<?xml version="1.0" encoding="utf-8" standalone="yes"?>
<item>
  <id>04078628</id>
  <dt>j</dt>
  <an>04078628</an>
  <augroup>
    <au>Galanis, S.</au>
    <au>Hadjidimos, A.</au>
    <au>Noustos, D.</au>
  </augroup>
  <ti>On the equivalence of the k-step iterative Euler methods and successive overrelaxation (SOR) methods for k-cyclic matrices.</ti>
  <so>Math. Comput. Simul. 30, No.3, 213-230 (1988).</so>
  <py>1988</py>
  <pu>Elsevier Science B.V. (North-Holland), Amsterdam</pu>
  <lagroup>
    <la>EN</la>
  </lagroup>
  <ccgroup>
  </ccgroup>
  <utgroup>
    <ut>consistently ordered k-cyclic matrices</ut>
    <ut>successive overrelaxation</ut>
    <ut>Jacobi matrix</ut>
    <ut>k-cyclic SOR method</ut>
    <ut>stationary k-step iterative method</ut>
    <ut>optimal relaxation factor</ut>
    <ut>region of convergence</ut>
  </utgroup>
  <cigroup>
    <ci>Zbl 0487.65018</ci>
  </cigroup>
  <ligroup>
    <li>doi:10.1016/0378-4754(88)90001-8</li>
  </ligroup>
  <abgroup>
    <ab>The authors establish the relationship $\omega\sp kT\sp k{\cal L}\sb{\omega}\sp{k-1}=[{\cal L}\sb{\omega}+(\omega +1)I]\sp k$ connecting the successive overrelaxation (SOR) matrix ${\cal L}\sb{\omega}$ and the Jacobi matrix T associated with a linear system $x=Tx+c$, where T is weakly cyclic of index k. Based on this result, they derive an equivalence between the k-cyclic SOR method and a certain stationary k- step iterative method [cf. {\it W. Niethammer} and {\it R. S. Varga}, Numer. Math. 41, 177-206 (1983; Zbl 0487.65018)]. By applying the theory of these stationary k-step methods, old and new results on the k-cyclic SOR method are shown, e.g., the optimal relaxation factor and the region of convergence are determined.</ab>
    <rv>M.Eiermann</rv>
  </abgroup>
</item>