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

Parameters

Compulsory Input Parameters

1: $uplo$ – string (length ≥ 1)

Specifies how

A

has been factorized.

$uplo ='U'$: $A = P U D U^{T} P^{T}$ , where $U$ is upper triangular.
$uplo ='L'$: $A = P L D L^{T} P^{T}$ , where $L$ is lower triangular.

Constraint:

uplo ='U'

'L'

2: $a (lda :)$ – complex array

The first dimension of the array a must be at least

\max (1, n)

The second dimension of the array a must be at least

\max (1, n)

Details of the factorization of

A

, as returned by nag_lapack_zsytrf (f07nr).

3: $ipiv (:)$ – int64int32nag_int array

The dimension of the array ipiv must be at least

\max (1, n)

Details of the interchanges and the block structure of

D

, as returned by nag_lapack_zsytrf (f07nr).

4: $anorm$ – double scalar

The

1

-norm of the original matrix

A

. anorm must be computed either before calling nag_lapack_zsytrf (f07nr) or else from a copy of the original matrix

A

Constraint:

anorm \geq 0.0

Optional Input Parameters

1: $n$ – int64int32nag_int scalar: Default: the first dimension of the array a and the second dimension of the arrays a, ipiv.
$n$ , the order of the matrix $A$ .

Constraint: $n \geq 0$ .

Output Parameters

1: $rcond$ – double scalar: An estimate of the reciprocal of the condition number of $A$ . rcond is set to zero if exact singularity is detected or the estimate underflows. If rcond is less than machine precision, $A$ is singular to working precision.
2: $info$ – int64int32nag_int scalar: $info = 0$ unless the function detects an error (see Error Indicators and Warnings).

Error Indicators and Warnings

$info < 0$: If $info = - i$ , argument $i$ had an illegal value. An explanatory message is output, and execution of the program is terminated.

Accuracy

The computed estimate rcond is never less than the true value

ρ

, and in practice is nearly always less than

10 ρ

, although examples can be constructed where rcond is much larger.

Further Comments

A call to nag_lapack_zsycon (f07nu) involves solving a number of systems of linear equations of the form

A x = b

; the number is usually

5

and never more than

11

. Each solution involves approximately

8 n^{2}

real floating-point operations but takes considerably longer than a call to nag_lapack_zsytrs (f07ns) with one right-hand side, because extra care is taken to avoid overflow when

A

is approximately singular.

The real analogue of this function is nag_lapack_dsycon (f07mg).

Example

This example estimates the condition number in the

1

-norm (or

\infty

-norm) of the matrix

A

, where

A = (\begin{array}{r} - 0.39 - 0.71 i & 5.14 - 0.64 i & - 7.86 - 2.96 i & 3.80 + 0.92 i \\ 5.14 - 0.64 i & 8.86 + 1.81 i & - 3.52 + 0.58 i & 5.32 - 1.59 i \\ - 7.86 - 2.96 i & - 3.52 + 0.58 i & - 2.83 - 0.03 i & - 1.54 - 2.86 i \\ 3.80 + 0.92 i & 5.32 - 1.59 i & - 1.54 - 2.86 i & - 0.56 + 0.12 i \end{array}) .

Here

A

is symmetric and must first be factorized by nag_lapack_zsytrf (f07nr). The true condition number in the

1

-norm is

32.92

Open in the MATLAB editor: f07nu_example

function f07nu_example


fprintf('f07nu example results\n\n');

% Complex symmetrix matrix A, lower triangle stored.
uplo = 'L';
a = [-0.39 - 0.71i,  0    + 0i,     0    + 0i,     0    + 0i;
      5.14 - 0.64i,  8.86 + 1.81i,  0    + 0i,     0    + 0i;
     -7.86 - 2.96i, -3.52 + 0.58i, -2.83 - 0.03i,  0    + 0i;
      3.80 + 0.92i,  5.32 - 1.59i, -1.54 - 2.86i, -0.56 + 0.12i];

%Factorize A
[af, ipiv, info] = f07nr( ...
                          uplo, a);

% Norm of A
an = a + a' - diag(diag(a));
anorm = norm(an,1);

% Condition number estimator
[rcond, info] = f07nu( ...
                       uplo, af, ipiv, anorm);

fprintf('Estimate of condition number = %9.2e\n', 1/rcond);

f07nu example results

Estimate of condition number =  2.06e+01

PDF version (NAG web site, 64-bit version, 64-bit version)

Chapter Contents

Chapter Introduction

NAG Toolbox