1999 New Complexity Analysis of the Primal-Dual Newton Method for Linear Optimization

1999 New Complexity Analysis of the Primal-Dual Newton Method for Linear Optimization

ID:41634312

大小:240.37 KB

页数:20页

时间:2019-08-29

1999 New Complexity Analysis of the Primal-Dual Newton Method for Linear Optimization_第1页
1999 New Complexity Analysis of the Primal-Dual Newton Method for Linear Optimization_第2页
1999 New Complexity Analysis of the Primal-Dual Newton Method for Linear Optimization_第3页
1999 New Complexity Analysis of the Primal-Dual Newton Method for Linear Optimization_第4页
1999 New Complexity Analysis of the Primal-Dual Newton Method for Linear Optimization_第5页
资源描述:

《1999 New Complexity Analysis of the Primal-Dual Newton Method for Linear Optimization》由会员上传分享,免费在线阅读,更多相关内容在学术论文-天天文库

1、AdvancesinComputationalMathematics0(1999)?{?1NewComplexityAnalysisofthePrimal-DualNewtonMethodforLinearOptimizationJ.PengC.RoosandT.TerlakyDelftUniversityofTechnology,FacultyofTechnicalMathematicsandInformatics,P.O.Box5031,2600GADelft,TheNetherlands.E-mail:J.Peng@its.tudelft.nl,C.Roos@its.t

2、udelft.nl,T.Terlaky@its.tudelft.nlWedealwiththeprimal-dualNewtonmethodforlinearoptimization(LO).Nowa-days,thismethodistheworkinghorseinallecientinteriorpointalgorithmsforLO,anditsanalysisisthebasicelementinallpolynomialityproofsofsuchal-gorithms.Atpresentthereisstillagapbetweenthepracticalb

3、ehaviorofthealgorithmsandthetheoreticalperformanceresults,infavorofthepracticalbehav-ior.Thisisespeciallytrueforso-calledlarge-updatemethods.Wepresentsomenewanalysistools,basedonaproximitymeasureintroducedbyJansenetal.,in1994,thatmayhelptoclosethisgap.Thisproximitymeasurehasnotbeenusedinthea

4、nalysisoflarge-updatemethodsbefore.Thenewanalysisdoesnotimprovetheknowncomplexityresultsbutprovidesauni edwayfortheanalysisofbothlarge-updateandsmall-updatemethods.Keywords:Linearoptimization,interior-pointmethod,primal-dualmethod,proximitymeasure,polynomialcomplexity.AMSSubjectclassi cation

5、:AMSSubjectClassi cation:90C051.IntroductionInteriorpointmethods(IPMs)areamongthemoste ectivemethodsforsolvingwideclassesofoptimizationproblems.SincetheseminalworkofKar-markar[9],manyresearchershaveproposedandanalyzedvariousIPMsforLinearOptimization(LO)andnumerousresultshavebeenreported.Fors

6、urvey,werefertorecentbooksonthesubject([22],[26],[28]).Aninterestingfactisthatalmostallknownpolynomial-timevariantsofIPMsusetheso-calledcentralpath[23]ThisresearchissupportedbytheprojectHighPerformanceMethodsforMathematicalOptimizationundertheSWON-grant613-304-200.2/asaguidelinetotheoptimal

7、set,andsomevariantofNewton'smethodisusedtofollowthecentralpathapproximately.Therefore,thetheoreticalanalysisofIPMsconsiststoalargeextentofanalyzingNewton'smethod.Atpresentthereisstillagapbetweenthepracticalbehaviorofthealgorithmsandthetheoreticalpe

当前文档最多预览五页,下载文档查看全文

此文档下载收益归作者所有

当前文档最多预览五页,下载文档查看全文
温馨提示:
1. 部分包含数学公式或PPT动画的文件,查看预览时可能会显示错乱或异常,文件下载后无此问题,请放心下载。
2. 本文档由用户上传,版权归属用户,天天文库负责整理代发布。如果您对本文档版权有争议请及时联系客服。
3. 下载前请仔细阅读文档内容,确认文档内容符合您的需求后进行下载,若出现内容与标题不符可向本站投诉处理。
4. 下载文档时可能由于网络波动等原因无法下载或下载错误,付费完成后未能成功下载的用户请联系客服处理。