• Aucun résultat trouvé

A Frequency Decomposition Waveform Relaxation Algorithm for Nonlinear Evolution Equations

N/A
N/A
Protected

Academic year: 2022

Partager "A Frequency Decomposition Waveform Relaxation Algorithm for Nonlinear Evolution Equations"

Copied!
16
0
0

Texte intégral

(1)

Article

Reference

A Frequency Decomposition Waveform Relaxation Algorithm for Nonlinear Evolution Equations

GANDER, Martin Jakob

Abstract

Semilinear evolution equations arise in many applications ranging from mathematical biology to chemical reactions (e.g., combustion). The significant difficulty in these equations is the nonlinearity, which combined with the discretized diffusion operator leads to large systems of nonlinear equations. To solve these equations, Newton's method or a variant thereof is often used, and to achieve convergence can require individual fine tuning for each case. This can be especially difficult if nothing is known about the solution behavior. In addition, one observes in many cases that not all frequency components are equally important for the solution; the frequency interaction is determined by the nonlinearity. It is therefore of interest to work in frequency space when analyzing the unknown behavior of such problems numerically. We propose in this paper an algorithm which reduces the dimensionality of the nonlinear problems to be solved to a size chosen by the user. The algorithm performs a decomposition in frequency space into subspaces, and an iteration is used to obtain the solution of the original problem from the [...]

GANDER, Martin Jakob. A Frequency Decomposition Waveform Relaxation Algorithm for Nonlinear Evolution Equations. Electronic Transactions on Numerical Analysis , 2004, vol.

17, p. 181-194

Available at:

http://archive-ouverte.unige.ch/unige:6280

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

1 / 1

(2)

! " # $Æ

%& $& $'$$($

%&)$$

$$$#

$Æ*& $$

+

$$ &*&'

*&

,&$$

$ ' $

$$

, $$

%&$ - $$ %&

& &% ,

&. /

&$&%$$ $$$

, .% 0 % $%

.1

23 45%23 33%23675

"#

"

$

%

$&

Æ

' #

" ())*"

$% 8% %$9

(3)

(+*

"

#

(+)*

", %

,

#

,(+-*

.

# /()0*/1()2*

3

&

,

, 4

(+2 5*

())5* % 6

"

# "

78(-* 4

, , (++

+9* (: 2*

, #

(;* %

#

#

()-* ()+* (+5* (+0* %

#

,

#

)

# #

-

"

, 5

, #

<

(4)

<

6

=,

0

#

! "#$%%#& '

#

, '

(+:*'

!

> ! ? !!

! ? )+!

7

9 +

! !

6?> 7

6?

! 7

#

6

!

99Æ+

!

!

Æ

>

! ))!

!

! ! 9 !

!

! (

(6 !

*

(6 !

"

?

!

# (+: +@2* A =,

#

(+:*

=BC

B D

!6?

!

#

(5)

! )

? !

B!

' "()&$ $%)(

)+!# ? !

# (9* '

7

E !

6?

!

E

?

4

,

?

4#

6?

½

6?

¾

¾

>+ $

%? '

'

F

6

F

?F

4

F

?

½

F

?

¾

?

!

<

F

%F

)+!

>

? F

! 9

9! ? F

-+!

?+)

6?F

'C

6

?

C

-)!

C

?C

½

?

¾

>

½

¾

C

?C

?

½

>

(6)

>

?+!?

AC

G

!6?

C

!

>

? F

G

! 9

9! ? F

--!

$

" "

+) #

! # " # G

.

#(+;* 4#F

#

>

? F

G

! 9

9! ? F

-5!

% H#

>

? F

G

! 9

9! ? F

-0!

())*,

% C

-0!

> ? I

! 9

9! ? -2!

?

C

F

-)!

I

!6?

C

F

G

F

F

F

F

F

!

-2! #

$ #

-2! Æ

-2! #

!

$%

! I

! I

!? !

(7)

* +$,$$ &#-( ( '

)+! '

!Æ9 9

' $?$ !

% $

% Æ

9

# (+@*

* .&## -)$ ( %&$( '

" #

. J ! K

J !?

&

%

&

!

!

6? !

6

( 9 + " !

'

6(9* "'

9!?9 (?9+ !

' !

+

)!

' )!>

'

)!! )

9

9 "

'

!

J + !!

J ( + !>+!

' !

" ?

!

?

%

´

½  ´½ µµ

·½

 ´´·½µ´½ µµ

´·½µ´½ µ

)

J + !!

>+

5+!

?

# (?9

(

' !

+

)!

' )! )>

J + !!

J ( + !>+!

)

)!

)' !

'

(0* '

*!

* )!

*? )!

+ ! 9

(8)

+ !7L&"(@*

+ !?

+ *!

*

*

' "

' !,

+

*!

' *! *> (!

5)!

, (!

,?

+ ! + !+ !

(!?

' !

J + !!

J ( + !>+!

+ + ( + !>+!

+ ! + ! (>+! + !>+!

+ - + !+ !

>+

$ 7 K

(@*

+ !?

J !J !

J >!

#, (!

'

,?

J + !!

J >+! + !!

(!?

' !

J + !!

J (>+! + !>+!

J (>+! + !>+!

J + !!

J (>!>+! + !>+!

>+

$ (!

K = ( '

(!

J (>+! + !>+!

J (>!>+! + !>+!

)

" 5)!

+

*!

#

' !,

' *! *>

I

(!

(9)

I

(!?

' !

J + !!

J (>+! + !>+!

)

J + !!

>+

$K=

' ! I

(!%

´·½µ´½ µ

A

( " 9+

' !

% Æ

)!

' )!>

'

)!! )

9

9Æ9

Æ

J + !

'

!

J + !

Æ

J + !

' !

# '

' !

% Æ

)! )

' !

>

'

!

!

% &?Æ + )#!

' ! +

Æ

Æ

%

&

&

' !

>

'

!

!

? +

Æ

J

Æ

+ !

' !

>

'

!

!

J

!K

J

!?

&

%

&

"

' !

J + !

Æ

' !

>

'

!

!

J + !

(10)

*! +$,$$$(#( '

-2! )+! %

%

>% ? I

! I

! 9

% 9! ? 9 5-!

=D I

) 9+!

I

!

I

!

I

! I

!

55!

)

Æ

$J + !

Æ

$J + !

"$?$ !Æ *$%

%

! .

% !

"

.?

$J + !

Æ

$J + ! +

) ** )

!! **"+

) )

"*%

#

% !?

%

I

)! )!!

I

)!

)!!! )

% "

% !

%

I

)! )!!

I

)!

)!! )

A=, I

=5+

% ! $

% Æ

)!

% )! >

%

)! ! )

$' !6?%

! =5-

D I

) 9

+! ** " )

$% 9 "

%

!

J + !!

J ( + !>+!

% !

(11)

"

?$

%

´

½  ´½ µµ

·½

 ´´·½µ´½ µµ

´·½µ´½ µ

)

$

J + !!

>+

?

$?$ ! *

# F 55

% ! $

% Æ

)!

% )! >

%

)! ! )

$' !6?% Æ

%

! =5)

$ K

(

Æ

/ %$ &# 0&%)#$( ' #

# (+*

/ "#$ $## $ &# $# "#$%

?

! 9+ 9

4

F

#

?

!

! 9+ 9

' ,

+99 " L -99

' ? + 9! ? + !

# ?- #

'

?9 40+

!

7

(12)

0 1 2 3 4 5 6 7 8 10 −9

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

Iteration k

Error, L2 in space, Linf in time

numerical experiment analytical bound

0 1 2 3 4 5 6 7 8

10 −14 10 −12 10 −10 10 −8 10 −6 10 −4 10 −2 10 0 10 2

Iteration k

Error, L2 in space, Linf in time

numerical experiment analytical bound

L

(

?+

?)

?-

?0

9 +;;)29) +;;)29) +;;)29) +;;)29)

+ +;5:092 +;5:092 500-;9: -:2959@

) +;5;+9@ +;5;+9@ +)0+2++ 55);9+-

- 59-92+- 59-92+- +;-50+0 590)-+;

½

55 $ ? + Æ ? /

$

?+#+9

? 9 4 0+

%

52

$

$# ,

'

$ 4

, '

# ? + 0+

4

6

(13)

"

4

#

?+

@;M '

?-@@0M

@@@M

())*(+*

/! %"( $# "#$% D

6 # ()9* Æ

, ())* ,

4

%

',

N?%

·

9+ 9

9!

" % O

4

" %

% #" ,

?) /! !/! !?+)

(+* %

E",

' 9!?+99 + ! + !

?+ ', ++++

7"L )9 '

0)

,?-5

$

7#

?+ ++

?0+-

?2-+

?++

(14)

L

(

?+

?0

?2

?++

9 +;);@9+ +;);@9+ +;);@9+ +;);@9+

+ 09;-+9- +9:059- +9:059- 0-95:95

) @):0295 ;59:;90 ;59:)90 +050)90

- :;;5090 -;5;)92 -;5;+92 :2)5-9;

5 +2;;:90 )5@2)9; )5@2+9; ):5009:

0 +0@;@92 +-+229: +-+209: +0@9)9@

½

L

(

?+

?2

?++

?+

?2

?++

9 -)0+09+ -)0+09+ -)0+09+ +0-@09+ +0-@09+ +0-@09+

+ -)9@09) +9;;;9) 222:99- +50+)9) 5;95)9- ):++;9-

) @@25@9- -;;059- )9:959- 5--:59- +0-2)9- ;@:)@95

- )--0-9- 2@@--95 --50095 @)99:95 )2+-995 ++@9:95

5 5292@95 +5);095 2+@-290 +255095 5;+9290 +@):;90

0 ;+09590 +@92)90 ;5+5+92 ))+-:90 000:092 )92@292

½

"

% %

#

@:M"

@@)M $

40-

?-0 E ?99+

52 $

6

?+ @)M

@:M 0-

#

P

1 #( ' ,,

# ())*#

(15)

#

%

',

=,

<

=, E

#

6 (9*

3

=, 7

# "

2.#$,%$(3 " "%

Q Q 3 %C

0;

@7A B !1%

8% %C555

@CA / % &*%$86$$$

&&$Æ$$"#$%&

$%DE!C"F25DG2DH%7HH4

@DA /IJ'"#( ) 1

%8$%&%7HH3

@EA ?B;; &'$ "#$%&"*%

73!D"F3K4G25D%7HHE

@3A ; $>&$$#B

$;B6$%*%3K!7HK"F25DG

2D5%7HH7

@2A $-&'&. $%

;0%$?B;;%$%+# *

% %; C7K%7HHK

@4A $% > 6% $ 0 - $ B

&'&.;B6>%1/IJ$% ;%$-,$B

$% $% * %

$$%7HHH

@KA $$ 6L$$

H4BE%;;1;%;%7HH4

@HA $ $ '* + " ! # = 6

%%7HK7

@75A $$ B&.

"#$%&",*%7H!2"FC57EGC5D7%7HHK

(16)

$6$NO% $LOP*%$L *%$%

!$-.%ECG37*8% 7HH4

@7CA $%1$1(B

1%$8%7HH4

@7DA ,6* $B$$% F;

$ $$ &' $ % ;B B

8Q%L% %7HH7

@7EA $=$&% ,/

=DD%C%E32BE4E%

7HH2

@73A $=$&% ,/

;%=74%7%7DDB733%7HH2

@72A O$ .&$$$

$ &*%<L'% O$ B

%1O.%$-,$$%$%0#"

%! %1$%1%7HH7

@74A -*1$B>$R1#+%CHFDCKGDE2%7HKH

@7KA 1'" ' $!

B=%&<*%7HKD

@7HA 1 #$$+

'!$ $

#2-.$1 %7HH3

@C5A LL*

%

/!$ "%D7%7%42BK5%7HKH

@C7A )$ 6 M'

#;%=72%3%75HCB775E%7HH3

@CCA LL ;$

$ :%=E%D%7HH2

@CDA =$&$ 60

;%=3E%E%D74BDD5%7HH3

@CEA <,%?B;;%$ L$$&'$

$%;0%$?B;;%$%+#*

%%37DG3C7 %; C7K%7HHK

@C3A ?$ $$ "#$%) %

DE!E"F3K7G27D% 7HHC

@C2A ?$M%$$$"#$%) %E5FK34G

H7E%7HHK

Références

Documents relatifs

Keywords: Fourier transform, Heisenberg group, frequency space, Hermite functions.. 2010 Mathematics Subject Classification:

This paper compares two extensions of the Matching Pursuit algorithm which introduce an explicit modeling of harmonic struc- tures for the analysis of audio signals.. A

method to solve the graph oloring problem by an exat way. This

It was shown that the state estimation problem can be solved using a convex relaxation based on semi-definite programming.. In the present paper, we propose a fast algorithm for

In this section, we briefly summarize the results of Ariki which give an interpretation of the decomposition matrix in terms of the canonical basis of a certain U q (c sl e

Hence the idea: If we saw servers or nodes of the graph as a node in an electrical circuit, and the &#34;distances&#34; between vertices as a resistor values in a circuit, then more

In particular, we want the Fourier transform to be a complex valued function defined on some explicit ‘frequency space’ that may be endowed with a structure of a locally compact

Structure of this paper. In Section 1, we recall some results about the Jacobian derivative and Darboux polynomials. In Section 2, we describe a reduction step which eases