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:Inacomposite-stepapproach,astepskiscomputedasthesumoftwocomponentsvkandhk.Thenormalcomponentvk,whichiscalledtheverticalstep,aimstoimprovethelinearizedfeasibility,whilethetangentialcomponenthk,whichisalsocalledhorizontalstep,concentratesonreducingamodelofthemeritfunctions.Asafiltermethod,itreducesboththeinfeasibilityandtheobjectivefunction.Thisisthesamepropertyofthesetwomethods.Inthispaper,oneconcernsthecomposite-steplikefilterapproach.Thatis,astepistangentialcomponenthkiftheinfeasibilityisreduced.Orelse,skisacompositestepcomposedofnormalcomponentVkandtangentialcomponenthk.
简介:SincethefoundingofthePeople’sRepublicofChina,theChinesegovernment,toensureequalityandunityamongethnicgroupsandachievetheircommondevelopment,hasformulatedaseriesofethnicandreligiouspoliciesonthebasisoftheactualsituationsofthevariousethnicgroupsandreligions,andthesepolicieshave
简介:Thispaperpresentsatrustregionalgorithmwithnullspacetechniquefornonlinearequalityconstrainedoptimization.Consideringinthenullspacemethodsthat,theconvergentrateofrangespacestepisfasterthanthenullspacestepforthemostcases,theproposedalgorithmcomputesnullstepsmoreoftenthanrangespacestep.Moreover,thenewalgorithmisbasedonthereducedHessianSQPmethod.Globalconvergenceoftheproposedalgorithmisproved.Theeffectivenessofthemethodisdemonstratedbysomenumericalexamples.
简介:Thispaperpresentsatrustregiontwo-phasemodelalgorithmforsolvingtheequalityandboundconstratinednonlinearoptimizationproblem.Aconceptofsubstationarypointisgiven.Undersutableassumptions.theglobalconvergenceofthisalgorithmisprovedwithoutassumingthelinearindependenceofthegradientofactiveconstraints.Anumericalexampleisalsopresented.
简介:Thispaperstudiesafamilyofthelocalconvergenceoftheimprovedsecantmethodsforsolvingthenonlinearequalityconstrainedoptimizationsubjecttoboundsonvariables.TheHessianoftheLagrangianisapproximatedusingtheDFPortheBFGSsecantupdates.Theimprovedsecantmethodsareusedtogenerateasearchdirection.Combiningwithasuitablestepsize,eachiterateswitchestotrialstepofstrictinteriorfeasibility.WhentheHessianisonlypositivedefiniteinanaffinenullsubspace,oneshowsthatthealgorithmsgeneratethesequencesconvergingq-linearlyandtwo-stepq-superlinearly.Furthermore,undersomesuitableassumptions,somesequencesgeneratedbythealgorithmsconvergelocallyone-stepq-superlinearly.Finally,somenumericalresultsarepresentedtoillustratetheeffectivenessoftheproposedalgorithms.