You cannot select more than 25 topics
Topics must start with a letter or number, can include dashes ('-') and can be up to 35 characters long.
776 lines
24 KiB
HTML
776 lines
24 KiB
HTML
<!DOCTYPE HTML PUBLIC "-//W3C//DTD HTML 4.0 Transitional//EN">
|
|
|
|
<!--Converted with LaTeX2HTML 2002-2-1 (1.71)
|
|
original version by: Nikos Drakos, CBLU, University of Leeds
|
|
* revised and updated by: Marcus Hennecke, Ross Moore, Herb Swan
|
|
* with significant contributions from:
|
|
Jens Lippmann, Marek Rouchal, Martin Wilck and others -->
|
|
<HTML>
|
|
<HEAD>
|
|
<TITLE>Multi-level Schwarz Preconditioners</TITLE>
|
|
<META NAME="description" CONTENT="Multi-level Schwarz Preconditioners">
|
|
<META NAME="keywords" CONTENT="userhtml">
|
|
<META NAME="resource-type" CONTENT="document">
|
|
<META NAME="distribution" CONTENT="global">
|
|
|
|
<META NAME="Generator" CONTENT="LaTeX2HTML v2002-2-1">
|
|
<META HTTP-EQUIV="Content-Style-Type" CONTENT="text/css">
|
|
|
|
<LINK REL="STYLESHEET" HREF="userhtml.css">
|
|
|
|
<LINK REL="next" HREF="node12.html">
|
|
<LINK REL="previous" HREF="node10.html">
|
|
<LINK REL="up" HREF="node10.html">
|
|
<LINK REL="next" HREF="node12.html">
|
|
</HEAD>
|
|
|
|
<BODY >
|
|
|
|
<DIV CLASS="navigation"><!--Navigation Panel-->
|
|
<A NAME="tex2html200"
|
|
HREF="node12.html">
|
|
<IMG WIDTH="37" HEIGHT="24" ALIGN="BOTTOM" BORDER="0" ALT="next"
|
|
SRC="file:/usr/share/latex2html/icons/next.png"></A>
|
|
<A NAME="tex2html196"
|
|
HREF="node10.html">
|
|
<IMG WIDTH="26" HEIGHT="24" ALIGN="BOTTOM" BORDER="0" ALT="up"
|
|
SRC="file:/usr/share/latex2html/icons/up.png"></A>
|
|
<A NAME="tex2html190"
|
|
HREF="node10.html">
|
|
<IMG WIDTH="63" HEIGHT="24" ALIGN="BOTTOM" BORDER="0" ALT="previous"
|
|
SRC="file:/usr/share/latex2html/icons/prev.png"></A>
|
|
<A NAME="tex2html198"
|
|
HREF="node2.html">
|
|
<IMG WIDTH="65" HEIGHT="24" ALIGN="BOTTOM" BORDER="0" ALT="contents"
|
|
SRC="file:/usr/share/latex2html/icons/contents.png"></A>
|
|
<BR>
|
|
<B> Next:</B> <A NAME="tex2html201"
|
|
HREF="node12.html">Smoothed Aggregation</A>
|
|
<B> Up:</B> <A NAME="tex2html197"
|
|
HREF="node10.html">Multi-level Domain Decomposition Background</A>
|
|
<B> Previous:</B> <A NAME="tex2html191"
|
|
HREF="node10.html">Multi-level Domain Decomposition Background</A>
|
|
<B> <A NAME="tex2html199"
|
|
HREF="node2.html">Contents</A></B>
|
|
<BR>
|
|
<BR></DIV>
|
|
<!--End of Navigation Panel-->
|
|
|
|
<H2><A NAME="SECTION00061000000000000000"></A><A NAME="sec:multilevel"></A>
|
|
<BR>
|
|
Multi-level Schwarz Preconditioners
|
|
</H2>
|
|
|
|
<P>
|
|
The Multilevel preconditioners implemented in MLD2P4 are obtained by combining
|
|
AS preconditioners with coarse-space corrections; therefore
|
|
we first provide a sketch of the AS preconditioners.
|
|
|
|
<P>
|
|
Given the linear system ,
|
|
where <!-- MATH
|
|
$A=(a_{ij}) \in \Re^{n \times n}$
|
|
-->
|
|
<SPAN CLASS="MATH"><IMG
|
|
WIDTH="137" HEIGHT="38" ALIGN="MIDDLE" BORDER="0"
|
|
SRC="img3.png"
|
|
ALT="$A=(a_{ij}) \in \Re^{n \times n}$"></SPAN> is a
|
|
nonsingular sparse matrix with a symmetric nonzero pattern,
|
|
let <SPAN CLASS="MATH"><IMG
|
|
WIDTH="93" HEIGHT="36" ALIGN="MIDDLE" BORDER="0"
|
|
SRC="img4.png"
|
|
ALT="$G=(W,E)$"></SPAN> be the adjacency graph of <SPAN CLASS="MATH"><IMG
|
|
WIDTH="18" HEIGHT="15" ALIGN="BOTTOM" BORDER="0"
|
|
SRC="img2.png"
|
|
ALT="$A$"></SPAN>, where <!-- MATH
|
|
$W=\{1, 2, \ldots, n\}$
|
|
-->
|
|
<SPAN CLASS="MATH"><IMG
|
|
WIDTH="138" HEIGHT="36" ALIGN="MIDDLE" BORDER="0"
|
|
SRC="img5.png"
|
|
ALT="$W=\{1, 2, \ldots, n\}$"></SPAN>
|
|
and <!-- MATH
|
|
$E=\{(i,j) : a_{ij} \neq 0\}$
|
|
-->
|
|
<SPAN CLASS="MATH"><IMG
|
|
WIDTH="162" HEIGHT="36" ALIGN="MIDDLE" BORDER="0"
|
|
SRC="img6.png"
|
|
ALT="$E=\{(i,j) : a_{ij} \neq 0\}$"></SPAN> are the vertex set and the edge set of <SPAN CLASS="MATH"><IMG
|
|
WIDTH="18" HEIGHT="15" ALIGN="BOTTOM" BORDER="0"
|
|
SRC="img7.png"
|
|
ALT="$G$"></SPAN>,
|
|
respectively. Two vertices are called adjacent if there is an edge connecting
|
|
them. For any integer <SPAN CLASS="MATH"><IMG
|
|
WIDTH="45" HEIGHT="34" ALIGN="MIDDLE" BORDER="0"
|
|
SRC="img8.png"
|
|
ALT="$\delta > 0$"></SPAN>, a <SPAN CLASS="MATH"><IMG
|
|
WIDTH="13" HEIGHT="15" ALIGN="BOTTOM" BORDER="0"
|
|
SRC="img9.png"
|
|
ALT="$\delta$"></SPAN>-overlap
|
|
partition of <SPAN CLASS="MATH"><IMG
|
|
WIDTH="24" HEIGHT="15" ALIGN="BOTTOM" BORDER="0"
|
|
SRC="img10.png"
|
|
ALT="$W$"></SPAN> can be defined recursively as follows.
|
|
Given a 0-overlap (or non-overlapping) partition of <SPAN CLASS="MATH"><IMG
|
|
WIDTH="24" HEIGHT="15" ALIGN="BOTTOM" BORDER="0"
|
|
SRC="img10.png"
|
|
ALT="$W$"></SPAN>,
|
|
i.e. a set of <SPAN CLASS="MATH"><IMG
|
|
WIDTH="20" HEIGHT="14" ALIGN="BOTTOM" BORDER="0"
|
|
SRC="img11.png"
|
|
ALT="$m$"></SPAN> disjoint nonempty sets <!-- MATH
|
|
$W_i^0 \subset W$
|
|
-->
|
|
<SPAN CLASS="MATH"><IMG
|
|
WIDTH="73" HEIGHT="39" ALIGN="MIDDLE" BORDER="0"
|
|
SRC="img12.png"
|
|
ALT="$W_i^0 \subset W$"></SPAN> such that
|
|
<!-- MATH
|
|
$\cup_{i=1}^m W_i^0 = W$
|
|
-->
|
|
<SPAN CLASS="MATH"><IMG
|
|
WIDTH="107" HEIGHT="39" ALIGN="MIDDLE" BORDER="0"
|
|
SRC="img13.png"
|
|
ALT="$\cup_{i=1}^m W_i^0 = W$"></SPAN>, a <SPAN CLASS="MATH"><IMG
|
|
WIDTH="13" HEIGHT="15" ALIGN="BOTTOM" BORDER="0"
|
|
SRC="img9.png"
|
|
ALT="$\delta$"></SPAN>-overlap
|
|
partition of <SPAN CLASS="MATH"><IMG
|
|
WIDTH="24" HEIGHT="15" ALIGN="BOTTOM" BORDER="0"
|
|
SRC="img10.png"
|
|
ALT="$W$"></SPAN> is obtained by considering the sets
|
|
<!-- MATH
|
|
$W_i^\delta \supset W_i^{\delta-1}$
|
|
-->
|
|
<SPAN CLASS="MATH"><IMG
|
|
WIDTH="97" HEIGHT="41" ALIGN="MIDDLE" BORDER="0"
|
|
SRC="img14.png"
|
|
ALT="$W_i^\delta \supset W_i^{\delta-1}$"></SPAN> obtained by including the vertices that
|
|
are adjacent to any vertex in <!-- MATH
|
|
$W_i^{\delta-1}$
|
|
-->
|
|
<SPAN CLASS="MATH"><IMG
|
|
WIDTH="48" HEIGHT="41" ALIGN="MIDDLE" BORDER="0"
|
|
SRC="img15.png"
|
|
ALT="$W_i^{\delta-1}$"></SPAN>.
|
|
|
|
<P>
|
|
Let <SPAN CLASS="MATH"><IMG
|
|
WIDTH="22" HEIGHT="39" ALIGN="MIDDLE" BORDER="0"
|
|
SRC="img16.png"
|
|
ALT="$n_i^\delta$"></SPAN> be the size of <SPAN CLASS="MATH"><IMG
|
|
WIDTH="30" HEIGHT="39" ALIGN="MIDDLE" BORDER="0"
|
|
SRC="img17.png"
|
|
ALT="$W_i^\delta$"></SPAN> and <!-- MATH
|
|
$R_i^{\delta} \in
|
|
\Re^{n_i^\delta \times n}$
|
|
-->
|
|
<SPAN CLASS="MATH"><IMG
|
|
WIDTH="93" HEIGHT="43" ALIGN="MIDDLE" BORDER="0"
|
|
SRC="img18.png"
|
|
ALT="$R_i^{\delta} \in
|
|
\Re^{n_i^\delta \times n}$"></SPAN> the restriction operator that maps
|
|
a vector <SPAN CLASS="MATH"><IMG
|
|
WIDTH="56" HEIGHT="34" ALIGN="MIDDLE" BORDER="0"
|
|
SRC="img19.png"
|
|
ALT="$v \in \Re^n$"></SPAN> onto the vector <!-- MATH
|
|
$v_i^{\delta} \in \Re^{n_i^\delta}$
|
|
-->
|
|
<SPAN CLASS="MATH"><IMG
|
|
WIDTH="70" HEIGHT="43" ALIGN="MIDDLE" BORDER="0"
|
|
SRC="img20.png"
|
|
ALT="$v_i^{\delta} \in \Re^{n_i^\delta}$"></SPAN>
|
|
containing the components of <SPAN CLASS="MATH"><IMG
|
|
WIDTH="13" HEIGHT="14" ALIGN="BOTTOM" BORDER="0"
|
|
SRC="img21.png"
|
|
ALT="$v$"></SPAN> corresponding to the vertices in
|
|
<SPAN CLASS="MATH"><IMG
|
|
WIDTH="30" HEIGHT="39" ALIGN="MIDDLE" BORDER="0"
|
|
SRC="img17.png"
|
|
ALT="$W_i^\delta$"></SPAN>. The transpose of <SPAN CLASS="MATH"><IMG
|
|
WIDTH="25" HEIGHT="39" ALIGN="MIDDLE" BORDER="0"
|
|
SRC="img22.png"
|
|
ALT="$R_i^{\delta}$"></SPAN> is a
|
|
prolongation operator from <!-- MATH
|
|
$\Re^{n_i^\delta}$
|
|
-->
|
|
<SPAN CLASS="MATH"><IMG
|
|
WIDTH="32" HEIGHT="24" ALIGN="BOTTOM" BORDER="0"
|
|
SRC="img23.png"
|
|
ALT="$\Re^{n_i^\delta}$"></SPAN> to <SPAN CLASS="MATH"><IMG
|
|
WIDTH="26" HEIGHT="15" ALIGN="BOTTOM" BORDER="0"
|
|
SRC="img24.png"
|
|
ALT="$\Re^n$"></SPAN>.
|
|
The matrix <!-- MATH
|
|
$A_i^\delta=R_i^\delta A (R_i^\delta)^T \in
|
|
\Re^{n_i^\delta \times n_i^\delta}$
|
|
-->
|
|
<SPAN CLASS="MATH"><IMG
|
|
WIDTH="201" HEIGHT="43" ALIGN="MIDDLE" BORDER="0"
|
|
SRC="img25.png"
|
|
ALT="$A_i^\delta=R_i^\delta A (R_i^\delta)^T \in
|
|
\Re^{n_i^\delta \times n_i^\delta}$"></SPAN> can be considered
|
|
as a restriction of <SPAN CLASS="MATH"><IMG
|
|
WIDTH="18" HEIGHT="15" ALIGN="BOTTOM" BORDER="0"
|
|
SRC="img2.png"
|
|
ALT="$A$"></SPAN> corresponding to the set <SPAN CLASS="MATH"><IMG
|
|
WIDTH="31" HEIGHT="39" ALIGN="MIDDLE" BORDER="0"
|
|
SRC="img26.png"
|
|
ALT="$W_i^{\delta}$"></SPAN>.
|
|
|
|
<P>
|
|
The <SPAN CLASS="textit">classical one-level AS</SPAN> preconditioner is defined by
|
|
<BR><P></P>
|
|
<DIV ALIGN="CENTER" CLASS="mathdisplay">
|
|
<!-- MATH
|
|
\begin{displaymath}
|
|
M_{AS}^{-1}= \sum_{i=1}^m (R_i^{\delta})^T
|
|
(A_i^\delta)^{-1} R_i^{\delta},
|
|
\end{displaymath}
|
|
-->
|
|
|
|
<IMG
|
|
WIDTH="206" HEIGHT="58" BORDER="0"
|
|
SRC="img27.png"
|
|
ALT="\begin{displaymath}
|
|
M_{AS}^{-1}= \sum_{i=1}^m (R_i^{\delta})^T
|
|
(A_i^\delta)^{-1} R_i^{\delta},
|
|
\end{displaymath}">
|
|
</DIV>
|
|
<BR CLEAR="ALL">
|
|
<P></P>
|
|
where <SPAN CLASS="MATH"><IMG
|
|
WIDTH="25" HEIGHT="39" ALIGN="MIDDLE" BORDER="0"
|
|
SRC="img28.png"
|
|
ALT="$A_i^\delta$"></SPAN> is assumed to be nonsingular. Its application
|
|
to a vector <SPAN CLASS="MATH"><IMG
|
|
WIDTH="56" HEIGHT="34" ALIGN="MIDDLE" BORDER="0"
|
|
SRC="img19.png"
|
|
ALT="$v \in \Re^n$"></SPAN> within a Krylov solver requires the following
|
|
three steps:
|
|
|
|
<OL>
|
|
<LI>restriction of <SPAN CLASS="MATH"><IMG
|
|
WIDTH="13" HEIGHT="14" ALIGN="BOTTOM" BORDER="0"
|
|
SRC="img21.png"
|
|
ALT="$v$"></SPAN> as <!-- MATH
|
|
$v_i = R_i^{\delta} v$
|
|
-->
|
|
<SPAN CLASS="MATH"><IMG
|
|
WIDTH="71" HEIGHT="39" ALIGN="MIDDLE" BORDER="0"
|
|
SRC="img29.png"
|
|
ALT="$v_i = R_i^{\delta} v$"></SPAN>, <SPAN CLASS="MATH"><IMG
|
|
WIDTH="97" HEIGHT="33" ALIGN="MIDDLE" BORDER="0"
|
|
SRC="img30.png"
|
|
ALT="$i=1,\ldots,m$"></SPAN>;
|
|
</LI>
|
|
<LI>solution of the linear systems <!-- MATH
|
|
$A_i^\delta w_i = v_i$
|
|
-->
|
|
<SPAN CLASS="MATH"><IMG
|
|
WIDTH="80" HEIGHT="39" ALIGN="MIDDLE" BORDER="0"
|
|
SRC="img31.png"
|
|
ALT="$A_i^\delta w_i = v_i$"></SPAN>,
|
|
<SPAN CLASS="MATH"><IMG
|
|
WIDTH="97" HEIGHT="33" ALIGN="MIDDLE" BORDER="0"
|
|
SRC="img30.png"
|
|
ALT="$i=1,\ldots,m$"></SPAN>;
|
|
</LI>
|
|
<LI>prolongation and sum of the <SPAN CLASS="MATH"><IMG
|
|
WIDTH="23" HEIGHT="31" ALIGN="MIDDLE" BORDER="0"
|
|
SRC="img32.png"
|
|
ALT="$w_i$"></SPAN>'s, i.e. <!-- MATH
|
|
$w = \sum_{i=1}^m (R_i^{\delta})^T w_i$
|
|
-->
|
|
<SPAN CLASS="MATH"><IMG
|
|
WIDTH="144" HEIGHT="39" ALIGN="MIDDLE" BORDER="0"
|
|
SRC="img33.png"
|
|
ALT="$w = \sum_{i=1}^m (R_i^{\delta})^T w_i$"></SPAN>.
|
|
</LI>
|
|
</OL>
|
|
Note that the linear systems at step 2 are usually solved approximately,
|
|
e.g. using incomplete LU factorizations such as ILU(<SPAN CLASS="MATH"><IMG
|
|
WIDTH="13" HEIGHT="31" ALIGN="MIDDLE" BORDER="0"
|
|
SRC="img34.png"
|
|
ALT="$p$"></SPAN>), MILU(<SPAN CLASS="MATH"><IMG
|
|
WIDTH="13" HEIGHT="31" ALIGN="MIDDLE" BORDER="0"
|
|
SRC="img34.png"
|
|
ALT="$p$"></SPAN>) and
|
|
ILU(<SPAN CLASS="MATH"><IMG
|
|
WIDTH="27" HEIGHT="31" ALIGN="MIDDLE" BORDER="0"
|
|
SRC="img35.png"
|
|
ALT="$p,t$"></SPAN>) [<A
|
|
HREF="node30.html#Saad_book">19</A>, Chapter 10].
|
|
|
|
<P>
|
|
A variant of the classical AS preconditioner that outperforms it
|
|
in terms of convergence rate and of computation and communication
|
|
time on parallel distributed-memory computers is the so-called <SPAN CLASS="textit">Restricted AS
|
|
(RAS)</SPAN> preconditioner [<A
|
|
HREF="node30.html#CAI_SARKIS">5</A>,<A
|
|
HREF="node30.html#EFSTATHIOU">13</A>]. It
|
|
is obtained by zeroing the components of <SPAN CLASS="MATH"><IMG
|
|
WIDTH="23" HEIGHT="31" ALIGN="MIDDLE" BORDER="0"
|
|
SRC="img32.png"
|
|
ALT="$w_i$"></SPAN> corresponding to the
|
|
overlapping vertices when applying the prolongation. Therefore,
|
|
RAS differs from classical AS by the prolongation operators,
|
|
which are substituted by <!-- MATH
|
|
$(\tilde{R}_i^0)^T \in \Re^{n_i^\delta \times n}$
|
|
-->
|
|
<SPAN CLASS="MATH"><IMG
|
|
WIDTH="118" HEIGHT="43" ALIGN="MIDDLE" BORDER="0"
|
|
SRC="img36.png"
|
|
ALT="$(\tilde{R}_i^0)^T \in \Re^{n_i^\delta \times n}$"></SPAN>,
|
|
where <SPAN CLASS="MATH"><IMG
|
|
WIDTH="25" HEIGHT="42" ALIGN="MIDDLE" BORDER="0"
|
|
SRC="img37.png"
|
|
ALT="$\tilde{R}_i^0$"></SPAN> is obtained by zeroing the rows of <SPAN CLASS="MATH"><IMG
|
|
WIDTH="25" HEIGHT="39" ALIGN="MIDDLE" BORDER="0"
|
|
SRC="img38.png"
|
|
ALT="$R_i^\delta$"></SPAN>
|
|
corresponding to the vertices in <!-- MATH
|
|
$W_i^\delta \backslash W_i^0$
|
|
-->
|
|
<SPAN CLASS="MATH"><IMG
|
|
WIDTH="66" HEIGHT="39" ALIGN="MIDDLE" BORDER="0"
|
|
SRC="img39.png"
|
|
ALT="$W_i^\delta \backslash W_i^0$"></SPAN>:
|
|
<BR><P></P>
|
|
<DIV ALIGN="CENTER" CLASS="mathdisplay">
|
|
<!-- MATH
|
|
\begin{displaymath}
|
|
M_{RAS}^{-1}= \sum_{i=1}^m (\tilde{R}_i^0)^T
|
|
(A_i^\delta)^{-1} R_i^{\delta}.
|
|
\end{displaymath}
|
|
-->
|
|
|
|
<IMG
|
|
WIDTH="216" HEIGHT="58" BORDER="0"
|
|
SRC="img40.png"
|
|
ALT="\begin{displaymath}
|
|
M_{RAS}^{-1}= \sum_{i=1}^m (\tilde{R}_i^0)^T
|
|
(A_i^\delta)^{-1} R_i^{\delta}.
|
|
\end{displaymath}">
|
|
</DIV>
|
|
<BR CLEAR="ALL">
|
|
<P></P>
|
|
Analogously, the AS variant called <SPAN CLASS="textit">AS with Harmonic extension (ASH)</SPAN>
|
|
is defined by
|
|
<BR><P></P>
|
|
<DIV ALIGN="CENTER" CLASS="mathdisplay">
|
|
<!-- MATH
|
|
\begin{displaymath}
|
|
M_{ASH}^{-1}= \sum_{i=1}^m (R_i^{\delta})^T
|
|
(A_i^\delta)^{-1} \tilde{R}_i^0.
|
|
\end{displaymath}
|
|
-->
|
|
|
|
<IMG
|
|
WIDTH="218" HEIGHT="58" BORDER="0"
|
|
SRC="img41.png"
|
|
ALT="\begin{displaymath}M_{ASH}^{-1}= \sum_{i=1}^m (R_i^{\delta})^T
|
|
(A_i^\delta)^{-1} \tilde{R}_i^0.
|
|
\end{displaymath}">
|
|
</DIV>
|
|
<BR CLEAR="ALL">
|
|
<P></P>
|
|
We note that for <SPAN CLASS="MATH"><IMG
|
|
WIDTH="45" HEIGHT="15" ALIGN="BOTTOM" BORDER="0"
|
|
SRC="img42.png"
|
|
ALT="$\delta=0$"></SPAN> the three variants of the AS preconditioner are
|
|
all equal to the block-Jacobi preconditioner.
|
|
|
|
<P>
|
|
As already observed, the convergence rate of the one-level Schwarz
|
|
preconditioned iterative solvers deteriorates as the number <SPAN CLASS="MATH"><IMG
|
|
WIDTH="20" HEIGHT="14" ALIGN="BOTTOM" BORDER="0"
|
|
SRC="img11.png"
|
|
ALT="$m$"></SPAN> of partitions
|
|
of <SPAN CLASS="MATH"><IMG
|
|
WIDTH="24" HEIGHT="15" ALIGN="BOTTOM" BORDER="0"
|
|
SRC="img10.png"
|
|
ALT="$W$"></SPAN> increases [<A
|
|
HREF="node30.html#dd1_94">7</A>,<A
|
|
HREF="node30.html#dd2_96">20</A>]. To reduce the dependency
|
|
of the number of iterations on the degree of parallelism we may
|
|
introduce a global coupling among the overlapping partitions by defining
|
|
a coarse-space approximation <SPAN CLASS="MATH"><IMG
|
|
WIDTH="29" HEIGHT="32" ALIGN="MIDDLE" BORDER="0"
|
|
SRC="img43.png"
|
|
ALT="$A_C$"></SPAN> of the matrix <SPAN CLASS="MATH"><IMG
|
|
WIDTH="18" HEIGHT="15" ALIGN="BOTTOM" BORDER="0"
|
|
SRC="img2.png"
|
|
ALT="$A$"></SPAN>.
|
|
In a pure algebraic setting, <SPAN CLASS="MATH"><IMG
|
|
WIDTH="29" HEIGHT="32" ALIGN="MIDDLE" BORDER="0"
|
|
SRC="img43.png"
|
|
ALT="$A_C$"></SPAN> is usually built with
|
|
a Galerkin approach. Given a set <SPAN CLASS="MATH"><IMG
|
|
WIDTH="32" HEIGHT="32" ALIGN="MIDDLE" BORDER="0"
|
|
SRC="img44.png"
|
|
ALT="$W_C$"></SPAN> of <SPAN CLASS="textit">coarse vertices</SPAN>,
|
|
with size <SPAN CLASS="MATH"><IMG
|
|
WIDTH="26" HEIGHT="31" ALIGN="MIDDLE" BORDER="0"
|
|
SRC="img45.png"
|
|
ALT="$n_C$"></SPAN>, and a suitable restriction operator
|
|
<!-- MATH
|
|
$R_C \in \Re^{n_C \times n}$
|
|
-->
|
|
<SPAN CLASS="MATH"><IMG
|
|
WIDTH="101" HEIGHT="38" ALIGN="MIDDLE" BORDER="0"
|
|
SRC="img46.png"
|
|
ALT="$R_C \in \Re^{n_C \times n}$"></SPAN>, <SPAN CLASS="MATH"><IMG
|
|
WIDTH="29" HEIGHT="32" ALIGN="MIDDLE" BORDER="0"
|
|
SRC="img43.png"
|
|
ALT="$A_C$"></SPAN> is defined as
|
|
<BR><P></P>
|
|
<DIV ALIGN="CENTER" CLASS="mathdisplay">
|
|
<!-- MATH
|
|
\begin{displaymath}
|
|
A_C=R_C A R_C^T
|
|
\end{displaymath}
|
|
-->
|
|
|
|
<IMG
|
|
WIDTH="109" HEIGHT="31" BORDER="0"
|
|
SRC="img47.png"
|
|
ALT="\begin{displaymath}
|
|
A_C=R_C A R_C^T
|
|
\end{displaymath}">
|
|
</DIV>
|
|
<BR CLEAR="ALL">
|
|
<P></P>
|
|
and the coarse-level correction matrix to be combined with a generic
|
|
one-level AS preconditioner <SPAN CLASS="MATH"><IMG
|
|
WIDTH="38" HEIGHT="32" ALIGN="MIDDLE" BORDER="0"
|
|
SRC="img48.png"
|
|
ALT="$M_{1L}$"></SPAN> is obtained as
|
|
<BR><P></P>
|
|
<DIV ALIGN="CENTER" CLASS="mathdisplay">
|
|
<!-- MATH
|
|
\begin{displaymath}
|
|
M_{C}^{-1}= R_C^T A_C^{-1} R_C,
|
|
\end{displaymath}
|
|
-->
|
|
|
|
<IMG
|
|
WIDTH="144" HEIGHT="32" BORDER="0"
|
|
SRC="img49.png"
|
|
ALT="\begin{displaymath}
|
|
M_{C}^{-1}= R_C^T A_C^{-1} R_C,
|
|
\end{displaymath}">
|
|
</DIV>
|
|
<BR CLEAR="ALL">
|
|
<P></P>
|
|
where <SPAN CLASS="MATH"><IMG
|
|
WIDTH="29" HEIGHT="32" ALIGN="MIDDLE" BORDER="0"
|
|
SRC="img43.png"
|
|
ALT="$A_C$"></SPAN> is assumed to be nonsingular. The application of <SPAN CLASS="MATH"><IMG
|
|
WIDTH="42" HEIGHT="41" ALIGN="MIDDLE" BORDER="0"
|
|
SRC="img50.png"
|
|
ALT="$M_{C}^{-1}$"></SPAN>
|
|
to a vector <SPAN CLASS="MATH"><IMG
|
|
WIDTH="13" HEIGHT="14" ALIGN="BOTTOM" BORDER="0"
|
|
SRC="img21.png"
|
|
ALT="$v$"></SPAN> corresponds to a restriction, a solution and
|
|
a prolongation step; the solution step, involving the matrix <SPAN CLASS="MATH"><IMG
|
|
WIDTH="29" HEIGHT="32" ALIGN="MIDDLE" BORDER="0"
|
|
SRC="img43.png"
|
|
ALT="$A_C$"></SPAN>,
|
|
may be carried out also approximately.
|
|
|
|
<P>
|
|
The combination of <SPAN CLASS="MATH"><IMG
|
|
WIDTH="32" HEIGHT="32" ALIGN="MIDDLE" BORDER="0"
|
|
SRC="img51.png"
|
|
ALT="$M_{C}$"></SPAN> and <SPAN CLASS="MATH"><IMG
|
|
WIDTH="38" HEIGHT="32" ALIGN="MIDDLE" BORDER="0"
|
|
SRC="img48.png"
|
|
ALT="$M_{1L}$"></SPAN> may be
|
|
performed in either an additive or a multiplicative framework.
|
|
In the former case, the <SPAN CLASS="textit">two-level additive</SPAN> Schwarz preconditioner
|
|
is obtained:
|
|
<BR><P></P>
|
|
<DIV ALIGN="CENTER" CLASS="mathdisplay">
|
|
<!-- MATH
|
|
\begin{displaymath}
|
|
M_{2LA}^{-1} = M_{C}^{-1} + M_{1L}^{-1}.
|
|
\end{displaymath}
|
|
-->
|
|
|
|
<IMG
|
|
WIDTH="165" HEIGHT="32" BORDER="0"
|
|
SRC="img52.png"
|
|
ALT="\begin{displaymath}
|
|
M_{2LA}^{-1} = M_{C}^{-1} + M_{1L}^{-1}.
|
|
\end{displaymath}">
|
|
</DIV>
|
|
<BR CLEAR="ALL">
|
|
<P></P>
|
|
Applying <SPAN CLASS="MATH"><IMG
|
|
WIDTH="59" HEIGHT="41" ALIGN="MIDDLE" BORDER="0"
|
|
SRC="img53.png"
|
|
ALT="$M_{2L-A}^{-1}$"></SPAN> to a vector <SPAN CLASS="MATH"><IMG
|
|
WIDTH="13" HEIGHT="14" ALIGN="BOTTOM" BORDER="0"
|
|
SRC="img21.png"
|
|
ALT="$v$"></SPAN> within a Krylov solver
|
|
corresponds to applying <SPAN CLASS="MATH"><IMG
|
|
WIDTH="42" HEIGHT="41" ALIGN="MIDDLE" BORDER="0"
|
|
SRC="img50.png"
|
|
ALT="$M_{C}^{-1}$"></SPAN>
|
|
and <SPAN CLASS="MATH"><IMG
|
|
WIDTH="42" HEIGHT="41" ALIGN="MIDDLE" BORDER="0"
|
|
SRC="img54.png"
|
|
ALT="$M_{1L}^{-1}$"></SPAN> to <SPAN CLASS="MATH"><IMG
|
|
WIDTH="13" HEIGHT="14" ALIGN="BOTTOM" BORDER="0"
|
|
SRC="img21.png"
|
|
ALT="$v$"></SPAN> independently and then summing up
|
|
the results.
|
|
|
|
<P>
|
|
In the multiplicative case, the combination can be
|
|
performed by first applying the smoother <SPAN CLASS="MATH"><IMG
|
|
WIDTH="42" HEIGHT="41" ALIGN="MIDDLE" BORDER="0"
|
|
SRC="img54.png"
|
|
ALT="$M_{1L}^{-1}$"></SPAN> and then
|
|
the coarse-level correction operator <SPAN CLASS="MATH"><IMG
|
|
WIDTH="42" HEIGHT="41" ALIGN="MIDDLE" BORDER="0"
|
|
SRC="img50.png"
|
|
ALT="$M_{C}^{-1}$"></SPAN>:
|
|
<BR><P></P>
|
|
<DIV ALIGN="CENTER" CLASS="mathdisplay">
|
|
<!-- MATH
|
|
\begin{displaymath}
|
|
\begin{array}{l}
|
|
w = M_{1L}^{-1} v, \\
|
|
z = w + M_{C}^{-1} (v-Aw);
|
|
\end{array}
|
|
\end{displaymath}
|
|
-->
|
|
|
|
<IMG
|
|
WIDTH="177" HEIGHT="51" BORDER="0"
|
|
SRC="img55.png"
|
|
ALT="\begin{displaymath}
|
|
\begin{array}{l}
|
|
w = M_{1L}^{-1} v, \\
|
|
z = w + M_{C}^{-1} (v-Aw);
|
|
\end{array}\end{displaymath}">
|
|
</DIV>
|
|
<BR CLEAR="ALL">
|
|
<P></P>
|
|
this corresponds to the following <SPAN CLASS="textit">two-level hybrid pre-smoothed</SPAN>
|
|
Schwarz preconditioner:
|
|
<BR><P></P>
|
|
<DIV ALIGN="CENTER" CLASS="mathdisplay">
|
|
<!-- MATH
|
|
\begin{displaymath}
|
|
M_{2LH-PRE}^{-1} = M_{C}^{-1} + \left( I - M_{C}^{-1}A \right) M_{1L}^{-1}.
|
|
\end{displaymath}
|
|
-->
|
|
|
|
<IMG
|
|
WIDTH="308" HEIGHT="33" BORDER="0"
|
|
SRC="img56.png"
|
|
ALT="\begin{displaymath}
|
|
M_{2LH-PRE}^{-1} = M_{C}^{-1} + \left( I - M_{C}^{-1}A \right) M_{1L}^{-1}.
|
|
\end{displaymath}">
|
|
</DIV>
|
|
<BR CLEAR="ALL">
|
|
<P></P>
|
|
On the other hand, by applying the smoother after the coarse-level correction,
|
|
i.e. by computing
|
|
<BR><P></P>
|
|
<DIV ALIGN="CENTER" CLASS="mathdisplay">
|
|
<!-- MATH
|
|
\begin{displaymath}
|
|
\begin{array}{l}
|
|
w = M_{C}^{-1} v , \\
|
|
z = w + M_{1L}^{-1} (v-Aw) ,
|
|
\end{array}
|
|
\end{displaymath}
|
|
-->
|
|
|
|
<IMG
|
|
WIDTH="177" HEIGHT="51" BORDER="0"
|
|
SRC="img57.png"
|
|
ALT="\begin{displaymath}
|
|
\begin{array}{l}
|
|
w = M_{C}^{-1} v , \\
|
|
z = w + M_{1L}^{-1} (v-Aw) ,
|
|
\end{array}\end{displaymath}">
|
|
</DIV>
|
|
<BR CLEAR="ALL">
|
|
<P></P>
|
|
the <SPAN CLASS="textit">two-level hybrid post-smoothed</SPAN>
|
|
Schwarz preconditioner is obtained:
|
|
<BR><P></P>
|
|
<DIV ALIGN="CENTER" CLASS="mathdisplay">
|
|
<!-- MATH
|
|
\begin{displaymath}
|
|
M_{2LH-POST}^{-1} = M_{1L}^{-1} + \left( I - M_{1L}^{-1}A \right) M_{C}^{-1}.
|
|
\end{displaymath}
|
|
-->
|
|
|
|
<IMG
|
|
WIDTH="317" HEIGHT="33" BORDER="0"
|
|
SRC="img58.png"
|
|
ALT="\begin{displaymath}
|
|
M_{2LH-POST}^{-1} = M_{1L}^{-1} + \left( I - M_{1L}^{-1}A \right) M_{C}^{-1}.
|
|
\end{displaymath}">
|
|
</DIV>
|
|
<BR CLEAR="ALL">
|
|
<P></P>
|
|
One more variant of two-level hybrid preconditioner is obtained by applying
|
|
the smoother before and after the coarse-level correction. In this case, the
|
|
preconditioner is symmetric if <SPAN CLASS="MATH"><IMG
|
|
WIDTH="18" HEIGHT="15" ALIGN="BOTTOM" BORDER="0"
|
|
SRC="img2.png"
|
|
ALT="$A$"></SPAN>, <SPAN CLASS="MATH"><IMG
|
|
WIDTH="38" HEIGHT="32" ALIGN="MIDDLE" BORDER="0"
|
|
SRC="img48.png"
|
|
ALT="$M_{1L}$"></SPAN> and <SPAN CLASS="MATH"><IMG
|
|
WIDTH="32" HEIGHT="32" ALIGN="MIDDLE" BORDER="0"
|
|
SRC="img51.png"
|
|
ALT="$M_{C}$"></SPAN> are symmetric.
|
|
|
|
<P>
|
|
As previously noted, on parallel computers the number of submatrices usually matches
|
|
the number of available processors. When the size of the system to be preconditioned
|
|
is very large, the use of many processors, i.e. of many small submatrices, often
|
|
leads to a large coarse-level system, whose solution may be computationally expensive.
|
|
On the other hand, the use of few processors often leads to local sumatrices that
|
|
are too expensive to be processed on single processors, because of memory and/or
|
|
computing requirements. Therefore, it seems natural to use a recursive approach,
|
|
in which the coarse-level correction is re-applied starting from the current
|
|
coarse-level system. The corresponding preconditioners, called <SPAN CLASS="textit">multi-level</SPAN>
|
|
preconditioners, can significantly reduce the computational cost of preconditioning
|
|
with respect to the two-level case (see [<A
|
|
HREF="node30.html#dd2_96">20</A>, Chapter 3]).
|
|
Additive and hybrid multilevel preconditioners
|
|
are obtained as direct extensions of the two-level counterparts.
|
|
For a detailed descrition of them, the reader is
|
|
referred to [<A
|
|
HREF="node30.html#dd2_96">20</A>, Chapter 3].
|
|
The algorithm for the application of a multi-level hybrid
|
|
post-smoothed preconditioner <SPAN CLASS="MATH"><IMG
|
|
WIDTH="23" HEIGHT="15" ALIGN="BOTTOM" BORDER="0"
|
|
SRC="img59.png"
|
|
ALT="$M$"></SPAN> to a vector <SPAN CLASS="MATH"><IMG
|
|
WIDTH="13" HEIGHT="14" ALIGN="BOTTOM" BORDER="0"
|
|
SRC="img21.png"
|
|
ALT="$v$"></SPAN>, i.e. for the
|
|
computation of <SPAN CLASS="MATH"><IMG
|
|
WIDTH="87" HEIGHT="21" ALIGN="BOTTOM" BORDER="0"
|
|
SRC="img60.png"
|
|
ALT="$w=M^{-1}v$"></SPAN>, is reported, for
|
|
example, in Figure <A HREF="#fig:mlhpost_alg">1</A>. Here the number of levels
|
|
is denoted by <SPAN CLASS="MATH"><IMG
|
|
WIDTH="37" HEIGHT="15" ALIGN="BOTTOM" BORDER="0"
|
|
SRC="img61.png"
|
|
ALT="$nlev$"></SPAN> and the levels are numbered in increasing order starting
|
|
from the finest one, i.e. the finest level is level 1; the coarse matrix
|
|
and the corresponding basic preconditioner at each level <SPAN CLASS="MATH"><IMG
|
|
WIDTH="10" HEIGHT="15" ALIGN="BOTTOM" BORDER="0"
|
|
SRC="img62.png"
|
|
ALT="$l$"></SPAN> are denoted by <SPAN CLASS="MATH"><IMG
|
|
WIDTH="22" HEIGHT="32" ALIGN="MIDDLE" BORDER="0"
|
|
SRC="img63.png"
|
|
ALT="$A_l$"></SPAN> and
|
|
<SPAN CLASS="MATH"><IMG
|
|
WIDTH="27" HEIGHT="32" ALIGN="MIDDLE" BORDER="0"
|
|
SRC="img64.png"
|
|
ALT="$M_l$"></SPAN>, respectively, with <SPAN CLASS="MATH"><IMG
|
|
WIDTH="61" HEIGHT="32" ALIGN="MIDDLE" BORDER="0"
|
|
SRC="img65.png"
|
|
ALT="$A_1=A$"></SPAN>.
|
|
|
|
<DIV ALIGN="CENTER"><A NAME="fig:mlhpost_alg"></A><A NAME="508"></A>
|
|
<TABLE>
|
|
<CAPTION ALIGN="BOTTOM"><STRONG>Figure 1:</STRONG>
|
|
Application of the multi-level hybrid post-smoothed preconditioner.</CAPTION>
|
|
<TR><TD>
|
|
<DIV ALIGN="CENTER">
|
|
<!-- MATH
|
|
$\framebox{
|
|
\begin{minipage}{.85\textwidth} {\small
|
|
\begin{tabbing}
|
|
\quad \=\quad \=\quad \=\quad \\[-1mm]
|
|
$v_1 = v$; \\[2mm]
|
|
\textbf{for $l=2, nlev$\ do}\\[1mm]
|
|
\> ! transfer $v_{l-1}$\ to the next coarser level\\
|
|
\> $v_l = R_lv_{l-1}$\ \\[1mm]
|
|
\textbf{endfor} \\[2mm]
|
|
! apply the coarsest-level correction\\[1mm]
|
|
$y_{nlev} = A_{nlev}^{-1} v_{nlev}$\\[2mm]
|
|
\textbf{for $l=nlev -1 , 1, -1$\ do}\\[1mm]
|
|
\> ! transfer $y_{l+1}$\ to the next finer level\\
|
|
\> $y_l = R_{l+1}^T y_{l+1}$;\\[1mm]
|
|
\> ! compute the residual at the current level\\
|
|
\> $r_l = v_l-A_l^{-1} y_l$;\\[1mm]
|
|
\> ! apply the basic Schwarz preconditioner to the residual\\
|
|
\> $r_l = M_l^{-1} r_l$\\[1mm]
|
|
\> ! update $y_l$\\
|
|
\> $y_l = y_l+r_l$\\
|
|
\textbf{endfor} \\[1mm]
|
|
$w = y_1$;
|
|
\end{tabbing}
|
|
}
|
|
\end{minipage}
|
|
}$
|
|
-->
|
|
<IMG
|
|
WIDTH="430" HEIGHT="435" ALIGN="BOTTOM" BORDER="0"
|
|
SRC="img66.png"
|
|
ALT="\framebox{
|
|
\begin{minipage}{.85\textwidth} {\small
|
|
\begin{tabbing}
|
|
\quad \=\quad...
|
|
...= y_l+r_l$\\
|
|
\textbf{endfor} [1mm]
|
|
$w = y_1$;
|
|
\end{tabbing}}
|
|
\end{minipage}}">
|
|
|
|
</DIV></TD></TR>
|
|
</TABLE>
|
|
</DIV>
|
|
|
|
<P>
|
|
|
|
<DIV CLASS="navigation"><HR>
|
|
<!--Navigation Panel-->
|
|
<A NAME="tex2html200"
|
|
HREF="node12.html">
|
|
<IMG WIDTH="37" HEIGHT="24" ALIGN="BOTTOM" BORDER="0" ALT="next"
|
|
SRC="file:/usr/share/latex2html/icons/next.png"></A>
|
|
<A NAME="tex2html196"
|
|
HREF="node10.html">
|
|
<IMG WIDTH="26" HEIGHT="24" ALIGN="BOTTOM" BORDER="0" ALT="up"
|
|
SRC="file:/usr/share/latex2html/icons/up.png"></A>
|
|
<A NAME="tex2html190"
|
|
HREF="node10.html">
|
|
<IMG WIDTH="63" HEIGHT="24" ALIGN="BOTTOM" BORDER="0" ALT="previous"
|
|
SRC="file:/usr/share/latex2html/icons/prev.png"></A>
|
|
<A NAME="tex2html198"
|
|
HREF="node2.html">
|
|
<IMG WIDTH="65" HEIGHT="24" ALIGN="BOTTOM" BORDER="0" ALT="contents"
|
|
SRC="file:/usr/share/latex2html/icons/contents.png"></A>
|
|
<BR>
|
|
<B> Next:</B> <A NAME="tex2html201"
|
|
HREF="node12.html">Smoothed Aggregation</A>
|
|
<B> Up:</B> <A NAME="tex2html197"
|
|
HREF="node10.html">Multi-level Domain Decomposition Background</A>
|
|
<B> Previous:</B> <A NAME="tex2html191"
|
|
HREF="node10.html">Multi-level Domain Decomposition Background</A>
|
|
<B> <A NAME="tex2html199"
|
|
HREF="node2.html">Contents</A></B> </DIV>
|
|
<!--End of Navigation Panel-->
|
|
<ADDRESS>
|
|
Salvatore Filippone
|
|
2008-07-23
|
|
</ADDRESS>
|
|
</BODY>
|
|
</HTML>
|