NAG Library Routine Document
F04YCF
1 Purpose
F04YCF estimates the -norm of a real matrix without accessing the matrix explicitly. It uses reverse communication for evaluating matrix-vector products. The routine may be used for estimating matrix condition numbers.
2 Specification
INTEGER |
ICASE, N, IWORK(N), IFAIL |
REAL (KIND=nag_wp) |
X(N), ESTNRM, WORK(N) |
|
3 Description
F04YCF computes an estimate (a lower bound) for the
-norm
of an
by
real matrix
. The routine regards the matrix
as being defined by a user-supplied ‘Black Box’ which, given an input vector
, can return either of the matrix-vector products
or
. A reverse communication interface is used; thus control is returned to the calling program whenever a matrix-vector product is required.
Note: this routine is
not recommended for use when the elements of
are known explicitly; it is then more efficient to compute the
-norm directly from formula
(1) above.
The main
use of the routine is for estimating , and hence the condition number
, without forming explicitly ( above).
If, for example, an factorization of is available, the matrix-vector products and required by F04YCF may be computed by back- and forward-substitutions, without computing .
The routine can also be used to estimate
-norms of matrix products such as
and
, without forming the products explicitly. Further applications are described by
Higham (1988).
Since , F04YCF can be used to estimate the -norm of by working with instead of .
The algorithm used is based on a method given by
Hager (1984) and is described by
Higham (1988). A comparison of several techniques for condition number estimation is given by
Higham (1987).
Note: F04YDF can also be used to estimate the norm of a real matrix.
F04YDF uses a more recent algorithm than F04YCF and it is recommended that
F04YDF be used in place of F04YCF.
4 References
Hager W W (1984) Condition estimates SIAM J. Sci. Statist. Comput. 5 311–316
Higham N J (1987) A survey of condition number estimation for triangular matrices SIAM Rev. 29 575–596
Higham N J (1988) FORTRAN codes for estimating the one-norm of a real or complex matrix, with applications to condition estimation ACM Trans. Math. Software 14 381–396
5 Parameters
Note: this routine uses
reverse communication. Its use involves an initial entry, intermediate exits and re-entries, and a final exit, as indicated by the parameter
ICASE. Between intermediate exits and re-entries,
all parameters other than X must remain unchanged.
- 1: ICASE – INTEGERInput/Output
On initial entry: must be set to .
On intermediate exit:
or
, and
, for
, contain the elements of a vector
. The calling program must
(a) |
evaluate (if ) or (if ), |
(b) |
place the result in X, and |
(c) |
call F04YCF once again, with all the other parameters unchanged. |
On final exit: .
- 2: N – INTEGERInput
On initial entry: , the order of the matrix .
Constraint:
.
- 3: X(N) – REAL (KIND=nag_wp) arrayInput/Output
On initial entry: need not be set.
On intermediate exit:
contains the current vector .
On intermediate re-entry: must contain (if ) or (if ).
On final exit: the array is undefined.
- 4: ESTNRM – REAL (KIND=nag_wp)Input/Output
On initial entry: need not be set.
On intermediate exit:
should not be changed.
On final exit: an estimate (a lower bound) for .
- 5: WORK(N) – REAL (KIND=nag_wp) arrayInput/Output
On initial entry: need not be set.
On final exit: contains a vector
such that
where
(
is not returned). If
and
ESTNRM is large, then
is an approximate null vector for
.
- 6: IWORK(N) – INTEGER arrayCommunication Array
- 7: IFAIL – INTEGERInput/Output
On initial entry:
IFAIL must be set to
,
. If you are unfamiliar with this parameter you should refer to
Section 3.3 in the Essential Introduction for details.
For environments where it might be inappropriate to halt program execution when an error is detected, the value
is recommended. If the output of error messages is undesirable, then the value
is recommended. Otherwise, because for this routine the values of the output parameters may be useful even if
on exit, the recommended value is
.
When the value is used it is essential to test the value of IFAIL on exit.
On final exit:
unless the routine detects an error or a warning has been flagged (see
Section 6).
6 Error Indicators and Warnings
If on entry
or
, explanatory error messages are output on the current error message unit (as defined by
X04AAF).
Errors or warnings detected by the routine:
On entry, .
7 Accuracy
In extensive tests on
random matrices of size up to
the estimate
ESTNRM has been found always to be within a factor eleven of
; often the estimate has many correct figures. However, matrices exist for which the estimate is smaller than
by an arbitrary factor; such matrices are very unlikely to arise in practice. See
Higham (1988) for further details.
8.1 Timing
The total time taken within F04YCF is proportional to . For most problems the time taken during calls to F04YCF will be negligible compared with the time spent evaluating matrix-vector products between calls to F04YCF.
The number of matrix-vector products required varies from to (or is if ). In most cases or products are required; it is rare for more than to be needed.
8.2 Overflow
It is your responsibility to guard against potential overflows during evaluation of the matrix-vector products. In particular, when estimating using a triangular factorization of , F04YCF should not be called if one of the factors is exactly singular – otherwise division by zero may occur in the substitutions.
8.3 Use in Conjunction with NAG Library Routines
To estimate the
-norm of the inverse of a matrix
, the following skeleton code can normally be used:
... code to factorize A ...
IF (A is not singular) THEN
ICASE = 0
10 CALL F04YCF (ICASE,N,X,ESTNRM,WORK,IWORK,IFAIL)
IF (ICASE.NE.0) THEN
IF (ICASE.EQ.1) THEN
... code to compute inv(A)*x ...
ELSE
... code to compute inv(transpose(A))*x ...
END IF
GO TO 10
END IF
END IF
To compute or , solve the equation or for , overwriting on . The code will vary, depending on the type of the matrix , and the NAG routine used to factorize .
Note that if
is any type of
symmetric matrix, then
, and the code following the call of F04YCF can be reduced to:
IF (ICASE.NE.0) THEN
... code to compute inv(A)*x ...
GO TO 10
END IF
The factorization will normally have been performed by a suitable routine from
Chapters F01,
F03 or
F07. Note also that many of the ‘Black Box’ routines in
Chapter F04 for solving systems of equations also return a factorization of the matrix. The example program in
Section 9 illustrates how F04YCF can be used in conjunction with NAG Library routines for two important types of matrix: full nonsymmetric matrices (factorized by
F07ADF (DGETRF)) and sparse nonsymmetric matrices (factorized by
F01BRF).
It is straightforward to use F04YCF for the following other types of matrix, using the named routines for factorization and solution:
For upper or lower triangular matrices, no factorization routine is needed:
and
may be computed by calls to
F06PJF (DTRSV) (or
F06PKF (DTBSV) if the matrix is banded, or
F06PLF (DTPSV) if the matrix is stored in packed form).
9 Example
For this routine two examples are presented. There is a single example program for F04YCF, with a main program and the code to solve the two example problems is given in Example 1 (EX1) and Example 2 (EX2).
Example 1 (EX1)
To estimate the condition number
of the matrix
given by
Example 2 (EX2)
To estimate the condition number
of the matrix
given by
9.1 Program Text
Program Text (f04ycfe.f90)
9.2 Program Data
Program Data (f04ycfe.d)
9.3 Program Results
Program Results (f04ycfe.r)