• Aucun résultat trouvé

The modulus o f c o n v e x i t y i n n o r m e d l i n e a r s p a c e s

N/A
N/A
Protected

Academic year: 2022

Partager "The modulus o f c o n v e x i t y i n n o r m e d l i n e a r s p a c e s "

Copied!
3
0
0

Texte intégral

(1)

ARKIV FOR MATEMATIK Band 4 nr 2

C o m m u n i c a t e d 12 N o v e m b e r 1958 b y LENNART CARLESON

The modulus o f c o n v e x i t y i n n o r m e d l i n e a r s p a c e s

B y Gi3TE NORDLANDER

W i t h 1 f i g u r e in t h e t e x t

The modulus of c o n v e x i t y 5 (e) of a n o r m e d linear space is defined as inf ( 1 - x + Y l ~

~(e)=

(Clarkson, D a y (1), L e m m a 5.1). F o r example, the modulus of convexity OE(~) of an a b s t r a c t Euclidean space is easily determined from the parallelogram identity a n d is found to be

die (e,) = 1 - r

d

E 2

4"

Theorem. The inequality

(e) < ~E (~)

holds /or every modulus o/ convexity r (e) in a spac~ with real scalars.

Proof. I t is sufficient to prove the theorem in t h e two-dimensional case. I n the usual w a y such a space m a y be t h o u g h t of as a plane in which the n o r m is determined b y a central-symmetric convex curve F with midpoint in O. The n o r m of a vector is determined b y the ratio of the length of the vector to the length of the parallel vector starting at O a n d ending at F. Thus F is the locus of the endpoints of the vectors from O of n o r m one.

If the two unit vectors x = OA a n d y = OB are r o t a t e d a r o u n d F, while their difference x - y = B A has c o n s t a n t l y the n o r m e, the endpoint M of the vector O M = 8 9 describes a curve F~ (Fig. 1).

l~ is a simple closed curve, as can be inferred from L e m m a 2.4 of D a y (1).

L e t Y be the area of the region inside 1 ~ a n d Y~ the area of the region inside F~.

L e m m a .

The lemma is a simple extension of a theorem of Holditch (de La Vall~e Poussin, p. 318), and is p r o v e d b y the same method.

15

(2)

G. :NORDLA.NDER, The modulus of convexity in normed linear spaces

C - . A

\ ,, .,,n / p

Fig. 1.

(The polar angle ~ is measured from an arbitrarily chosen direction.)

L e t (xl, Yl) a n d (x~, y~) b e t h e c o o r d i n a t e s of A a n d B in a n a r b i t r a r y rec- t a n g u l a r c o o r d i n a t e system. T h e n

Y = f yl d Xl ~ f y2 d x2

where t h e v a r i a b l e s xl, Yl, x2 a n d Y2 a r e t h o u g h t of as d e p e n d e n t of a p a r a - m e t e r t, v a r y i n g f r o m ~ t o /~ a s t h e curve is d e s c r i b e d once in t h e n e g a t i v e direction. The c o o r d i n a t e s of M are t h e n ' xl§ Yl~-Y~ a n d

[

\

2 ' 2 ]

Y~ = 88 f (y, § y~) d (x, § x~).

H e n c e Y - Y~ = ~ f (YI - Y~) d (X 1 - - X2).

The i n t e g r a l in t h e r i g h t m e m b e r expresses t h e a r e a of t h e region inside t h e curve d e s c r i b e d b y t h e e n d p o i n t of t h e v e c t o r AB, if t h i s v e c t o r is l a i d off from a fixed p o i n t . This curve is similar to F in t h e r a t i o e. H e n c e

Y - Y~= 88 Y, which p r o v e s t h e l e m m a .

The n o r m of t h e p a r t MC of t h e u n i t v e c t o r OC t h r o u g h M is d e n o t e d A (e, ~).

W e h a v e

(~ (e) = inf A (s, ~).

r

If t h e areas Y a n d :Y, are e x p r e s s e d in p o l a r c o o r d i n a t e s t h e l e m m a t a k e s t h e f o r m 16

(3)

ARKIV FOR MATEMATIK. Bd 4 nr 2

2~ 2~

o 0

where r(~) is the l e n g t h of the vector OC. Hence

which implies

because r 2 (~) > 0.

2~

0

V ~2

b(e)=infA(e,~)~<l- 1-~=~E(e),

Theorem 4.1 of D a y (2) shows t h a t if there is e q u a l i t y i n our t h e o r e m for all e, 0 < ~ ~< 2, the space is a b s t r a c t E u c l i d e a n . I have n o t been able to decide whether the same conclusion holds u n d e r the sole a s s u m p t i o n of e q u a l i t y for a certain ~0, 0 < e o < 2.

R E F E R E N C E S

CLARKSO~-, J . A . : U n i f o r m l y c o n v e x spaces. T r a n s . A m e r . M a t h . Soc. 40, 396-414 (1936).

DAY, M. M.: (1) U n i f o r m c o n v e x i t y in f a c t o r a n d c o n j u g a t e spaces. A n n a l s of M a t h . (2), 45, 375-385 (1944).

- - - - (2) S o m e c h a r a c t e r i z a t i o n s of i n n e r - p r o d u c t spaces. T r a n s . A m e r . M a t h . Soc. 62, 320-337 (1947).

DE LA VALL~E POUSSI~r CH.-J.: Cours d ' A n a l y s e I n f i n i t 6 s i m a l e , I.

Tryck~ den 2 upril 1959

Uppsala 1959. Almqvist & Wiksells Boktryckeri AB

i t 17

Références

Documents relatifs

It was however thought desirable to adhere to the familiar expressions since, as will be shown (a) each &#34;vector&#34; is uniquely determined by certain components, and

For stable spaces (~hat is, for spaces isomorphic to its square) one has the following stronger result.. isomorphic for every

There exists a coarse moduli scheme M for the equivalence classes of semistable parabolic vector bundles (on an irreducible nonsingular complete curve X}.. of

construct a reflexive sheaf from Y using the Serre construction, and then to describe this sheaf by means of a monad, where the family of monads involved

Nous 6nonr ~t la fin du second paragraphe, le r6sultat d'approximation dans le eas particulier des points minimaux, et, dans le troisi6me paragraphe, nous l'appli- quons pour

The following theorem implies Theorem 5 o f Rothschild--Stein [1] if one takes for Yi left invariant vector fields from the appropriate nilpotent Lie group...

N o u s nous placerons dans un espace vectoriel rrel, 6ventuellement muni d'une topologie vectorielle (ce qui sera prrcis6 dans chaque 6noncr) et adopterons les

As an approach to the problem of characterising and classifying Banach spaces in terms of their geometric structure, consideration has been given to the following