File:  [local] / rpl / lapack / lapack / dggglm.f
Revision 1.19: download - view: text, annotated - select for diffs - revision graph
Mon Aug 7 08:38:51 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 DGGGLM
    2: *
    3: *  =========== DOCUMENTATION ===========
    4: *
    5: * Online html documentation available at
    6: *            http://www.netlib.org/lapack/explore-html/
    7: *
    8: *> \htmlonly
    9: *> Download DGGGLM + dependencies
   10: *> <a href="http://www.netlib.org/cgi-bin/netlibfiles.tgz?format=tgz&filename=/lapack/lapack_routine/dggglm.f">
   11: *> [TGZ]</a>
   12: *> <a href="http://www.netlib.org/cgi-bin/netlibfiles.zip?format=zip&filename=/lapack/lapack_routine/dggglm.f">
   13: *> [ZIP]</a>
   14: *> <a href="http://www.netlib.org/cgi-bin/netlibfiles.txt?format=txt&filename=/lapack/lapack_routine/dggglm.f">
   15: *> [TXT]</a>
   16: *> \endhtmlonly
   17: *
   18: *  Definition:
   19: *  ===========
   20: *
   21: *       SUBROUTINE DGGGLM( N, M, P, A, LDA, B, LDB, D, X, Y, WORK, LWORK,
   22: *                          INFO )
   23: *
   24: *       .. Scalar Arguments ..
   25: *       INTEGER            INFO, LDA, LDB, LWORK, M, N, P
   26: *       ..
   27: *       .. Array Arguments ..
   28: *       DOUBLE PRECISION   A( LDA, * ), B( LDB, * ), D( * ), WORK( * ),
   29: *      $                   X( * ), Y( * )
   30: *       ..
   31: *
   32: *
   33: *> \par Purpose:
   34: *  =============
   35: *>
   36: *> \verbatim
   37: *>
   38: *> DGGGLM solves a general Gauss-Markov linear model (GLM) problem:
   39: *>
   40: *>         minimize || y ||_2   subject to   d = A*x + B*y
   41: *>             x
   42: *>
   43: *> where A is an N-by-M matrix, B is an N-by-P matrix, and d is a
   44: *> given N-vector. It is assumed that M <= N <= M+P, and
   45: *>
   46: *>            rank(A) = M    and    rank( A B ) = N.
   47: *>
   48: *> Under these assumptions, the constrained equation is always
   49: *> consistent, and there is a unique solution x and a minimal 2-norm
   50: *> solution y, which is obtained using a generalized QR factorization
   51: *> of the matrices (A, B) given by
   52: *>
   53: *>    A = Q*(R),   B = Q*T*Z.
   54: *>          (0)
   55: *>
   56: *> In particular, if matrix B is square nonsingular, then the problem
   57: *> GLM is equivalent to the following weighted linear least squares
   58: *> problem
   59: *>
   60: *>              minimize || inv(B)*(d-A*x) ||_2
   61: *>                  x
   62: *>
   63: *> where inv(B) denotes the inverse of B.
   64: *> \endverbatim
   65: *
   66: *  Arguments:
   67: *  ==========
   68: *
   69: *> \param[in] N
   70: *> \verbatim
   71: *>          N is INTEGER
   72: *>          The number of rows of the matrices A and B.  N >= 0.
   73: *> \endverbatim
   74: *>
   75: *> \param[in] M
   76: *> \verbatim
   77: *>          M is INTEGER
   78: *>          The number of columns of the matrix A.  0 <= M <= N.
   79: *> \endverbatim
   80: *>
   81: *> \param[in] P
   82: *> \verbatim
   83: *>          P is INTEGER
   84: *>          The number of columns of the matrix B.  P >= N-M.
   85: *> \endverbatim
   86: *>
   87: *> \param[in,out] A
   88: *> \verbatim
   89: *>          A is DOUBLE PRECISION array, dimension (LDA,M)
   90: *>          On entry, the N-by-M matrix A.
   91: *>          On exit, the upper triangular part of the array A contains
   92: *>          the M-by-M upper triangular matrix R.
   93: *> \endverbatim
   94: *>
   95: *> \param[in] LDA
   96: *> \verbatim
   97: *>          LDA is INTEGER
   98: *>          The leading dimension of the array A. LDA >= max(1,N).
   99: *> \endverbatim
  100: *>
  101: *> \param[in,out] B
  102: *> \verbatim
  103: *>          B is DOUBLE PRECISION array, dimension (LDB,P)
  104: *>          On entry, the N-by-P matrix B.
  105: *>          On exit, if N <= P, the upper triangle of the subarray
  106: *>          B(1:N,P-N+1:P) contains the N-by-N upper triangular matrix T;
  107: *>          if N > P, the elements on and above the (N-P)th subdiagonal
  108: *>          contain the N-by-P upper trapezoidal matrix T.
  109: *> \endverbatim
  110: *>
  111: *> \param[in] LDB
  112: *> \verbatim
  113: *>          LDB is INTEGER
  114: *>          The leading dimension of the array B. LDB >= max(1,N).
  115: *> \endverbatim
  116: *>
  117: *> \param[in,out] D
  118: *> \verbatim
  119: *>          D is DOUBLE PRECISION array, dimension (N)
  120: *>          On entry, D is the left hand side of the GLM equation.
  121: *>          On exit, D is destroyed.
  122: *> \endverbatim
  123: *>
  124: *> \param[out] X
  125: *> \verbatim
  126: *>          X is DOUBLE PRECISION array, dimension (M)
  127: *> \endverbatim
  128: *>
  129: *> \param[out] Y
  130: *> \verbatim
  131: *>          Y is DOUBLE PRECISION array, dimension (P)
  132: *>
  133: *>          On exit, X and Y are the solutions of the GLM problem.
  134: *> \endverbatim
  135: *>
  136: *> \param[out] WORK
  137: *> \verbatim
  138: *>          WORK is DOUBLE PRECISION array, dimension (MAX(1,LWORK))
  139: *>          On exit, if INFO = 0, WORK(1) returns the optimal LWORK.
  140: *> \endverbatim
  141: *>
  142: *> \param[in] LWORK
  143: *> \verbatim
  144: *>          LWORK is INTEGER
  145: *>          The dimension of the array WORK. LWORK >= max(1,N+M+P).
  146: *>          For optimum performance, LWORK >= M+min(N,P)+max(N,P)*NB,
  147: *>          where NB is an upper bound for the optimal blocksizes for
  148: *>          DGEQRF, SGERQF, DORMQR and SORMRQ.
  149: *>
  150: *>          If LWORK = -1, then a workspace query is assumed; the routine
  151: *>          only calculates the optimal size of the WORK array, returns
  152: *>          this value as the first entry of the WORK array, and no error
  153: *>          message related to LWORK is issued by XERBLA.
  154: *> \endverbatim
  155: *>
  156: *> \param[out] INFO
  157: *> \verbatim
  158: *>          INFO is INTEGER
  159: *>          = 0:  successful exit.
  160: *>          < 0:  if INFO = -i, the i-th argument had an illegal value.
  161: *>          = 1:  the upper triangular factor R associated with A in the
  162: *>                generalized QR factorization of the pair (A, B) is
  163: *>                singular, so that rank(A) < M; the least squares
  164: *>                solution could not be computed.
  165: *>          = 2:  the bottom (N-M) by (N-M) part of the upper trapezoidal
  166: *>                factor T associated with B in the generalized QR
  167: *>                factorization of the pair (A, B) is singular, so that
  168: *>                rank( A B ) < N; the least squares solution could not
  169: *>                be computed.
  170: *> \endverbatim
  171: *
  172: *  Authors:
  173: *  ========
  174: *
  175: *> \author Univ. of Tennessee
  176: *> \author Univ. of California Berkeley
  177: *> \author Univ. of Colorado Denver
  178: *> \author NAG Ltd.
  179: *
  180: *> \ingroup doubleOTHEReigen
  181: *
  182: *  =====================================================================
  183:       SUBROUTINE DGGGLM( N, M, P, A, LDA, B, LDB, D, X, Y, WORK, LWORK,
  184:      $                   INFO )
  185: *
  186: *  -- LAPACK driver routine --
  187: *  -- LAPACK is a software package provided by Univ. of Tennessee,    --
  188: *  -- Univ. of California Berkeley, Univ. of Colorado Denver and NAG Ltd..--
  189: *
  190: *     .. Scalar Arguments ..
  191:       INTEGER            INFO, LDA, LDB, LWORK, M, N, P
  192: *     ..
  193: *     .. Array Arguments ..
  194:       DOUBLE PRECISION   A( LDA, * ), B( LDB, * ), D( * ), WORK( * ),
  195:      $                   X( * ), Y( * )
  196: *     ..
  197: *
  198: *  ===================================================================
  199: *
  200: *     .. Parameters ..
  201:       DOUBLE PRECISION   ZERO, ONE
  202:       PARAMETER          ( ZERO = 0.0D+0, ONE = 1.0D+0 )
  203: *     ..
  204: *     .. Local Scalars ..
  205:       LOGICAL            LQUERY
  206:       INTEGER            I, LOPT, LWKMIN, LWKOPT, NB, NB1, NB2, NB3,
  207:      $                   NB4, NP
  208: *     ..
  209: *     .. External Subroutines ..
  210:       EXTERNAL           DCOPY, DGEMV, DGGQRF, DORMQR, DORMRQ, DTRTRS,
  211:      $                   XERBLA
  212: *     ..
  213: *     .. External Functions ..
  214:       INTEGER            ILAENV
  215:       EXTERNAL           ILAENV
  216: *     ..
  217: *     .. Intrinsic Functions ..
  218:       INTRINSIC          INT, MAX, MIN
  219: *     ..
  220: *     .. Executable Statements ..
  221: *
  222: *     Test the input parameters
  223: *
  224:       INFO = 0
  225:       NP = MIN( N, P )
  226:       LQUERY = ( LWORK.EQ.-1 )
  227:       IF( N.LT.0 ) THEN
  228:          INFO = -1
  229:       ELSE IF( M.LT.0 .OR. M.GT.N ) THEN
  230:          INFO = -2
  231:       ELSE IF( P.LT.0 .OR. P.LT.N-M ) THEN
  232:          INFO = -3
  233:       ELSE IF( LDA.LT.MAX( 1, N ) ) THEN
  234:          INFO = -5
  235:       ELSE IF( LDB.LT.MAX( 1, N ) ) THEN
  236:          INFO = -7
  237:       END IF
  238: *
  239: *     Calculate workspace
  240: *
  241:       IF( INFO.EQ.0) THEN
  242:          IF( N.EQ.0 ) THEN
  243:             LWKMIN = 1
  244:             LWKOPT = 1
  245:          ELSE
  246:             NB1 = ILAENV( 1, 'DGEQRF', ' ', N, M, -1, -1 )
  247:             NB2 = ILAENV( 1, 'DGERQF', ' ', N, M, -1, -1 )
  248:             NB3 = ILAENV( 1, 'DORMQR', ' ', N, M, P, -1 )
  249:             NB4 = ILAENV( 1, 'DORMRQ', ' ', N, M, P, -1 )
  250:             NB = MAX( NB1, NB2, NB3, NB4 )
  251:             LWKMIN = M + N + P
  252:             LWKOPT = M + NP + MAX( N, P )*NB
  253:          END IF
  254:          WORK( 1 ) = LWKOPT
  255: *
  256:          IF( LWORK.LT.LWKMIN .AND. .NOT.LQUERY ) THEN
  257:             INFO = -12
  258:          END IF
  259:       END IF
  260: *
  261:       IF( INFO.NE.0 ) THEN
  262:          CALL XERBLA( 'DGGGLM', -INFO )
  263:          RETURN
  264:       ELSE IF( LQUERY ) THEN
  265:          RETURN
  266:       END IF
  267: *
  268: *     Quick return if possible
  269: *
  270:       IF( N.EQ.0 ) THEN
  271:          DO I = 1, M
  272:             X(I) = ZERO
  273:          END DO
  274:          DO I = 1, P
  275:             Y(I) = ZERO
  276:          END DO
  277:          RETURN
  278:       END IF
  279: *
  280: *     Compute the GQR factorization of matrices A and B:
  281: *
  282: *          Q**T*A = ( R11 ) M,    Q**T*B*Z**T = ( T11   T12 ) M
  283: *                   (  0  ) N-M                 (  0    T22 ) N-M
  284: *                      M                         M+P-N  N-M
  285: *
  286: *     where R11 and T22 are upper triangular, and Q and Z are
  287: *     orthogonal.
  288: *
  289:       CALL DGGQRF( N, M, P, A, LDA, WORK, B, LDB, WORK( M+1 ),
  290:      $             WORK( M+NP+1 ), LWORK-M-NP, INFO )
  291:       LOPT = INT( WORK( M+NP+1 ) )
  292: *
  293: *     Update left-hand-side vector d = Q**T*d = ( d1 ) M
  294: *                                               ( d2 ) N-M
  295: *
  296:       CALL DORMQR( 'Left', 'Transpose', N, 1, M, A, LDA, WORK, D,
  297:      $             MAX( 1, N ), WORK( M+NP+1 ), LWORK-M-NP, INFO )
  298:       LOPT = MAX( LOPT, INT( WORK( M+NP+1 ) ) )
  299: *
  300: *     Solve T22*y2 = d2 for y2
  301: *
  302:       IF( N.GT.M ) THEN
  303:          CALL DTRTRS( 'Upper', 'No transpose', 'Non unit', N-M, 1,
  304:      $                B( M+1, M+P-N+1 ), LDB, D( M+1 ), N-M, INFO )
  305: *
  306:          IF( INFO.GT.0 ) THEN
  307:             INFO = 1
  308:             RETURN
  309:          END IF
  310: *
  311:          CALL DCOPY( N-M, D( M+1 ), 1, Y( M+P-N+1 ), 1 )
  312:       END IF
  313: *
  314: *     Set y1 = 0
  315: *
  316:       DO 10 I = 1, M + P - N
  317:          Y( I ) = ZERO
  318:    10 CONTINUE
  319: *
  320: *     Update d1 = d1 - T12*y2
  321: *
  322:       CALL DGEMV( 'No transpose', M, N-M, -ONE, B( 1, M+P-N+1 ), LDB,
  323:      $            Y( M+P-N+1 ), 1, ONE, D, 1 )
  324: *
  325: *     Solve triangular system: R11*x = d1
  326: *
  327:       IF( M.GT.0 ) THEN
  328:          CALL DTRTRS( 'Upper', 'No Transpose', 'Non unit', M, 1, A, LDA,
  329:      $                D, M, INFO )
  330: *
  331:          IF( INFO.GT.0 ) THEN
  332:             INFO = 2
  333:             RETURN
  334:          END IF
  335: *
  336: *        Copy D to X
  337: *
  338:          CALL DCOPY( M, D, 1, X, 1 )
  339:       END IF
  340: *
  341: *     Backward transformation y = Z**T *y
  342: *
  343:       CALL DORMRQ( 'Left', 'Transpose', P, 1, NP,
  344:      $             B( MAX( 1, N-P+1 ), 1 ), LDB, WORK( M+1 ), Y,
  345:      $             MAX( 1, P ), WORK( M+NP+1 ), LWORK-M-NP, INFO )
  346:       WORK( 1 ) = M + NP + MAX( LOPT, INT( WORK( M+NP+1 ) ) )
  347: *
  348:       RETURN
  349: *
  350: *     End of DGGGLM
  351: *
  352:       END

CVSweb interface <joel.bertrand@systella.fr>