• 検索結果がありません。

INVERSE PROBLEM ON ISOMORPHISM THEOREM OF $A^{p}(G)$-ALGEBRAS ${1}\le{p}\le{2}$ (Nonlinear Analysis and Convex Analysis)

N/A
N/A
Protected

Academic year: 2021

シェア "INVERSE PROBLEM ON ISOMORPHISM THEOREM OF $A^{p}(G)$-ALGEBRAS ${1}\le{p}\le{2}$ (Nonlinear Analysis and Convex Analysis)"

Copied!
6
0
0

読み込み中.... (全文を見る)

全文

(1)

Proceedingsof RIMSWorkshop

onNACA2015 Kyoto University

INVERSE PROBLEM ON ISOMORPHISM THEOREM OF

A^{p}(G)

‐ALGEBRAS 1\leq p\leq 2

CHENG‐TELIU,JIN‐CHIRNGLEEAND HANG‐CHIN LAI*

Abstract

Let

G_{1}

and

G_{2}

be

locally

compact Hausdorff groups,

E(G_{1})

and

E(G_{2})

the function

spaces

(Banach

algebras

orBanach

spaces)

on

G_{1}

and

G_{2}

respectively.

Then it isknown

that if

G_{1}\simeq G_{2}

,

implies

E(G_{1})

and

E(G_{2})

are

isomorphic. Naturally,

aninverse

problem

arises that

(P)

Whether an

algebaic isomorphism

$\Phi$ :

E(G_{1})\rightarrow E(G_{2})

could deduce

G_{1}\simeq G_{2}

?

In this paper, wewould solve Problem

(P)

for

A^{p}(G)

‐algebras,

1\leq p\leq 2.

1. PRELIMINARIES

(1)

1948,

Y. Kawada

[7]

solved this

problem

under

bipositive isomorphism

$\Phi$:L^{1}(G_{1})\rightarrow L^{1}(G_{2})

.

(2)

1952,

Wendel

[13]

proved

(P)

under the

isomorphism

$\Phi$ from the

algebra

L^{1}(G_{1})

onto

L^{1}(G_{2})

by assuming

$\Phi$ is a norm

nonincreasing.

(3)

1965,

Edwards

[2]

considered the groups

G_{i}(i=1,2)

are

compact,

and if there exists a

bipositive isomorphism

of

IP(G_{1})

onto

L^{p}(G_{2})

to

get,then

G_{1}\simeq G_{2}

. He

asked whether the compact groups

G_{1}

and

G_{2}

are

necessarily homeomorphic,

if

bipositive

is

replaced by isometry?

(4)

1966,

The affirmative answer to this

question

in

[2]

by

positive

replaced

isometry

was

given

by

Strichartz

[12].

(5)

1968,

Further,

Parrott

[11]

proved

the

question

inEdwards

[2]

for

general locally

compact

groups

G_{1}

and

G_{2}

if there is an isomertictransformation of

L^{p}(G_{1})

onto

IP(G_{2})(1\leq p<\infty, p\neq 2)

.

(2)

CHENG‐TELIU,JIN‐CHIRNG LEE AND HANG‐CHIN LAI*

(6)

1973,

Lai/Lien[10]

solved the

problem

(P)

by

assume that if there exists an

injective bipositive

linear

mapping

from the Banach space

L^{p}(G_{1})

onto Banach

space

L^{p}(G_{2})

,then

G_{1}\simeq G_{2}

is deduced.

(7)

Someother

isomorphism problems

weresolved

by

Johnson

[6],

Gaudry

[4]

and

Figa

Talamanca

[3]

in different view

points.

(8)

In this articlewewould solve

problem

(P)

onthe Banach

algebra

A^{\mathrm{p}}(G)

,

1\leq p\leq 2.

2.

A^{p}(G)

—ALGEBRAS,

1\leq p<\infty

In thispaper,wewould consider the

isomorphism

theorem for

A^{p}(G)

—algebras.

Let G bea LCA group with dual group

\hat{G}

. The space

A^{p}(G)

is defined

by

A^{p}(G)= {

f\in L^{1}(G)

; Fouriertransform

\hat{f}\in L^{p}(\hat{G})

},

1\leq p<\infty

.

(1)

Then

A^{p}(G)

isa commutativeBanach

algebra

under convolution

product

with thenorm

given by

\Vert f\Vert^{p}=\Vert f\Vert_{1}+\Vert\hat{f}\Vert_{\mathrm{p}}

, foreach p,

1\leq p<\infty

for

f\in A^{p}(G)

.

(2)

Thenorm

\Vert\Vert^{p}

is

equivalent

to\displaystyle \max

f\Vert_{1},

\Vert\hat{f}\Vert_{p}

).

Since

f\in A^{p}(G)\Rightarrow\hat{f}\in L^{P}(\hat{G})\cap C_{0}(\hat{G})

, thus

\hat{f}\in L^{r}(\hat{G})

for

r>p>1

, but

such

f\not\in A^{r}

for

1\leq p\leq 2\leq q<r<\infty.

By

this

fact,

weknow that

A^{p}(G)

cannot include all Fourier transforms of

C_{\mathrm{c}}(G)\cap A^{r}.

And

A^{1}(G)\supset A^{p}(G)\supset A^{2}(G)\supset A^{q}(G)\supset C_{0}(G)

, where

A^{1}(G)=\displaystyle \bigcup_{1<p\leq 2}A^{p}(G)

is the

closure of such union sets.

We then conclude that

1\leq p\leq 2,

C_{c}\cap A^{p}(G)

isdense in

A^{p}(G)

with

respect

tothe A^{p}‐norm.

Thus,

if

f\in A^{p}(G)

, then

\hat{f}\in L^{p}(\hat{G})

and

\hat{f}\in L^{q}(\hat{G})

for

p\leq 2<q,

f\not\in A^{q}(G)

,

andso

\forall p,

1\leq p\leq 2\leq q<\infty,

\displaystyle \frac{1}{p}+\frac{1}{\mathrm{q}}=1,

A^{p}(G)\cap A^{q}(G)=\emptyset.

Hence the indexp,

only

taken in the interval

1\leq p\leq 2

could

get

T(A^{p})\subset A^{p}

by

a

(3)

1\leq p\leq 2

which could

get

T(A^{p})\subset A^{p}

. Therefore in later

part,

all

A^{p}(G)

we discuss will

take

1\leq p\leq 2.

3. MULTIPLIERS OF

A^{p}(G)

A

multiplier

T of

$\Lambda$^{p}(G)

isacontinuous linear

mapping

of

A^{p}(G)

,

1\leq p\leq 2

into

itself,

such that

T(f*g)=T(f)*g=f*T(g)

, for all

f,

g\in A^{p}(G)

.

In order to solve

problem

(P)

on

A^{p}(G)

‐algebras.

We use a

technique by passing

the

multiplier

of

A^{p}(G)

, thus we

subscrip

the definition of

A^{p}(G)

, as follows. Let

\mathfrak{L}(A^{p})

be

thespaceof all bounded linearoperatorof

A^{p}(G)

,

1\leq p\leq 2

intoitself.

Definition 1. An

operator

T\in \mathfrak{L}(A^{p}(G))

is saidto be a

multiplier of

A^{p}(G)

if

T(f*g)=Tf*g=f*Tg

for f,

g\in A^{\mathrm{p}}(G)

.

(3)

The

concept

of

multiplier

T,one canconsult

\mathrm{L}\mathrm{a}\mathrm{i}/\mathrm{L}\mathrm{e}\mathrm{e}/

Liu

[9,

Theorem

1.1].

It deduces

thespace

\mathfrak{M}(A^{\mathrm{p}})

of

multipliers

of

A^{p}(G)

is

isometrically isomorphic

to

M(G)

, the space

of all

regular

measures of G, that is

\mathfrak{M}(A^{p})\cong \mathfrak{M}(L^{1})\cong M(G) , 1\leq p\leq 2

.

(4)

On the other

hand,

it is known that

A^{p}(G)

is essential

L^{1}(G)

‐module,

since

L^{1}(G)

has bounded

approximate

identity

ofnorm 1

[9,

Theorem

2.1].

It is remarkable that

A^{p}(G)

hasnoA^{p}‐uniform bounded

approximate

identity

[8, p.574].

A^{p}*L^{1}=A^{p}

, and

||f*g||^{p}\leq||f||^{p}||g||

, for

f\in A^{p},

g\in L^{1}

.

(5)

Thus the space

\mathfrak{M}(A^{p}, L^{1})

of

multiplier

A^{p} into

L^{1}

is identical to

\mathfrak{M}(A^{p})

. Hence there

exists a

unique

$\mu$\in M(G)

such that

Tf= $\mu$*f

for all

f\in A^{p}(G)

(6)

forany

T\in \mathfrak{M}(A^{p}, L^{1})\cong \mathfrak{M}(A^{p})

.

By

the

property

of

A^{p}(G)

‐algebras,

wewill show the

(4)

CHENG‐TELIU,JIN‐CHIRNG LEE AND HANG‐CHIN LAI*

Theorem 2. Let

G_{1}

and

G_{2}

be

locally

compact

abeliangroupsand $\Phi$ an

algebaic

isomor‐

phism of

A^{P}(G_{1})

onto

A^{p}(G_{2})

,

1\leq p\leq 2

.

Suppose

thatone

of

\hat{G_{1}}

and

\hat{G_{2}}

is

connected,

then $\Phi$ induces a

topological isomorphism

$\tau$

carrying G_{2}

onto

G_{1}

.

Furthermore,

$\Phi$ f(x)=c\hat{x}(x)f( $\tau$ x)

for

f\in A^{p}(G_{1})

, and

x\in G_{2},

where

\hat{x}(x)

is a

fixed

character on

G_{2}

andc a constant

depending only

on the choice

of

Haarmeasurein

G_{2}.

Outline of the

proof

for the main Theorem is

given

as follows: Since the

isomorphism

$\Phi$:A^{p}(G_{1})\rightarrow^{onto}A^{p}(G_{2})

,

\Rightarrow $\Phi$ mapsthe Maximal idealspaces

\mathfrak{M}at(A^{p}(G_{1}))

of

A^{p}(G_{1})

onto

\mathfrak{M}a\mathfrak{x}(A^{p}(G_{2}))

of

A^{p}(G_{2})

,

\Rightarrow $\Phi$ :

\mathfrak{M}at(A^{p}(G_{1}))\rightarrow \mathfrak{M}a\mathfrak{x}(A^{p}(G_{2}))

\Vert

1

(7)

= $\Phi$ :

\hat{G_{1}}

onto

\hat{G_{2}}

Thereasonof

(7)

isthat since

A^{p}(G)

is a

semisimple

commutativeBanach

algebra,

then thespace

\mathfrak{M}\mathfrak{a}\mathfrak{x}(A^{p}(G))

ischaracterized

by

\hat{G}.

Hence ifoneof

\hat{G_{1}}

and

\hat{G_{2}}

is

connected,

then both of

\hat{G_{1}}

and

\hat{G_{2}}

areconnected. There‐ fore

G_{1}

and

G_{2}

arenon‐compact.

Sincethe theorem in

[3]

is

applicable,

wenote that

operator

T commuteswith convo‐ lutionon

A^{p}(G_{1})

is

represented uniquely by

$\mu$\in M(G_{1})

Tf= $\mu$*f=0

for all

f\in A^{p}(G_{1})

\Rightarrow $\mu$=0

.

(8)

Thus wetake

\mathrm{v}\in M(G_{2})

for any

f\in A^{p}(G_{1})

, it candefine this operator

T:A^{p}(G_{1})\rightarrow A^{p}(G_{2})

by

(5)

It is well‐defined

by

(8)

since

A^{p}(G_{1})

is

semisimple,

and

by

Loomis

[book:

p.76

Theorem],

one seesthat $\Phi$ is bicontinuous and hence T isa

multiplier

of

A^{p}(G_{1})

,thus

\exists! $\mu$\in M(G_{1})

such that

$\mu$*f=$\Phi$^{-1}( $\nu$* $\Phi$ f)=Tf.

This $\mu$is

uniquely

determined

by

$\nu$, we definea

mapping

$\Psi$ of

M(G_{2})

into

M(G_{1})

by

$\Psi \nu$*f=$\Phi$^{-1}( $\nu$* $\Phi$ f)

.

It is not hard to prove that $\Psi$ is an

isomorphism

of

M(G_{2})

onto

M(G_{1})

. Since both

measure

algebras

M(G_{1})

and

M(G_{2})

are

semi‐simple

and

commutative,

$\Psi$ is bicontinuous and onecanshow that

$\Psi$|_{A^{p}(G_{2})}

onthe

algebra

A^{p}(G_{2})

is dense in

L^{1}

(G2),

hence

$\Psi$|L^{1}(G_{2})

becomesan

isomorphism

of

L^{1}(G_{2})

onto

L^{1}(G_{1}) [See

Rudin’s book The‐

orem

6.6.4].

Hence

by

Helsen

[5],

the theorem is

complete.

\square

The fullpaperabout

Isomorphism

Theorem of

A^{p}(G)

‐algebras

willappearinelsewhere. REFERENCES

[1] B.Brainerd and R. E.Edwards,Linearoperatorwhichcommuteswithtranslation, The JournalofAustrahan Math.

Soc.,6(1966),289‐327.

[2] R.E. Edwards, Bipositiveand isometricisomorphismof certain convolutionalgebras, Canad. J.Math, 17 (1965), 839‐846.

[3] A.Figa‐Talamancaand G.I.Gaudry, Multipliersandsetsofuniquenessof L^{p},MichiganMath.J.,17(1970),170‐191.

[4] G.I.Gaudry, IsomorphismsofMultipliers algebras,Canad. J. Math.,20(1969),1165‐1172.

[5] H.Helsen, Isomorphismof abeliangroupalgebras,Ark. Math.Bd.,2(1953),475‐487.

[6] B. E.Johnson,Isometricisomorphismsof Measurealgebras,Proc. Amer. Math.Soc.,15 (1964),186188.

[7] Y.Kawada,On the groupringofatopologicalgroup, Math.Japonica.,1(1948),1‐5.

[S] Hang‐Chin Lai,OnsomepropertiesofA^{p}(G)‐algebras,Proc. JapanAcad.,45(1969),572‐576.

[9] Hang‐Chin Lai, Jin‐ChirngLee andCheng‐TeLiu,Multipliersof Banach‐valued functionspacesonLCA group, J. Nonlinear ConvexAnal., 16(2015),1949‐1963.

[10] Hang‐ChinLaiandMing‐Chao Lien, Isomorphasm ofthespacesL^{\mathrm{p}}(G),ChineseJ.Math.1(1973)2,167‐173.

[11] S.K.Parrott,Isometricmultiplier, PacificJ.Math.,5(1968),158‐166.

[12] R. S.Strichartz, Isomorphismofgroupalgebrasp,Proc. Amer. Math. Soc.,17(1966),858‐862. [13] I.G.Wendel,Left centralizers andisomorphismof groupalgebras, PacificJ. Math.,2(1952),251‐261.

(6)

CHENG‐TELIU,JIN‐CHIRNG LEE AND HANG‐CHIN LA1*

[Cheng‐Te Liu]

Department

of

Applied

Mathematics

Chung

YuanChristian

University

Taoyuan

32023,

TAIWAN

E‐‐mail address:

[email protected]

[Jin‐Chirng Lee]

Department

of

Applied

Mathematics

Chung

Yuan Christian

University

Taoyuan

32023,

TAIWAN

E‐‐mail address:

[email protected]

[Hang‐Chin Lai]

Department

of Mathematics

National

Tsing

Hua

University

Hsinchu

30013,

TAIWAN

参照

関連したドキュメント

8.1 In § 8.1 ∼ § 8.3, we give some explicit formulas on the Jacobi functions, which are key to the proof of the Parseval-Plancherel type formula of branching laws of

In this paper, we take some initial steps towards illuminating the (hypothetical) p-adic local Langlands functoriality principle relating Galois representations of a p-adic field L

Greenberg and G.Stevens, p-adic L-functions and p-adic periods of modular forms, Invent.. Greenberg and G.Stevens, On the conjecture of Mazur, Tate and

In addition to the basic facts just stated on existence and uniqueness of solutions for our problems, the analysis of the approximation scheme, based on a minimization of the

In this work, our main purpose is to establish, via minimax methods, new versions of Rolle's Theorem, providing further sufficient conditions to ensure global

Nakanishi, “Exact WKB analysis and cluster algebras II: simple poles, orbifold points, and generalized cluster algebras”, arXiv:1401.7094.. 13

This work was supported by the Open Fund (PLN1003) of State Key Laboratory of Oil and Gas Reservoir Geology and Exploitation (Southwest Petroleum University), the Scientific

A groupoid G is said to be principal if all the isotropy groups are trivial, and a topological groupoid is said to be essentially principal if the points with trivial isotropy