F04BFF (PDF version)
F04 Chapter Contents
F04 Chapter Introduction
NAG Library Manual

NAG Library Routine Document

F04BFF

Note:  before using this routine, please read the Users' Note for your implementation to check the interpretation of bold italicised terms and other implementation-dependent details.

+ Contents

    1  Purpose
    7  Accuracy

1  Purpose

F04BFF computes the solution to a real system of linear equations AX=B, where A is an n by n symmetric positive definite band matrix of band width 2k+1, and X and B are n by r matrices. An estimate of the condition number of A and an error bound for the computed solution are also returned.

2  Specification

SUBROUTINE F04BFF ( UPLO, N, KD, NRHS, AB, LDAB, B, LDB, RCOND, ERRBND, IFAIL)
INTEGER  N, KD, NRHS, LDAB, LDB, IFAIL
REAL (KIND=nag_wp)  AB(LDAB,*), B(LDB,*), RCOND, ERRBND
CHARACTER(1)  UPLO

3  Description

The Cholesky factorization is used to factor A as A=UTU, if UPLO='U', or A=LLT, if UPLO='L', where U is an upper triangular band matrix with k superdiagonals, and L is a lower triangular band matrix with k subdiagonals. The factored form of A is then used to solve the system of equations AX=B.

4  References

Anderson E, Bai Z, Bischof C, Blackford S, Demmel J, Dongarra J J, Du Croz J J, Greenbaum A, Hammarling S, McKenney A and Sorensen D (1999) LAPACK Users' Guide (3rd Edition) SIAM, Philadelphia http://www.netlib.org/lapack/lug
Higham N J (2002) Accuracy and Stability of Numerical Algorithms (2nd Edition) SIAM, Philadelphia

5  Parameters

1:     UPLO – CHARACTER(1)Input
On entry: if UPLO='U', the upper triangle of the matrix A is stored.
If UPLO='L', the lower triangle of the matrix A is stored.
Constraint: UPLO='U' or 'L'.
2:     N – INTEGERInput
On entry: the number of linear equations n, i.e., the order of the matrix A.
Constraint: N0.
3:     KD – INTEGERInput
On entry: the number of superdiagonals k (and the number of subdiagonals) of the band matrix A.
Constraint: KD0.
4:     NRHS – INTEGERInput
On entry: the number of right-hand sides r, i.e., the number of columns of the matrix B.
Constraint: NRHS0.
5:     AB(LDAB,*) – REAL (KIND=nag_wp) arrayInput/Output
Note: the second dimension of the array AB must be at least max1,N.
On entry: the n by n symmetric band matrix A. The upper or lower triangular part of the symmetric matrix is stored in the first KD+1 rows of the array. The jth column of A is stored in the jth column of the array AB as follows:
  • if UPLO='U', AB k+1+i-j ,j =aij for max1,j-kij;
  • if UPLO='L', AB 1+i-j,j =aij for jiminn,j+k.
See Section 8 below for further details.
On exit: if IFAIL=0 or N+1, the factor U or L from the Cholesky factorization A=UTU or A=LLT, in the same storage format as A.
6:     LDAB – INTEGERInput
On entry: the first dimension of the array AB as declared in the (sub)program from which F04BFF is called.
Constraint: LDABKD+1.
7:     B(LDB,*) – REAL (KIND=nag_wp) arrayInput/Output
Note: the second dimension of the array B must be at least max1,NRHS.
On entry: the n by r matrix of right-hand sides B.
On exit: if IFAIL=0 or N+1, the n by r solution matrix X.
8:     LDB – INTEGERInput
On entry: the first dimension of the array B as declared in the (sub)program from which F04BFF is called.
Constraint: LDBmax1,N.
9:     RCOND – REAL (KIND=nag_wp)Output
On exit: if IFAIL=0 or N+1, an estimate of the reciprocal of the condition number of the matrix A, computed as RCOND=1/A1A-11.
10:   ERRBND – REAL (KIND=nag_wp)Output
On exit: if IFAIL=0 or N+1, an estimate of the forward error bound for a computed solution x^, such that x^-x1/x1ERRBND, where x^ is a column of the computed solution returned in the array B and x is the corresponding column of the exact solution X. If RCOND is less than machine precision, then ERRBND is returned as unity.
11:   IFAIL – INTEGERInput/Output
On entry: IFAIL must be set to 0, -1​ or ​1. 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 -1​ or ​1 is recommended. If the output of error messages is undesirable, then the value 1 is recommended. Otherwise, if you are not familiar with this parameter, the recommended value is 0. When the value -1​ or ​1 is used it is essential to test the value of IFAIL on exit.
On exit: IFAIL=0 unless the routine detects an error or a warning has been flagged (see Section 6).

6  Error Indicators and Warnings

If on entry IFAIL=0 or -1, explanatory error messages are output on the current error message unit (as defined by X04AAF).
Errors or warnings detected by the routine:
IFAIL<0 and IFAIL-999
If IFAIL=-i, the ith argument had an illegal value.
IFAIL=-999
Allocation of memory failed. The integer allocatable memory required is N, and the real allocatable memory required is 3×N. Allocation failed before the solution could be computed.
IFAIL>0 and IFAILN
If IFAIL=i, the leading minor of order i of A is not positive definite. The factorization could not be completed, and the solution has not been computed.
IFAIL=N+1
RCOND is less than machine precision, so that the matrix A is numerically singular. A solution to the equations AX=B has nevertheless been computed.

7  Accuracy

The computed solution for a single right-hand side, x^, satisfies an equation of the form
A+E x^=b,
where
E1=Oε A1
and ε is the machine precision. An approximate error bound for the computed solution is given by
x^-x1 x1 κA E1 A1 ,
where κA=A-11A1, the condition number of A with respect to the solution of the linear equations. F04BFF uses the approximation E1=εA1 to estimate ERRBND. See Section 4.4 of Anderson et al. (1999) for further details.

8  Further Comments

The band storage scheme for the array AB is illustrated by the following example, when n=6, k=2, and UPLO='U':
On entry:
* * a13 a24 a35 a46 * a12 a23 a34 a45 a56 a11 a22 a33 a44 a55 a66
On exit:
* * u13 u24 u35 u46 * u12 u23 u34 u45 u56 u11 u22 u33 u44 u55 u66
Similarly, if UPLO='L' the format of AB is as follows:
On entry:
a11 a22 a33 a44 a55 a66 a21 a32 a43 a54 a65 * a31 a42 a53 a64 * *
On exit:
l11 l22 l33 l44 l55 l66 l21 l32 l43 l54 l65 * l31 l42 l53 l64 * *
Array elements marked * need not be set and are not referenced by the routine.
Assuming that nk, the total number of floating point operations required to solve the equations AX=B is approximately nk+12 for the factorization and 4nkr for the solution following the factorization. The condition number estimation typically requires between four and five solves and never more than eleven solves, following the factorization.
In practice the condition number estimator is very reliable, but it can underestimate the true condition number; see Section 15.3 of Higham (2002) for further details.
The complex analogue of F04BFF is F04CFF.

9  Example

This example solves the equations
AX=B,
where A is the symmetric positive definite band matrix
A= 5.49 2.68 0 0 2.68 5.63 -2.39 0 0 -2.39 2.60 -2.22 0 0 -2.22 5.17   and   B= 22.09 5.10 9.31 30.81 -5.24 -25.82 11.83 22.90 .
An estimate of the condition number of A and an approximate error bound for the computed solutions are also printed.

9.1  Program Text

Program Text (f04bffe.f90)

9.2  Program Data

Program Data (f04bffe.d)

9.3  Program Results

Program Results (f04bffe.r)


F04BFF (PDF version)
F04 Chapter Contents
F04 Chapter Introduction
NAG Library Manual

© The Numerical Algorithms Group Ltd, Oxford, UK. 2012