f01maf
f01maf
© Numerical Algorithms Group, 2002.
Purpose
F01MAF Incomplete Cholesky factorization of real sparse symmetric
positive-definite matrix
Synopsis
[a,irn,icn,wkeep,ikeep,inform,droptl,densw,ifail] = f01maf(n,nz,a,irn,icn<,...
droptl,densw,abort,ifail>)
Description
F01MAF computes an incomplete Cholesky factorization
T T
C=PLDL P , WAW=C+E
for the sparse symmetric positive-definite matrix A, where P is a
permutation matrix, L is a unit lower triangular matrix, D is a
diagonal matrix with positive diagonal elements, E is an error
matrix representing elements dropped during the factorization and
diagonal elements that have been modified to ensure that C is
positive-definite, and W is a diagonal matrix, chosen to make the
diagonal elements of WAW unity.
-1 -1
W CW is a pre-conditioning matrix for A, and the factorization
of C is intended to be used by F04MAF to solve systems of linear
equations Ax=b.
The permutation matrix P is chosen to reduce the amount of fill-
in that occurs in L and the user-supplied parameter DROPTL can
also be used to control the amount of fill-in that occurs.
Parameters
f01maf
Required Input Arguments:
n integer
nz integer
a (:) real
irn (:) integer
icn (:) integer
Optional Input Arguments: <Default>
droptl real 0.1
densw real 0.8
abort (3) logical [1;1;1]
ifail integer -1
Output Arguments:
a (:) real
irn (:) integer
icn (:) integer
wkeep (n,3) real
ikeep (n,2) integer
inform (4) integer
droptl real
densw real
ifail integer