<?xml version="1.0" encoding="utf-8" standalone="yes"?>
<item>
  <id>03924521</id>
  <dt>j</dt>
  <an>03924521</an>
  <augroup>
    <au>Yuan, Y.</au>
  </augroup>
  <ti>On the superlinear convergence of a trust region algorithm for nonsmooth optimization.</ti>
  <so>Math. Program. 31, 269-285 (1985).</so>
  <py>1985</py>
  <pu>Springer-Verlag, Berlin</pu>
  <lagroup>
    <la>EN</la>
  </lagroup>
  <ccgroup>
  </ccgroup>
  <utgroup>
    <ut>trust region algorithms</ut>
    <ut>nonsmooth optimization</ut>
    <ut>"second order correction" method</ut>
    <ut>superlinear convergent</ut>
  </utgroup>
  <cigroup>
    <ci>Zbl 0555.65037</ci>
    <ci>Zbl 0476.65048</ci>
  </cigroup>
  <ligroup>
    <li>doi:10.1007/BF02591949</li>
  </ligroup>
  <abgroup>
    <ab>It was shown by the author [IMA J. Numer. Anal. 4, 327-335 (1984; Zbl 0555.65037)] that some trust region algorithms for nonsmooth optimization may converge only linearly. In this paper we study a "second order correction" method proposed by {\it R. Fletcher} [in: Numerical analysis, Lect. Notes Math. 912, 85-114 (1982; Zbl 0476.65048)]. Under some mild conditions it is proved that the method is superlinear convergent.</ab>
    <rv></rv>
  </abgroup>
</item>