• Aucun résultat trouvé

On isometries of Product of normed linear spaces

N/A
N/A
Protected

Academic year: 2021

Partager "On isometries of Product of normed linear spaces"

Copied!
8
0
0

Texte intégral

(1)

HAL Id: hal-01086201

https://hal.archives-ouvertes.fr/hal-01086201

Preprint submitted on 23 Nov 2014

HAL is a multi-disciplinary open access

archive for the deposit and dissemination of

sci-entific research documents, whether they are

pub-lished or not. The documents may come from

teaching and research institutions in France or

L’archive ouverte pluridisciplinaire HAL, est

destinée au dépôt et à la diffusion de documents

scientifiques de niveau recherche, publiés ou non,

émanant des établissements d’enseignement et de

recherche français ou étrangers, des laboratoires

On isometries of Product of normed linear spaces

Mohammed Bachir

To cite this version:

(2)

Mohammed Ba hir

November23, 2014

Laboratoire SAMM4543, UniversitéParis1Panthéon-Sorbonne, CentreP.M.F. 90rue

Tolbia 75634Paris edex13

Email: Mohammed.Ba hiruniv-paris1.fr,

Abstra t. We give a ondition on norms under whi h twove tornormed spa es

X

and

Y

areisometri allyisomorphi ifandonlyif

X

× R

and

Y

× R

areisometri allyisomorphi . We also provethat this resultfail forarbitrarynorms even if

X

= Y = R

2

bybuilding ageneri

ounterexamples.

Keyword,phrase: Normedve torspa eandisometries.

1 Introdu tion

We areinterestedin this paperin the followingquestion. Let

X

and

Y

beve torspa es and let

N

X

and

N

Y

betwonormson

(X × R, N

X

)

and

(Y × R, N

Y

)

respe tively. Thenorm

N

X

on

X

(andinasimilar way

N

Y

)denotes

N

X

(x, 0)

forall

x

∈ X

.

Problem. It is true that

(X × R, N

X

)

and

(Y × R, N

Y

)

are isometri ally isomorphi if and onlyif

(X, N

X

)

and

(Y, N

Y

)

areisometri allyisomorphi ?

Webeginbyshowingthat inthegeneral asetheanswertothisquestionisnoforarbitrary

norms

N

X

and

N

Y

,evenwhen

X

and

Y

aretwodimensionalve torspa es,by onstru tinga generi ounterexamples(SeeTheorem1). WeprovetheninTheorem2thattheresultistrue

forallnorms

(N

X

, N

Y

)

satisfyingthefollowingproperty

(P )

.

Denition1 Let

X

and

Y

betwove torspa es. Let

N

X

and

N

Y

betwo normson

X

× R

and

X

× R

respe tively. Wesaythethe pair

(N

X

, N

Y

)

satisfy theproperty

(P )

iffor all

x

∈ X

and all

y

∈ Y

:

N

X

(x, 0) = N

Y

(y, 0) ⇒ N

X

(x, λ) = N

Y

(y, λ), ∀λ ∈ R.

In allthe arti leweidentify

X

with

X

× {0}

andthe norm

N

X

on

X

denotes

N

X

(x, 0)

for all

x

∈ X

.

Exemples 1 Let

(X, k.k

X

)

and

(Y, k.k

Y

)

betwo normedve torspa es. Let

p

∈ [1, +∞[

and

N

X,p

(x, t) := (kxk

p

X

+ |t|

p

)

1

p

,

N

X,∞

(x, t) := max(kxk

X

,

|t|),

for all

(x, t) ∈ X × R.

Inasimilarwaywedene

N

Y,p

and

N

Y,∞

. Then thepairs

(N

X,p

, N

Y,p

)

and

(N

X,∞

, N

Y,∞

)

satisesthe property

(P )

.

(3)

Proposition1 Let

N

R

2

beanynormon

R

2

su hthat

N

X

(x, t) := N

R

2

(kxk

X

,

|t|)

forall

(x, t) ∈

X

× R

denedanormon

X

× R

(Similarly we dene

N

Y

on

Y

× R

). Then

(N

X

, N

Y

)

satisfy the property

(P )

.

Proof. Let

x

∈ X

and

y

∈ Y

be su h that

N

X

(x, 0) = N

Y

(y, 0)

. Then

N

R

2

(kxk

X

,

0) =

N

R

2

(kyk

Y

,

0)

andso

kxk

X

N

R

2

(1, 0) = kyk

Y

N

R

2

(1, 0)

,whi himpliesthat

kxk

X

= kyk

Y

. It fol-lowsthat

N

R

2

(kxk

X

,

|λ|) = N

R

2

(kyk

Y

,

|λ|)

for all

λ

∈ R

. Inotherwords

N

X

(x, λ) = N

Y

(y, λ)

forall

λ

∈ R

.

Theproblemmentionedabovewasmotivatedatthersttimein[1℄byquestions onne ted

to the Bana h-Stone theorem, and solvedpositively only for the parti ular norms

N

X,p

and

N

Y,p

when

p

∈ [1, +∞[\ {2}

. The te hnique used in [1℄ did not in lude the ase p=2. The property

(P )

hear is moregeneraland allowed to in lude variednorms. We givein se tion4. othersimpleexamplesofappli ationsofTheorem2.

2 A generi ounterexample.

Theorem 1 Let

X

= Y = R

2

. For ea h norm

k.k

X

on

X

there existsa norm

k.k

Y

on

Y

, a norm

N

X

on

X

× R

andanorm

N

Y

on

Y

× R

su hthat:

(1) (X, k.k

X

)

is notisometri ally isomorphi to

(Y, k.k

Y

)

.

(2) (X × R, N

X

)

isisometri allyisomorphi to

(Y × R, N

Y

)

.

(3)

the restri tionof

N

X

to

X

oin idewith

k.k

X

andthe restri tion of

N

Y

to

Y

oin ide with

k.k

Y

.

Proof. Let

p

∈ [1, +∞[

. Letusdene

N

X

and

N

Y

as follow:

N

X

(x

1

, x

2

, t) := (k(x

1

, x

2

)k

p

X

+ |t|

p

)

1

p

,

∀(x

1

, x

2

, t) ∈ X × R

and

N

Y

(y

1

, y

2

, s) := (|y

2

|

p

+

k(y

1

, s)k

p

X

a

p

)

1

p

,

∀(y

1

, y

2

, s) ∈ Y × R.

Where

a

= k(1, 0)k

X

. Letus dene thenorm

k.k

Y,p

on

Y

as follows

k(y

1

, y

2

)k

Y,p

:= (|y

1

|

p

+

|y

2

|

p

)

1

p

forall

(y

1

, y

2

) ∈ Y

. Clearly,

N

X

(x

1

, x

2

,

0) = k(x

1

, x

2

)k

X

,

∀(x

1

, x

2

) ∈ X

and

N

Y

(y

1

, y

2

,

0) = (|y

1

|

p

+ |y

2

|

p

)

1

p

:= k(y

1

, y

2

)k

Y,p

,

∀(y

1

, y

2

) ∈ Y.

(Sin e

k(y

1

,0)k

p

X

a

p

= |y

1

|

k(1,0)k

p

X

a

p

= |y

1

|

). Onthe otherhand,the followingmap isan isometri isomorphism:

Θ : (X × R, N

X

)

→ (Y × R, N

Y

)

(x

1

, x

2

, t) 7→ (ax

1

, t, ax

2

).

Now,there existto ases:

Case 1: Ifeverypointof thesphere

S

X

of

X

isanextremepoint,we hoose

p

= 1

andso

S

Y

hasanon extremepointsin einthis ase

k(y

1

, y

2

)k

Y,1

= |y

1

| + |y

2

|

(Forexample

(

1

2

,

1

2

)

isnot extremefor

k.k

Y,1

). Consequently

X

and

Y

annotbeisometri allyisomorphi .

Case 2: Ifthereexistssomepointofthesphere

S

X

whi hisnotextremepointthenwe hoose

p

= 2

andsoeverypointsof

S

Y

isanextremepointsin e

k(y

1

, y

2

)k

Y,2

= (|y

1

|

2

+ |y

2

|

2

)

1

2

isthe eu lideannorm. Also

X

and

Y

annotbeisometri allyisomorphi .

(4)

Theorem 2 Let

X

and

Y

beave torspa es. Supposethat

(N

X

, N

Y

)

satisfythe property

(P )

. Then

(X × R, N

X

)

and

(Y × R, N

Y

)

areisometri ally isomorphi if and only if

(X, N

X

)

and

(Y, N

Y

)

areisometri ally isomorphi .

Theproofoftheabovetheoremisgivenin se tion3.2aftersomelemmas.

3.1 Notations and lemmas.

Weneedsomenotationsandlemmas. Let

Θ : (X × R, N

X

) → (Y × R, N

Y

)

beanisomorphism isometri . Weset

(a, u) = Θ

−1

(0, 1)

and

(b, v) = Θ(0, 1)

. Let us denethe linear ontinuous map

χ

X

as follow:

χ

X

: X × R

−→ R

(x, t)

7→

t

Wedeneanalogouslythemap

χ

Y

by

χ

Y

: Y × R

−→ R

(y, t)

7→

t

Weobtainthefollowinglinearmapon

Y

× {0}

:

χ

X

◦ Θ

−1

: Y × {0} −→ R

Analogouslywehavealsothelinearmapon

X

× {0}

:

χ

Y

◦ Θ : X × {0} −→ R

Letusset

X

0

:= Ker (χ

Y

◦ Θ)

and

Y

0

:= Ker χ

X

◦ Θ

−1



.

Remark1 The linear spa es

X

0

and

Y

0

are not ne essarily losed sin e

χ

X

and

χ

Y

are not ne essarily ontinuous.

Lemma1

X

0

and

Y

0

are isometri ally isomorphi . Morepre isely,the map

Θ : (X

0

, N

X

)

→ (Y

0

, N

Y

)

(z, 0) 7→ Θ(z, 0)

(1)

isan isomorphism isometri .

Proof. Sin e

Θ

is anisomorphismisometri ,itsu esto showthattherestri tionof

Θ

to

X

0

is onto. Indeed, let

(y, 0) ∈ Y

0

. Clearly,

(z, 0) := Θ

−1

(y, 0) ∈ X

0

sin e

χ

Y

◦ Θ(Θ

−1

(y, 0)) =

χ

Y

(y, 0) = 0

andwehave

(y, 0) = Θ(z, 0)

.

Lemma2 Wehave onlytwo ases.

Case1:

u

6= 0

. In this ase,wehave

X

× {0} = X

0

. Case2:

u

= 0

. In this asewe have

Θ

−1

(0, 1) = (a, 0)

and

X

× {0} = X

0

⊕ R(a, 0)

. Similarly wehave,

Case1:

v

6= 0

. In this ase,wehave

Y

× {0} = Y

0

.

(5)

Proof. Forall

x

∈ X

thereexists

(y

x

, λ

x

) ∈ Y × R

su hthat

(x, 0) = Θ

−1

(y

x

, λ

x

)

= Θ

−1

(y

x

,

0) + λΘ

−1

(0, 1)

= Θ

−1

(y

x

,

0) + λ

x

(a, u)

(2)

= Θ

−1

(y

x

,

0) + (λ

x

a, λ

x

u)

Sin e

Θ

−1

(y

x

,

0) ∈ X

0

⊂ X × {0}

andalso

(x, 0) ∈ X × {0}

,thenfromtheaboveequationwe obtainthat

x

a, λ

x

u) ∈ X × {0}

whi h impliesthat

λ

x

u

= 0

. Sowehave:

Case1:

u

6= 0

. In this ase,

X

× {0} = X

0

. Indeed, if

u

6= 0

then

λ

x

= 0

and so

(x, 0) =

Θ

−1

(y

x

,

0) ∈ X

0

,forall

x

∈ X

i.e

X

×{0} ⊂ X

0

. Ontheotherhandweknowthat

X

0

⊂ X ×{0}

. Case2:

u

= 0

. Inthis asewehave

Θ

−1

(0, 1) = (a, 0)

andso

X

= X

0

⊕ R(a, 0)

. Indeed,We have

X

0

∩ R(a, 0) = (0, 0)

, sin eif

α

is areal numbersu h that

α(a, 0) ∈ X

0

then

0 = χ

Y

Θ (α(a, 0)) = αχ

Y

(0, 1) = α

. Inotherwordsfrom(2),forall

x

∈ X

,thereexist

(y

x

, λ

x

) ∈ Y ×R

su hthat

(x, 0)

= Θ

−1

(y

x

,

0) + λ

x

(a, 0).

whith

Θ

−1

(y

x

,

0) ∈ X

0

. Thus

X

× {0} ⊂ X

0

⊕ R(a, 0) ⊂ X × {0}

andso

X

× {0} = X

0

⊕ R(a, 0)

.

Inasimilarwayweobtainthese ondpartofthelemma.

Lemma3 Wehave,

u

= 0

ifandonly if

v

= 0

.

Proof. Suppose that

v

= 0

. Thenforall

(x, t) ∈ X × R

,wehave

Θ(x, t) = Θ(x, 0) + Θ(0, t) =

Θ(x, 0) + tΘ(0, 1) = Θ(x, 0) + t(b, v) = Θ(x, 0) + (tb, 0)

. Now,wearegoingtoprovethat

u

= 0

. Supposethatthe ontraryhold,that is

u

6= 0

. Then

X

× {0} = X

0

(Seethe ase1. inLemma 2). So

Θ(x, 0) ∈ Θ(X × {0}) = Θ(X

0

) = Y

0

, sin e

Θ

is an isomorphism isometri from

X

0

onto

Y

0

(See the formula(1)). Nowsin e

Y

0

⊂ Y × {0}

, then

Θ(x, 0) + t(b, 0) ∈ Y × {0}

. In other words,

Θ(x, t) ∈ Y × {0}

for all

(x, t) ∈ X × R

. So

Θ(X × R) ⊂ Y × {0}

. But

Θ

isan isomorphismbetween

X

× R

and

Y

× R

. Thisimpliesthat

Y

× {0} = Y × R

whi hisimpossible. Thus

u

= 0

. Inasimilarwayweobtainthe onverse.

3.2 Proof of Theorem 2 and some orollaries.

Wegivenowtheproofofthemain result.

Proof of Theorem 2. For the if part, let

T

: (X, N

X

) → (Y, N

Y

)

be an isomorphism iso-metri . Let usdene

Θ : (X × R, N

X

) → (Y × R, N

Y

)

by

Θ(x, λ) = (T (x), λ)

. Then, learly

Θ

is an isomorphismandby theproperty

(P )

itis alsoisometri . We provenowthe onlyif part. By ombiningLemma2and Lemma3wehavethat:

Case1. If

u

6= 0

and

v

6= 0

,then

X

× {0} = X

0

and

Y

× {0} = Y

0

. SobyLemma1we on lude that

X

× {0}

and

Y

× {0}

areisometri allyisomorphi forthenorms

N

X

and

N

Y

. So

(X, N

X

)

and

(Y, N

Y

)

areisometri allyisomorphi .

Case2. If

u

= 0

and

v

= 0

, using Lemma 2we havethat

Θ

−1

(0, 1) = (a, 0)

and

X

× {0} =

X

0

⊕ R(a, 0)

and

Θ(0, 1) = (b, 0)

and

Y

× {0} = Y

0

⊕ R(b, 0)

. Nowweprovethat themap

ψ

: X × {0} = X

0

⊕ R(a, 0) → Y × {0} = Y

0

⊕ R(b, 0)

(z, 0) + λ(a, 0)

7→ Θ(z, 0) + λ(b, 0)

is anisomorphismisometri . Indeed, thefa t that

ψ

islinearand onto map is learby using Lemma 1. Letus provethat

ψ

isisometri forthenorms

N

X

and

N

Y

. Sin e

(z, 0) ∈ X

0

, by (1)there exist

(y, 0) ∈ Y

0

su hthat

Θ(z, 0) = (y, 0)

. Sin e

(z, 0) + λ(a, 0)

= Θ

−1

(Θ(z, 0)) + λΘ

−1

(0, 1)

= Θ

−1

(Θ(z, 0) + (0, λ))

= Θ

−1

(y, λ)

(6)

then,using thefa t that

Θ

−1

isisometri wehave

N

X

((z, 0) + λ(a, 0)) =

N

X

−1

(y, λ))

=

N

Y

(y, λ).

(3)

On theother hand weknown that

(b, 0) = Θ(0, 1)

so

Θ(z, 0) + λ(b, 0) = Θ(z, 0) + λΘ(0, 1) =

Θ(z, λ)

. Thus,usingthefa tthat

Θ

isisometri wehave,

N

Y

(ψ ((z, 0) + λ(a, 0))) =

N

Y

(Θ(z, 0) + λ(b, 0))

=

N

Y

(Θ(z, λ))

=

N

X

(z, λ).

(4)

But

N

X

(z, 0) = N

Y

(y, 0)

sin e

Θ(z, 0) = (y, 0)

and

Θ

is isometri . Sin e

(N

X

, N

Y

)

satisfythe property

(P )

then

N

X

(z, λ) = N

Y

(y, λ)

. Thus,using theformulas(3)and (4)weobtainthat

ψ

is isometri .

Remark2 Byindu tion, we an easily extend the above theorem to

X

× R

n

(

n

∈ N

) if we

assume that

(N

X

, N

Y

)

isapair ofnorms satisfyingthe following property

(P

n

)

: for all

x

∈ X

all

y

∈ Y

, all

i

∈ {1, 2, ...n}

andall

(s

1

, s

2

, ..., s

i

); (s

1

, s

2

, ..., s

i

) ∈ R

i

: if

N

X

(x, s

1

, s

2

, ..., s

i

,

0, 0, ..., 0) = N

Y

(y, s

1

, s

2

, ..., s

i

,

0, 0, ..., 0)

then

N

X

(x, s

1

, s

2

, ..., s

i

, λ,

0, ..., 0) = N

Y

(y, s

1

, s

2

, ..., s

i

, λ,

0, ..., 0), ∀λ ∈ R

.

Exemples 2 Let

p

∈ [1, +∞[

, and

N

X,p

(x, s

1

, ..., s

n

) = (kxk

p

X

+

n

X

k=1

|s

k

|

p

)

1

p

,

N

X,∞

(x, s

1

, ..., s

n

) = max(kxk

X

,

|s

1

|, ..., |s

n

|)

for all

(x, s

1

, ..., s

n

) ∈ X × R

n

. In a similar way we dene

N

Y,p

and

N

Y,∞

. Then the pairs

(N

X,p

, N

Y,p

)

and

(N

X,∞

, N

Y,∞

)

satisesthe property

(P

n

)

.

Corollary 1 Let

X

and

Y

be ave torspa es. Let

n

∈ N

andsupposethat

(N

X

, N

Y

)

satisfy

(P

n

)

. Then

(X ×R

n

, N

X

)

and

(Y ×R

n

, N

Y

)

areisometri allyisomorphi ifandonlyif

(X, N

X

)

and

(Y, N

Y

)

areisometri ally isomorphi .

As a remark we have the following orollary for inner produ t spa es. Note that a non

omplete inner produ t spa e has no orthonormal basis in general (See [5℄). The symbol

=

meansisometri allyisomorphi .

Corollary 2 Let

(H, k.k

H

)

and

(L, k.k

L

)

betwo innerprodu tspa e(notne essary omplete). Then

(H, k.k

H

) ∼

= (L, k.k

L

)

ifandonlyifforallnitedimensionalsubspa es

E

⊂ H

and

F

⊂ L

su hthat

dim(E) = dim(F )

wehavethat

(E

,

k.k

H

) ∼

= (F

,

k.k

L

)

. Where

E

and

F

denotes

the orthogonal of

E

and

F

respe tively.

Proof. Let

E

⊂ H

and

F

⊂ L

su h that

dim(E) = dim(F ) = n

for

n

∈ N

. Bythe lassi al proje tion theorem on a omplete ve tor subspa e of an inner produ t spa e, we have

H

=

E

⊕ E

and

L

= F

⊕ F

. On theother hand itis learthat

(H, k.k

H

) ∼

= (E

× R

n

, N

E

,2

)

and

(L, k.k

L

) ∼

= (F

× R

n

, N

F

,2

)

, where

N

E

,2

and

N

F

,2

are dened as in theExample 2

with

p

= 2

. Sin e

(N

E

,2

, N

F

,2

)

satisfy

(P

n

)

thenfrom Corollary1weobtain

(E

,

k.k

H

) ∼

=

(F

,

k.k

(7)

We give in this se tion two appli ations of Theorem 2. We denote by

(C

1

[0, 1], N

C

1

[0,1]

)

the spa e of ontinuously dierentiable fun tions on

[0, 1]

endowed with thenorm

N

C

1

[0,1]

(f ) :=

N

R

2

(kf

k

,

|f (0)|)

,where

N

R

2

denotesanynormsatisfyingProposition 1. Let

(X, k.k

X

)

bea Bana hspa e. Wedenoteby

N

X

thenormdenedon

X

× R

by

N

X

(x, t) := N

R

2

(kxk

X

,

|t|)

for all

(x, t) ∈ X × R

. Finally, wedenote by

(C[0, 1], k.k

)

thespa e of ontinuous fun tions on

[0, 1]

endowedwiththesupremumnorm.

Proposition2 Wehave

(X×R, N

X

) ∼

= (C

1

[0, 1], N

C

1

[0,1]

)

,ifandonlyif

(X, k.k

X

) ∼

= (C[0, 1], k.k

)

.

Proof. Let us dene thenorm

N

C[0,1]

on

C[0, 1] × R

by

N

C[0,1]

(g, t) := N

R

2

(kgk

,

|t|)

for all

(g, t) ∈ C[0, 1] × R

. Letus onsider themap

χ

: (C

1

[0, 1], N

C

1

[0,1]

) → (C[0, 1] × R, N

C[0,1]

)

f

7→ (f

, f(0))

Clearly,

χ

is an isomorphism isometri . So we have

(X × R, N

X

) ∼

= (C[0, 1] × R, N

C[0,1]

)

. Sin e

(N

X

, N

C[0,1]

)

satisfy the property

(P )

by Proposition 1 then using Theorem 2 we ob-tain that

(X, k.k

X

) ∼

= (C[0, 1], k.k

)

, sin e

N

X

(kxk

X

,

0) = kxk

X

N

R

2

(1, 0)

and

N

C[0,1]

(g, 0) =

kgk

N

R

2

(1, 0)

.

Let us re all some notions. Let

K

and

C

be onvex subsets of ve tor spa es. A fun tion

T

: K → C

is said to be ane if for all

x, y

∈ K

and

0 ≤ λ ≤ 1

,

T

(λx + (1 − λ)y) =

λT

(x) + (1 − λ)T (y)

. Thesetofall ontinuousreal-valuedanefun tions ona onvexsubset

K

ofatopologi alve torspa ewillbedenotedby

Af f

(K)

. Clearly,alltranslatesof ontinuous linearfun tionalsareelementsofA

(K)

,but the onversein nottruein general(see[4℄page 22.). However,wedohavethefollowingrelationship.

Proposition3 ([4 ℄,Proposition4.5)Assumethat

K

isa ompa t onvexsubsetofaseparated lo ally onvexspa e

X

then

n

a

∈ Af f (K) : a = r + x

|K

f or some x

∈ X

and some r

∈ R

o

isdense in

(Af f (K), k.k

)

, where

k.k

denotesthe normof uniform onvergen e. Butintheparti ular asewhen

X

isaBana hspa eand

K

= (B

X

, w

)

istheunitballof

thedualspa e

X

endowedwiththeweakstartopology,thewellknownresultduetoBana h

andDieudonné statesthat:

Theorem 3 (Bana h-Dieudonné). Thespa e

(Af f

0

(B

X

), k.k

)

isisometri ally identied to

(X, k.k)

. In other words,

Af f

0

(B

X

) =

 ˆ

z

|B

X∗

: z ∈ X .

Where

Af f

0

(B

X

)

denotes the spa e of all ane weak star ontinuous fun tions that vanish at

0

and

z

ˆ

: p 7→ p(z)

for all

p

∈ X

and

z

ˆ

|B

X∗

denotes the restri tionof

z

ˆ

to

B

X

.

Now,let

X

and

Y

betwoBana hspa esandletusendowedthespa e

Af f

(B

X

)

(andin a similar way the spa e

Af f

(B

Y

)

) with the norm

N

(f ) := N

R

2

(kf − f (0)k

,

|f (0)|)

for all

f

∈ Af f (B

X

)

, where

N

R

2

denotes anynorm on

R

2

satisfying Proposition 1. We obtainthe

followingversionoftheBana h-Stonetheoremforanefun tions(Formoreinformationabout

theBana h-Stonetheoremsee[2℄and[3℄).

Proposition4 Let

X

and

Y

betwo Bana h spa es. Then thefollowingassertionsare equiva-lent.

(1) (Af f (B

X

), N )

and

(Af f (B

Y

), N )

areisometri ally isomorphi .

(8)

Proof. Let

N

˜

be the norm on

Af f

0

(B

X

) × R

dened by

N

˜

(f

0

, t) := N

R

2

(kf

0

k

,

|t|)

for all

(f

0

, t) ∈ Af f

0

(B

X

) × R

. Letus onsiderthemap,

χ

: (Af f (B

X

), N ) → (Af f

0

(B

X

) × R, ˜

N

)

f

7→ (f − f (0), f (0))

Clearly,

χ

isanisometri isomorphism. ThususingTheorem1wehavethat

(Af f (B

X

), N )

and

(Af f (B

X

), N )

areisometri allyisomorphi ifandonlyif

(Af f

0

(B

X

), k.k

)

and

(Af f

0

(B

Y

), k.k

)

areisometri allyisomorphi ,whi hisequivalentbyTheorem3tothefa t that

(X, k.k

X

)

and

(Y, k.k

Y

)

areisometri allyisomorphi .

A knowledgments. Theauthor thanksProfessorJean-BernardBaillon andProfessorGilles

Godefroyforthediverseenri hingdis ussions.

Referen es

[1℄ M.Ba hir,Theinf- onvolutionasalawofmonoid.AnanaloguetoBana h-Stonetheorem

J.Math. Anal.Appl420,(2014)No.1,145-166.

[2℄ M. Ba hir, Surla dierentiabilité génériqueetle théorèmede Bana h-Stone C.R.A ad.

S i. Paris,330(2000),687-690.

[3℄ M.I. Garrido,J.A.Jaramillo, Variationsonthe Bana h-Stone Theorem,Extra taMath.

Vol.17,N. 3,(2002)351-383.

[4℄ R.Phelps,Le turesonChoquet'sTheorem,Se ondEdition,Le tureNotesinMath.Berlin,

(1997).

[5℄ P.Halmos,AHilbert Spa eProblem Book,Springer-Verlag,NewYork-Heidelberg-Berlin,

Références

Documents relatifs

To test whether the vesicular pool of Atat1 promotes the acetyl- ation of -tubulin in MTs, we isolated subcellular fractions from newborn mouse cortices and then assessed

Néanmoins, la dualité des acides (Lewis et Bronsted) est un système dispendieux, dont le recyclage est une opération complexe et par conséquent difficilement applicable à

Cette mutation familiale du gène MME est une substitution d’une base guanine par une base adenine sur le chromosome 3q25.2, ce qui induit un remplacement d’un acide aminé cystéine

En ouvrant cette page avec Netscape composer, vous verrez que le cadre prévu pour accueillir le panoramique a une taille déterminée, choisie par les concepteurs des hyperpaysages

Chaque séance durera deux heures, mais dans la seconde, seule la première heure sera consacrée à l'expérimentation décrite ici ; durant la seconde, les élèves travailleront sur

A time-varying respiratory elastance model is developed with a negative elastic component (E demand ), to describe the driving pressure generated during a patient initiated

The aim of this study was to assess, in three experimental fields representative of the various topoclimatological zones of Luxembourg, the impact of timing of fungicide

Attention to a relation ontology [...] refocuses security discourses to better reflect and appreciate three forms of interconnection that are not sufficiently attended to