SUBROUTINE DGELS( TRANS, M, N, NRHS, A, LDA, B, LDB, WORK, LWORK,
$ INFO )
*
* -- LAPACK driver routine (version 3.1) --
* Univ. of Tennessee, Univ. of California Berkeley and NAG Ltd..
* November 2006
*
* .. Scalar Arguments ..
CHARACTER TRANS
INTEGER INFO, LDA, LDB, LWORK, M, N, NRHS
* ..
* .. Array Arguments ..
DOUBLE PRECISION A( LDA, * ), B( LDB, * ), WORK( * )
* ..
*
* Purpose
* =======
*
* DGELS solves overdetermined or underdetermined real linear systems
* involving an M-by-N matrix A, or its transpose, using a QR or LQ
* factorization of A. It is assumed that A has full rank.
*
* The following options are provided:
*
* 1. If TRANS = 'N' and m >= n: find the least squares solution of
* an overdetermined system, i.e., solve the least squares problem
* minimize || B - A*X ||.
*
* 2. If TRANS = 'N' and m < n: find the minimum norm solution of
* an underdetermined system A * X = B.
*
* 3. If TRANS = 'T' and m >= n: find the minimum norm solution of
* an undetermined system A**T * X = B.
*
* 4. If TRANS = 'T' and m < n: find the least squares solution of
* an overdetermined system, i.e., solve the least squares problem
* minimize || B - A**T * X ||.
*
* 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.
*
* Arguments
* =========
*
* TRANS (input) CHARACTER*1
* = 'N': the linear system involves A;
* = 'T': the linear system involves A**T.
*
* 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.
*
* A (input/output) DOUBLE PRECISION array, dimension (LDA,N)
* On entry, the M-by-N matrix A.
* On exit,
* if M >= N, A is overwritten by details of its QR
* factorization as returned by DGEQRF;
* if M < N, A is overwritten by details of its LQ
* factorization as returned by DGELQF.
*
* LDA (input) INTEGER
* The leading dimension of the array A. LDA >= max(1,M).
*
* B (input/output) DOUBLE PRECISION array, dimension (LDB,NRHS)
* On entry, the matrix B of right hand side vectors, stored
* columnwise; B is M-by-NRHS if TRANS = 'N', or N-by-NRHS
* if TRANS = 'T'.
* On exit, if INFO = 0, B is overwritten by the solution
* vectors, stored columnwise:
* if TRANS = 'N' and m >= n, rows 1 to n of B contain the least
* squares solution vectors; the residual sum of squares for the
* solution in each column is given by the sum of squares of
* elements N+1 to M in that column;
* if TRANS = 'N' and m < n, rows 1 to N of B contain the
* minimum norm solution vectors;
* if TRANS = 'T' and m >= n, rows 1 to M of B contain the
* minimum norm solution vectors;
* if TRANS = 'T' and m < n, rows 1 to M of B contain the
* least squares solution vectors; the residual sum of squares
* for the solution in each column is given by the sum of
* squares of elements M+1 to N in that column.
*
* LDB (input) INTEGER
* The leading dimension of the array B. LDB >= MAX(1,M,N).
*
* WORK (workspace/output) DOUBLE PRECISION array, dimension (MAX(1,LWORK))
* On exit, if INFO = 0, WORK(1) returns the optimal LWORK.
*
* LWORK (input) INTEGER
* The dimension of the array WORK.
* LWORK >= max( 1, MN + max( MN, NRHS ) ).
* For optimal performance,
* LWORK >= max( 1, MN + max( MN, NRHS )*NB ).
* where MN = min(M,N) and NB is the optimum block size.
*
* If LWORK = -1, then a workspace query is assumed; the routine
* only calculates the optimal size of the WORK array, returns
* this value as the first entry of the WORK array, and no error
* message related to LWORK is issued by XERBLA.
*
* INFO (output) INTEGER
* = 0: successful exit
* < 0: if INFO = -i, the i-th argument had an illegal value
* > 0: if INFO = i, the i-th diagonal element of the
* triangular factor of A is zero, so that A does not have
* full rank; the least squares solution could not be
* computed.
*
* =====================================================================
*
* .. Parameters ..
DOUBLE PRECISION ZERO, ONE
PARAMETER ( ZERO = 0.0D0, ONE = 1.0D0 )
* ..
* .. Local Scalars ..
LOGICAL LQUERY, TPSD
INTEGER BROW, I, IASCL, IBSCL, J, MN, NB, SCLLEN, WSIZE
DOUBLE PRECISION ANRM, BIGNUM, BNRM, SMLNUM
* ..
* .. Local Arrays ..
DOUBLE PRECISION RWORK( 1 )
* ..
* .. External Functions ..
LOGICAL LSAME
INTEGER ILAENV
DOUBLE PRECISION DLAMCH, DLANGE
EXTERNAL LSAME, ILAENV, DLABAD, DLAMCH, DLANGE
* ..
* .. External Subroutines ..
EXTERNAL DGELQF, DGEQRF, DLASCL, DLASET, DORMLQ, DORMQR,
$ DTRTRS, XERBLA
* ..
* .. Intrinsic Functions ..
INTRINSIC DBLE, MAX, MIN
* ..
* .. Executable Statements ..
*
* Test the input arguments.
*
INFO = 0
MN = MIN( M, N )
LQUERY = ( LWORK.EQ.-1 )
IF( .NOT.( LSAME( TRANS, 'N' ) .OR. LSAME( TRANS, 'T' ) ) ) THEN
INFO = -1
ELSE IF( M.LT.0 ) THEN
INFO = -2
ELSE IF( N.LT.0 ) THEN
INFO = -3
ELSE IF( NRHS.LT.0 ) THEN
INFO = -4
ELSE IF( LDA.LT.MAX( 1, M ) ) THEN
INFO = -6
ELSE IF( LDB.LT.MAX( 1, M, N ) ) THEN
INFO = -8
ELSE IF( LWORK.LT.MAX( 1, MN+MAX( MN, NRHS ) ) .AND. .NOT.LQUERY )
$ THEN
INFO = -10
END IF
*
* Figure out optimal block size
*
IF( INFO.EQ.0 .OR. INFO.EQ.-10 ) THEN
*
TPSD = .TRUE.
IF( LSAME( TRANS, 'N' ) )
$ TPSD = .FALSE.
*
IF( M.GE.N ) THEN
NB = ILAENV( 1, 'DGEQRF', ' ', M, N, -1, -1 )
IF( TPSD ) THEN
NB = MAX( NB, ILAENV( 1, 'DORMQR', 'LN', M, NRHS, N,
$ -1 ) )
ELSE
NB = MAX( NB, ILAENV( 1, 'DORMQR', 'LT', M, NRHS, N,
$ -1 ) )
END IF
ELSE
NB = ILAENV( 1, 'DGELQF', ' ', M, N, -1, -1 )
IF( TPSD ) THEN
NB = MAX( NB, ILAENV( 1, 'DORMLQ', 'LT', N, NRHS, M,
$ -1 ) )
ELSE
NB = MAX( NB, ILAENV( 1, 'DORMLQ', 'LN', N, NRHS, M,
$ -1 ) )
END IF
END IF
*
WSIZE = MAX( 1, MN+MAX( MN, NRHS )*NB )
WORK( 1 ) = DBLE( WSIZE )
*
END IF
*
IF( INFO.NE.0 ) THEN
CALL XERBLA( 'DGELS ', -INFO )
RETURN
ELSE IF( LQUERY ) THEN
RETURN
END IF
*
* Quick return if possible
*
IF( MIN( M, N, NRHS ).EQ.0 ) THEN
CALL DLASET( 'Full', MAX( M, N ), NRHS, ZERO, ZERO, B, LDB )
RETURN
END IF
*
* Get machine parameters
*
SMLNUM = DLAMCH( 'S' ) / DLAMCH( 'P' )
BIGNUM = ONE / SMLNUM
CALL DLABAD( SMLNUM, BIGNUM )
*
* Scale A, B if max element outside range [SMLNUM,BIGNUM]
*
ANRM = DLANGE( 'M', M, N, A, LDA, RWORK )
IASCL = 0
IF( ANRM.GT.ZERO .AND. ANRM.LT.SMLNUM ) THEN
*
* Scale matrix norm up to SMLNUM
*
CALL DLASCL( 'G', 0, 0, ANRM, SMLNUM, M, N, A, LDA, INFO )
IASCL = 1
ELSE IF( ANRM.GT.BIGNUM ) THEN
*
* Scale matrix norm down to BIGNUM
*
CALL DLASCL( 'G', 0, 0, ANRM, BIGNUM, M, N, A, LDA, INFO )
IASCL = 2
ELSE IF( ANRM.EQ.ZERO ) THEN
*
* Matrix all zero. Return zero solution.
*
CALL DLASET( 'F', MAX( M, N ), NRHS, ZERO, ZERO, B, LDB )
GO TO 50
END IF
*
BROW = M
IF( TPSD )
$ BROW = N
BNRM = DLANGE( 'M', BROW, NRHS, B, LDB, RWORK )
IBSCL = 0
IF( BNRM.GT.ZERO .AND. BNRM.LT.SMLNUM ) THEN
*
* Scale matrix norm up to SMLNUM
*
CALL DLASCL( 'G', 0, 0, BNRM, SMLNUM, BROW, NRHS, B, LDB,
$ INFO )
IBSCL = 1
ELSE IF( BNRM.GT.BIGNUM ) THEN
*
* Scale matrix norm down to BIGNUM
*
CALL DLASCL( 'G', 0, 0, BNRM, BIGNUM, BROW, NRHS, B, LDB,
$ INFO )
IBSCL = 2
END IF
*
IF( M.GE.N ) THEN
*
* compute QR factorization of A
*
CALL DGEQRF( M, N, A, LDA, WORK( 1 ), WORK( MN+1 ), LWORK-MN,
$ INFO )
*
* workspace at least N, optimally N*NB
*
IF( .NOT.TPSD ) THEN
*
* Least-Squares Problem min || A * X - B ||
*
* B(1:M,1:NRHS) := Q' * B(1:M,1:NRHS)
*
CALL DORMQR( 'Left', 'Transpose', M, NRHS, N, A, LDA,
$ WORK( 1 ), B, LDB, WORK( MN+1 ), LWORK-MN,
$ INFO )
*
* workspace at least NRHS, optimally NRHS*NB
*
* B(1:N,1:NRHS) := inv(R) * B(1:N,1:NRHS)
*
CALL DTRTRS( 'Upper', 'No transpose', 'Non-unit', N, NRHS,
$ A, LDA, B, LDB, INFO )
*
IF( INFO.GT.0 ) THEN
RETURN
END IF
*
SCLLEN = N
*
ELSE
*
* Overdetermined system of equations A' * X = B
*
* B(1:N,1:NRHS) := inv(R') * B(1:N,1:NRHS)
*
CALL DTRTRS( 'Upper', 'Transpose', 'Non-unit', N, NRHS,
$ A, LDA, B, LDB, INFO )
*
IF( INFO.GT.0 ) THEN
RETURN
END IF
*
* B(N+1:M,1:NRHS) = ZERO
*
DO 20 J = 1, NRHS
DO 10 I = N + 1, M
B( I, J ) = ZERO
10 CONTINUE
20 CONTINUE
*
* B(1:M,1:NRHS) := Q(1:N,:) * B(1:N,1:NRHS)
*
CALL DORMQR( 'Left', 'No transpose', M, NRHS, N, A, LDA,
$ WORK( 1 ), B, LDB, WORK( MN+1 ), LWORK-MN,
$ INFO )
*
* workspace at least NRHS, optimally NRHS*NB
*
SCLLEN = M
*
END IF
*
ELSE
*
* Compute LQ factorization of A
*
CALL DGELQF( M, N, A, LDA, WORK( 1 ), WORK( MN+1 ), LWORK-MN,
$ INFO )
*
* workspace at least M, optimally M*NB.
*
IF( .NOT.TPSD ) THEN
*
* underdetermined system of equations A * X = B
*
* B(1:M,1:NRHS) := inv(L) * B(1:M,1:NRHS)
*
CALL DTRTRS( 'Lower', 'No transpose', 'Non-unit', M, NRHS,
$ A, LDA, B, LDB, INFO )
*
IF( INFO.GT.0 ) THEN
RETURN
END IF
*
* B(M+1:N,1:NRHS) = 0
*
DO 40 J = 1, NRHS
DO 30 I = M + 1, N
B( I, J ) = ZERO
30 CONTINUE
40 CONTINUE
*
* B(1:N,1:NRHS) := Q(1:N,:)' * B(1:M,1:NRHS)
*
CALL DORMLQ( 'Left', 'Transpose', N, NRHS, M, A, LDA,
$ WORK( 1 ), B, LDB, WORK( MN+1 ), LWORK-MN,
$ INFO )
*
* workspace at least NRHS, optimally NRHS*NB
*
SCLLEN = N
*
ELSE
*
* overdetermined system min || A' * X - B ||
*
* B(1:N,1:NRHS) := Q * B(1:N,1:NRHS)
*
CALL DORMLQ( 'Left', 'No transpose', N, NRHS, M, A, LDA,
$ WORK( 1 ), B, LDB, WORK( MN+1 ), LWORK-MN,
$ INFO )
*
* workspace at least NRHS, optimally NRHS*NB
*
* B(1:M,1:NRHS) := inv(L') * B(1:M,1:NRHS)
*
CALL DTRTRS( 'Lower', 'Transpose', 'Non-unit', M, NRHS,
$ A, LDA, B, LDB, INFO )
*
IF( INFO.GT.0 ) THEN
RETURN
END IF
*
SCLLEN = M
*
END IF
*
END IF
*
* Undo scaling
*
IF( IASCL.EQ.1 ) THEN
CALL DLASCL( 'G', 0, 0, ANRM, SMLNUM, SCLLEN, NRHS, B, LDB,
$ INFO )
ELSE IF( IASCL.EQ.2 ) THEN
CALL DLASCL( 'G', 0, 0, ANRM, BIGNUM, SCLLEN, NRHS, B, LDB,
$ INFO )
END IF
IF( IBSCL.EQ.1 ) THEN
CALL DLASCL( 'G', 0, 0, SMLNUM, BNRM, SCLLEN, NRHS, B, LDB,
$ INFO )
ELSE IF( IBSCL.EQ.2 ) THEN
CALL DLASCL( 'G', 0, 0, BIGNUM, BNRM, SCLLEN, NRHS, B, LDB,
$ INFO )
END IF
*
50 CONTINUE
WORK( 1 ) = DBLE( WSIZE )
*
RETURN
*
* End of DGELS
*
END