ebooks logo journals logo reference works logo abstract databases logo
bullet  SIGN IN Register | Why Register? | Got a Voucher? alerts   marked lists   shopping cart 

informaworld

HOME   |   SEARCH   |   BROWSE
    Issues List       Latest Issue       Forthcoming Articles       Volume 53 Issue 5 & 6       Subscribe       Article       References       Related articles      
<< firstfirst   < prevprev   Table of contentstoc   next >next   last >>last
Publisher Logo Publication Cover
Search within this journal

Singular Riemannian barrier methods and gradient-projection dynamical systems for constrained optimization 

Authors: H. Attouch a;  J. Bolte a;  P. Redont a; M. Teboulle b
Affiliations:   a ACSIOM-CNRS UMR 5149, Deacutepartment de Matheacutematiques, case 51, Universiteacute Montpellier II, Place Eugegravene Bataillon, 34095 Montpellier cedex 5, France
b School of Mathematical Sciences, Tel-Aviv University, Ramat-Aviv 69978, Israel
DOI: 10.1080/02331930412331327184
Publication Frequency: 8 issues per year
Published in: journal Optimization, Volume 53, Issue 5 & 6 October 2004 , pages 435 - 454
Number of References: 14
Formats available: HTML (English) : PDF (English)
Previously published as: Mathematische Operationsforschung und Statistik. Series Optimization (0323-3898) until 1985
Article Requests: Order Reprints : Request Permissions


Abstract

This work is devoted to the dynamical system (SRB): d (∂ h(x(t)))/dt+∇ Φ (x(t))∋ 0, with h a proper lower semicontinuous convex function. Existence and uniqueness of solutions are examined. Systems (SRB) include the class of gradient systems with respect to a Hessian Riemannian metric induced by a convex Legendre function h:  GOPT0435math001. Moreover, class (SRB) is closed in a variational sense: links are made with regularized Lotka-Volterra systems and the limit equations obtained by letting the barrier parameter go to 0. Of particular interest is the case h(x)= (1/2)|x|2C(x): this way, one obtains a new gradient-projection method. System (SRB) bears a direct relation with the minimization of Φ over the domain of ∂ h; the asymptotic behaviour of solutions, as time goes to infinity, is a real issue, which is addressed for a convex Φ and a h of the form h=kC, with k convex  GOPT0435math002 and C a finite-dimensional polyhedron.
Keywords: Dynamical systems; Continuous gradient method; Asymptotic analysis; Barrier methods in constrained optimization; Singular Hessian Riemannian metric; Convex Legendre functions; Bregman distances; AMS Subject Classification: 34A60; 37Bxx; 37Lxx; 37N40; 49K24
view references (14)
Bookmark with:
  • CiteULike
  • Del.icio.us
  • BibSonomy
  • Connotea
  • More bookmarks
Privacy Policy | Terms & Conditions | Accessibility | RSS
FAQs in: English . Français . Español . 中文(简体和繁體)
© 2010 Informa plc