首页 | 本学科首页   官方微博 | 高级检索  
     检索      

A VARIATION OF GOLDFARB'S METHOD FOR LINEARLY CONSTRAINED OPTIMIZATIONPROBLEMS AND ITS SUPERLINEAR CONVERGENCE RATE
作者姓名:SHI  Zhenjun  ZHANG  Yuzhong
作者单位:SHI Zhenjun ZHANG Yuzhong(Institute Of Operations Research,Qufu Normal University,Qse 273165,China)LI Qun(Shandong Financial College,Jinan 25000,China)
摘    要:1.IntroductionRosen'sgradientprojectionmethodisaveryinterestingtechniqueinoptimization(see1,2]),butitsglobalconvergenceisalong-standingopenproblemintheliterature.Thefirstconvergentversionwasfoundbyp.l.k3].ButPolak'sversionistoocomplicated.Du4]alsofoundasimplerconvergentversionbydeletingPolak'sspecialprocedure.CombiningRosen'smethodwithvariablemetricmethods,Goldfarbsl,MurtaghandSargentl'lobtainedtwoefficientalgorithms.Theconvergenceofsuchalgorithmsisstillanopenproblem.However,severalcon…


A VARIATION OF GOLDFARB'S METHOD FOR LINEARLY CONSTRAINED OPTIMIZATION PROBLEMS AND ITS SUPERLINEAR CONVERGENCE RATE
SHI Zhenjun ZHANG Yuzhong.A VARIATION OF GOLDFARB'S METHOD FOR LINEARLY CONSTRAINED OPTIMIZATIONPROBLEMS AND ITS SUPERLINEAR CONVERGENCE RATE[J].Journal of Systems Science and Complexity,1998(4).
Authors:SHI Zhenjun ZHANG Yuzhong
Abstract:This paper presents a variant algorithm of Goldfarb's method for linearlyconstrained optimization problems. In the variant algorithm, we introduce a concept calledconjugate projection, which differs from orthogonal projection. The variant algorithm hasglobal convergence, superlinear convergence rate.
Keywords:Conjugate projection  superlinear convergence rate  global convergence
本文献已被 CNKI 等数据库收录!
设为首页 | 免责声明 | 关于勤云 | 加入收藏

Copyright©北京勤云科技发展有限公司  京ICP备09084417号