1. Trang chủ
  2. » Kỹ Thuật - Công Nghệ

Tài liệu Lược đồ sai phân của nghiệm một lớp phương trình vi phân ellip phi tuyến. pptx

7 456 0
Tài liệu đã được kiểm tra trùng lặp

Đang tải... (xem toàn văn)

THÔNG TIN TÀI LIỆU

Thông tin cơ bản

Định dạng
Số trang 7
Dung lượng 2,9 MB

Các công cụ chuyển đổi và chỉnh sửa cho tài liệu này

Nội dung

that in many applied problems the data are nonregular.. The approximate methods for the problems of no linear differential equations with data belo ging the So olev spaces W i, G are p

Trang 1

c hi Tin va Di'eu khi T.17, S.1 (200 1 ) , 1 - 1

Abstract It is known (see [1], [2], etc.) that in many applied problems the data are nonregular The approximate methods for the problems of no linear differential equations with data belo ging the So olev spaces W i, ( G ) are presented in [3 - 5] In this paper the finite -difference schemes of generalized solutio s

for a class of elliptic n nlinear differential equations are considered The theorem for the convergence of

appro imate solutio to generalized o e and error norm estimations is proved in the class of equatio s with the right-hand side defined by a continuo s linear functional in W J- (G)

Torn tlit N e u ba toan th u'c t~n du'oc dfin v'e gid cac bai toan d i vo'i phu' n trlnh vi p an r ieng voi

d ir kien kh6ng tro'n (xem [1 ]' [2)) Phuo-ng ph ap xfi p xl giai mot so b ai toan doi vo'i cac phtro'ng trlnh vi

ph an phi t uyen vci ve phdi thucc cac 1 >ham kh d tich kh ac n hau (cac kho g gian Sobolev WI ; ( G l du'o'c ngh ien cu'u trong c.ic cong trlnh [3- 5] Bai nay xet luo'c do sai ph an, n hien crru su' h9i tu va din gii sai

so cd a ngh iem bai to an d i vo'i mot 161> ph ong trln vi phan phi t uyeri lcai ellip v 'i ve p rii kh ng twn d9

cao kie'u cac phiem ham tu en tinh lien tuc (cac khOng gian WJ - (G)

Let G be a retan le with the.boundary aGo Consider the following problem

6 u T X, U ,-,- =-f x ) xEa, u(x) = , xE aG,

where the given f (x) E W2-1(G) - the space of continuo s linear functio als on the space

being a nonegative integer, the function T(x, a), a = (ao, aI, a2 ) satsfies the conditions:

2

[ T( x,a ) - T( x, b) ao - bo) ~ e l 2 )ai - b i ) 2,

= 0

W~(G),1

[ T(x, a ) - T(x, b) [ < c , [2 :)ai - bi)2] ,

i = O

( 2)

where e1, J= 1, 2, are the p siive co stants (see [3,chap 3, sec 4 ) ).

We shall use the same notatio s a in [6] Consider the generalized solutio u(x) of the problem

o (1) in the space W ~ (G) satisfying the following equality:

P(u , v) = J J [6 u + T(x , u, : 1 ' :x:)] v(x)dx = - JJ f(x)v(x)dx, (3)

where v( x ) is a functio in the space D ( G ) of Schwartz basic functo s [7

One has v (x) E WHG). Then, by [3] (chap 3, sc 4), if the function ri-, u, - satisfie

aXl aX2

the condi ons (2), f x) E L (G) there exists uniquely a solution of integral equation (3) u(x) E

W~ ( G ) n W ~ (G)

• This w o rk i s pa rti a ll s u pported by the N a ti na l B asics R esea r c h Pr ogra m in N atu r a l S c ienc es

Trang 2

2 CONSTRUCTION OF DIFFERENCE SCHEMES

We first consider the case where J(x) E L ( G ) and let G be the unit square G = {x = (Xl, X2 )

° <X" < 1, n.= 1, 2}

Let us introduce in the region G a grid w with interior a d b undary grid points denoted by w

and, respectively [ 6

To co struct the differe c schemes one may ta e the test functio s v ( x ) in the form:

{ _lk-kexp {- I x k l2k } x·E e,

( 4)

wh re e= e ( x ) = { ~= ( ~1'~2 ) : k " - xn l < O h ", n = 1, 2} ,h" b ing the steple gths, k being a natural number

Let e ery gridpoint x E w be corresponding to a mesh e ( x ) The gen ralize solution (denoted

by the GS) u ( x ) ofth problem (1) in esatsfies the following integral equation:

:£ 1 + O , Sh l x 2 + 0 5h 2

J J [ ~U( ~ )+T( ~'U'U ( , :~,:~ ) ]a ( l ) dl

(6)

One may rewrite the equation (S) as follows

(7)

where

1

SiU ( X ) = h :

t

x,+O,5h,

J U ( Xl, ··· ,li , ,x,, ) d1 i u(± O Gi) ( x) -- U Xl,( · ··,Xt .±O 1Sh· tl • , Xn' )

Now, to obtain th difference schemes of the oper ator (7) pr e (u , a) o e ma approximate the mean integral operators S , by the quadrature formula of average retangles and th partial derivatives

by difference q otients as in [ 6 (see 2.1) Hence, one get th following difference approximations

corresponding to (7), (3):

K ( y ) = 1P l ( y, a ) = L ( a iY x , ) x, - SlS2 L aXi ( x ) Y x , + SlS2a ( dT ( I, y ( x ) , Yx " YX2) = -<p, x E w,

y ( x ) = o , xE"

(8)

and (ef [3, chap 3, sec 4]

2

L(y) ==2P~ (y , a ) = L Xi Xi + SlS2a ( T ( I, y ( x ) , fi x " fi x 2) = -<p, x Ew,

y ( x ) = 0, x E/,

Trang 3

( ± 1,) - ( ± I ( ) - ( ±h, ) > 1

u == u x - u Xl) "" X l t, ,X n , 1 , _ ,

Note that b [3] (see chap.3, sec.4) there exists uniquely a solution of the operator equation

2P, : (y ) = - ' P and, then, of the equation IP: : (y , a )

3 ESTIMATION OF THE CONVERGENCE RATE

Estimate n w the method error and the approximate one of the scheme (8) and (9)

3 1 Consider the difference scheme ( 9 ) with ' P defined by (10), (7). Den te the metho error by

z = y - u, where y being the solutio of the problem ( 9 ) It follows from (9) that

L z = - t P( x ), x E w ; z ( x ) = 0, x E /, (11)

where t P( x ) is the approximatio error of the scheme ( 9 )

\{I( x ) = ' P+ Lu

From (10), ( 7) and by formulas (10), (11) in [6,sec 2], for the sufficiently small mesh sizes hi

and h2, o e has

"' [ (a U)- O 5 , )] (", a a U) ' P = - L S - aa;: x + S 1S 2 L 7 J 7 J.

( a aU)

- S1S 2 ~ , u( ~ ), - , - , x Ew

Thus,

'" [ ( au ) -0.5,) ] ~ aa a u

\{I=L Ux ,-S3- ; aa;: x + SI S L a7J

- SIS2 [T( ~, u(d , a au , ~) - T( ~ , u(x), UX 1(x), UX2(x))].

~ I a 2

By ( 9 ) one has

2

L Y = = L x ,x, = - SI S 2[ T(~ 'Y (X),y l' Yx ,)] -'P= 'Po, x E w.

(13)

Then,

L o X = L o y - L ou ==- \ {I o ( x ), x E w; z( x ) = 0, x ET

From (12) - (14) it follows that

(14)

'" " [ ( a U) ( O.5,)] '" aa a u

\{Io = L u - 'Po = L U X i X, - L S 3-i aa:; x + S1S2 L 7 J.7J

+ S I S2T(~, y ( x ), YX l ' Y x 2 ) S1S 2 T ( ~, u(d , ~ , aU)

a~ 1 a 2

Trang 4

DIFFERENCE SCHEMES OF GENERALIZED SOLUTIONS 13

(16)

-L ( zx , x" = L (l7i x , , z) + ( AO,Z ) +( Bo ,z) ,

i=l i = l

: E

(17)

IIZ t ] 1i2 -== (zx 1, ZXi 1i'

Nl N 2 -1 ~

( a, Z] l = L L a ( Jlhl,J 2h 2 )Z(Jlhl ,J 2h2)hlh 2 ,

Y1 =1 J ' 2= 1

N 1- l N 2

( a, zb = L L a ( Jlhl,J2h2)Z ( Jlhl,hh2 )hl h2 ,

VI = 1 ] 2 = 1

N , N 2

Il a l 2 = L L a2 ( Jl h l ,J2h2)hlh 2'

lt := 1i 2= 1

Then

(18)

where the constant C is independent of h ( l h l2 = h I +h~ ) and z ( x),

II I1 7.w = li z ll L + I V 'zI12, I l zllo w == I zll·

Now, we fist co sider the funct io al nj r ]defned by (16)

1

1 71 ( x ) = U X I - h

This expressio coincides with the one of 17 1( x ) (19) in [6] Hence, by (23) in [6]we have

117t x ) I ~ M l h l ( hlh2 ) - ~Il u I1 el,

where e is the followin mesh of the grid

Trang 5

ei = ei (x) == {I = ( 11 , 12 ) : Xi - hi < Ii < Xi, 1 13 - i - X 3 l <0, 5h3- d,

II U II " , " l == II U ll w; n(c l ) = ( L J I Duu l2dx ) 1 /2.

l(l~"Lf:

The functio al T / dx ) is estimated similarly Then,

x

The expressio of Ao coincides with the o e of T/ o (15) in 1 ].Then, b (26) in 1 ]we have

where H( h ) - + °as hl ' h - + 0

Consider now f3 (~ in (18) The difference of the form f 3(~ is estimated in 1 ] see chap.3, see 4) ,

one has

11f 3( ~ < C l h III U I12( ; From the last inequality and (16) it follows that

Finaly, combining (18) (2 ) we get

3.2. Consider the following difference scheme

1

2

where y = ~ ( y + 1 ) y and 11are defined (8) and (9) respectively Then,

2

M y = Z [ (1 +a i ) Y x , L , - Z3132 L ax , y , +

+ Z3132 a ( \ "lT( I ,Y ( X ) , Y x "Y x 2 ) + T( I , Y( X )'Y x "Y x , ) J

(23)

Thus,

2

MoY == L [(1 + ai )Y x , L ,

2

= 3132 L a x , Yx, - 3132 [ aT ( ' y ( x ) , Y l' Y x , ) + T ( s" , y ( x ) , u z , Y X2 )] - 2 p

- <Po, X E w,

y ( x ) = 0, X ET

From the last equality, (7) and (12) one has

Trang 6

DIFFERENCE SCHEMES OF GENERALIZED SOLUTIONS 1

\II(X) = - L [(1 +a ; z x , L , = L [ 7) i + f l )x, + ) 0 + (30 + q o,

i = 1 i = 1

(25)

By (24), (25), in the same way as in 3.1 o e has

2

[[Z[ w < C ( L (1 1 7)il l + Ilflil l i) + 11 ) 011 + 1 1(30 11 + Il q O II )

=

(26)

In (26) 7)i has the form (16), then one has the estimatio (19) for 7) i

The expression of fl i coincides with the one of I i (31) in [6)' then by (39) in [6) o e has

(27)

where H i ( h ) t = 1, 2 tend to z ro as h - O

) has the form (33) of ~ in [6)' then by (44) in [6)'

( 28 ) ( 30 has the form (16), then by (21) o e has

(29)

Consider the last summand q ill (26) The form of q is analogous to (3() and o e may easiy

verify that

Il q ll < C ! h H(h) lll u llv;

Now, combining (19), (2) ( 3 0) yields

I li I + f j - 2 ll lw :::: C : lhl ", - l iul l rn ( ;, m= 2 ,3

( 3 0)

(31)

Finally, by (22) and (31) we get the estmation of metho error for the difference scheme (8):

( 32 )

R ema rk In a manner analogous to the proof of the inequal ies (22) and (32), o e may verify that

these ineq alites are also valid if in the formula of the GS u(x) (5), (7), v(x) ( = a (hx) ) is a Schwarz

basic functio

3.3. The estimates (22) an (3 ) are obtained with the assumptio f E L2( G ) n w wesh w that the

results may be generalized to the eq atio s with rg t-hand side f E W J - I ) (G) , W J- I ) G ) being the

is the Dirac delta functio 6

Indeed, by our assumpto , f ( x ) E D'(G) , D'(G) bein the space of Schwartz distributions

Therefore, b the theorem on local structure of the distrib to s (see [7,chap 3, sec 6) there exists

Trang 7

f(x) = D~ D~ g ( x ) , (33)

where x Ee,the set eis compact in G ER", D i = a / aXi

Let v ( x ) E D( e ) , By (S) and (33) one has

// [6u( x ) +T(x, u, : [ , :XU2 ) ] v ( x ) dx = - IIg ( x ) v ( x ) dx, (34)

where

We see that v ( x ) is also a test functio : v ( x ) E D ( e ) c W~ ( e ) an g (x) E L 2 ( e ). Thus, the

equatio (34) has the form (S) Hence, one may repeat the procedure used above for the difference

schemes (8), (9) and obtaines the following

Theorem. L e t in the probl e m ( 1 ) the [uric o T ( ) s ati s fy the condition s (2 ) and the right - hand

s id e f E W ~ - I) ( G) Then the s olu t on y of the diffe r ence s cheme (8) or ( 9 ) ( y = y or 1j) conve r ges to

th e GS (S) u( x ) of the problem ( 1 ) In the grid norm Wi(w) with t h~ rate O ( I ~ I ) ' that I S, one has the

follo wing err o r e timation

Il - U ll l W :s: C lhlll u ll 2 ( ;,

w here the con s tan t C is independe n t of h and u ( x )

REFEREN C E S

[ G.1 Marchuk, Mathematical M odelling in the Env i ronme n t Problems , Nau ka, Moscow, 1982 (Russian)

[2] V S Vlad irniro , Generalized F unction s in M athematica l Ph qsic s , Mir, Moscow, 197

[3] A A Sam arsk ii,R D Laz arov ,V 1.Makarov, D ifference Scheme s for Gene r alized Solution s of Diffe r ential Equation s, Vus Univ., Moscow, 1987

[4] C Padr a, A posterior error estimators for nonconforming approximation of some quasi-Ne wto-nian flows, S I AM J Num e r , Anal 34 (4) (1997) 1600-161S

[S] C N Davson ,M.F Wheeler, C S Woodward, A two-grd finite difference scheme for non-linear parabolc equations, S I AM 1.Nurner Anal 35 (2) (1998) 43S-4S2

[6] Hoang Dinh Dung, Difference schemes for generalized soluto s of some elliptic differental

equations, I, Journal of Compute r Science and Cuberneiics 15 (1) (1999) 49-61

[7] L Schwartz, Th eorie de s Di s t ri but i ons, Hermann, Paris, 1 78

R eceived M a r h 2 , 20 00

Revi s ed January 5, 2001

I n s t itu t e o f M a th em atic s, N CS T o f Vi e t n am

Ngày đăng: 27/02/2014, 06:20

TỪ KHÓA LIÊN QUAN

TÀI LIỆU CÙNG NGƯỜI DÙNG

TÀI LIỆU LIÊN QUAN

🧩 Sản phẩm bạn có thể quan tâm

w