• 検索結果がありません。

ofp-Hyponormal Powers

N/A
N/A
Protected

Academic year: 2022

シェア "ofp-Hyponormal Powers"

Copied!
6
0
0

読み込み中.... (全文を見る)

全文

(1)

Photocopying permitted bylicenseonly theGordon and Breach Science Publishersimprint.

Printed in Malaysia.

Powers ofp-Hyponormal Operators

ARIYADASA ALUTHGEaand DERMINGWANG

aDepartmentofMathematics,Marshall University, Huntington, VVV25-755-2560,USA,"bDepartmentof Mathematics, CaliforniaStateUniversity,LongBeach,CA90840-1001,USA (Received20 November 1997; Revised 9 July1998)

ApplyingFuruta’s andHansen’sinequalities, it is shownthatifTis ap-hyponormal operator, thenT is(p/n)-hyponormal. Applicationsare obtained.

Keywords: p-Hyponormal operator;Furuta’sinequality;Hansen’sinequality 1991AMS SubjectClassifications: Primary47B20, 47A30

1

INTRODUCTION

Let H be a complex Hilbert space and

L(H)

be the algebra of bounded linear operators on H.

An

operator TE

L(H)

is said to be p-hyponormal,p

>

0,if

(T’T)

p

>_ (TT*) p. A

p-hyponormaloperatoris said to be hyponormal if p 1; semi-hyponormal if p

1/2.

The well

known L6wner-Heinz inequality implies that every p-hyponormal operator is q-hyponormal for any 0

<

q

<p.

Hyponormal operators have beenstudiedbymanyauthors,such as Halmos

[7],

Stampfli

[10,11]

andXia

[13].

Semi-hyponormalitywasintroducedbyXia

[12].

See

[13]

for properties of semi-hyponormal operators. For p-hyponormal operators,see

[1,2].

Throughoutthis paperweassume0

<

p

<

and use acapital letter to denote an operator in

L(H).

In

[7,

Problem 164], Halmos gave an example of a hyponormal operator A whose square A2 is not

*Correspondingauthor.

279

(2)

hyponormal.

Here

weuse

Furuta’s [5]

and

Hansen’s [6]

inequalitiesto show that ifTisp-hyponormal, then T2is(p/2)-hyponormal.

In

fact, wewillshow that for any positive integer n, the operator Tn is (p/n)- hyponormal. Applicationsofourresultarealso obtained.

2

THE RESULT

LEMMA (Furuta’s

inequality

[5]) If

A

>_

B

>_

0, then the inequalities

(BrAPBr)

1/q B(p+2r)/q

and

A(p+2r)/q

>_ (ArBPAr)

1/q

hold

for

p,r

>_

0,q

_>

with

(1 +

2r)q

>_

p

+

2r.

LEMMA

2

(Hansen’s

inequality

[6]) IfA >_

0and

[[B][ _<

1, then

(B*AB)

p

>_ B*APB

forO<_p< 1.

THEOREM Let Tbeap-hyponormaloperator. Theinequalities

(T

n*

Tn)P/n >_ (T’T)

p

>_ (TT*)

p

>_ (TnTn*)

p/n

hold

for

allpositive integern.

Proof

Let T=

UITI

be the polar decomposition of T. For each positive integer n, let

An=(Tn*Tn)

p/n and

Bn=(TnTn*)

p/n. We will useinductiontoestablishthe inequalities

An >_ A1 >_ B1 >_ Bn. (1)

The inequalities

(1)

clearlyhold forn 1.

Assume (1)

holdforn k.

The induction hypothesis and the assumption that Tisp-hyponormal imply

U*AkU >_ U’A1

U

>_ A1.

(3)

Let

Ck u*Akk/P U) p/k. Hansen’s

inequalityimplies

Ck >_ U*AkU >_

A

.

Thus

Ak+l (T

*+

Tk+l)

(p/k+l)

T*

T*kT

k) T)

(p/k+l)

([TIU*Akk/PUIT[)

(p/k+l)

(A1/2prk/p

""k

All/2p)(p/k+I)

>_A

by

Furuta’s

inequality.

On

the other hand, the induction hypothesis implies

Bk

<_B1 <_A1.

Thus

where the inequality follows from

Furuta’s

inequality.Therefore,

Ak+l _ A1 >_ B1 >_ Bk+l

and hence, by induction, inequalities

(1)

hold forn

>

1. The proofis complete.

COROLLARY

If

the operator Tis p-hyponormal, then

T"

is (p/n)- hyponormal.

Concrete examplesofnon-hyponormal p-hyponormaloperatorsare hard to come by.

In [12],

Xia gave an example ofa singularintegral operator which is semi-hyponormal but nothyponormal. Corollary allowsustogive anotherexampleofasemi-hyponormal operator which

(4)

is nothyponormal. Let Abe the operatorinHalmos’[7, Problem

164].

Thus,

A

ishyponormalbutA2is nothyponormal.

By

Corollary 1,A2is semi-hyponormal.

Moreover,

Anis(1/2n)-hyponormal.

3

APPLICATIONS

In [10, Theorem

5],

StampfliprovedthatifTishyponormaland Tnis normal forsomepositive integern,then Tisnormal. Stampfli’s result had been extended by Ando

[3]

to the case where Tis paranormal.

Althoughnot as broad asAndo’s extension, Theorem caneasily be usedtoextend Stampfli’s resulttop-hyponormaloperatorsasfollows.

COIOILARY 2 Let the operator Tbep-hyponormal.

If

T isnormal,

then Tisnormal.

Proof By

Theorem andthe assumption that

T"

isnormal,

Whence

T’T-

TT*.Theproofiscomplete.

In

[9,

Theorem

7],

Putnam provedthatifTishyponormal,andr

>

0 is such that r

Er(T*T),

then there is a z

Ea(T)

such that

Izl

=r.

Recently, Ch6 and Itoh

[4,

Theorem

4]

generalized

Putnam’s

result to thecase where the operator Tisp-hyponormal. Theorem can be utilizedtogiveageneralization of the result of Ch6 and Itohasfollows.

THEOIEM 2. Let Tbe ap-hyponormal operator and n be apositive integer.

If

r

>_

0 is such that r2

cr(T

n*

Tn),

then there is a z

or(T)

such that

Izl"-r.

Proof

Theorem implies T is (p/n)-hyponormal. Therefore, by

[4,

Theorem

4],

there is a w

or(T")

such that

Iwl

=r. Since

{z":

z

a(T)},

thereis a z E

or(T)

suchthatz

"-

w. Clearly

Izl"-

rand

theproofiscomplete.

As

an extension of the well-known Putnam’s area inequality for hyponormal operators [8], Xia [13, Theorem

XI.5.1]

proved the following Theorem 3 for the case in which Tis p-hyponormal with p

_> 1/2

andn 1.In [4,Theorem

5],

Ch6 andItoh extendedXia’sresult

top-hyponormaloperators with 0

<

p

< 1/2.

(5)

THEOrtEM 3 Let T be p-hyponormal.

If or(T)

C_

{rei:

0

<

0

<

2Trim}

for

some positive integerm, then

ii(Tn, zn)p/n (TnTn.)p/nll <_ nPTr JJ

(r)

p2p-1

dpdO

for

positive integers n

<

m.

Proof By

Theorem 1, Tn is (p/n)-hyponormal. It follows from [4,Theorem

5]

that

I](Tn*Zn)p/n (ZnZ"*)P/nll <- PnTf JY"

r2(p/n)-I drdqS.

o’(Z

Since

cr(T {pneinO:

peiOE

or(T)),

the result follows by the substitu- tionsr

pn

and

b

nO.

Acknowledgment

Thispaperwaswritten whilethefirstauthor, onsabbaticalleave from Marshall University,was avisiting adjunctprofessoratthe

Department

of Mathematics, CaliforniaState University,

Long

Beach, Fall, 1997.

Hewishes toexpresshisdeep gratitudetohis host forwarm hospitality and support.

References

[1] A.Aluthge, On p-hyponormal operatorsfor 0<p<1,Integr.Equat. Oper. Th.,13 (1990),307-315.

[2] A.Aluthge, Some generalized theoremsonp-hyponormal operators,Integr.Equat.

Oper.Th., 24(1996),497-501.

[3] T. Ando,Operatorswith a normcondition,ActaSci.Math.;33(1972),169-178.

[4] M.Ch6 andM.Itoh,Putnam’sinequality forp-hyponormal operators,Proc. Amer.

Math.Soc.,123(1995),2435-2440.

[5] T. Furuta, A>B>0 assures (BrAPBr)TM>_B(p+2r)/q for r>0, p>0, q> with (1

+

2r)q>p+2r, Proc.Amer.Math.Soc.,101(1987),85-88.

[6] F.Hansen, Anoperator inequality,Math.Ann.,246(1980),249-250.

[7] P.R.Halmos,AHilbertSpaceProblemBook,VanNostrand, Princeton,NewJersey, 1967.

[8] C.R.Putnam, Aninequality forthe areaofhyponormal spectra, Math.Z.,116(1970), 323-330.

[9] C.R. Putnam, Spectraofpolar factors of hyponormal operators, Trans.Amer.Math.

Soc.,188(1974),419-428.

(6)

[10] J.G.Stampfli,Hyponormal operators,PacificJ.Math., 12(1962),1453-1458.

[11] J.G.Stampfli, Hyponormal operators andspectral density, Trans. Amer. Math.Soc., 117(1965),469-476.

[12] D.Xia, On the nonnormal operators-semihyponormal operators,Sci. Sininca,23 (1980),700-713.

[13] D.Xia, Spectral TheoryofHyponormalOperators,BirkhiuserVerlag,Boston,1983.

参照

関連したドキュメント

In [5], Qi studied a very interesting integral inequality and proved the following result Theorem 1... Now consider the quotient of both sides

In [5], Qi studied a very interesting integral inequality and proved the following result..

This is the well-known Hahn-Banach theorem, that is, the extension theorem for bounded lin- ear functionals on normed linear spaces.. The following theorem is Hahn’s result

Recently, Chen and Kimball [1], studied a very interesting Qi-type integral inequality and proved the following result..

Rao [4] developed the Hajek-Renyi inequality for associated sequences and proved the following theorem... 3196 SLLN for associated

Recently, Chen and Kimball [1], studied a very interesting Qi-type integral inequality and proved the following result..

Yamazaki [18] showed that “If $T$ is an invertible log-hyponormal operator, then $T^{n}$.. is $al_{8}o$ log-hyponormal for any positive integer $n$ , “ and also he

But until now, we have not obtained any property of Aluthge transform of a class A operator which is a weaker class than the class of $w$ -hyponormal operators, so it was