简介:Inthispaper,LCPisconvertedtoanequivalentnonsmoothnonlinearequationsystemH(x,y)=0byusingthefamousNCPfunction-Fischer-Burmeisterfunction.NotethatsomeequationsinH(x,y)=0arenonsmoothandnonlinearhencedifficulttosolvewhiletheothersarelinearhenceeasytosolve.ThenwefurtherconvertthenonlinearequationsystemH(x,y)=0toanoptimizationproblemwithlinearequalityconstraints.AfterthatwestudytheconditionsunderwhichtheK-TpointsoftheoptimizationproblemarethesolutionsoftheoriginalLCPandproposeamethodtosolvetheoptimizationproblem.Inthisalgorithm,thesearchdirectionisobtainedbysolvingastrictconvexprogrammingateachiterativepoint,However,ouralgorithmisessentiallydifferentfromtraditionalSQPmethod.Theglobalconvergenceofthemethodisprovedundermildconditions.Inaddition,wecanprovethatthealgorithmisconvergentsuperlinearlyundertheconditions:MisP0matrixandthelimitpointisastrictcomplementaritysolutionofLCP.Preliminarynumericalexperimentsarereportedwiththismethod.
简介:Inthispaper,weproposeanewartificialantalgorithmforsolvingconstrainedoptimizationproblemsingeneralcase.Numericalexamplesshowthatitiseffectiveandrobust.
简介:AbstractInthispaper,anewSQPmethodforinequalityconstrainedoptimizationisproposedandtheglobalconvergenceisobtainedunderverymildconditions.
简介:Inordertosolvetheconstrainedglobaloptimizationproblem,weusepenaltyfunctionsnotonlyonconstraintsbutalsoonobjectivefunction.Thenwithintheframeworkofintervalanalysis,anintervalBranch-and-Boundalgorithmisgiven,whichdoesnotneedtosolveasequenceofunconstrainedproblems.Globalconvergenceisproved.Numericalexamplesshowthatthisalgorithmisefficient.
简介:Inordertomeettheneedsofpracticaldesign,aninterpolationtechniqueisemployedtoconstraintheshapeofsurfaces.Themethodofpreservingpositivityontheinterpolationsurfaceandconstraintoninterpolatingdataisalsodeveloped.Theadvantageofthisnewmethodisthatitcanbeusedtoconstraintheshapeofaninterpolatingsurfaceonlybyselectingsuitableparameters,andnumericalexamplesarepresentedtoshowtheperformanceofthemethod.
简介:Resolventmethodsarepresentedforgeneratingsystematicallyiterativenumericalalgorithmsforconstrainedproblemsinmechanics.Theabstractframeworkcorrespondstoageneralmixedfiniteelementsubdif-ferentialmodel,withdualandprimalevolutionversions,whichisshowntoapplytoproblemsoffluiddynamics,transportphenomenaandsolidmechanics,amongothers.Inthismanner,Uzawa’stypemethodsandpenalization-dualityschemes,aswellasmacro-hybridformulations,aregeneralizedtononnecessarilypotentialnanlinearmechanicalproblems.
简介:Animprovedconstrained(IC)steeringlawforsinglegimbalcontrolmomentgyros(SGCMGs)withdeformedpyramidconfiguration(DPC)isproposed.Firstofall,theoriginalsystemwithfivepyramidconfiguration(FPC)whosetwoadjacentgyrosareinfailurestateisreconfiguredasadegradedsystemwithDPC.Then,thesingularangularmomentumhypersurfacesoftheoriginalandthedegradedsystemsareplottedviathesingularangularmomentumequationofSGCMGs.Basedonsingularsurfaces,thedifferencesbetweenFPCandDPCinsingularityandmomentumenvelopeareobtaineddirectly,whichprovideanimportantreferenceforsteeringlawdesignofDPC.Finally,anICsteeringlawisdesignedandappliedtoDPC.ThesimulationresultsdemonstratethattheICsteeringlawhasadvantagesinsimplicityofcalculation,avoidanceofsingularityandexactnessofoutputtorque,whichendowthedegradedsystemwithfinecontrollabilityinarestrictedworkspace.
简介:Fordecades,nacrehasinspiredresearchersbecauseofitssophisticatedhierarchicalstructureandremarkablemechanicalproperties,especiallyitsextremefracturetoughnesscomparedwiththatofitspredominantconstituent,CaCO3,intheformofaragonite.Crackdeflectionhasbeenextensivelyreportedandregardedastheprincipaltougheningmechanismfornacre.Inthispaper,ourattentionisfocusedoncrackevolutioninnacreunderaquasi-staticstate.Weusethenotchedthree-pointbendingtestofdehydratednacreinsituinascanningelectronmicroscope(SEM)tomonitortheevolutionofdamagemechanismsaheadofthecracktip.Theobservationsshowthatthecrackdeflectionactuallyoccursbyconstrainedmicrocracking.Onthebasisofourfindings,acrackpropagationmodelisproposed,whichwillcontributetouncoveringtheunderlyingmechanismsofnacre’sfracturetoughnessanditsdamageevolution.Theseinvestigationswouldbeofgreatvaluetothedesignandsynthesisofnovelbiomimeticmaterials.
简介:Thispaperpresentsthestabilityofdifferenceapproximationsofanoptimalcontrolproblemforaquasilinearparabolicequationwithcontrolsinthecoefficients,boundaryconditionsandadditionalrestrictions.Theoptimalcontrolproblemhasbeenconveredtooneoftheoptimizationproblemusingapenaltyfunctiontechnique.Thedifferenceapproximationsproblemfortheconsideredproblemisobtained.Theestimationsofstabilityofthesolutionofdifferenceapproximationsproblemareproved.Thestabilityestimationofthesolutionofdifferenceapproximationsproblembythecontrolsisobtained.
简介:Akindofnondecreasingsubgradientalgorithmwithappropriatestoppingrulehasbeenproposedfornonsmoothconstrainedminimizationproblem.Thedualtheoryisinvokedindealingwiththestoppingruleandgeneralglobalminimiizingalgorithmisemployedasasubroutineofthealgorithm.Themethodisexpectedtotacklealargeclassofnonsmoothconstrainedminimizationproblem.
简介:Thispaperproposesamultiple-constraints-guaranteedmidcourseguidancelawfortheinterceptionofthehypersonictargets.Intraditionalmidcourselawdesign,theconstraintsoftheaero-thermalheatingarerarelytakenintoconsideration.Theperformanceoftheinfrareddetectionsystemmaybedegradedandtheinstabilityoftheflightcontrolsystemmaybeinduced.Toaddressthisproblem,astate-constrainedmodelpredictivestaticprogrammingmethodisintroducedsuchthatbothterminalconstraints(positionandangle)andoptimalenergyconsumptioncanbeensured.Asaresult,asub-optimalmidcourseguidance,guaranteeingtheaforementionedmultiple-constraintstobeneverviolated,issynthesized.Simulationresultsdemonstratetheeffectivenessoftheproposedmethod.
简介:ASUCCESSIVEAPPROXIMATIONMETHODFORSOLVINGPROBABILISTICCONSTRAINEDPROGRAMSWANGJINDE(王金德)(DepartmentofMathematics,NanjingUnivers...
简介:Inthispaper,atrustregionmethodforequalityconstrainedoptlmizationbasedonnondiferentiableexactpenaltyisproposed.Inthisalgorithin,thetrailstepischaracterizedbycomputationofitsnormalcomponentbeingseparatedfromcomputationofitstangentialcomponent,i.e.,onlythetangentialcomponentofthetrailstepisconstrainedbytrustradiuswhilethenormalcomponentandtrailstepitselfhavenoconstraints.Theothermaincharacteristicofthealgorithmisthedecisionoftrustregionradius.Here,thedecisionoftrustregionradiususestheinformationofthegradientofobjectivefunctionandreducedHessian.However,Maratoseffectwilloccurwhenweusethenondifferentiableexactpenaltyfunctionasthemeritfunction.Inordertoobtainthesuperlinearconvergenceofthealgorithm,weusethetwiceordercorrectiontechnique.Becauseofthespecialityoftheadaptivetrustregionmethod,weusetwiceordercorrectionwhenp=0(thedefinitionisasinSection2)andthisisdifferentfromthetraditionaltrustregionmethodsforequalityconstrainedopthnization.Sothecomputationofthealgorithminthispaperisreduced.Whatismore,wecanprovethatthealgorithmisgloballyandsuperlinearlyconvergent.
简介:InthispaperthegeneralizedBianchi’sidentitiesforthevariantconstrainedsystem(GBIVOS)withnon-invariantactionintegralandconstraintconditionswasderived,andthestrongandweakconservationlawsforsuchsystemwasdeduced.ThepreliminaryapplicationsoftheGBIVCStothecaseforsomemodelsoffieldtheorieswasgiven.TheDiracconstraintofsuchsystemwasdiscussed.
简介:Apotentialreductionalgorithmisproposedforoptimizationofaconvexfunctionsubjecttolinearconstraints.Ateachstepofthealgorithm,asystemoflinearequationsissolvedtogetasearchdirectionandtheArmijo'sruleisusedtodetermineastepsize.Itisprovedthatthealgorithmisgloballyconvergent.Computationalresultsarereported.