A family of variable metric proximal methods

نویسندگان

  • J. Frédéric Bonnans
  • Jean Charles Gilbert
  • Claude Lemaréchal
  • Claudia A. Sagastizábal
چکیده

We consider conceptual optimization methods combining two ideas: the Moreau-Yosida regularization in convex analysis, and quasi-Newton approximations of smooth functions. We outline several approaches based on this combination, and establish their global convergence. Then we study theoretically the local convergence properties of one of these approaches, which uses quasi-Newton updates of the objective function itself. Also, we obtain a globally and superlinearly convergent BFGS proximal method. At each step of our study, we single out the assumptions that are useful to derive the result concerned. Une famille de m ethodes de quasi-Newton proximales R esum e : Nous consid erons des m ethodes conceptuelles d'optimisation combinant deux id ees: la r egularisation de Moreau-Yosida en analyse convexe et les approximations quasi-Newtoniennes des fonctions r eguli eres. Nous d eveloppons quelques approches bas ees sur cette combinaison et etablissons leur convergence globale. Nous etudions ensuite d'un point de vue th eorique les propri et es de convergence locale d'une de ces approches, dans laquelles les mises-a-jour utilisent la fonction originale. Nous pr esentons aussi une m ethode de BFGS proximale qui converge globalement et superlin eairement. A chaque etape de notre d eve-loppement, nous pr ecisons les hypoth eses minimales n ecessaires a l'obtention des r esultats.

برای دانلود رایگان متن کامل این مقاله و بیش از 32 میلیون مقاله دیگر ابتدا ثبت نام کنید

ثبت نام

اگر عضو سایت هستید لطفا وارد حساب کاربری خود شوید

منابع مشابه

A Proximal Point Algorithm for Finding a Common Zero of a Finite Family of Maximal Monotone Operators

In this paper, we consider a proximal point algorithm for finding a common zero of a finite family of maximal monotone operators in real Hilbert spaces. Also, we give a necessary and sufficient condition for the common zero set of finite operators to be nonempty, and by showing that in this case, this iterative sequence converges strongly to the metric projection of some point onto the set of c...

متن کامل

OPTML 2017:Variable Metric Proximal Gradient Method with Diagonal Borzilai-Borwein Stepsize

We propose a diagonal metric selection for variable metric proximal gradient method (VMPG). The proposed metric better captures the local geometry of the problem and provides improved convergence compared to the standard proximal gradient (PG) methods with Barzilai-Borwein (BB) stepsize selection. Further, we provide convergence guarantees for the proposed method and illustrate its advantages o...

متن کامل

W-convergence of the proximal point algorithm in complete CAT(0) metric spaces

‎In this paper‎, ‎we generalize the proximal point algorithm to complete CAT(0) spaces and show‎ ‎that the sequence generated by the proximal point algorithm‎ $w$-converges to a zero of the maximal‎ ‎monotone operator‎. ‎Also‎, ‎we prove that if $f‎: ‎Xrightarrow‎ ‎]-infty‎, +‎infty]$ is a proper‎, ‎convex and lower semicontinuous‎ ‎function on the complete CAT(0) space $X$‎, ‎then the proximal...

متن کامل

Optimal coincidence best approximation solution in non-Archimedean Fuzzy Metric Spaces

In this paper, we introduce the concept of best proximal contraction theorems in non-Archimedean fuzzy metric space for two mappings and prove some proximal theorems. As a consequence, it provides the existence of an optimal approximate solution to some equations which contains no solution. The obtained results extend further the recently development proximal contractions in non-Archimedean fuz...

متن کامل

Non-Archimedean fuzzy metric spaces and Best proximity point theorems

In this paper, we introduce some new classes of proximal contraction mappings and establish  best proximity point theorems for such kinds of mappings in a non-Archimedean fuzzy metric space. As consequences of these results, we deduce certain new best proximity and fixed point theorems in partially ordered non-Archimedean fuzzy metric spaces. Moreover, we present an example to illustrate the us...

متن کامل

New best proximity point results in G-metric space

Best approximation results provide an approximate solution to the fixed point equation $Tx=x$, when the non-self mapping $T$ has no fixed point. In particular, a well-known best approximation theorem, due to Fan cite{5}, asserts that if $K$ is a nonempty compact convex subset of a Hausdorff locally convex topological vector space $E$ and $T:Krightarrow E$ is a continuous mapping, then there exi...

متن کامل

ذخیره در منابع من


  با ذخیره ی این منبع در منابع من، دسترسی به آن را برای استفاده های بعدی آسان تر کنید

عنوان ژورنال:
  • Math. Program.

دوره 68  شماره 

صفحات  -

تاریخ انتشار 1995