A Simple Sufficient Descent Method for Unconstrained Optimization英文文献资料.docVIP

HindawiPublishingCorporation MathematicalProblemsinEngineering Volume2010,ArticleID684705,9pages doi:10.1155/2010/684705 ResearchArticle ASimpleSuf?cientDescentMethodfor Unconstrained Optimization Ming-LiangZhang,Yun-HaiXiao,andDangzhenZhou Institute of Applied Mathematics, College of Mathematics and Information Science, Henan University, Kaifeng475004,China Correspondence shouldbeaddressedtoYun-HaiXiao,yunhai816@163.com Received18April2010;Revised26August2010;Accepted 26October2010 Academic Editor:JoaquimJ.Ju′dice Copyright q2010Ming-Liang Zhangetal. This is an open access article distributed under the Creative Commons Attribution License, which permits unrestricted use, distribution, and reproduction inanymedium, providedtheoriginalworkisproperlycited. Wedevelop asu?cientdescent method forsolving large-scale unconstrained optimization prob- lems. Ateach iteration, thesearch direction isalinear combination ofthegradient atthecurrent and the previous steps. An attractive property of this method is that the generated directions are always descent. Under some appropriate conditions, we show that the proposed method converges globally.Numerical experiments onsome unconstrained minimization problems from CUTErlibraryarereported,whichillustratethattheproposedmethodispromising. 1.Introduction Inthispaper,weconsidertheunconstrainedoptimizationproblem minfx, x∈ n, 1.1 where f : n → is a continuously di?erentiable function, and its gradient at point xk isdenotedbygxk,orgk forthesakeofsimplicity.nisthenumberofvariables,whichis automaticallyassumedtobelarge.Large-scaleoptimizationisoneoftheimportantresearch areasbothinoptimizationtheoryandalgorithmdesign.Thereexistsomekindsofe?ective methodsavailableforsolving1.1,asforinstance,inexactNewton,limitedmemoryquasi- Newton,conjugategradient,spectralgradient,andsubspacemethods1–4 . Theiterativeformulathemethodisgivenby xk1 xkαkdk, 1.2  2 MathematicalProblemsinEngineering whereαk isasteplength,anddk