Вы находитесь на странице: 1из 9

ISSN 2304-1579.i . . -. . i .2013 ..18, . 4(20)..

2230

Mathematical Subject Classification: 15A23, 65F08


519.635.4
. . , . .
. .




i . ., Ii I. . i .
i i i. i .
ii .
i : , ,
i .
. ., . . . .

.
.
: , , .
Verbitsky V., Ivanisheva I. An incomplete generalized Cholesky factorization
preconditioner for a saddle-point matrix. The existence of an incomplete generalized
Cholesky factorization of a saddle matrix is proved. The factorization is applied to construct
saddle point problem preconditioner. The preconditioner allows to use the conjugate gradient method for solving the preconditioned system.
Key words: an incomplete generalized Cholesky factorization, preconditioner, saddle point
problems.

.
[
]

=
,
(1)
0
R , R
( ) ,
, , [1].
, MINRES
(the minimal residual method), GMRES (the generalized minimal residual method),

c . ., . ., 2013

23

CG (conjugate gradients) [2,3,4]. [4].


[3]. [5] ,
(1) .
(1)
.
.
1.
0
=

(2)

0 = 0 .
,
0 + (, 0 ),

(3)

(, 0 ) span{0 , 0 , . . . , 1 0 }

, 0 .
, ,
dim + (, 0 ) + ,
. .
1 (, 0 ) (, 0 ) = = + (, 0 ).
, (, 0 )
.

, .
= 0 + (, 0 ),

(4)

.
(, 0 )

= > 0 = (, 0 )

()

det =
0 = (, 0 ),

()

(3), (4), .

24

. ., . .

()
MINRES CG, () GMRES.

CG. , - ,
[4].
, (1) . , .
2.
, .
[4]. . (2)
1 = 1 ,

(5)

, :
1) = ;
2) 1 , , , .
. ,

= + ,
(6)
,
. = 0, = 0, = 0
= 0, = 0. (6),

= .
, (6)
. , ,
[5].

[
] [
] [
] [
]
1
0

0
1 1
1
=
,
(7)

0
0
(1

1 )
1 ( = 1 1 ),
1

1
1
= (1

1 ) (1 ) = 1 1 =

25

. =
, (7)
[

[
=

(1
1 )

] [

1
0

1
1

]
.

(8)

(8) [6].
,
. ,

11

= ...
1

1
.. ,
.

..
.

11

..

, . = diag().
= 2 2 . = 2 2 +
, (7)
[

[
=

(1
1 )

0
2

] [

1
1
2

1
0

[
+

0
0

]
.

(9)

(9) .
(2), . (2)
1 = 1 ,
(10)
= ,
[
=

(1
1 )

0
2

]
,

1
0

1
1
2

]
.

1. 1 (10)
, 1
1, 2 2
1

, ,
,

||||1 < ().


(11)

26

. ., . .

. ,
[
]

0
1


=
.

0 + 1
2 2
.

{, } 1
2 2 ,

1
2 2 = ,

= 2 2 ,
(2 2 )1 = ,

= 0,

=
2 = 0,
= (2 2 ) =

0.


||||1 ||||2
.
|| ||||2 = ||(2 2 )1 ||2 = ||1 ||2 6 ||||2 ||1 ||2 ||||2 6

()
,
|| 6

||||1
.

()

, (11) ,

1
1
2 2
1


.
, 1, (10) CG.
, , , [7]. ,
,
, , . , ,

.
. [6]
. ,
.
, ,
:

11
12
13
14

1,21
1,2

12
22
23
24

2,21
2,2

13

3,2

33
34
3,21
23

24
34
44

4,21
4,2
= 14
. (12)
..

..
..
..
..
..
..
.

.
.
.
.
.
.

1,21 2,21
3,21 4,21 21,21 21,2

1,2
02,2
3,2
4,2
21,2
2,2
.

27

1.
[
]

1,2 , 2,2
, 3,2
, 4,2
, , 21,2
,
2,2 , = 1, , .
. (12) , , -, (12) . , ,
,
2,2 , .
2. ,
, .
. (12) . .
. (0) = .

(0)
(0)
(0)
11
13

1,21

(0)
(0)
(0)
33

3,21
(13 )
(0)

=
..
..
..
..

.
.
.
(0)

(1,21 )

(0)

(3,21 )

(0)

21,21
(0)

. 11 =
11 11 .
[
]
(1)
(1)
(1)
(1)
(1)
11 , 12 , 13 , 14 , , 1,21 , 1,2 ,

(1)

(0)

1,2 = 1
11 1,2 ,

(1)

(0)

1,21 = 1
11 1,21 ,

= 2, ,


]
[
(1)
(1)
(1)
(1)
(1)
11 , 12 , 13 , 14 , , 1,2 , 1,21 .

:

(0)
(0)
(1)
33

3,21
(13 )

(0.5)
..
..
..
..

=
.
.
.
.


(1)
(0)
(0)

(1,21 )
(3,21 )
21,21

(1)

(13 )
..
.
(1)

(1,21 )

(13)

(1)
2,21

(0)
2,21

(1)
(1)
(1,2 ) 1,,21 ,

= 1, , = + 1, ;

28

. ., . .
(1)

(0)

(1)

(1)

21,2 = 21,2 (1,21 ) 1,,2 ,


(1)

(1)

(1)

2,2 = diag((1,2 ) 1,2 ),

= 2, , = , ;

(14)

= 1, ;

2,2 , = , . , (1)
2,21 ,
(0)

(0.5)

2,21 .

(0)
, . 1
(1)
2,2 ,
(1)

2,2 = 22 22 .
,
[
]
(1)
(1)
22 , 23 , 24 , , 2,21 , 2,2 ,

[
(1)

(1)

(1)

22 , 23 , 24 , , 2,21 , 2,2

(1)

2,21 = 1
22 2,21 , = 2, .
:

(1)
(23 )

(1)
(0.5)
..
= +
.

(1)
(2,21 )

(1)

(23 )
..
.
(1)
(2,21 )

(0.5)

(15)

(1)


,
.
, . , ( + 1)- .

()
()
2+1,2+1
2+1,21

()
..
..
..

=
.
.
.

()
()

(21,2+1 )
21,21
,
, (13) (15), . ,
( + 1)- . ( + 1)- ,

(
(+1)
(1)
(1)
(2)
(2)
2+2,2+2 = diag (1,2+2 ) 1,2+2 + (3,2+2 ) 2,2+2 + +
(+1)

(+1)

+(2+1,2+2 ) 2+1,2+2

29

. - ,
-
]
[
(1)
(2)
(+1)

(1,2+2 ) , 2,2+2
, (3,2+2 ) , , (2+1,2+2 )
. 1
[
]

1,2+2 , 2,2+2
, 3,2+2
, , 2+1,2+2
, , 1 + 1,
(21,2+2 )*, = 0,

1
< ,

(21,2+2 )*, = 0.

, (14), ,
()

(21,2+2 )*, = 0.
3.

2 =

= = 0 ,
= {(, ) : 0 < < 1, 0 < < 1 },
, (2)
(1), .
,
(, ) = 2 2 (1 )2 . , .
. (1) , , . (12) ,
, .

1.

. .
10 10 198
66
26
16
20 20 693
231
123
59
30 30 1488 496
*
140
40 40 2583 861
*
259
50 50 3978 1346
*
393
CG,
|| ||2 /||||2 106
, [6], , .

30

. ., . .

. .
. . ()
MATLAB. 1 .
, (
[6]).
, , , .
. . , ,
, . ,
,

. ,

, . ,

.

1.

Brezzi F. Mixed and Hybrid Finite Element Methods / F. Brezzi, M. Fortin. Springer
Verlag, 1991. 350 p.

2.

Benzi M. Numerical solution of saddle point problems / M. Benzi, G. H. Golub, J.


Liesen // Acta Numerica. 2005. P. 1137.

3.

Benzi M. Some Preconditioning Techniques for Saddle Point Problems / M. Benzi, A.


J. Wathen // Mathematics in Industry. 2008. V. 13. P. 195211.

4.

Saad Y. Iterative Methods for Sparse Linear Systems, second edn / Y. Saad. SIAM,
Philadelphia, PA, 2003. 547 p.

5.

Manteuffel T. A. An incomplete factorization technique for positive difinite linear


systems / T. A. Manteuffel // Mathematics of computation. 1980. V. 34(150). P.
473497.

6.

..
/ . . // . 1989.
. 29, 1. . 6774.

7.

Botta E. F. F. Matrix renumbering ILU: An effective algebraic multilevel ILU


preconditioner for sparce matrices / E. F. F. Botta , F. W. Wubs // SIAM J. Matrix
Anal. Appl. 1999. V. 20, Is. 4. P. 10071026.

Оценить