dggsvd (l) - Linux Manuals

dggsvd: computes the generalized singular value decomposition (GSVD) of an M-by-N real matrix A and P-by-N real matrix B

NAME

DGGSVD - computes the generalized singular value decomposition (GSVD) of an M-by-N real matrix A and P-by-N real matrix B

SYNOPSIS

SUBROUTINE DGGSVD(
JOBU, JOBV, JOBQ, M, N, P, K, L, A, LDA, B, LDB, ALPHA, BETA, U, LDU, V, LDV, Q, LDQ, WORK, IWORK, INFO )

    
CHARACTER JOBQ, JOBU, JOBV

    
INTEGER INFO, K, L, LDA, LDB, LDQ, LDU, LDV, M, N, P

    
INTEGER IWORK( * )

    
DOUBLE PRECISION A( LDA, * ), ALPHA( * ), B( LDB, * ), BETA( * ), Q( LDQ, * ), U( LDU, * ), V( LDV, * ), WORK( * )

PURPOSE

DGGSVD computes the generalized singular value decomposition (GSVD) of an M-by-N real matrix A and P-by-N real matrix B:
 Uaq*A*Q = D1*( 0 R ),    Vaq*B*Q = D2*( 0 R )
where U, V and Q are orthogonal matrices, and Zaq is the transpose of Z. Let K+L = the effective numerical rank of the matrix (Aaq,Baq)aq, then R is a K+L-by-K+L nonsingular upper triangular matrix, D1 and D2 are M-by-(K+L) and P-by-(K+L) "diagonal" matrices and of the following structures, respectively:
If M-K-L >= 0,

              K  L

 D1 =     K ( I  0 )

          L ( 0  C )

      M-K-L ( 0  0 )

            K  L

 D2 =   L ( 0  S )

      P-L ( 0  0 )

          N-K-L  K    L

  ( 0 R ) = K (  0   R11  R12 )

      L (  0    0   R22 )
where

  C = diag( ALPHA(K+1), ... , ALPHA(K+L) ),

  S = diag( BETA(K+1),  ... , BETA(K+L) ),

  C**2 + S**2 = I.

  R is stored in A(1:K+L,N-K-L+1:N) on exit.
If M-K-L < 0,

            K M-K K+L-M

 D1 =   K ( I  0    0   )

      M-K ( 0  C    0   )

              K M-K K+L-M

 D2 =   M-K ( 0  S    0  )

      K+L-M ( 0  0    I  )

        P-L ( 0  0    0  )

             N-K-L  K   M-K  K+L-M

  ( 0 R ) =     K ( 0    R11  R12  R13  )

        M-K ( 0     0   R22  R23  )

      K+L-M ( 0     0    0   R33  )
where

  C = diag( ALPHA(K+1), ... , ALPHA(M) ),

  S = diag( BETA(K+1),  ... , BETA(M) ),

  C**2 + S**2 = I.

  (R11 R12 R13 ) is stored in A(1:M, N-K-L+1:N), and R33 is stored
  ( 0  R22 R23 )

  in B(M-K+1:L,N+M-K-L+1:N) on exit.
The routine computes C, S, R, and optionally the orthogonal transformation matrices U, V and Q.
In particular, if B is an N-by-N nonsingular matrix, then the GSVD of A and B implicitly gives the SVD of A*inv(B):

               A*inv(B) = U*(D1*inv(D2))*Vaq.
If ( Aaq,Baq)aq has orthonormal columns, then the GSVD of A and B is also equal to the CS decomposition of A and B. Furthermore, the GSVD can be used to derive the solution of the eigenvalue problem:
               Aaq*A x = lambda* Baq*B x.
In some literature, the GSVD of A and B is presented in the form
           Uaq*A*X = ( 0 D1 ),   Vaq*B*X = ( 0 D2 )
where U and V are orthogonal and X is nonsingular, D1 and D2 are ``diagonalaqaq. The former GSVD form can be converted to the latter form by taking the nonsingular matrix X as

               X = Q*( I   0    )

                     ( 0 inv(R) ).

ARGUMENTS

JOBU (input) CHARACTER*1
= aqUaq: Orthogonal matrix U is computed;
= aqNaq: U is not computed.
JOBV (input) CHARACTER*1

= aqVaq: Orthogonal matrix V is computed;
= aqNaq: V is not computed.
JOBQ (input) CHARACTER*1

= aqQaq: Orthogonal matrix Q is computed;
= aqNaq: Q is not computed.
M (input) INTEGER
The number of rows of the matrix A. M >= 0.
N (input) INTEGER
The number of columns of the matrices A and B. N >= 0.
P (input) INTEGER
The number of rows of the matrix B. P >= 0.
K (output) INTEGER
L (output) INTEGER On exit, K and L specify the dimension of the subblocks described in the Purpose section. K + L = effective numerical rank of (Aaq,Baq)aq.
A (input/output) DOUBLE PRECISION array, dimension (LDA,N)
On entry, the M-by-N matrix A. On exit, A contains the triangular matrix R, or part of R. See Purpose for details.
LDA (input) INTEGER
The leading dimension of the array A. LDA >= max(1,M).
B (input/output) DOUBLE PRECISION array, dimension (LDB,N)
On entry, the P-by-N matrix B. On exit, B contains the triangular matrix R if M-K-L < 0. See Purpose for details.
LDB (input) INTEGER
The leading dimension of the array B. LDB >= max(1,P).
ALPHA (output) DOUBLE PRECISION array, dimension (N)
BETA (output) DOUBLE PRECISION array, dimension (N) On exit, ALPHA and BETA contain the generalized singular value pairs of A and B; ALPHA(1:K) = 1,
BETA(1:K) = 0, and if M-K-L >= 0, ALPHA(K+1:K+L) = C,
BETA(K+1:K+L) = S, or if M-K-L < 0, ALPHA(K+1:M)=C, ALPHA(M+1:K+L)=0
BETA(K+1:M) =S, BETA(M+1:K+L) =1 and ALPHA(K+L+1:N) = 0
BETA(K+L+1:N) = 0
U (output) DOUBLE PRECISION array, dimension (LDU,M)
If JOBU = aqUaq, U contains the M-by-M orthogonal matrix U. If JOBU = aqNaq, U is not referenced.
LDU (input) INTEGER
The leading dimension of the array U. LDU >= max(1,M) if JOBU = aqUaq; LDU >= 1 otherwise.
V (output) DOUBLE PRECISION array, dimension (LDV,P)
If JOBV = aqVaq, V contains the P-by-P orthogonal matrix V. If JOBV = aqNaq, V is not referenced.
LDV (input) INTEGER
The leading dimension of the array V. LDV >= max(1,P) if JOBV = aqVaq; LDV >= 1 otherwise.
Q (output) DOUBLE PRECISION array, dimension (LDQ,N)
If JOBQ = aqQaq, Q contains the N-by-N orthogonal matrix Q. If JOBQ = aqNaq, Q is not referenced.
LDQ (input) INTEGER
The leading dimension of the array Q. LDQ >= max(1,N) if JOBQ = aqQaq; LDQ >= 1 otherwise.
WORK (workspace) DOUBLE PRECISION array,
dimension (max(3*N,M,P)+N)
IWORK (workspace/output) INTEGER array, dimension (N)
On exit, IWORK stores the sorting information. More precisely, the following loop will sort ALPHA for I = K+1, min(M,K+L) swap ALPHA(I) and ALPHA(IWORK(I)) endfor such that ALPHA(1) >= ALPHA(2) >= ... >= ALPHA(N).
INFO (output) INTEGER
= 0: successful exit
< 0: if INFO = -i, the i-th argument had an illegal value.
> 0: if INFO = 1, the Jacobi-type procedure failed to converge. For further details, see subroutine DTGSJA.

PARAMETERS

TOLA DOUBLE PRECISION
TOLB DOUBLE PRECISION TOLA and TOLB are the thresholds to determine the effective rank of (Aaq,Baq)aq. Generally, they are set to TOLA = MAX(M,N)*norm(A)*MAZHEPS, TOLB = MAX(P,N)*norm(B)*MAZHEPS. The size of TOLA and TOLB may affect the size of backward errors of the decomposition. Further Details =============== 2-96 Based on modifications by Ming Gu and Huan Ren, Computer Science Division, University of California at Berkeley, USA

Pages related to dggsvd

  • dggsvd (3)
  • dggsvp (l) - computes orthogonal matrices U, V and Q such that N-K-L K L Uaq*A*Q = K ( 0 A12 A13 ) if M-K-L >= 0
  • dggbak (l) - forms the right or left eigenvectors of a real generalized eigenvalue problem A*x = lambda*B*x, by backward transformation on the computed eigenvectors of the balanced pair of matrices output by DGGBAL
  • dggbal (l) - balances a pair of general real matrices (A,B)
  • dgges (l) - computes for a pair of N-by-N real nonsymmetric matrices (A,B),
  • dggesx (l) - computes for a pair of N-by-N real nonsymmetric matrices (A,B), the generalized eigenvalues, the real Schur form (S,T), and,
  • dggev (l) - computes for a pair of N-by-N real nonsymmetric matrices (A,B)
  • dggevx (l) - computes for a pair of N-by-N real nonsymmetric matrices (A,B)