## MB03XD

### Computing the eigenvalues of a Hamiltonian matrix

[Specification] [Arguments] [Method] [References] [Comments] [Example]

Purpose

```  To compute the eigenvalues of a Hamiltonian matrix,

[  A   G  ]         T        T
H  =  [       T ],   G = G,   Q = Q,                  (1)
[  Q  -A  ]

where A, G and Q are real n-by-n matrices.

Due to the structure of H all eigenvalues appear in pairs
(lambda,-lambda). This routine computes the eigenvalues of H
using an algorithm based on the symplectic URV and the periodic
Schur decompositions as described in [1],

T       [  T   G  ]
U H V =  [       T ],                                    (2)
[  0   S  ]

where U and V are 2n-by-2n orthogonal symplectic matrices,
S is in real Schur form and T is upper triangular.

The algorithm is backward stable and preserves the eigenvalue
pairings in finite precision arithmetic.

Optionally, a symplectic balancing transformation to improve the
conditioning of eigenvalues is computed (see MB04DD). In this
case, the matrix H in decomposition (2) must be replaced by the
balanced matrix.

The SLICOT Library routine MB03ZD can be used to compute invariant
subspaces of H from the output of this routine.

```
Specification
```      SUBROUTINE MB03XD( BALANC, JOB, JOBU, JOBV, N, A, LDA, QG, LDQG,
\$                   T, LDT, U1, LDU1, U2, LDU2, V1, LDV1, V2, LDV2,
\$                   WR, WI, ILO, SCALE, DWORK, LDWORK, INFO )
C     .. Scalar Arguments ..
CHARACTER          BALANC, JOB, JOBU, JOBV
INTEGER            ILO, INFO, LDA, LDQG, LDT, LDU1, LDU2, LDV1,
\$                   LDV2, LDWORK, N
C     .. Array Arguments ..
DOUBLE PRECISION   A(LDA,*), DWORK(*), QG(LDQG,*), SCALE(*),
\$                   T(LDT,*), U1(LDU1,*), U2(LDU2,*), V1(LDV1,*),
\$                   V2(LDV2,*), WI(*), WR(*)

```
Arguments

Mode Parameters

```  BALANC  CHARACTER*1
Indicates how H should be diagonally scaled and/or
permuted to reduce its norm.
= 'N': Do not diagonally scale or permute;
= 'P': Perform symplectic permutations to make the matrix
closer to Hamiltonian Schur form. Do not diagonally
scale;
= 'S': Diagonally scale the matrix, i.e., replace A, G and
Q by D*A*D**(-1), D*G*D and D**(-1)*Q*D**(-1) where
D is a diagonal matrix chosen to make the rows and
columns of H more equal in norm. Do not permute;
= 'B': Both diagonally scale and permute A, G and Q.
Permuting does not change the norm of H, but scaling does.

JOB     CHARACTER*1
Indicates whether the user wishes to compute the full
decomposition (2) or the eigenvalues only, as follows:
= 'E': compute the eigenvalues only;
= 'S': compute matrices T and S of (2);
= 'G': compute matrices T, S and G of (2).

JOBU    CHARACTER*1
Indicates whether or not the user wishes to compute the
orthogonal symplectic matrix U of (2) as follows:
= 'N': the matrix U is not computed;
= 'U': the matrix U is computed.

JOBV    CHARACTER*1
Indicates whether or not the user wishes to compute the
orthogonal symplectic matrix V of (2) as follows:
= 'N': the matrix V is not computed;
= 'V': the matrix V is computed.

```
Input/Output Parameters
```  N       (input) INTEGER
The order of the matrix A. N >= 0.

A       (input/output) DOUBLE PRECISION array, dimension (LDA,N)
On entry, the leading N-by-N part of this array must
contain the matrix A.
On exit, this array is overwritten. If JOB = 'S' or
JOB = 'G', the leading N-by-N part of this array contains
the matrix S in real Schur form of decomposition (2).

LDA     INTEGER
The leading dimension of the array A.  LDA >= max(1,N).

QG      (input/output) DOUBLE PRECISION array, dimension
(LDQG,N+1)
On entry, the leading N-by-N+1 part of this array must
contain in columns 1:N the lower triangular part of the
matrix Q and in columns 2:N+1 the upper triangular part
of the matrix G.
On exit, this array is overwritten. If JOB = 'G', the
leading N-by-N+1 part of this array contains in columns
2:N+1 the matrix G of decomposition (2).

LDQG    INTEGER
The leading dimension of the array QG.  LDQG >= max(1,N).

T       (output) DOUBLE PRECISION array, dimension (LDT,N)
On exit, if JOB = 'S' or JOB = 'G', the leading N-by-N
part of this array contains the upper triangular matrix T
of the decomposition (2). Otherwise, this array is used as
workspace.

LDT     INTEGER
The leading dimension of the array T.  LDT >= MAX(1,N).

U1      (output) DOUBLE PRECISION array, dimension (LDU1,N)
On exit, if JOBU = 'U', the leading N-by-N part of this
array contains the (1,1) block of the orthogonal
symplectic matrix U of decomposition (2).

LDU1    INTEGER
The leading dimension of the array U1.  LDU1 >= 1.
LDU1 >= N,    if JOBU = 'U'.

U2      (output) DOUBLE PRECISION array, dimension (LDU2,N)
On exit, if JOBU = 'U', the leading N-by-N part of this
array contains the (2,1) block of the orthogonal
symplectic matrix U of decomposition (2).

LDU2    INTEGER
The leading dimension of the array U2.  LDU2 >= 1.
LDU2 >= N,    if JOBU = 'U'.

V1      (output) DOUBLE PRECISION array, dimension (LDV1,N)
On exit, if JOBV = 'V', the leading N-by-N part of this
array contains the (1,1) block of the orthogonal
symplectic matrix V of decomposition (2).

LDV1    INTEGER
The leading dimension of the array V1.  LDV1 >= 1.
LDV1 >= N,    if JOBV = 'V'.

V2      (output) DOUBLE PRECISION array, dimension (LDV2,N)
On exit, if JOBV = 'V', the leading N-by-N part of this
array contains the (2,1) block of the orthogonal
symplectic matrix V of decomposition (2).

LDV2    INTEGER
The leading dimension of the array V2.  LDV2 >= 1.
LDV2 >= N,    if JOBV = 'V'.

WR      (output) DOUBLE PRECISION array, dimension (N)
WI      (output) DOUBLE PRECISION array, dimension (N)
On exit, the leading N elements of WR and WI contain the
real and imaginary parts, respectively, of N eigenvalues
that have nonnegative imaginary part. Their complex
conjugate eigenvalues are not stored. If imaginary parts
are zero (i.e., for real eigenvalues), only positive
eigenvalues are stored.

ILO     (output) INTEGER
ILO is an integer value determined when H was balanced.
The balanced A(i,j) = 0 if I > J and J = 1,...,ILO-1.
The balanced Q(i,j) = 0 if J = 1,...,ILO-1 or
I = 1,...,ILO-1.

SCALE   (output) DOUBLE PRECISION array, dimension (N)
On exit, if BALANC <> 'N', the leading N elements of this
array contain details of the permutation and/or scaling
factors applied when balancing H, see MB04DD.
This array is not referenced if BALANC = 'N'.

```
Workspace
```  DWORK   DOUBLE PRECISION array, dimension (LDWORK)
On exit, if INFO = 0,  DWORK(1)  returns the optimal
value of LDWORK, and   DWORK(2)  returns the 1-norm of the
scaled (if BALANC = 'S' or 'B') Hamiltonian matrix.
On exit, if  INFO = -25,  DWORK(1)  returns the minimum
value of LDWORK.

LDWORK  (input) INTEGER
The dimension of the array DWORK. LDWORK >= max( 2, 8*N ).
Moreover:
If JOB = 'E' or 'S' and JOBU = 'N' and JOBV = 'N',
LDWORK >= 7*N+N*N.
If JOB = 'G' and JOBU = 'N' and JOBV = 'N',
LDWORK >= max( 7*N+N*N, 2*N+3*N*N ).
If JOB = 'G' and JOBU = 'U' and JOBV = 'N',
LDWORK >= 7*N+2*N*N.
If JOB = 'G' and JOBU = 'N' and JOBV = 'V',
LDWORK >= 7*N+2*N*N.
If JOB = 'G' and JOBU = 'U' and JOBV = 'V',
LDWORK >= 7*N+N*N.
For good performance, LDWORK must generally be larger.

If LDWORK = -1, then a workspace query is assumed;
the routine only calculates the optimal size of the
DWORK array, returns this value as the first entry of
the DWORK array, and no error message related to LDWORK
is issued by XERBLA.

```
Error Indicator
```  INFO     (output) INTEGER
= 0:  successful exit;
< 0:  if INFO = -i, the i-th argument had an illegal
value;
> 0:  if INFO = i, the periodic QR algorithm failed to
compute all the eigenvalues, elements i+1:N of WR
and WI contain eigenvalues which have converged.

```
References
```  [1] Benner, P., Mehrmann, V., and Xu, H.
A numerically stable, structure preserving method for
computing the eigenvalues of real Hamiltonian or symplectic
pencils.
Numer. Math., Vol. 78(3), pp. 329-358, 1998.

[2] Benner, P., Mehrmann, V., and Xu, H.
A new method for computing the stable invariant subspace of a
real Hamiltonian matrix,  J. Comput. Appl. Math., vol. 86,
pp. 17-43, 1997.

```
```  None
```
Example

Program Text

```*     MB03XD EXAMPLE PROGRAM TEXT
*     Copyright (c) 2002-2017 NICONET e.V.
*
*     .. Parameters ..
DOUBLE PRECISION ZERO, ONE
PARAMETER        ( ZERO = 0.0D0, ONE = 1.0D0 )
INTEGER          NIN, NOUT
PARAMETER        ( NIN = 5, NOUT = 6 )
INTEGER          NMAX
PARAMETER        ( NMAX = 100 )
INTEGER          LDA, LDQG, LDRES, LDT, LDU1, LDU2, LDV1, LDV2,
\$                 LDWORK
PARAMETER        ( LDA = NMAX, LDQG = NMAX, LDRES = NMAX,
\$                   LDT = NMAX, LDU1 = NMAX, LDU2 = NMAX,
\$                   LDV1 = NMAX, LDV2 = NMAX,
\$                   LDWORK = 3*NMAX*NMAX + 7*NMAX )
*     .. Local Scalars ..
CHARACTER*1      BALANC, JOB, JOBU, JOBV
INTEGER          I, ILO, INFO, J, N
DOUBLE PRECISION TEMP
*     .. Local Arrays ..
DOUBLE PRECISION A(LDA, NMAX), DWORK(LDWORK), QG(LDQG, NMAX+1),
\$                 RES(LDRES,3*NMAX+1), SCALE(NMAX), T(LDT,NMAX),
\$                 U1(LDU1,NMAX), U2(LDU2, NMAX), V1(LDV1,NMAX),
\$                 V2(LDV2, NMAX), WI(NMAX), WR(NMAX)
*     .. External Functions ..
LOGICAL          LSAME
DOUBLE PRECISION DLANGE, DLAPY2, MA02JD
EXTERNAL         DLANGE, DLAPY2, LSAME, MA02JD
*     .. External Subroutines ..
EXTERNAL         DGEMM, DLACPY, MB03XD, MB04DD
*     .. Executable Statements ..
WRITE ( NOUT, FMT = 99999 )
*     Skip the heading in the data file and read the data.
READ ( NIN, FMT = '()' )
READ ( NIN, FMT = * )  N, BALANC, JOB, JOBU, JOBV
IF ( N.LE.0 .OR. N.GT.NMAX ) THEN
WRITE ( NOUT, FMT = 99988 ) N
ELSE
READ ( NIN, FMT = * ) ( ( A(I,J), J = 1,N ), I = 1,N )
CALL DLACPY( 'All', N, N, A, LDA, RES(1,N+1), LDRES )
READ ( NIN, FMT = * ) ( ( QG(I,J), J = 1,N+1 ), I = 1,N )
CALL DLACPY( 'All', N, N+1, QG, LDQG, RES(1,2*N+1), LDRES )
INFO = 0
CALL MB03XD( BALANC, JOB, JOBU, JOBV, N, A, LDA, QG, LDQG,
\$                T, LDT, U1, LDU1, U2, LDU2, V1, LDV1, V2, LDV2,
\$                WR, WI, ILO, SCALE, DWORK, LDWORK, INFO )
IF ( INFO.NE.0 ) THEN
WRITE ( NOUT, FMT = 99998 ) INFO
ELSE
WRITE ( NOUT, FMT = 99997 )
DO 20  I = 1, N
WRITE ( NOUT, FMT = 99996 ) I, WR(I), WI(I)
20          CONTINUE
IF ( LSAME( JOB, 'S' ).OR.LSAME( JOB, 'G' ) ) THEN
WRITE ( NOUT, FMT = 99995 )
DO 30  I = 1, N
WRITE ( NOUT, FMT = 99990 ) ( A(I,J), J = 1,N )
30             CONTINUE
WRITE ( NOUT, FMT = 99994 )
DO 40  I = 1, N
WRITE ( NOUT, FMT = 99990 ) ( T(I,J), J = 1,N )
40             CONTINUE
END IF
IF ( LSAME( JOB, 'G' ) ) THEN
WRITE ( NOUT, FMT = 99993 )
DO 50  I = 1, N
WRITE ( NOUT, FMT = 99990 ) ( QG(I,J+1), J = 1,N )
50             CONTINUE
END IF
C
IF ( LSAME( JOB, 'G' ).AND.LSAME( JOBU, 'U' ).AND.
\$           LSAME( JOBV, 'V' ) ) THEN
CALL MB04DD( BALANC, N, RES(1,N+1), LDRES, RES(1,2*N+1),
\$                      LDRES, I, DWORK, INFO )
CALL DGEMM( 'No Transpose', 'No Transpose', N, N, N, ONE,
\$                     RES(1,N+1), LDRES, V1, LDV1, ZERO, RES,
\$                     LDRES )
CALL DSYMM ( 'Left', 'Upper', N, N, -ONE, RES(1,2*N+2),
\$                      LDRES, V2, LDV2, ONE, RES, LDRES )
CALL DGEMM( 'No Transpose', 'No Transpose', N, N, N,
\$                     -ONE, U1, LDU1, T, LDT, ONE, RES, LDRES )
TEMP = DLANGE( 'Frobenius', N, N, RES, LDRES, DWORK )
CALL DGEMM( 'No Transpose', 'No Transpose', N, N, N, ONE,
\$                     RES(1,N+1), LDRES, V2, LDV2, ZERO, RES,
\$                     LDRES )
CALL DSYMM( 'Left', 'Upper', N, N, ONE, RES(1,2*N+2),
\$                     LDRES, V1, LDV1, ONE, RES, LDRES )
CALL DGEMM( 'No Transpose', 'No Transpose', N, N, N,
\$                     -ONE, U1, LDU1, QG(1,2), LDQG, ONE, RES,
\$                     LDRES )
CALL DGEMM( 'No Transpose', 'Transpose', N, N, N,
\$                     -ONE, U2, LDU2, A, LDA, ONE, RES, LDRES )
TEMP = DLAPY2( TEMP, DLANGE( 'Frobenius', N, N, RES,
\$                                      LDRES, DWORK ) )
CALL DSYMM( 'Left', 'Lower', N, N, ONE, RES(1,2*N+1),
\$                     LDRES, V1, LDV1, ZERO, RES, LDRES )
CALL DGEMM( 'Transpose', 'No Transpose', N, N, N, ONE,
\$                     RES(1,N+1), LDRES, V2, LDV2, ONE, RES,
\$                     LDRES )
CALL DGEMM( 'No Transpose', 'No Transpose', N, N, N, ONE,
\$                     U2, LDU2, T, LDT, ONE, RES, LDRES )
TEMP = DLAPY2( TEMP, DLANGE( 'Frobenius', N, N, RES,
\$                                      LDRES, DWORK ) )

CALL DSYMM( 'Left', 'Lower', N, N, ONE, RES(1,2*N+1),
\$                     LDRES, V2, LDV2, ZERO, RES, LDRES )
CALL DGEMM( 'Transpose', 'No Transpose', N, N, N, -ONE,
\$                     RES(1,N+1), LDRES, V1, LDV1, ONE, RES,
\$                     LDRES )
CALL DGEMM( 'No Transpose', 'No Transpose', N, N, N, ONE,
\$                     U2, LDU2, QG(1,2), LDQG, ONE, RES, LDRES )
CALL DGEMM( 'No Transpose', 'Transpose', N, N, N,
\$                     -ONE, U1, LDU1, A, LDA, ONE, RES, LDRES )
TEMP = DLAPY2( TEMP, DLANGE( 'Frobenius', N, N, RES,
\$                                      LDRES, DWORK ) )
WRITE ( NOUT, FMT = 99987 ) TEMP
END IF
C
IF ( LSAME( JOBU, 'U' ) ) THEN
WRITE ( NOUT, FMT = 99992 )
DO 60  I = 1, N
WRITE ( NOUT, FMT = 99990 )
\$               ( U1(I,J), J = 1,N ), ( U2(I,J), J = 1,N )
60             CONTINUE
DO 70  I = 1, N
WRITE ( NOUT, FMT = 99990 )
\$               ( -U2(I,J), J = 1,N ), ( U1(I,J), J = 1,N )
70             CONTINUE
WRITE ( NOUT, FMT = 99986 ) MA02JD( .FALSE., .FALSE., N,
\$                 U1, LDU1, U2, LDU2, RES, LDRES )
END IF
IF ( LSAME( JOBV, 'V' ) ) THEN
WRITE ( NOUT, FMT = 99991 )
DO 80  I = 1, N
WRITE ( NOUT, FMT = 99990 )
\$               ( V1(I,J), J = 1,N ), ( V2(I,J), J = 1,N )
80             CONTINUE
DO 90  I = 1, N
WRITE ( NOUT, FMT = 99990 )
\$               ( -V2(I,J), J = 1,N ), ( V1(I,J), J = 1,N )
90             CONTINUE
WRITE ( NOUT, FMT = 99985 ) MA02JD( .FALSE., .FALSE., N,
\$                 V1, LDV1, V2, LDV2, RES, LDRES )
END IF
IF ( LSAME( BALANC, 'S' ).OR.LSAME( BALANC, 'B' ) ) THEN
WRITE ( NOUT, FMT = 99989 )
DO 100  I = 1, N
WRITE ( NOUT, FMT = 99996 ) I, SCALE(I)
100            CONTINUE
END IF
END IF
END IF
*
99999 FORMAT (' MB03XD EXAMPLE PROGRAM RESULTS',/1X)
99998 FORMAT (' INFO on exit from MB03XD = ',I2)
99997 FORMAT (' The eigenvalues are',//'   i',6X,
\$        'WR(i)',6X,'WI(i)',/)
99996 FORMAT (I4,3X,F8.4,3X,F8.4)
99995 FORMAT (/' The matrix S of the reduced matrix is')
99994 FORMAT (/' The matrix T of the reduced matrix is')
99993 FORMAT (/' The matrix G of the reduced matrix is')
99992 FORMAT (/' The orthogonal symplectic factor U is')
99991 FORMAT (/' The orthogonal symplectic factor V is')
99990 FORMAT (20(1X,F19.16))
99989 FORMAT (/' The diagonal scaling factors are ',//'   i',6X,
\$        'SCALE(i)',/)
99988 FORMAT (/' N is out of range.',/' N = ',I5)
99987 FORMAT (/' Residual: || H*V - U*R ||_F = ',G7.2)
99986 FORMAT (/' Orthogonality of U: || U^T U - I ||_F = ',G7.2)
99985 FORMAT (/' Orthogonality of V: || V^T V - I ||_F = ',G7.2)
END
```
Program Data
```MB03XD EXAMPLE PROGRAM DATA
5	N	G	U	V
3.7588548168313685e-001  9.1995720669587144e-001  1.9389317998466821e-001  5.4878212553858818e-001  6.2731478808399666e-001
9.8764628987858052e-003  8.4472150190817474e-001  9.0481233416635698e-001  9.3158335257969060e-001  6.9908013774533750e-001
4.1985780631021896e-001  3.6775288246828447e-001  5.6920574967174709e-001  3.3519743020639464e-001  3.9718395379261456e-001
7.5366962581358721e-001  6.2080133182114383e-001  6.3178992922175603e-001  6.5553105501201447e-001  4.1362889533818031e-001
7.9387177473231862e-001  7.3127726446634478e-001  2.3441295540825388e-001  3.9190420688900335e-001  6.5521294635567051e-001
1.8015558545989005e-001  4.1879254941592853e-001  2.7203760737317784e-001  2.8147214090719214e-001  1.7731904815580199e-001  3.4718672159409536e-001
2.7989257702981651e-001  3.5042861661866559e-001  2.5565572408444881e-001  4.3977750345993827e-001  2.8855026075967616e-001  2.1496327083014577e-001
1.7341073886969158e-001  3.9913855375815932e-001  4.0151317011596516e-001  4.0331887464437133e-001  2.6723538667317948e-001  3.7110275606849241e-001
3.7832182695699140e-001  3.3812641389556752e-001  8.4360396433341395e-002  4.3672540277019672e-001  7.0022228267365608e-002  3.8210230186291916e-001
1.9548216143135175e-001  2.9055490787446736e-001  4.7670819669167425e-001  1.4636498713707141e-001  2.7670398401519275e-001  2.9431082727794898e-002
```
Program Results
``` MB03XD EXAMPLE PROGRAM RESULTS

The eigenvalues are

i      WR(i)      WI(i)

1     3.1941     0.0000
2     0.1350     0.3179
3    -0.1350     0.3179
4     0.0595     0.2793
5    -0.0595     0.2793

The matrix S of the reduced matrix is
-3.1844761777714723  0.1612357243439330 -0.0628592203751083  0.2449004200921966  0.1974400149992633
0.0000000000000000 -0.1510667773167795  0.4260444411622876 -0.1775026035208666  0.3447278421198404
0.0000000000000000 -0.1386140422054281 -0.3006779624777419  0.2944143257134117  0.3456440339120381
0.0000000000000000  0.0000000000000000  0.0000000000000000 -0.2710128384740589  0.0933189808067083
0.0000000000000000  0.0000000000000000  0.0000000000000000  0.4844146572359630  0.2004347508746724

The matrix T of the reduced matrix is
3.2038208121776366  0.1805955192510636  0.2466389119377562 -0.2539149302433404 -0.0359238844381174
0.0000000000000000 -0.7196686433290822  0.0000000000000000  0.2428659121580382 -0.0594190100670709
0.0000000000000000  0.0000000000000000 -0.1891741194498124 -0.3309578443491325 -0.0303520731950498
0.0000000000000000  0.0000000000000000  0.0000000000000000 -0.4361574461961496  0.0000000000000000
0.0000000000000000  0.0000000000000000  0.0000000000000000  0.0000000000000000  0.1530894573304220

The matrix G of the reduced matrix is
-0.0370982242678458  0.0917788436945730 -0.0560402416315236  0.1345152517579191  0.0256668227276665
0.0652183678916926 -0.0700457231988328  0.0350041175858816 -0.2233868768749277 -0.1171980260782820
-0.0626428681377074  0.2327575351902838 -0.1251515732208133 -0.0177816046663209  0.3696921118421109
0.0746042309265569 -0.0828007611045243  0.0217427473546003 -0.1157775118548850 -0.3161183681200607
0.1374372236164838  0.1002727885506978  0.4021556774753987 -0.0431072263235625  0.1067394572547804

Residual: || H*V - U*R ||_F = .38E-14

The orthogonal symplectic factor U is
0.3806883009357249 -0.0347810363019666 -0.5014665065895627  0.5389691288472425  0.2685446895251499 -0.1795922007470744  0.1908329820840935  0.0868799433942041  0.3114741142062469 -0.2579907627915116
0.4642712665555327 -0.5942766860716397  0.4781179763952658  0.2334370556238072  0.0166790369048881 -0.2447897730222852 -0.1028403314750053 -0.1157840914576285 -0.1873268885694422  0.1700708002861556
0.2772789197782788 -0.0130145392695854 -0.2123817030594153 -0.2550292626960007 -0.5049268366774478 -0.2243335325285329  0.3180998613802498  0.3315380214794935  0.1977859924739816  0.5072476567310036
0.4209268575081797  0.1499593172661210 -0.1925590746592206 -0.5472292877802408  0.4543329704184014 -0.2128397588651423 -0.2740560593051884  0.1941418870268831 -0.3096684962457407 -0.0581576193198820
0.3969669479129447  0.6321903535930841  0.3329156356041941  0.0163533225344391 -0.2638879466190056 -0.2002027567371933 -0.0040094115506845 -0.3979373387545256  0.1520881534833996 -0.2010804514091296
0.1795922007470744 -0.1908329820840935 -0.0868799433942041 -0.3114741142062469  0.2579907627915116  0.3806883009357249 -0.0347810363019666 -0.5014665065895627  0.5389691288472425  0.2685446895251499
0.2447897730222852  0.1028403314750053  0.1157840914576285  0.1873268885694422 -0.1700708002861556  0.4642712665555327 -0.5942766860716397  0.4781179763952658  0.2334370556238072  0.0166790369048881
0.2243335325285329 -0.3180998613802498 -0.3315380214794935 -0.1977859924739816 -0.5072476567310036  0.2772789197782788 -0.0130145392695854 -0.2123817030594153 -0.2550292626960007 -0.5049268366774478
0.2128397588651423  0.2740560593051884 -0.1941418870268831  0.3096684962457407  0.0581576193198820  0.4209268575081797  0.1499593172661210 -0.1925590746592206 -0.5472292877802408  0.4543329704184014
0.2002027567371933  0.0040094115506845  0.3979373387545256 -0.1520881534833996  0.2010804514091296  0.3969669479129447  0.6321903535930841  0.3329156356041941  0.0163533225344391 -0.2638879466190056

Orthogonality of U: || U^T U - I ||_F = .28E-14

The orthogonal symplectic factor V is
0.4447147692018332 -0.6830166755147445 -0.0002576861753461  0.5781954611783305 -0.0375091627893695  0.0000000000000000  0.0000000000000000  0.0000000000000000  0.0000000000000000  0.0000000000000000
0.5121756358795815  0.0297197140254867  0.4332229148788674 -0.3240527006890555  0.5330850295256576  0.0299719306696789 -0.2322624725320732 -0.0280846899680330 -0.3044255686880015 -0.1077641482535463
0.3664711365265602  0.3288511296455134  0.0588396016404466  0.1134221597062252  0.1047567336850027 -0.0069083614679702  0.3351358347080169 -0.4922707032978909  0.4293545450291777  0.4372821269061838
0.4535357098437908  0.1062866148880810 -0.3964092656837799 -0.2211800890450648  0.0350667323996154  0.0167847133528844  0.2843629278945263  0.5958979805231206  0.3097336757510830 -0.2086733033047175
0.4450432900616098  0.2950206358263727 -0.1617837757183794 -0.0376369332204945 -0.6746752660482708  0.0248567764822071 -0.2810759958040465 -0.1653113624869875 -0.3528780198620394 -0.0254898556119200
0.0000000000000000  0.0000000000000000  0.0000000000000000  0.0000000000000000  0.0000000000000000  0.4447147692018332 -0.6830166755147445 -0.0002576861753461  0.5781954611783305 -0.0375091627893695
-0.0299719306696789  0.2322624725320732  0.0280846899680330  0.3044255686880015  0.1077641482535463  0.5121756358795815  0.0297197140254867  0.4332229148788674 -0.3240527006890555  0.5330850295256576
0.0069083614679702 -0.3351358347080169  0.4922707032978909 -0.4293545450291777 -0.4372821269061838  0.3664711365265602  0.3288511296455134  0.0588396016404466  0.1134221597062252  0.1047567336850027
-0.0167847133528844 -0.2843629278945263 -0.5958979805231206 -0.3097336757510830  0.2086733033047175  0.4535357098437908  0.1062866148880810 -0.3964092656837799 -0.2211800890450648  0.0350667323996154
-0.0248567764822071  0.2810759958040465  0.1653113624869875  0.3528780198620394  0.0254898556119200  0.4450432900616098  0.2950206358263727 -0.1617837757183794 -0.0376369332204945 -0.6746752660482708

Orthogonality of V: || V^T V - I ||_F = .25E-14
```