1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60
61
62
63
64
65
66
67
68
69
70
71
72
73
74
75
76
77
78
79
80
81
82
83
84
85
86
87
88
89
90
91
92
93
94
95
96
97
98
99
100
101
102
103
104
105
106
107
108
109
110
111
112
113
114
115
116
117
118
119
120
121
122
123
124
125
126
127
128
129
130
131
132
133
134
135
136
137
138
139
140
141
142
143
144
145
146
147
148
149
150
151
152
153
154
155
156
157
158
159
160
161
162
163
164
165
166
167
168
169
170
171
172
173
174
175
176
177
178
179
180
181
182
183
184
185
186
187
188
189
190
191
192
193
194
195
196
197
198
199
200
201
202
203
204
205
206
207
208
209
210
211
212
213
214
215
216
217
218
219
220
221
222
223
224
225
226
227
228
229
230
231
232
233
234
235
236
237
238
239
240
241
242
243
244
245
246
247
248
249
250
251
252
253
254
255
256
257
258
259
260
261
262
263
264
265
266
267
268
269
270
271
272
273
274
275
276
277
278
279
280
281
282
283
284
285
286
287
288
289
290
291
292
293
294
295
296
297
298
299
300
301
302
303
304
305
306
307
308
309
310
311
312
313
314
315
316
317
318
319
320
321
322
323
324
325
326
327
328
329
330
331
332
333
334
335
336
337
338
339
340
341
342
343
344
345
346
347
348
349
350
351
352
353
354
355
356
357
358
359
360
361
362
363
364
365
366
367
368
369
370
371
372
373
374
375
376
377
378
379
380
381
382
383
384
|
<HTML>
<HEAD><TITLE>MB04PB - SLICOT Library Routine Documentation</TITLE>
</HEAD>
<BODY>
<H2><A Name="MB04PB">MB04PB</A></H2>
<H3>
Computation of the Paige/Van Loan (PVL) form of a Hamiltonian matrix (block algorithm)
</H3>
<A HREF ="#Specification"><B>[Specification]</B></A>
<A HREF ="#Arguments"><B>[Arguments]</B></A>
<A HREF ="#Method"><B>[Method]</B></A>
<A HREF ="#References"><B>[References]</B></A>
<A HREF ="#Comments"><B>[Comments]</B></A>
<A HREF ="#Example"><B>[Example]</B></A>
<P>
<B><FONT SIZE="+1">Purpose</FONT></B>
<PRE>
To reduce a Hamiltonian matrix,
[ A G ]
H = [ T ] ,
[ Q -A ]
where A is an N-by-N matrix and G,Q are N-by-N symmetric matrices,
to Paige/Van Loan (PVL) form. That is, an orthogonal symplectic U
is computed so that
T [ Aout Gout ]
U H U = [ T ] ,
[ Qout -Aout ]
where Aout is upper Hessenberg and Qout is diagonal.
Blocked version.
</PRE>
<A name="Specification"><B><FONT SIZE="+1">Specification</FONT></B></A>
<PRE>
SUBROUTINE MB04PB( N, ILO, A, LDA, QG, LDQG, CS, TAU, DWORK,
$ LDWORK, INFO )
C .. Scalar Arguments ..
INTEGER ILO, INFO, LDA, LDQG, LDWORK, N
C .. Array Arguments ..
DOUBLE PRECISION A(LDA,*), CS(*), DWORK(*), QG(LDQG,*), TAU(*)
</PRE>
<A name="Arguments"><B><FONT SIZE="+1">Arguments</FONT></B></A>
<P>
</PRE>
<B>Input/Output Parameters</B>
<PRE>
N (input) INTEGER
The order of the matrix A. N >= 0.
ILO (input) INTEGER
It is assumed that A is already upper triangular and Q is
zero in rows and columns 1:ILO-1. ILO is normally set by a
previous call to MB04DD; otherwise it should be set to 1.
1 <= ILO <= N, if N > 0; ILO = 1, if 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, the leading N-by-N part of this array contains
the matrix Aout and, in the zero part of Aout,
information about the elementary reflectors used to
compute the PVL factorization.
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 the lower triangular part of the matrix Q and
the upper triangular part of the matrix G.
On exit, the leading N-by-N+1 part of this array contains
the diagonal of the matrix Qout, the upper triangular part
of the matrix Gout and, in the zero parts of Qout,
information about the elementary reflectors used to
compute the PVL factorization.
LDQG INTEGER
The leading dimension of the array QG. LDQG >= MAX(1,N).
CS (output) DOUBLE PRECISION array, dimension (2N-2)
On exit, the first 2N-2 elements of this array contain the
cosines and sines of the symplectic Givens rotations used
to compute the PVL factorization.
TAU (output) DOUBLE PRECISION array, dimension (N-1)
On exit, the first N-1 elements of this array contain the
scalar factors of some of the elementary reflectors.
</PRE>
<B>Workspace</B>
<PRE>
DWORK DOUBLE PRECISION array, dimension (LDWORK)
On exit, if INFO = 0, DWORK(1) returns the optimal
value of LDWORK, 8*N*NB + 3*NB, where NB is the optimal
block size determined by the function UE01MD.
On exit, if INFO = -10, DWORK(1) returns the minimum
value of LDWORK.
LDWORK INTEGER
The length of the array DWORK. LDWORK >= MAX(1,N-1).
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.
</PRE>
<B>Error Indicator</B>
<PRE>
INFO INTEGER
= 0: successful exit;
< 0: if INFO = -i, the i-th argument had an illegal
value.
</PRE>
<A name="Method"><B><FONT SIZE="+1">Method</FONT></B></A>
<PRE>
The matrix U is represented as a product of symplectic reflectors
and Givens rotations
U = diag( H(1),H(1) ) G(1) diag( F(1),F(1) )
diag( H(2),H(2) ) G(2) diag( F(2),F(2) )
....
diag( H(n-1),H(n-1) ) G(n-1) diag( F(n-1),F(n-1) ).
Each H(i) has the form
H(i) = I - tau * v * v'
where tau is a real scalar, and v is a real vector with
v(1:i) = 0 and v(i+1) = 1; v(i+2:n) is stored on exit in
QG(i+2:n,i), and tau in QG(i+1,i).
Each F(i) has the form
F(i) = I - nu * w * w'
where nu is a real scalar, and w is a real vector with
w(1:i) = 0 and w(i+1) = 1; w(i+2:n) is stored on exit in
A(i+2:n,i), and nu in TAU(i).
Each G(i) is a Givens rotation acting on rows i+1 and n+i+1,
where the cosine is stored in CS(2*i-1) and the sine in
CS(2*i).
</PRE>
<A name="Numerical Aspects"><B><FONT SIZE="+1">Numerical Aspects</FONT></B></A>
<PRE>
The algorithm requires O(N**3) floating point operations and is
strongly backward stable.
</PRE>
<A name="References"><B><FONT SIZE="+1">References</FONT></B></A>
<PRE>
[1] C. F. VAN LOAN:
A symplectic method for approximating all the eigenvalues of
a Hamiltonian matrix.
Linear Algebra and its Applications, 61, pp. 233-251, 1984.
[2] D. KRESSNER:
Block algorithms for orthogonal symplectic factorizations.
BIT, 43 (4), pp. 775-790, 2003.
</PRE>
<A name="Comments"><B><FONT SIZE="+1">Further Comments</FONT></B></A>
<PRE>
None
</PRE>
<A name="Example"><B><FONT SIZE="+1">Example</FONT></B></A>
<P>
<B>Program Text</B>
<PRE>
* MB04PB/MB04WP EXAMPLE PROGRAM TEXT
*
* .. Parameters ..
DOUBLE PRECISION ZERO, ONE, TWO
PARAMETER ( ZERO = 0.0D0, ONE = 1.0D0, TWO = 2.0D0 )
INTEGER NIN, NOUT
PARAMETER ( NIN = 5, NOUT = 6 )
INTEGER NMAX, NBMAX
PARAMETER ( NMAX = 7, NBMAX = 3 )
INTEGER LDA, LDQG, LDRES, LDU1, LDU2, LDWORK
PARAMETER ( LDA = NMAX, LDQG = NMAX, LDRES = NMAX,
$ LDU1 = NMAX, LDU2 = NMAX,
$ LDWORK = 8*NBMAX*NMAX + 3*NBMAX )
* .. Local Scalars ..
INTEGER I, INFO, J, N
* .. Local Arrays ..
DOUBLE PRECISION A(LDA, NMAX), CS(2*NMAX), DWORK(LDWORK),
$ QG(LDQG, NMAX+1), RES(LDRES,3*NMAX+1), TAU(NMAX),
$ U1(LDU1,NMAX), U2(LDU2, NMAX)
* .. External Functions ..
DOUBLE PRECISION MA02ID, MA02JD
EXTERNAL MA02ID, MA02JD
* .. External Subroutines ..
EXTERNAL DGEMM, DLACPY, DLASET, DSCAL, DSYMM, DSYR,
$ DSYR2K, DTRMM, MB04PB, MB04WP
* .. Executable Statements ..
WRITE ( NOUT, FMT = 99999 )
* Skip the heading in the data file and read the data.
READ ( NIN, FMT = '()' )
READ ( NIN, FMT = * ) N
IF( N.LE.0 .OR. N.GT.NMAX ) THEN
WRITE ( NOUT, FMT = 99992 ) 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 )
CALL MB04PB( N, 1, A, LDA, QG, LDQG, CS, TAU, DWORK, LDWORK,
$ INFO )
INFO = 0
IF ( INFO.NE.0 ) THEN
WRITE ( NOUT, FMT = 99998 ) INFO
ELSE
CALL DLACPY( 'Lower', N, N, A, LDA, U1, LDU1 )
CALL DLACPY( 'Lower', N, N, QG, LDQG, U2, LDU2 )
CALL MB04WP( N, 1, U1, LDU1, U2, LDU2, CS, TAU, DWORK,
$ LDWORK, INFO )
IF ( INFO.NE.0 ) THEN
WRITE ( NOUT, FMT = 99997 ) INFO
ELSE
IF ( N.GT.2 )
$ CALL DLASET( 'Lower', N-2, N-2, ZERO, ZERO, A(3,1),
$ LDA )
IF ( N.GT.1 )
$ CALL DLASET( 'Lower', N-1, N-1, ZERO, ZERO, QG(2,1),
$ LDQG )
WRITE ( NOUT, FMT = 99996 )
DO 10 I = 1, N
WRITE (NOUT, FMT = 99993)
$ ( U1(I,J), J = 1,N ), ( U2(I,J), J = 1,N )
10 CONTINUE
DO 20 I = 1, N
WRITE (NOUT, FMT = 99993)
$ ( -U2(I,J), J = 1,N ), ( U1(I,J), J = 1,N )
20 CONTINUE
WRITE ( NOUT, FMT = 99991 ) MA02JD( .FALSE., .FALSE., N,
$ U1, LDU1, U2, LDU2, RES, LDRES )
WRITE ( NOUT, FMT = 99995 )
DO 30 I = 1, N
WRITE (NOUT, FMT = 99993) ( A(I,J), J = 1,N )
30 CONTINUE
WRITE ( NOUT, FMT = 99994 )
DO 40 I = 1, N
WRITE (NOUT, FMT = 99993) ( QG(I,J), J = 1,N+1 )
40 CONTINUE
C
CALL DGEMM( 'No Transpose', 'No Transpose', N, N, N, ONE,
$ U1, LDU1, A, LDA, ZERO, RES, LDRES )
CALL DGEMM( 'No Transpose', 'Transpose', N, N, N, -ONE,
$ RES, LDRES, U1, LDU1, ONE, RES(1,N+1),
$ LDRES )
CALL DGEMM( 'No Transpose', 'Transpose', N, N, N, ONE,
$ U2, LDU2, A, LDA, ZERO, RES, LDRES )
CALL DGEMM( 'No Transpose', 'Transpose', N, N, N, ONE,
$ RES, LDRES, U2, LDU2, ONE, RES(1,N+1),
$ LDRES )
CALL DSYMM ( 'Right', 'Upper', N, N, ONE, QG(1,2), LDQG,
$ U1, LDU1, ZERO, RES, LDRES )
CALL DGEMM( 'No Transpose', 'Transpose', N, N, N, -ONE,
$ RES, LDRES, U2, LDU2, ONE, RES(1,N+1),
$ LDRES )
CALL DLACPY( 'All', N, N, U2, LDU2, RES, LDRES )
DO 50 I = 1, N
CALL DSCAL( N, QG(I,I), RES(1,I), 1 )
50 CONTINUE
CALL DGEMM( 'No Transpose', 'Transpose', N, N, N, -ONE,
$ RES, LDRES, U1, LDU1, ONE, RES(1,N+1),
$ LDRES )
CALL DGEMM( 'No Transpose', 'No Transpose', N, N, N, ONE,
$ U2, LDU2, A, LDA, ZERO, RES, LDRES )
CALL DSYR2K( 'Lower', 'No Transpose', N, N, ONE, RES,
$ LDRES, U1, LDU1, ONE, RES(1,2*N+1), LDRES )
CALL DSCAL( N, ONE/TWO, QG(1,2), LDQG+1 )
CALL DLACPY( 'Full', N, N, U2, LDU2, RES, LDRES )
CALL DTRMM( 'Right', 'Upper' , 'No Transpose',
$ 'Not unit', N, N, ONE, QG(1,2), LDQG,
$ RES, LDRES )
CALL DSYR2K( 'Lower', 'No Transpose', N, N, ONE, RES,
$ LDRES, U2, LDU2, ONE, RES(1,2*N+1), LDRES )
DO 60 I = 1, N
CALL DSYR( 'Lower', N, -QG(I,I), U1(1,I), 1,
$ RES(1,2*N+1), LDRES )
60 CONTINUE
CALL DGEMM( 'No Transpose', 'No Transpose', N, N, N, ONE,
$ U1, LDU1, A, LDA, ZERO, RES, LDRES )
CALL DSYR2K( 'Upper', 'No Transpose', N, N, ONE, RES,
$ LDRES, U2, LDU2, ONE, RES(1,2*N+2), LDRES )
CALL DLACPY( 'Full', N, N, U1, LDU1, RES, LDRES )
CALL DTRMM( 'Right', 'Upper' , 'No Transpose',
$ 'Not unit', N, N, ONE, QG(1,2), LDQG,
$ RES, LDRES )
CALL DSYR2K( 'Upper', 'No Transpose', N, N, -ONE, RES,
$ LDRES, U1, LDU1, ONE, RES(1,2*N+2), LDRES )
DO 70 I = 1, N
CALL DSYR( 'Upper', N, QG(I,I), U2(1,I), 1,
$ RES(1,2*N+2), LDRES )
70 CONTINUE
C
WRITE ( NOUT, FMT = 99990 ) MA02ID( 'Hamiltonian',
$ 'Frobenius', N, RES(1,N+1), LDRES, RES(1,2*N+1),
$ LDRES, DWORK )
END IF
END IF
END IF
*
99999 FORMAT (' TMB04PB EXAMPLE PROGRAM RESULTS',/1X)
99998 FORMAT (' INFO on exit from MB04PB = ',I2)
99997 FORMAT (' INFO on exit from MB04WP = ',I2)
99996 FORMAT (' The symplectic orthogonal factor U is ')
99995 FORMAT (/' The reduced matrix A is ')
99994 FORMAT (/' The reduced matrix QG is ')
99993 FORMAT (20(1X,F9.4))
99992 FORMAT (/' N is out of range.',/' N = ',I5)
99991 FORMAT (/' Orthogonality of U: || U''*U - I ||_F = ',G7.2)
99990 FORMAT (/' Residual: || H - U*R*U'' ||_F = ',G7.2)
END
</PRE>
<B>Program Data</B>
<PRE>
MB04PB EXAMPLE PROGRAM DATA
5
0.9501 0.7621 0.6154 0.4057 0.0579
0.2311 0.4565 0.7919 0.9355 0.3529
0.6068 0.0185 0.9218 0.9169 0.8132
0.4860 0.8214 0.7382 0.4103 0.0099
0.8913 0.4447 0.1763 0.8936 0.1389
0.3869 0.4055 0.2140 1.0224 1.1103 0.7016
1.3801 0.7567 1.4936 1.2913 0.9515 1.1755
0.7993 1.7598 1.6433 1.0503 0.8839 1.1010
1.2019 1.1956 0.9346 0.6824 0.7590 1.1364
0.8780 0.9029 1.6565 1.1022 0.7408 0.3793
</PRE>
<B>Program Results</B>
<PRE>
TMB04PB EXAMPLE PROGRAM RESULTS
The symplectic orthogonal factor U is
1.0000 0.0000 0.0000 0.0000 0.0000 0.0000 0.0000 0.0000 0.0000 0.0000
0.0000 -0.0927 0.2098 0.5594 -0.0226 0.0000 0.5538 0.3184 0.2519 -0.4031
0.0000 -0.2435 0.4745 -0.6362 -0.2542 0.0000 0.3207 -0.2455 0.0595 -0.2819
0.0000 -0.1950 -0.1770 -0.1519 -0.2857 0.0000 0.4823 0.4122 -0.2060 0.6173
0.0000 -0.3576 -0.0480 0.2302 0.4512 0.0000 0.3523 -0.6047 -0.3110 0.1635
0.0000 0.0000 0.0000 0.0000 0.0000 1.0000 0.0000 0.0000 0.0000 0.0000
0.0000 -0.5538 -0.3184 -0.2519 0.4031 0.0000 -0.0927 0.2098 0.5594 -0.0226
0.0000 -0.3207 0.2455 -0.0595 0.2819 0.0000 -0.2435 0.4745 -0.6362 -0.2542
0.0000 -0.4823 -0.4122 0.2060 -0.6173 0.0000 -0.1950 -0.1770 -0.1519 -0.2857
0.0000 -0.3523 0.6047 0.3110 -0.1635 0.0000 -0.3576 -0.0480 0.2302 0.4512
Orthogonality of U: || U'*U - I ||_F = .77E-15
The reduced matrix A is
0.9501 -1.5494 0.5268 0.3187 -0.6890
-2.4922 2.0907 -1.3598 0.5682 0.5618
0.0000 -1.7723 0.3960 -0.2624 -0.3709
0.0000 0.0000 -0.2648 0.2136 -0.3226
0.0000 0.0000 0.0000 -0.2308 0.2319
The reduced matrix QG is
0.3869 0.4055 0.0992 0.5237 -0.4110 -0.4861
0.0000 -3.7784 -4.1609 0.3614 0.3606 -0.0696
0.0000 0.0000 1.2192 -0.0848 0.2007 0.3735
0.0000 0.0000 0.0000 -0.8646 0.1538 -0.1970
0.0000 0.0000 0.0000 0.0000 -0.4527 0.0743
Residual: || H - U*R*U' ||_F = .33E-14
</PRE>
<HR>
<p>
<A HREF=..\libindex.html><B>Return to index</B></A></BODY>
</HTML>
|
