Purpose
To compute a solution, optionally corresponding to specified free elements, to a real linear least squares problem: minimize || A * X - B || using a complete orthogonal factorization of the M-by-N matrix A, which may be rank-deficient. Several right hand side vectors b and solution vectors x can be handled in a single call; they are stored as the columns of the M-by-NRHS right hand side matrix B and the N-by-NRHS solution matrix X.Specification
SUBROUTINE MB02QD( JOB, INIPER, M, N, NRHS, RCOND, SVLMAX, A, LDA, $ B, LDB, Y, JPVT, RANK, SVAL, DWORK, LDWORK, $ INFO ) C .. Scalar Arguments .. CHARACTER INIPER, JOB INTEGER INFO, LDA, LDB, LDWORK, M, N, NRHS, RANK DOUBLE PRECISION RCOND, SVLMAX C .. Array Arguments .. INTEGER JPVT( * ) DOUBLE PRECISION A( LDA, * ), B( LDB, * ), DWORK( * ), $ SVAL( 3 ), Y ( * )Arguments
Mode Parameters
JOB CHARACTER*1 Specifies whether or not a standard least squares solution must be computed, as follows: = 'L': Compute a standard least squares solution (Y = 0); = 'F': Compute a solution with specified free elements (given in Y). INIPER CHARACTER*1 Specifies whether an initial column permutation, defined by JPVT, must be performed, as follows: = 'P': Perform an initial column permutation; = 'N': Do not perform an initial column permutation.Input/Output Parameters
M (input) INTEGER The number of rows of the matrix A. M >= 0. N (input) INTEGER The number of columns of the matrix A. N >= 0. NRHS (input) INTEGER The number of right hand sides, i.e., the number of columns of the matrices B and X. NRHS >= 0. RCOND (input) DOUBLE PRECISION RCOND is used to determine the effective rank of A, which is defined as the order of the largest leading triangular submatrix R11 in the QR factorization with pivoting of A, whose estimated condition number is less than 1/RCOND. 0 <= RCOND <= 1. SVLMAX (input) DOUBLE PRECISION If A is a submatrix of another matrix C, and the rank decision should be related to that matrix, then SVLMAX should be an estimate of the largest singular value of C (for instance, the Frobenius norm of C). If this is not the case, the input value SVLMAX = 0 should work. SVLMAX >= 0. A (input/output) DOUBLE PRECISION array, dimension (LDA,N) On entry, the leading M-by-N part of this array must contain the given matrix A. On exit, the leading M-by-N part of this array contains details of its complete orthogonal factorization: the leading RANK-by-RANK upper triangular part contains the upper triangular factor T11 (see METHOD); the elements below the diagonal, with the entries 2 to min(M,N)+1 of the array DWORK, represent the orthogonal matrix Q as a product of min(M,N) elementary reflectors (see METHOD); the elements of the subarray A(1:RANK,RANK+1:N), with the next RANK entries of the array DWORK, represent the orthogonal matrix Z as a product of RANK elementary reflectors (see METHOD). LDA INTEGER The leading dimension of the array A. LDA >= max(1,M). B (input/output) DOUBLE PRECISION array, dimension (LDB,NRHS) On entry, the leading M-by-NRHS part of this array must contain the right hand side matrix B. On exit, the leading N-by-NRHS part of this array contains the solution matrix X. If M >= N and RANK = N, the residual sum-of-squares for the solution in the i-th column is given by the sum of squares of elements N+1:M in that column. If NRHS = 0, this array is not referenced, and the routine returns the effective rank of A, and its QR factorization. LDB INTEGER The leading dimension of the array B. LDB >= max(1,M,N). Y (input) DOUBLE PRECISION array, dimension ( N*NRHS ) If JOB = 'F', the elements Y(1:(N-RANK)*NRHS) are used as free elements in computing the solution (see METHOD). The remaining elements are not referenced. If JOB = 'L', or NRHS = 0, this array is not referenced. JPVT (input/output) INTEGER array, dimension (N) On entry with INIPER = 'P', if JPVT(i) <> 0, the i-th column of A is an initial column, otherwise it is a free column. Before the QR factorization of A, all initial columns are permuted to the leading positions; only the remaining free columns are moved as a result of column pivoting during the factorization. If INIPER = 'N', JPVT need not be set on entry. On exit, if JPVT(i) = k, then the i-th column of A*P was the k-th column of A. RANK (output) INTEGER The effective rank of A, i.e., the order of the submatrix R11. This is the same as the order of the submatrix T11 in the complete orthogonal factorization of A. SVAL (output) DOUBLE PRECISION array, dimension ( 3 ) The estimates of some of the singular values of the triangular factor R11: SVAL(1): largest singular value of R(1:RANK,1:RANK); SVAL(2): smallest singular value of R(1:RANK,1:RANK); SVAL(3): smallest singular value of R(1:RANK+1,1:RANK+1), if RANK < MIN( M, N ), or of R(1:RANK,1:RANK), otherwise. If the triangular factorization is a rank-revealing one (which will be the case if the leading columns were well- conditioned), then SVAL(1) will also be an estimate for the largest singular value of A, and SVAL(2) and SVAL(3) will be estimates for the RANK-th and (RANK+1)-st singular values of A, respectively. By examining these values, one can confirm that the rank is well defined with respect to the chosen value of RCOND. The ratio SVAL(1)/SVAL(2) is an estimate of the condition number of R(1:RANK,1:RANK).Workspace
DWORK DOUBLE PRECISION array, dimension (LDWORK) On exit, if INFO = 0, DWORK(1) returns the optimal value of LDWORK, and the entries 2 to min(M,N) + RANK + 1 contain the scalar factors of the elementary reflectors used in the complete orthogonal factorization of A. Among the entries 2 to min(M,N) + 1, only the first RANK elements are useful, if INIPER = 'N'. LDWORK INTEGER The length of the array DWORK. LDWORK >= max( min(M,N)+3*N+1, 2*min(M,N)+NRHS ) For optimum performance LDWORK should be larger.Error Indicator
INFO INTEGER = 0: successful exit; < 0: if INFO = -i, the i-th argument had an illegal value.Method
If INIPER = 'P', the routine first computes a QR factorization with column pivoting: A * P = Q * [ R11 R12 ] [ 0 R22 ] with R11 defined as the largest leading submatrix whose estimated condition number is less than 1/RCOND. The order of R11, RANK, is the effective rank of A. If INIPER = 'N', the effective rank is estimated during a truncated QR factorization (with column pivoting) process, and the submatrix R22 is not upper triangular, but full and of small norm. (See SLICOT Library routines MB03OD or MB03OY, respectively, for further details.) Then, R22 is considered to be negligible, and R12 is annihilated by orthogonal transformations from the right, arriving at the complete orthogonal factorization: A * P = Q * [ T11 0 ] * Z [ 0 0 ] The solution is then X = P * Z' [ inv(T11)*Q1'*B ] [ Y ] where Q1 consists of the first RANK columns of Q, and Y contains free elements (if JOB = 'F'), or is zero (if JOB = 'L').Numerical Aspects
The algorithm is backward stable.Further Comments
Significant gain in efficiency is possible for small-rank problems using truncated QR factorization (option INIPER = 'N').Example
Program Text
* MB02QD EXAMPLE PROGRAM TEXT * Copyright (c) 2002-2017 NICONET e.V. * * .. Parameters .. INTEGER NIN, NOUT PARAMETER ( NIN = 5, NOUT = 6 ) INTEGER NMAX, MMAX, NRHSMX PARAMETER ( NMAX = 20, MMAX = 20, NRHSMX = 20 ) INTEGER LDA, LDB PARAMETER ( LDA = MMAX, LDB = MAX( MMAX, NMAX ) ) INTEGER LDWORK PARAMETER ( LDWORK = MAX( MIN( MMAX, NMAX) + 3*NMAX + 1, $ 2*MIN( MMAX, NMAX) + NRHSMX ) ) * .. Local Scalars .. DOUBLE PRECISION RCOND, SVLMAX INTEGER I, INFO, J, M, N, NRHS, RANK CHARACTER*1 INIPER, JOB * .. Local Arrays .. INTEGER JPVT(NMAX) DOUBLE PRECISION A(LDA,NMAX), B(LDB,NRHSMX), DWORK(LDWORK), $ SVAL(3), Y(NMAX*NRHSMX) * .. External Functions .. LOGICAL LSAME EXTERNAL LSAME * .. External Subroutines .. EXTERNAL MB02QD * .. Intrinsic Functions .. INTRINSIC MAX, MIN * .. Executable Statements .. * WRITE ( NOUT, FMT = 99999 ) * Skip the heading in the data file and read the data. READ ( NIN, FMT = '()' ) READ ( NIN, FMT = * ) M, N, NRHS, RCOND, SVLMAX, JOB, INIPER IF ( M.LT.0 .OR. M.GT.MMAX ) THEN WRITE ( NOUT, FMT = 99994 ) M ELSE IF ( N.LT.0 .OR. N.GT.NMAX ) THEN WRITE ( NOUT, FMT = 99993 ) N ELSE IF ( NRHS.LT.0 .OR. NRHS.GT.NRHSMX ) THEN WRITE ( NOUT, FMT = 99992 ) NRHS ELSE READ ( NIN, FMT = * ) ( ( A(I,J), J = 1,N ), I = 1,M ) READ ( NIN, FMT = * ) ( ( B(I,J), J = 1,NRHS ), I = 1,M ) IF ( LSAME( JOB, 'F' ) ) $ READ ( NIN, FMT = * ) ( Y(I), I = 1,N*NRHS ) IF ( LSAME( INIPER, 'P' ) ) $ READ ( NIN, FMT = * ) ( JPVT(I), I = 1,N ) * Find the least squares solution. CALL MB02QD( JOB, INIPER, M, N, NRHS, RCOND, SVLMAX, A, $ LDA, B, LDB, Y, JPVT, RANK, SVAL, DWORK, $ LDWORK, INFO ) * IF ( INFO.NE.0 ) THEN WRITE ( NOUT, FMT = 99998 ) INFO ELSE WRITE ( NOUT, FMT = 99997 ) RANK, SVAL WRITE ( NOUT, FMT = 99996 ) DO 10 I = 1, N WRITE ( NOUT, FMT = 99995 ) ( B(I,J), J = 1,NRHS ) 10 CONTINUE END IF END IF END IF END IF STOP * 99999 FORMAT (' MB02QD EXAMPLE PROGRAM RESULTS',/1X) 99998 FORMAT (' INFO on exit from MB02QD =',I2) 99997 FORMAT (' The effective rank of A =',I2,/ $ ' Estimates of the singular values SVAL = '/3(1X,F8.4)) 99996 FORMAT (' The least squares solution is') 99995 FORMAT (20(1X,F8.4)) 99994 FORMAT (/' M is out of range.',/' M = ',I5) 99993 FORMAT (/' N is out of range.',/' N = ',I5) 99992 FORMAT (/' NRHS is out of range.',/' NRHS = ',I5) ENDProgram Data
MB02QD EXAMPLE PROGRAM DATA 4 3 2 2.3D-16 0.0 L N 2.0 2.0 -3.0 3.0 3.0 -1.0 4.0 4.0 -5.0 -1.0 -1.0 -2.0 1.0 0.0 0.0 0.0 0.0 0.0 0.0 1.0Program Results
MB02QD EXAMPLE PROGRAM RESULTS The effective rank of A = 2 Estimates of the singular values SVAL = 7.8659 2.6698 0.0000 The least squares solution is -0.0034 -0.1054 -0.0034 -0.1054 -0.0816 -0.1973