1題ThevariabletypesinMATLABareAintegervariableBfloating-pointtypesvariableCcellarrayDalloftheabove第2題MATLABprovidesavarietyofcurvedrawingfunctions,whichcandrawonlywithgivenfunctionformulaandvariables.第3題ThenewversionofMATLABprovidesdeeplearningtoolkitanddirectprogrammingfunctionforsomehardware.第4題ThedatainMATLABarestoredintheformofmatrix,soithasthequickandefficientmatrixandthemathematicsoperatingability.第5題MATLABprogramstatementsmustendwithsemicolons.第6題ThegraphicaluserinterfacecannotbedesignedinMATLAB.第7題TheendmarkofMATLABloopandstatementselectionis“end”.第8題Theresultofpointmultiplicationoftwomatriceswiththesamemagnitudeistheproductoftheelementsinthesameposition,whichisdifferentfromthemultiplicationofmathematicalmatrices.第9題In1:2:10,thethirdelementis第10題ThereisnomoduleprovidingsignalsourceinSIMULINk,whichneedstobegivenfromtheworkspace.TestQuestionsforChapter10第1題Inthecontrolsystemblockdiagramestablishedinthefuzzycontrolexperiment,thepurposeofconnectingthesignalgeneratortothevirtualoscilloscopeisAtoimprovethestabilityofthesystemBtoprovidefeedbacksignalCtherequirementoffuzzycontrollerDtoprovidethecontrastingfunctionbetweentheoriginalsignalandtheoutputtrackingsignaloffuzzycontrol第2題TherulessetintheexpertPIDcontrolexperimentcanchangeitsstrategiesandparametersaccordingtotheactualneeds.第3題Thecontrolledobjectintheblockdiagramoffuzzycontrolsystemcanbeopenedbydouble-clickingandreset.第4題Intheexperimentofthefuzzycontrolsystem,thevariablesaretheerrorandthechangerateoftheerror,andtheoutputofthefuzzycontrolleristher