• Aucun résultat trouvé

An iterative method for conformal mappings of multiply-connected domains

N/A
N/A
Protected

Academic year: 2022

Partager "An iterative method for conformal mappings of multiply-connected domains"

Copied!
4
0
0

Texte intégral

(1)

A R K I V F O R M A T E M A T I K B a n d 4 n r 44

Communicated 24 J a n u a r y 1962 by L. CA~LESON

An iterative method for conformal mappings of multiply-connected domains

By INGEMAR LIND

Preliminary remarks

Let D be a connected domain in the z-plane containing the points z = 0 and z = oo.

Suppose t h a t there exists a function w = F (z), analytic a n d univalent in D such t h a t F (co) = 0% F ' (oo) = 1 and F (0) = 0, eonformally m a p p i n g D onto a domain of prescribed canonical type. There are four fundamental, harmonic functions in D in certain meanings measuring a point's deviation from its original position under the mapping:

Re {F (z) - z}, I m {F (z) - z}, log F (z) and arg F (z)

Z Z

B y studying the behaviour of these very simple types of functions under iterative mapping processes, described in detail below, existence proofs can be obtained for the simplest types of canonical domains - - the parallel slit domain, the circular slit domain, the radial slit domain a n d the logarithmic spiral slit domain.

There are of course several simple proofs of these results. The iterative proof given here m a y have a certain value; it is in certain cases suitable for actual calculation of the mapping function, and it gives an explanation to the simplicity of these mappings.

The uniqueness proofs - - as exemplified in [1], p. 57 (the second proof) for a paral- lel slit domain - - are also based on simple properties of the n a m e d types of har- monic functions.

Preparations and denotations

The following paper demonstrates in detail an existence proof for conformal m a p - ping onto a parallel slit domain where the slits are parallel to the imaginary axis.

To be precise the following well-known theorem is reformulated:

Every domain D in the z-plane o/connectivity k can be con]ormally "mapped onto a parallel slit domain in the w-plane, where the k slits are parallel to the imaginary axis.

Two arbitrary points ~ and fl o / D can be carried into the origin and the point o] in- /inity, respectively. 1/ w = F (z) denotes the mapping/unction in question, F (z) may be normalized by the requirement that its residue at z = fi be equal to 1. This normali- zation determines the ]unction uniquely.

3 7 : 6 5 5 7

(2)

INGEMAR LIND,

An iterative method for conformal mappings of multiply-connected domains

D (n) :

k

L (~)= U F(~):

v = l

:For t h e uniqueness proof one is referred to [1]. The existence proof is based on induction. The above t h e o r e m is t r u e for k = 1 (Riemann) a n d hereafter it is assumed true for k - 1. W i t h this a s s u m p t i o n there is no loss of generality to assume t h a t t h e given k-connected d o m a i n is b o u n d e d b y k - 1 rectilinear slits parallel to t h e imagi- n a r y axis a n d a n o t h e r non-rectilinear, analytic slit, a n d t h a t a is chosen as t h e ori- gin a n d fl as the point of infinity. The following notations are used:

a d o m a i n of the a b o v e described t y p e in t h e zn-plane;

t h e b o u n d a r y of D(~); F(~ ~), v = 1, 2 ... k, denote t h e separate boun- d a r y c o m p o n e n t s one of which is non-rectilinear. Occasionally this c o m p o n e n t will be d e n o t e d b y F (n) (lower index omitted) a n d its corresponding rectilinear b o u n d a r y c o m p o n e n t in t h e z~.l-plane (see below) b y y(~-l);

D [F~ ~)] or D ~ ) : t h a t d o m a i n of connectivity k - 1 in the

zn-plane,

which b o u n d a r y can be written L <~) - 1~ n) (1 ~< ~ ~< k).

The iterative process

Assume i. e. t h a t F (1) = F(21). Map eonformally D(1 ~) in the zl-plane onto D~ 2) in the z2- plane where all t h e b o u n d a r y slits F(~ ), v = 2, 3 ...k are rectilinear slits parallel to t h e i m a g i n a r y axis. The possibility of this m a p p i n g is assured b y t h e induction assumption. ~(1)= F(11) in D~ 1) is h e r e b y carried into a non-rectilinear, analytic slit F(12)= F (2) in D(12). N e x t m a p in t h e same w a y D(2 2) in t h e z2-plane o n t o D(2 ~) in t h e

l r ~ ( k + l )

z3-plane, t h e n D(3 a) onto D(a 4) a n d so on. T h e k : t h step is to m a p D(k ~) o n t o ~ k 9 Next, m a p D(1 k+l) o n t o D(1 k+2), t h e n D(2 k+2) onto D(2 ~+3) a n d so on. The m a p p i n g pro- cess is n o w iterated infinitely in t h e suggested cyclic manner. T h e analytic function performing t h e described m a p p i n g between t h e actual d o m a i n in the Z~_l-plane a n d t h a t in t h e z~-plane is d e n o t e d b y z~ = F~ (Z~-l), a n d it is required t h a t Fn (0) = 0, F~ (cr = ~ a n d 2'~ ( ~ ) = 1 (n = 2, 3 ...), t h e possibility of this again being assured b y t h e induction assumption. L e t y~(zl) = F~ { F ~ - I ( 9 9 (Zl) 9 .)} denote t h e composed function, analytic in D (~) a n d m a p p i n g D (D conformally onto D (~).

Induction statement and proof

yJ(zl) =~i=m ~n(zl) exists, is analytic in D <1), maps D ~> con/ormally onto a parallel slit domain with k rectilinear slits parallel to the imaginary axis, and v(O) = O, ~ ( ~ )

= ~ , ~ ' ( ~ ) = 1.

Proo/:

The functions

un(z~)=Re

{ F ~ + l ( z ~ ) - z ~ } , n = 1, 2 ... (1) are harmonic in D [y(n)], u~ ( 0 ) = 0, a n d since F . + l ( Z ~ ) - z~ is analytic a n d b o u n d e d in D [~,(n)], u~ (z~) attains its m a x i m u m a n d m i n i m u m values on F (~). W i t h the n o t a t i o n

J~ = sup ] R e {z~ - z~'}[ = sup [ u n ( z ~ ) - u~(z~')[

it thus follows for all z~ t h a t

l un (zn) I ~< J~. (2)

558

(3)

ARKIV F6R MATEMATIK. B d 4 nr 44 Since u,~(z,) is n o t constant, it follows from the m a x i m u m principle t h a t there exists a n u m b e r qn, 0 < q, < 1, so t h a t

sup lU.(Z~)--U.(Z~*)[<~Jnq,.

* * * c F ( n ) z n , z

Combining [4], l e m m a 1, p. 282 with [2], "Verzerrungssats V " , p. 235 it m o s t easily follows t h a t it is possible to fix a n u m b e r q, 0 < q < 1, so t h a t q, c a n be chosen

~< q for all n. T h u s

sup [u,~(z~)-u,~(z;*)[ < J ~ q . B u t this is equivalent to

J , + l = sup [Re (z~ +x - z~'+l} [ ~< J~ q.

~:,+r z~'+l~r(" +1) Thus

Jn ~ J l q n-i for all n. (3)

The functions Um (z.) = Re {Z.+m -- Z.} are harmonic in D (n) a n d

m 1

Um (Zn) = ~ U.§

(4)

is valid. F r o m (3) it follows for e v e r y point z~ t h a t

q n - 1

[Um(zn)[ ~<J1 1 _ q < e if n>n~. (5) T h u s R e {y)n (zl) - zl} converges u n i f o r m l y to a harmonic function in D (1), and, since the c o n j u g a t e functions are single-valued a n d well determined, the convergence is easily extended to the sequence of analytic functions {y)n (Zl) - zl}. F u r t h e r (3) states t h a t t h e limit d o m a i n is of t h e desired type. Obviously ~p(0)=0, 7)(oo)= oo a n d v 2' (oo) = 1 are valid. The i n d u c t i o n s t a t e m e n t is proved.

C o m p l e m e n t a r y r e m a r k s

I t has just been p r o v e d t h a t w = ~0 (zl) m a p s D (1) eonformally onto a parallel slit d o m a i n where u -- c., ~ = 1, 2 ... k, on the separate slits (w = u + iv = Re~~ c. are cer- t a i n real constants). Obvious modifications of t h e given proof yield the existence of t h e following m a p p i n g functions:

a) w=~p,~(Zl); o~u+flv=c~ on the slits, t g v a = -~---

8'

b) w = P (zl); R = c~ on the slits, (circular slit domain);

e) w = Q (zx); 0 = c, on t h e slits (radial slit domain);

d) w = R (Zl); ~ log R + 80 = c~ on the slits (logathmie spiral slit domain), ~, 8 ~= 0.

Various kinds of iterative processes were a. o. studied b y Koebe. A n interesting example giving the existence of a conformal m a p p i n g onto t h e circle d o m a i n is found in [3].

559

(4)

I~GEMAR LIND, An iterative method for conformaI mappings of multiply-connected domains

R E F E R E N C E S

1. I:~. COURANT, D i r i c h l e t ' s P r i n c i p l e , C o n f o r m a l M a p p i n g , a n d M i n i m a l S u r f a c e s , I n t e r s c i e n c e P u b l i s h e r s , I n c . , :New Y o r k (1950).

2. P . KOEBE, A b h a n d l u n g e n z u r T h e o r i e d e r k o n f o r m e n A b b i l d u n g V . , M a t h . Z. 2. B a n d (1918), p p . 198-236.

3. - - , A b h a n d l u n g e n z u r T h e o r i e d e r k o n f o r m e n A b b i l d u n g VI., M a t h . Z. 7. B a n d (1920), pp.

235-301.

4. L. SARIo, A l i n e a r o p e r a t o r m e t h o d o n a r b i t r a r y R i e m a n n s u r f a c e s , T r a n s . A m e r . M a t h . Sot., Vol. 72, :No 2 (1952), p p . 281-295.

560

Tryckt den 23 november 1962

Uppsala 1962. Almqvist & Wiksells Boktryckeri AB

Références

Documents relatifs

Rafael Mahieu, MD, MPH; 1 Viviane Cassisa, MPH; 2 Diane Sanderink, MD, MPH; 1 Rachel Chenouard, MD; 1,2 Hélène Pailhoriès, PharmD, PhD; 1,2 Marie Kempf, PharmD, PhD; 1,2

We introduce an iterative method for finding a common fixed point of a semigroup of nonexpansive mappings in a Hilbert space, with respect to a sequence of left regular means defined

Given a directed and connected graph G = (V, E) and a collection {G ij } (i,j)∈E of matrices in R d×d , the problem is to find a collection of transitively consistent (or

restricted for a time to one-dimensional systems (E) with hyperbolic layer equations (E)o (see [6]), but these limitations were overcome in 1976 by using approximate

Toute utilisation commerciale ou impression systématique est constitutive d’une infrac- tion pénale.. Toute copie ou impression de ce fichier doit contenir la présente mention

Toute utilisation commerciale ou impression systématique est constitutive d’une infrac- tion pénale.. Toute copie ou impression de ce fichier doit contenir la présente mention

For these domains the BKM/AM results are several orders of magnitude more accurate than those obtained by Rabinowitz(1966), using the Szegö kernel function method with

proposed an exact solution method based on the difference of convex func- tions to solve cardinality constrained portfolio optimization problems.. To our knowledge, all existing