• Aucun résultat trouvé

An estimate on convergence for an homogeneisation problem for variational inequalities

N/A
N/A
Protected

Academic year: 2022

Partager "An estimate on convergence for an homogeneisation problem for variational inequalities"

Copied!
17
0
0

Texte intégral

(1)

R ENDICONTI

del

S EMINARIO M ATEMATICO

della

U NIVERSITÀ DI P ADOVA

M ARCO B IROLI

An estimate on convergence for an homogeneisation problem for variational inequalities

Rendiconti del Seminario Matematico della Università di Padova, tome 58 (1977), p. 9-24

<http://www.numdam.org/item?id=RSMUP_1977__58__9_0>

© Rendiconti del Seminario Matematico della Università di Padova, 1977, tous droits réservés.

L’accès aux archives de la revue « Rendiconti del Seminario Matematico della Università di Padova » (http://rendiconti.math.unipd.it/) implique l’accord avec les conditions générales d’utilisation (http://www.numdam.org/conditions).

Toute utilisation commerciale ou impression systématique est constitutive d’une infraction pénale. Toute copie ou impression de ce fichier doit conte- nir la présente mention de copyright.

Article numérisé dans le cadre du programme Numérisation de documents anciens mathématiques

http://www.numdam.org/

(2)

convergence for homogeneisation problem for variational inequalities.

MARCO BIROLI

(*)

1. - Intro duction

an d results.

Let be Q c I~n an open bounded set with smooth

boundary 1’,

Y =

~n [o, ail (y)

E Loo

( Y) i, j

-

1, ... ,

~, such that

i=l

(where

, &#x3E; is the

duality

between

Hà (Q)

and H-1

(Q)

and

aij (Y)

are

prolonged

to

by periodicity)

and

where qil

are constants defined as in

[3], [9], [10].

(*)

Indirizzo dell’A : Istituto di Matematica del Politecnico di Mi- lano - Via Bonardi, 9 Milano.

Lavoro

eseguito

nell’ambito del gruppo G.N.A.F.A. del CNR.

(3)

Let be

f

E

(Q),

we have :

Th. 1 - We consider the

problems

let be Us and u the solution

of (1, Ie)

and

(1, 1),

we have

Moreover it

we have

The first

part

of Th. 1 is shown in

[7]

for the

symmetric

case

and in

[8], [12]

in the

general

case

by

«local energy &#x3E;&#x3E;

method,

the

second

part

of Th. 1 is shown

by « multiple

scales » method in

[8]

[9] (*).

We consider now the

problems

where 1p

is a measurable

function,

such

that I v

E

HÓ (S~), v y~ ~ ~ ~.

Th. 2 Let be Us and u solution

of (1,3~)

and

(1,3)

we have

(*) Cf. also [1], [2].

(4)

The first

part

of Th. 2 is shown in

[11] by

the method of « local energy »

(cfr.

also

[5]

and

[6]),

the second

part

of Th. 2 is shown in

[4] by penalisation

method and H61der-estimates. The aim of this work is show an estimate

analogous

to

(1, 2)

in the case of varia-

tional

inequalities.

Th. 3.

Let be aij (y) 8

C’-

(.Rn) , f e

Lx

(Q) , Us (u)

the solution

The Th. 2 is shown

by penalisation

and

«multiple

scales » method.

In §

2 we show some

preliminary results, in §

3 we

study

the conver-

gence for the

penalised equations

and

in §

4 we

give

the

proof

of

Th. 3.

Rem. 1

(a)

The

hypothesis

of the

part (I)

are verified

if y

= 0 :

/

x

(b)

The result of Th. 3 can be

generalized

to the case

t E

8

(c)

We don’t know if the estimates in Th. 3 are

optimal ;

2. -

Some preliminary

results.

We consider the

penalised problem

(5)

LEMMA 1. The

problem (2.1)

has a

unique

solution

u-1;

E

Loo

(Q)

we have

From

(2,1)

we have

then

From

(2,5)

we have uA e C

(Q).

We show now

(2,2)

ab absurdum)).

We write

Let be

Q’ and

S~~

are open set and

,~~

c

,~~

c S2’.

Let be an open set with smooth

boundary

such that

,~;~

c

c c

Q"

c ’0’.

(6)

We have

then from the maximum

principle, [13],

From

(2,8)

we have

then from

(2,7) (2,9)

we have a contradiction.

From

(2,1) (2,2)

we have also

(2,3).

LEMMA 2.

If .

we have

Moreover

if A

1p E Loo

(0)

we have

We show at first

(2.11).

We observe that if

A y

E

(S~)

from the lemma 1 we have

IIA

I and

repeating

the calculation of

[6]

pg. 76-88 or

[13]

177-197 we have

where C

depend only

on

C.

From

(1,66) (2,1)

we have

(7)

then

From

(2,12)

we have

We

prolongate ul

and u

by

0 and we denote

again

and

u these new functions.

From

(2,13)

we have

then

then there

is x

E B

(x, e)

such that

From

(2,14)

we have

We choose then

We show now

(2,10).

(8)

From the lemma 2.1 of

[7] (the hypothesis

of

symmetry

is unne-

cessary for this

lemma) if y

E

(,S2), n

p _ + oo , y there is

a sequence such that

where

C3

is a constant

dependent only

on

I/1J’/lwl,P.

Let

(ur¡)

the solution of the

problem

From

[12]

pg.

22, part III,

we frave

From

(2,16) (2,21) (2,22)

we have

We choose

Rem. 1 - We observe that the constants in the lemma 2

depend

only

on norm of the coefficients of A.

(9)

Lemma 3. We consider the

problems

we have

We consider 0

== 2013Zt-i/ ,

0 E W2,p

(Q). Vp,

for the lemma 2.1 of

[7]

there is a sequence

0~

such that

Let

be us (un)

the solution of the

problem

We have

We

choose q

- 8 and we have the result.

3. - The

penalised equation.

We suppose for the moment

f

= 0.

We consider the

penalised problems

(10)

Lemma 4. We have

We use for the

proof

the

« multiple

scales » method.

x

Let

be y -

we

have, [8] [9],

E

where

Let be

we have

(11)

Let be

where X2 (x ; y)

is not

given

for the moment.

We consider w

= ~c8 - v~ ,

We

choose x2 (x ; y)

such that

X2

(0153 ; y) Y-periodic

in y.

We observe that we have

then the

problem (3,4)

has a solution.

We observe that

(12)

From

(3,5)

we have

then

From

(3,4) (3,5) (3,7)

we have

and, by multiplication

for we can

easily

show

We observe now that

From

(3,6) (3,7) (3,8) (3,9) (3,10)

we have

where

(13)

By

the same method used in

[10], [13]

for the

proof

of

- estimate in

elliptic problems

we have

then from

(3,12)

We consider now the case Let be

we have

We observe that the

problems

are

equivalent

to the

problems

where

(14)

We consider now the

problem

From

[12]

pg 22.

part

III we have

and from

(3,12)

we have

then

We consider the

problem

and the

corresponding penalised problems

(15)

We suppose for the moment

We have

then

We choose then

Let be now 1p e

(Q),

p &#x3E; n ; there is

y~~ , 1pfJ (~

&#x3E;

0).

such

that, [7],

(16)

We consider the

problems

From the abowe we have

then

We choose then

.Rem. 1 - We observe that the

dependence

on n of the estimate

of th. 3 derive from the

dependence

on n of the estimate in the lemma

2 ;

then from

(2,13) repeating

our calculations we have

(17)

REFERENCES

[1] BABUSKA I., Solutions

of problems witj interfaces

and

singularities,

Int. Fluid

Dyn. Appl.

Math.,

April

1974.

[2] BABUSKA I., Solution

of interferences problems by homogeneisation,

Inst. Fluid

Dyn. Sppl.

Math., March 1974.

mès

asymptotiques

stationaires, C. R. Acad. Sc. Paris.

[3]

BENSOUSSAN A., LIONS J. L., PAPANICOLAU G., Sur

quelques phéno-

mès

asymptotiques

stationaires, C. R. Acad. Sc. Paris.

[4]

BIROLI M., Sur la

G-convergence

pour des

inéquations quasi-varia-

tionelles, C. R. Acad. Sc. Paris to appear.

[5] BOCCARDO L., CAPUZZO-DOLCETTA I.,

G-Convergenza

per un pro- blema di Dirichlet unilatero, Boll. U.M.I. 4, 12, 115-123, 1975.

[6]

BOCCARDO L., MARCELLINI P., Sulla convergenza delle soluzioni di

disequazioni quasi-variazionali,

Inst. Mat. Ulisse Dini, Firenze, 1975.

[7]

DE GIORGI, SPAGNOLO S., Sulla convergenza

degli integrali dell’energia

per

operatori

ellittici del ordine, Boll. UMI, 4, 8, 391-411, 1973.

[8] LIONS J. L.,

Homogeneisation,

Cours au

College

de France

1975-76,

Paris.

[9] LIONS J. L.,

Asymptotic behavoir of

solution

of

variational

inequalities

with

highly oscillating coefficients, Applications

of method of Fun- ctional

Analysis

to

problems

in Mechanics, Lec. Notes in Math., 503, 30-56.

[10]

LADY-ZENSKAIA O. A., URAL’CEVA N. N.,

Equations

aux derivées

partielles

du type

elliptique,

Dunod 1968.

[11] MURAT F., Sur

l’homogeneisation d’inequations elliptiques

.... Lab.

ss. 189 Paris VI, n. 76013.

[12]

PUEL J. P., Thèse, 1975 Paris.

[13] STAMPACCHIA G.,

Equations elliptiques

du second ordre à

coefficients

discontinus, Les presses de l’Univ. de Montreal 1965.

[14] TARTAR L., Cours Peccot 1977, Avril.

Manoscritto

pervenuto

in redazione il 20

aprile

1977.

Références

Documents relatifs

L’accès aux archives de la revue « Rendiconti del Seminario Matematico della Università di Padova » ( http://rendiconti.math.unipd.it/ ) implique l’accord avec les

L’accès aux archives de la revue « Rendiconti del Seminario Matematico della Università di Padova » ( http://rendiconti.math.unipd.it/ ) implique l’accord avec les

L’accès aux archives de la revue « Rendiconti del Seminario Matematico della Università di Padova » ( http://rendiconti.math.unipd.it/ ) implique l’accord avec les

L’accès aux archives de la revue « Rendiconti del Seminario Matematico della Università di Padova » ( http://rendiconti.math.unipd.it/ ) implique l’accord avec les

L’accès aux archives de la revue « Rendiconti del Seminario Matematico della Università di Padova » ( http://rendiconti.math.unipd.it/ ) implique l’accord avec les

L’accès aux archives de la revue « Rendiconti del Seminario Matematico della Università di Padova » ( http://rendiconti.math.unipd.it/ ) implique l’accord avec les

L’accès aux archives de la revue « Rendiconti del Seminario Matematico della Università di Padova » ( http://rendiconti.math.unipd.it/ ) implique l’accord avec les

L’accès aux archives de la revue « Rendiconti del Seminario Matematico della Università di Padova » ( http://rendiconti.math.unipd.it/ ) implique l’accord avec les