• Aucun résultat trouvé

Overlapping Schwarz Waveform Relaxation for the Heat Equation in n-Dimensions

N/A
N/A
Protected

Academic year: 2022

Partager "Overlapping Schwarz Waveform Relaxation for the Heat Equation in n-Dimensions"

Copied!
18
0
0

Texte intégral

(1)

Article

Reference

Overlapping Schwarz Waveform Relaxation for the Heat Equation in n-Dimensions

GANDER, Martin Jakob, ZHAO, H.

GANDER, Martin Jakob, ZHAO, H. Overlapping Schwarz Waveform Relaxation for the Heat Equation in n-Dimensions. BIT Numerical Mathematics , 2002, vol. 42, no. 4, p. 779-795

Available at:

http://archive-ouverte.unige.ch/unige:6289

Disclaimer: layout of this document may differ from the published version.

1 / 1

(2)

!"

!" #$!#!

%

&&'()'*!" + $!,!,

! !"# $% " % % &'%

(%#(( "% % # ')'%#

%( (' )' %# %( )% (( %

%( ( ! % % ( " % *

('% ( ( % '#)" (')#( )'( % ( #% )

% (( (% (%% ('% " ! " % )#( !*

% (' ('%(%" % '#)"(')#(

'( ! !"#$% ( % ( ( " *

)'(%%"% '#)"(')#(+"% % #(%

('# '#$#%(,#'(( (

)-() !'('%()$%%#))#(

./ 01+./ //

# #(%+ "# $%+ ! % "

2)2)#(+'3

!"#

!!!$# !%# !&# !'#

(3)

( ) * +,-. *-

+,

* $%$"# $&# %#

!/#-0*- +,* -

1

* $%$"# $# &#- 2

( ) *3+,-

*- *-

4

1

$$# 5 $'#

6 5 !7!8# 2 ( 6 3

!/# *-

*-

)

9 "#

: !%#

)

* -

) (

*-

2 : ; !'#

*-

( ) 2

( (

< (

2 = '7#

2

) >

*- *-

*-

!'# 2

'7# 2

4

(4)

?

?

?

?

@

@

Æ Æ

<$!. +

Æ

> ( )

> )

? Ê

?

A BC*Ü- Ü? D

*Ü- A *Ü- Ü? D

*ÜD- A

*Ü- Ü?

*$!-

>

*Ü- *Ü-

*Ü-

) !D%D# =

4 *Ü- *$!-.

*Ü-D

*Ü- ?

D # *Ü-

1

$D!$8#

? 4

? ? ? A

!$ < $! >

? ? @

? ?

? @ ?

Æ @ < $!

*$!-

3

?

(5)

?

?

?

?

?

@

Æ Æ

<$$. 2 ? ?

?

? E @

A! ,

Ü@

?

Æ

.

?

?

F

<$$ +

4 ?

C! >

@

? ?

>4 A ?

C!

3

A B 3

C*Ü- Ü? D

3

*Ü- A

*Ü- Ü@

D

3

*Ü- A *Ü- Ü@

D

3

*ÜD- A

*Ü- Ü?

*&!-

A!$

6

:

2

( >

(6)

3

*Ü- ( *Ü-

*$!- 3

*Ü-*&!-4 )

3

A B 3

Ü? D

3

*Ü- A

*Ü- Ü@

D

3

*Ü- A D Ü@

D

3

*ÜD- A D Ü?

*&$-

>4 ) ?

?

.A <4

.A. A

+4 *Ü-.? D-

Ê

*-

.A

Ü4

*Ü-

33

*-

*Æ-

*-

*&&-

*Æ-

)

1 $!# > 4

4

.A

*-

G

3

D A BG 3

Ü?

G

3

*Ü- A

Ü@

G

3

*Ü- A D Ü@

*&%-

3

H G 3

4

G

3

? G 3

4

? H

G 3

G

3

G

3

>4

3

.A

Ü4 G

3

*Æ-

*Æ-! H

G 3

(7)

?

G

3

D A BG 3

Ü?

G

3

*Ü- A

Ü@

G

3

*Ü- A 3

Ü@

*&"-

3

3

*Æ-

G 3

? 4 3

3

4 3

*Æ-

*Æ-

*Æ- !

9 4 CC!

?

D A BG 33

Ü?

G

33

*Ü- A

Ü@

G

33

*Ü- A 3

Ü@

*&'-

G

33

? +4 33

4 33

3

*Æ-

*Æ-

*Æ-

3

*Æ- !

H CC$ 33

) 33

3

*Æ-

CC! +4

*Æ-.A

3

*Æ-

H *Æ-) 2

*Æ- *Æ-

536

*-

**Æ--

*-

*&7-

*Æ-!! "

1&!

H

?

(

!

> 1

(8)

*&!-

>

3

*&$- > 4 4

*Ü- @ ?

*-

7

.A

Ü7

*Ü-

>4

#

3

?

3

7

$ * Æ

$

-

7

1

*-.A *

*-

7

- 9

*Ü- Æ

@ F

4

A B Ü*Æ- D

*Ü- A

*- Ü*Æ- D

*ÜD- A D Ü*Æ-

*%!-

*Æ- Æ Ê

*Æ- Ê

Æ

*Æ-

G *Æ-

G

A BG Ü*Æ- D

G

*Ü- A

*- Ü*Æ- D

G

*ÜD- A D Ü*Æ-

*%$-

, G *Æ-

Æ @ F G

*Æ-

) $

A B ÜÊ

D D

*Ü- A

*- ÜÊ

AD D

*ÜD- A D ÜÊ

D

*%&-

4 ÜÊ

*%&- 8#

*Ü-A

*

-

*-

*%%-

(9)

*-A

$

*%"-

Æ @ $

*%%-

3

3

*-.A* 3

*-

7

-$

* Æ

-

*-

*%'-

H

*- D

.A Æ

$

*-

3

*-$ * Æ

$

-

7

*Æ*$

--

H

H 1 %!

D -

*-AD >

#

?

7 *$ -

* Æ

$

-

7

9 )*%'- 1%!

*- *$ -

* Æ

-

* Æ

-

7

1

1 2 1

!#

* Æ

-A

Æ

(10)

1

Æ

9 4

*-*$ -

* Æ

-

7

I 1%!

+4 *Ü-.? D -Ê

*-

.A

Ü4

*Ü-

*-

*$ -

* Æ

$

-

*-

*%7-

1%$

H

$&#

* -

J A

!

$ C!*

-

353 566

*%7-

*

-A

C!*

-

3 5 6

" # $%

> <

)$$

<

4

>

)

(11)

$ % & '

> )

A

C"

56

5

6

D! D

*D- A D D

*!- A

D

*D- A

D!

*"!-

?A D!# D # ?

A D"# D #

?

A #!# D # K

*#"-*D%D'-*D%"D""-*D%8D"$-

> &$

% ()

* ?

A D"# D #

?

A #!# D #

8

*-

#*!"-

"*!#-

8

*-

&$

F !%#

F 4 (

B ADD! , BADD! <

"!

A&

="!

(

=

DDD"#

>

%&

()

* ?

A D"# D #

?

A #!# D # D #

8

*-

*

*"#-

$

-

8

*-

%&

H$

(12)

0 2 4 6 8 10 12 14 16 18 20 10 −8

10 −7 10 −6 10 −5 10 −4 10 −3 10 −2 10 −1 10 0 10 1

iteration k

maximum error

<"!.

(

> 4

*#"- *D%D'-*D%"D""- <

"$

=

F

DD"#

> *#"-A *D%D'-

< "&

4

4

$ & '

>)

AB *

- D!# D!# D #

*"$-

)

> ) )

(13)

0 1 2 3 4 5 6 10 −4

10 −3 10 −2 10 −1 10 0

iteration k

maximum error

< "$.

(

0 5 10 15

10 −9 10 −8 10 −7 10 −6 10 −5 10 −4 10 −3 10 −2 10 −1 10 0

iteration k

maximum error

<"&.

4

(14)

0 2 4 6 8 10 12 14 16 18 20 10 −3

10 −2 10 −1 10 0

Iteration

maximum error

0.1 overlap 0.06 overlap

< "%.

A&

"!. H (

( >

$$ && %% ''

A" !$ !" !/ $8

ADD! " " " '

) >

<

Æ ÆD!DD' >

4 ( BADD$

, <"%

A&

< ""

ADD" H (

$" ( + (

>4

> ) *"$-

>

B A!&D "!

(

(15)

0 2 4 6 8 10 12 14 16 18 20 10 −10

10 −9 10 −8 10 −7 10 −6 10 −5 10 −4 10 −3 10 −2 10 −1 10 0

iteration k

maximum error

0.1 overlap 0.06 overlap

< "".

ADD"

&$ <

/#

%& "!

< "'

%%

A"

ADD!

&

F )

)

(16)

0 0.2

0.4 0.6

0.8 1

0 0.2 0.4 0.6 0.8 1 0 0.2 0.4 0.6 0.8 1

x2 x1

u1

0 0.2

0.4 0.6

0.8 1

0 0.2 0.4 0.6 0.8 1 0 0.2 0.4 0.6 0.8 1

x2 x1

u1

0 0.2

0.4 0.6

0.8 1

0 0.2 0.4 0.6 0.8 1 0 0.2 0.4 0.6 0.8 1

x2 x1

u2

0 0.2

0.4 0.6

0.8 1

0 0.2 0.4 0.6 0.8 1 0 0.2 0.4 0.6 0.8 1

x2 x1

u2

0 0.2

0.4 0.6

0.8 1

0 0.2 0.4 0.6 0.8 1 0 0.2 0.4 0.6 0.8 1

x2 x1

u3

0 0.2

0.4 0.6

0.8 1

0 0.2 0.4 0.6 0.8 1 0 0.2 0.4 0.6 0.8 1

x2 x1

u3

<"'.

A "

ADD!

H

(17)

Æ

2

!&# ( )

)

( ) !'#

'() > 6 ;

F

K,<,K,H=,

! !"#$% & #,

> I :

3 (!/'%

$ '% ( - ./ .

FHL !!*!//&-/"M!!%

& ')*+ 0 )

92&"*!//"-$/!M$/'

% , ' 1 ( * 2 % &3

'* I 32 HN!//"

" , '-( ./"#$- 0*+##- %1 1# 1 .

3 92&'

*!//'-"%M7'

' 23 0 () .3

*HL'D*!//!-%!M'!

7 2 () 2FL 6

=!"*!//%-"87M'D&

8 .1 % .& , '*, L

F F >3 = !/8%

/ 4 % #" &

FH!//%=I 3!//%'!M!%&

!D 4% 1 &3 '* 1 3

2 , =(!/'%

!! . 5% % () H 2

= + + L 3,9EO

L, + ;!//7 /7M!D%

!$ % ()

2 = + + L6L

= < P==FL = L $!8

!//8

(18)

!& 0

3 * H1 F F ' *!//8-

!$"M!%"

!% . 5% % ## ( .

*2FL6 4=

!/*!//8-$D!%M$D&!

!" . 5% % 6 ( % ()

3 F!%

F QR L ; RS ; L

I !//7%$M"!

!' 75%%6,(

K /7% =

=K3==!//7

!7 ..++%8%"2

4 . 2FL6 H

F &&*!//'-%"'M%7%

!8 2 4

. 2FL6= !7*!//'-!&&M!""

!/ 1.#+%'5

2FL6= !'*!//"-%DM%/

$D 5 ! ( % 1 &3 '* >

43 =!//'

$! 3 + % () # < 2

+ + L3+(

,) K: :::FL 63R

33F!/882FL!M%$

$$ !% 9+#2.

*92$7*!/87-$!'M$&%

$& : // % 9 0 6

2FL 6 = 8 *!/87-

%"/M%8$

$% 9 0 - 1.! 7 92 $/

*!/8/-&$8M&%'

$" - 1.! 7 92$/*!/8/-"&"M

"'$

$' 8%" % 5 6# .

. =

"%*!//"-&!7M&&D

Références

Documents relatifs

When using overlapping subdomains and “classical” Schwarz waveform relaxation -by a Dirichlet exchange of informations on the boundary- the so defined algorithm converges in a

This algorithm is like a classical overlapping additive Schwarz, but on subdomains, a time dependent problem is solved, like in waveform relaxation.. Thus the name overlapping

There are two classical convergence results for waveform relaxation algorithms for ODEs: (i) for linear systems of ODEs on unbounded time intervals one can show linear convergence

Note how the zeroth order approximation, which leads to a Robin condition instead of a Dirich- let one in the Schwarz algorithm, reduces the convergence rate already from 0.. 05 and

It is shown for a scalar reaction diffusion equation with variable diffusion coefficient that the new waveform relaxation algorithm converges superlinearly for bounded time

Overlapping Schwarz Waveform Relaxation for Convection Reaction Diffusion Problems.. DAOUD, S., GANDER,

Using Theorem 2.1 the non overlapping Schwarz waveform relaxation algorithm for the wave equation with a discontinuity in the wave speed converges in I steps where I denotes the

There are five main techniques to prove the convergence of Domain Decomposition Algorithms and they are: Orthorgonal Projection used for a linear Laplace equation (see [16]), Fourier