File:  [local] / rpl / lapack / lapack / dgelsx.f
Revision 1.18: download - view: text, annotated - select for diffs - revision graph
Mon Aug 7 08:38:48 2023 UTC (8 months, 3 weeks ago) by bertrand
Branches: MAIN
CVS tags: rpl-4_1_35, rpl-4_1_34, HEAD
Première mise à jour de lapack et blas.

    1: *> \brief <b> DGELSX solves overdetermined or underdetermined systems for GE matrices</b>
    2: *
    3: *  =========== DOCUMENTATION ===========
    4: *
    5: * Online html documentation available at
    6: *            http://www.netlib.org/lapack/explore-html/
    7: *
    8: *> \htmlonly
    9: *> Download DGELSX + dependencies
   10: *> <a href="http://www.netlib.org/cgi-bin/netlibfiles.tgz?format=tgz&filename=/lapack/lapack_routine/dgelsx.f">
   11: *> [TGZ]</a>
   12: *> <a href="http://www.netlib.org/cgi-bin/netlibfiles.zip?format=zip&filename=/lapack/lapack_routine/dgelsx.f">
   13: *> [ZIP]</a>
   14: *> <a href="http://www.netlib.org/cgi-bin/netlibfiles.txt?format=txt&filename=/lapack/lapack_routine/dgelsx.f">
   15: *> [TXT]</a>
   16: *> \endhtmlonly
   17: *
   18: *  Definition:
   19: *  ===========
   20: *
   21: *       SUBROUTINE DGELSX( M, N, NRHS, A, LDA, B, LDB, JPVT, RCOND, RANK,
   22: *                          WORK, INFO )
   23: *
   24: *       .. Scalar Arguments ..
   25: *       INTEGER            INFO, LDA, LDB, M, N, NRHS, RANK
   26: *       DOUBLE PRECISION   RCOND
   27: *       ..
   28: *       .. Array Arguments ..
   29: *       INTEGER            JPVT( * )
   30: *       DOUBLE PRECISION   A( LDA, * ), B( LDB, * ), WORK( * )
   31: *       ..
   32: *
   33: *
   34: *> \par Purpose:
   35: *  =============
   36: *>
   37: *> \verbatim
   38: *>
   39: *> This routine is deprecated and has been replaced by routine DGELSY.
   40: *>
   41: *> DGELSX computes the minimum-norm solution to a real linear least
   42: *> squares problem:
   43: *>     minimize || A * X - B ||
   44: *> using a complete orthogonal factorization of A.  A is an M-by-N
   45: *> matrix which may be rank-deficient.
   46: *>
   47: *> Several right hand side vectors b and solution vectors x can be
   48: *> handled in a single call; they are stored as the columns of the
   49: *> M-by-NRHS right hand side matrix B and the N-by-NRHS solution
   50: *> matrix X.
   51: *>
   52: *> The routine first computes a QR factorization with column pivoting:
   53: *>     A * P = Q * [ R11 R12 ]
   54: *>                 [  0  R22 ]
   55: *> with R11 defined as the largest leading submatrix whose estimated
   56: *> condition number is less than 1/RCOND.  The order of R11, RANK,
   57: *> is the effective rank of A.
   58: *>
   59: *> Then, R22 is considered to be negligible, and R12 is annihilated
   60: *> by orthogonal transformations from the right, arriving at the
   61: *> complete orthogonal factorization:
   62: *>    A * P = Q * [ T11 0 ] * Z
   63: *>                [  0  0 ]
   64: *> The minimum-norm solution is then
   65: *>    X = P * Z**T [ inv(T11)*Q1**T*B ]
   66: *>                 [        0         ]
   67: *> where Q1 consists of the first RANK columns of Q.
   68: *> \endverbatim
   69: *
   70: *  Arguments:
   71: *  ==========
   72: *
   73: *> \param[in] M
   74: *> \verbatim
   75: *>          M is INTEGER
   76: *>          The number of rows of the matrix A.  M >= 0.
   77: *> \endverbatim
   78: *>
   79: *> \param[in] N
   80: *> \verbatim
   81: *>          N is INTEGER
   82: *>          The number of columns of the matrix A.  N >= 0.
   83: *> \endverbatim
   84: *>
   85: *> \param[in] NRHS
   86: *> \verbatim
   87: *>          NRHS is INTEGER
   88: *>          The number of right hand sides, i.e., the number of
   89: *>          columns of matrices B and X. NRHS >= 0.
   90: *> \endverbatim
   91: *>
   92: *> \param[in,out] A
   93: *> \verbatim
   94: *>          A is DOUBLE PRECISION array, dimension (LDA,N)
   95: *>          On entry, the M-by-N matrix A.
   96: *>          On exit, A has been overwritten by details of its
   97: *>          complete orthogonal factorization.
   98: *> \endverbatim
   99: *>
  100: *> \param[in] LDA
  101: *> \verbatim
  102: *>          LDA is INTEGER
  103: *>          The leading dimension of the array A.  LDA >= max(1,M).
  104: *> \endverbatim
  105: *>
  106: *> \param[in,out] B
  107: *> \verbatim
  108: *>          B is DOUBLE PRECISION array, dimension (LDB,NRHS)
  109: *>          On entry, the M-by-NRHS right hand side matrix B.
  110: *>          On exit, the N-by-NRHS solution matrix X.
  111: *>          If m >= n and RANK = n, the residual sum-of-squares for
  112: *>          the solution in the i-th column is given by the sum of
  113: *>          squares of elements N+1:M in that column.
  114: *> \endverbatim
  115: *>
  116: *> \param[in] LDB
  117: *> \verbatim
  118: *>          LDB is INTEGER
  119: *>          The leading dimension of the array B. LDB >= max(1,M,N).
  120: *> \endverbatim
  121: *>
  122: *> \param[in,out] JPVT
  123: *> \verbatim
  124: *>          JPVT is INTEGER array, dimension (N)
  125: *>          On entry, if JPVT(i) .ne. 0, the i-th column of A is an
  126: *>          initial column, otherwise it is a free column.  Before
  127: *>          the QR factorization of A, all initial columns are
  128: *>          permuted to the leading positions; only the remaining
  129: *>          free columns are moved as a result of column pivoting
  130: *>          during the factorization.
  131: *>          On exit, if JPVT(i) = k, then the i-th column of A*P
  132: *>          was the k-th column of A.
  133: *> \endverbatim
  134: *>
  135: *> \param[in] RCOND
  136: *> \verbatim
  137: *>          RCOND is DOUBLE PRECISION
  138: *>          RCOND is used to determine the effective rank of A, which
  139: *>          is defined as the order of the largest leading triangular
  140: *>          submatrix R11 in the QR factorization with pivoting of A,
  141: *>          whose estimated condition number < 1/RCOND.
  142: *> \endverbatim
  143: *>
  144: *> \param[out] RANK
  145: *> \verbatim
  146: *>          RANK is INTEGER
  147: *>          The effective rank of A, i.e., the order of the submatrix
  148: *>          R11.  This is the same as the order of the submatrix T11
  149: *>          in the complete orthogonal factorization of A.
  150: *> \endverbatim
  151: *>
  152: *> \param[out] WORK
  153: *> \verbatim
  154: *>          WORK is DOUBLE PRECISION array, dimension
  155: *>                      (max( min(M,N)+3*N, 2*min(M,N)+NRHS )),
  156: *> \endverbatim
  157: *>
  158: *> \param[out] INFO
  159: *> \verbatim
  160: *>          INFO is INTEGER
  161: *>          = 0:  successful exit
  162: *>          < 0:  if INFO = -i, the i-th argument had an illegal value
  163: *> \endverbatim
  164: *
  165: *  Authors:
  166: *  ========
  167: *
  168: *> \author Univ. of Tennessee
  169: *> \author Univ. of California Berkeley
  170: *> \author Univ. of Colorado Denver
  171: *> \author NAG Ltd.
  172: *
  173: *> \ingroup doubleGEsolve
  174: *
  175: *  =====================================================================
  176:       SUBROUTINE DGELSX( M, N, NRHS, A, LDA, B, LDB, JPVT, RCOND, RANK,
  177:      $                   WORK, INFO )
  178: *
  179: *  -- LAPACK driver routine --
  180: *  -- LAPACK is a software package provided by Univ. of Tennessee,    --
  181: *  -- Univ. of California Berkeley, Univ. of Colorado Denver and NAG Ltd..--
  182: *
  183: *     .. Scalar Arguments ..
  184:       INTEGER            INFO, LDA, LDB, M, N, NRHS, RANK
  185:       DOUBLE PRECISION   RCOND
  186: *     ..
  187: *     .. Array Arguments ..
  188:       INTEGER            JPVT( * )
  189:       DOUBLE PRECISION   A( LDA, * ), B( LDB, * ), WORK( * )
  190: *     ..
  191: *
  192: *  =====================================================================
  193: *
  194: *     .. Parameters ..
  195:       INTEGER            IMAX, IMIN
  196:       PARAMETER          ( IMAX = 1, IMIN = 2 )
  197:       DOUBLE PRECISION   ZERO, ONE, DONE, NTDONE
  198:       PARAMETER          ( ZERO = 0.0D0, ONE = 1.0D0, DONE = ZERO,
  199:      $                   NTDONE = ONE )
  200: *     ..
  201: *     .. Local Scalars ..
  202:       INTEGER            I, IASCL, IBSCL, ISMAX, ISMIN, J, K, MN
  203:       DOUBLE PRECISION   ANRM, BIGNUM, BNRM, C1, C2, S1, S2, SMAX,
  204:      $                   SMAXPR, SMIN, SMINPR, SMLNUM, T1, T2
  205: *     ..
  206: *     .. External Functions ..
  207:       DOUBLE PRECISION   DLAMCH, DLANGE
  208:       EXTERNAL           DLAMCH, DLANGE
  209: *     ..
  210: *     .. External Subroutines ..
  211:       EXTERNAL           DGEQPF, DLAIC1, DLASCL, DLASET, DLATZM, DORM2R,
  212:      $                   DTRSM, DTZRQF, XERBLA
  213: *     ..
  214: *     .. Intrinsic Functions ..
  215:       INTRINSIC          ABS, MAX, MIN
  216: *     ..
  217: *     .. Executable Statements ..
  218: *
  219:       MN = MIN( M, N )
  220:       ISMIN = MN + 1
  221:       ISMAX = 2*MN + 1
  222: *
  223: *     Test the input arguments.
  224: *
  225:       INFO = 0
  226:       IF( M.LT.0 ) THEN
  227:          INFO = -1
  228:       ELSE IF( N.LT.0 ) THEN
  229:          INFO = -2
  230:       ELSE IF( NRHS.LT.0 ) THEN
  231:          INFO = -3
  232:       ELSE IF( LDA.LT.MAX( 1, M ) ) THEN
  233:          INFO = -5
  234:       ELSE IF( LDB.LT.MAX( 1, M, N ) ) THEN
  235:          INFO = -7
  236:       END IF
  237: *
  238:       IF( INFO.NE.0 ) THEN
  239:          CALL XERBLA( 'DGELSX', -INFO )
  240:          RETURN
  241:       END IF
  242: *
  243: *     Quick return if possible
  244: *
  245:       IF( MIN( M, N, NRHS ).EQ.0 ) THEN
  246:          RANK = 0
  247:          RETURN
  248:       END IF
  249: *
  250: *     Get machine parameters
  251: *
  252:       SMLNUM = DLAMCH( 'S' ) / DLAMCH( 'P' )
  253:       BIGNUM = ONE / SMLNUM
  254:       CALL DLABAD( SMLNUM, BIGNUM )
  255: *
  256: *     Scale A, B if max elements outside range [SMLNUM,BIGNUM]
  257: *
  258:       ANRM = DLANGE( 'M', M, N, A, LDA, WORK )
  259:       IASCL = 0
  260:       IF( ANRM.GT.ZERO .AND. ANRM.LT.SMLNUM ) THEN
  261: *
  262: *        Scale matrix norm up to SMLNUM
  263: *
  264:          CALL DLASCL( 'G', 0, 0, ANRM, SMLNUM, M, N, A, LDA, INFO )
  265:          IASCL = 1
  266:       ELSE IF( ANRM.GT.BIGNUM ) THEN
  267: *
  268: *        Scale matrix norm down to BIGNUM
  269: *
  270:          CALL DLASCL( 'G', 0, 0, ANRM, BIGNUM, M, N, A, LDA, INFO )
  271:          IASCL = 2
  272:       ELSE IF( ANRM.EQ.ZERO ) THEN
  273: *
  274: *        Matrix all zero. Return zero solution.
  275: *
  276:          CALL DLASET( 'F', MAX( M, N ), NRHS, ZERO, ZERO, B, LDB )
  277:          RANK = 0
  278:          GO TO 100
  279:       END IF
  280: *
  281:       BNRM = DLANGE( 'M', M, NRHS, B, LDB, WORK )
  282:       IBSCL = 0
  283:       IF( BNRM.GT.ZERO .AND. BNRM.LT.SMLNUM ) THEN
  284: *
  285: *        Scale matrix norm up to SMLNUM
  286: *
  287:          CALL DLASCL( 'G', 0, 0, BNRM, SMLNUM, M, NRHS, B, LDB, INFO )
  288:          IBSCL = 1
  289:       ELSE IF( BNRM.GT.BIGNUM ) THEN
  290: *
  291: *        Scale matrix norm down to BIGNUM
  292: *
  293:          CALL DLASCL( 'G', 0, 0, BNRM, BIGNUM, M, NRHS, B, LDB, INFO )
  294:          IBSCL = 2
  295:       END IF
  296: *
  297: *     Compute QR factorization with column pivoting of A:
  298: *        A * P = Q * R
  299: *
  300:       CALL DGEQPF( M, N, A, LDA, JPVT, WORK( 1 ), WORK( MN+1 ), INFO )
  301: *
  302: *     workspace 3*N. Details of Householder rotations stored
  303: *     in WORK(1:MN).
  304: *
  305: *     Determine RANK using incremental condition estimation
  306: *
  307:       WORK( ISMIN ) = ONE
  308:       WORK( ISMAX ) = ONE
  309:       SMAX = ABS( A( 1, 1 ) )
  310:       SMIN = SMAX
  311:       IF( ABS( A( 1, 1 ) ).EQ.ZERO ) THEN
  312:          RANK = 0
  313:          CALL DLASET( 'F', MAX( M, N ), NRHS, ZERO, ZERO, B, LDB )
  314:          GO TO 100
  315:       ELSE
  316:          RANK = 1
  317:       END IF
  318: *
  319:    10 CONTINUE
  320:       IF( RANK.LT.MN ) THEN
  321:          I = RANK + 1
  322:          CALL DLAIC1( IMIN, RANK, WORK( ISMIN ), SMIN, A( 1, I ),
  323:      $                A( I, I ), SMINPR, S1, C1 )
  324:          CALL DLAIC1( IMAX, RANK, WORK( ISMAX ), SMAX, A( 1, I ),
  325:      $                A( I, I ), SMAXPR, S2, C2 )
  326: *
  327:          IF( SMAXPR*RCOND.LE.SMINPR ) THEN
  328:             DO 20 I = 1, RANK
  329:                WORK( ISMIN+I-1 ) = S1*WORK( ISMIN+I-1 )
  330:                WORK( ISMAX+I-1 ) = S2*WORK( ISMAX+I-1 )
  331:    20       CONTINUE
  332:             WORK( ISMIN+RANK ) = C1
  333:             WORK( ISMAX+RANK ) = C2
  334:             SMIN = SMINPR
  335:             SMAX = SMAXPR
  336:             RANK = RANK + 1
  337:             GO TO 10
  338:          END IF
  339:       END IF
  340: *
  341: *     Logically partition R = [ R11 R12 ]
  342: *                             [  0  R22 ]
  343: *     where R11 = R(1:RANK,1:RANK)
  344: *
  345: *     [R11,R12] = [ T11, 0 ] * Y
  346: *
  347:       IF( RANK.LT.N )
  348:      $   CALL DTZRQF( RANK, N, A, LDA, WORK( MN+1 ), INFO )
  349: *
  350: *     Details of Householder rotations stored in WORK(MN+1:2*MN)
  351: *
  352: *     B(1:M,1:NRHS) := Q**T * B(1:M,1:NRHS)
  353: *
  354:       CALL DORM2R( 'Left', 'Transpose', M, NRHS, MN, A, LDA, WORK( 1 ),
  355:      $             B, LDB, WORK( 2*MN+1 ), INFO )
  356: *
  357: *     workspace NRHS
  358: *
  359: *     B(1:RANK,1:NRHS) := inv(T11) * B(1:RANK,1:NRHS)
  360: *
  361:       CALL DTRSM( 'Left', 'Upper', 'No transpose', 'Non-unit', RANK,
  362:      $            NRHS, ONE, A, LDA, B, LDB )
  363: *
  364:       DO 40 I = RANK + 1, N
  365:          DO 30 J = 1, NRHS
  366:             B( I, J ) = ZERO
  367:    30    CONTINUE
  368:    40 CONTINUE
  369: *
  370: *     B(1:N,1:NRHS) := Y**T * B(1:N,1:NRHS)
  371: *
  372:       IF( RANK.LT.N ) THEN
  373:          DO 50 I = 1, RANK
  374:             CALL DLATZM( 'Left', N-RANK+1, NRHS, A( I, RANK+1 ), LDA,
  375:      $                   WORK( MN+I ), B( I, 1 ), B( RANK+1, 1 ), LDB,
  376:      $                   WORK( 2*MN+1 ) )
  377:    50    CONTINUE
  378:       END IF
  379: *
  380: *     workspace NRHS
  381: *
  382: *     B(1:N,1:NRHS) := P * B(1:N,1:NRHS)
  383: *
  384:       DO 90 J = 1, NRHS
  385:          DO 60 I = 1, N
  386:             WORK( 2*MN+I ) = NTDONE
  387:    60    CONTINUE
  388:          DO 80 I = 1, N
  389:             IF( WORK( 2*MN+I ).EQ.NTDONE ) THEN
  390:                IF( JPVT( I ).NE.I ) THEN
  391:                   K = I
  392:                   T1 = B( K, J )
  393:                   T2 = B( JPVT( K ), J )
  394:    70             CONTINUE
  395:                   B( JPVT( K ), J ) = T1
  396:                   WORK( 2*MN+K ) = DONE
  397:                   T1 = T2
  398:                   K = JPVT( K )
  399:                   T2 = B( JPVT( K ), J )
  400:                   IF( JPVT( K ).NE.I )
  401:      $               GO TO 70
  402:                   B( I, J ) = T1
  403:                   WORK( 2*MN+K ) = DONE
  404:                END IF
  405:             END IF
  406:    80    CONTINUE
  407:    90 CONTINUE
  408: *
  409: *     Undo scaling
  410: *
  411:       IF( IASCL.EQ.1 ) THEN
  412:          CALL DLASCL( 'G', 0, 0, ANRM, SMLNUM, N, NRHS, B, LDB, INFO )
  413:          CALL DLASCL( 'U', 0, 0, SMLNUM, ANRM, RANK, RANK, A, LDA,
  414:      $                INFO )
  415:       ELSE IF( IASCL.EQ.2 ) THEN
  416:          CALL DLASCL( 'G', 0, 0, ANRM, BIGNUM, N, NRHS, B, LDB, INFO )
  417:          CALL DLASCL( 'U', 0, 0, BIGNUM, ANRM, RANK, RANK, A, LDA,
  418:      $                INFO )
  419:       END IF
  420:       IF( IBSCL.EQ.1 ) THEN
  421:          CALL DLASCL( 'G', 0, 0, SMLNUM, BNRM, N, NRHS, B, LDB, INFO )
  422:       ELSE IF( IBSCL.EQ.2 ) THEN
  423:          CALL DLASCL( 'G', 0, 0, BIGNUM, BNRM, N, NRHS, B, LDB, INFO )
  424:       END IF
  425: *
  426:   100 CONTINUE
  427: *
  428:       RETURN
  429: *
  430: *     End of DGELSX
  431: *
  432:       END

CVSweb interface <joel.bertrand@systella.fr>