• 検索結果がありません。

AN APPLICATION OF ALUTHGE TRANSFORM TO PUTNAM INEQUALITY FOR LOG-HYPONORMAL OPERATORS (Operator Inequalities and related topics)

N/A
N/A
Protected

Academic year: 2021

シェア "AN APPLICATION OF ALUTHGE TRANSFORM TO PUTNAM INEQUALITY FOR LOG-HYPONORMAL OPERATORS (Operator Inequalities and related topics)"

Copied!
4
0
0

読み込み中.... (全文を見る)

全文

(1)

AN APPLICATION OF ALUTHGE TRANSFORM TO

PUTNAM INEQUALITY FOR LOG-HYPONORMAL OPERATORS

$\mathrm{M}\wedge \mathrm{S}\Lambda \mathrm{T}\mathrm{o}\mathrm{s}\mathrm{I}\mathrm{I}\mathrm{I}\Gamma^{\mathrm{I}}\mathrm{U}\mathrm{J}[\mathrm{I}$

藤井正俊

A$\mathrm{D}\mathrm{S}\uparrow \mathrm{R}\mathrm{A}\mathrm{C}\mathrm{T}$. In this note, we give a shorl proof t,o the Putnam inequality for log-hyponormal operators due to Tanahashi: If7’ is $\mathrm{i}\mathrm{n}\mathrm{v}\mathrm{o}\mathrm{l}\mathrm{t}.\mathrm{i}\mathrm{b}[\mathrm{e}$ and $\log$-hyponormal, i.e., $\log T^{\cdot}T\geq\log TT$ , $\mathrm{t}$hen

$|| \log?^{\mathrm{s}_{7’-}}’\log T\tau^{\mathrm{s}}||\leq\frac{1}{\pi}\int\int_{\sigma(T)}T^{-\mathrm{l}}drd\theta$,

where $\sigma(T)$ is $\mathrm{t}$he spectrum

or

$T$. I$\mathrm{t}$, is based on his original idea that. the log-hyponormality

is regarded as $0$-liyponorrnalil,$\mathrm{y}$.

1. Introduction. After $\mathrm{t},1\mathrm{l}\mathrm{e}$ Furuta

$\mathrm{i}\mathrm{n}\mathrm{e}\mathrm{q}\mathrm{u}_{\mathrm{c}}\backslash 1\mathrm{i}\{,\mathrm{y}$ was originated by Furuta [12],

see

also

[5,13,16,18], we $\mathrm{i}\mathrm{n}\mathrm{i}\mathrm{t}\mathrm{i}_{\mathrm{e}}\gamma$t,ed to study it under {,he chaotic order $A>>B$, i.e., $\log A\geq\log B$, for

positiveinverl,ible operat,ors $\Lambda_{\mathrm{c}}’\iota 11\mathrm{d}f\mathit{3}[11]$. Wefinallycllaracterized $A>>B$ by

a

Furuta-type inequalit,$\mathrm{y}[6](\urcorner \mathrm{n}\mathrm{d}[7,81$: For $A,$$D>0,$ $\Lambda>>\mathcal{B}\mathrm{i}[’\iota\subset \mathrm{n}\mathrm{d}$ only if

(1) $(\Lambda^{\gamma}f\mathit{3}\mathrm{p}\Lambda^{\mathcal{T}})^{\frac{2r}{\rho\{2r}}\leq A^{2r}$ $11\mathrm{o}\mathrm{l}\mathrm{e}1_{\mathrm{S}}$ for all

$p,$$\uparrow\cdot\geq 0$.

Furt hermore we $\mathrm{i}\mathrm{n}\mathrm{t}\mathrm{e}\mathrm{r}\mathrm{p}\mathrm{o}\mathrm{l}\mathrm{t}\gamma\iota,\mathrm{e}\mathrm{d}$ between $\dagger,1$

)$\mathrm{e}$ Furuta inequality and Theorem A as follows

[9.10]:

Tlleorem A. $\Gamma’or$ a

fixed

$\delta>0,$ $A^{\delta}\geq B^{\delta}$

for

$\Lambda,$ $B\geq 0$

if

and only $i,f$

for

each $r\geq 0$

(2) $\Lambda^{\mathrm{L}_{\frac{+2r}{\eta}}}\geq(\Lambda^{r}B^{\rho}A^{r})^{\frac{1}{q}}$ holds

for

$p\geq 0$ and $q\geq 1$ with

(.3) $(\delta-\vdash 2?\cdot)q\geq|^{y\dashv-2}r$.

$1l’\mathrm{e}$ nole lhat (1) is $\mathrm{e}\mathrm{q}_{11\mathrm{i}\mathrm{V}_{\mathrm{C}}\gamma}1_{\mathrm{C}\mathrm{n}}\{,$ $\mathrm{t},0$ the case $\delta=0$ in Theorem A and the Furuta inequality

is.

$|_{11\mathrm{S}}\mathrm{t}$ lhe case $\delta=1$. The ($1_{\mathrm{o}\mathrm{m}_{\mathrm{c}}\urcorner}\mathrm{i}\cap$ given by (3) is explained by the figure below: 数理解析研究所講究録

(2)

From the viewpoint of this, Tanahashi [19] defined the $\log$-hyponormality for invertible

operators by $|T|>>|T^{*}|$, where $|X|$ is the squere root of$X^{*}X$, and constructed the Putnam

inequality for $\log$-hyponormal operators:

Theorem T.

If

$T$ is

an

invertible $log$-hyponormal operator, $i.e.,$ $\log T*T-\log\tau\tau*\geq 0$,

then

(4) $|| \log\tau^{*\tau}-\log T\tau^{*}||\leq\frac{1}{\pi}\int\int_{\sigma(T)}r-1drd\theta$,

where $\sigma(T)$ is the spectrum

of

$T$.

It was conjectured from the Putnam inequality for p–hyponormal operators by

Cho

and Itoh [3]:

If$T$ is a $\mathrm{p}$-hyponorlnal operator, i.e., $(T^{*}T)^{p}-(TT*)^{\mathrm{P}}\geq 0$, then

(5) $||( \tau*\tau)^{p}-(T\tau*)^{p}||\leq\frac{p}{\pi}\int\int_{\sigma(}\tau)r^{2p-}1drd\theta$ . As

a

matter offact, he understood (5)

as

follows:

(6) $|| \frac{(T^{*}T)p-(T\tau^{*})^{p}}{p}||\leq\frac{1}{\pi}\int\int_{\sigma(\tau)}r^{2_{\mathrm{P}}1}-drd\theta$.

By taking $parrow\infty$, he constructed Theorem $\mathrm{T}$ and proved it by the idea developed in (5).

Thepurpose ofthisnoteis to continuehisconsideration directly. That is,

we

here propose a straightand simple proofof Theorem$\mathrm{T}$which might be along with his

$\mathrm{i}\mathrm{n}\mathrm{t}\mathrm{e}\mathrm{n}\mathrm{t}\mathrm{i}_{0}\mathrm{n};\mathrm{T}\mathrm{a}\mathrm{n}\mathrm{a}\mathrm{h}\mathrm{a}\mathrm{S}\mathrm{h}\mathrm{i}$

might regards the $\log$-hyponormality

as

the $0$-hyponormality.

Our

tool in this note is the

Aluthge transform which is

grown

up the p–hyponormality,

see

[1,2,4,14,15,20]. 2. Preliminary.

For the sakeofconvenience,

we

cite the following characterization of chaotic orderwhich implies (1) by the help ofthe Furuta inequality [7],

see

also [8] and [9].

Theorem B. For $A,$ $B>0,$ $A>>B,$ $i.e.,$ $\log A\geq\log B$,

if

and only

if for

any $\delta\in(0,1]$

there exists

an

$\alpha=\alpha_{\delta}>0$ such that

$(e^{\delta}A)^{\alpha}>B^{\alpha}$

.

The essential part of Theorem $\mathrm{B}$ is

as follows:

If $A$ and $B$

are

selfadjoint and $A>B$ ,

then there exists an $\alpha\in(0,1]$ such that

$(^{*})$ $e^{\alpha A}>e^{\alpha B}$

.

It has the following simple proof: The assumption $A>B$

means

that $A-B\geq\epsilon>0$ for

some $\epsilon$. We here take $0<\alpha<\epsilon/(e^{||A||}+e^{||B||})$ and $\alpha\leq 1$

.

Then we have

$e^{\alpha A}-e^{\alpha B}= \alpha(A-B)+\sum_{n=2}\frac{\alpha^{n}}{n!}(A\infty)n-B^{n}$

$\geq\alpha\epsilon+\alpha^{2}\sum_{=n2}^{\infty}\frac{\alpha^{n-2}}{n!}(A^{n}-B^{n})$

$\geq\alpha\epsilon-\alpha|2|\sum_{n}\infty=2\frac{\alpha^{n-2}}{n!}(A^{n}-Bn)||$

$\geq\alpha\epsilon-\alpha^{2}n=\sum^{\infty}\frac{1}{n!}(||A||^{n}+||B||^{n})2$

$\geq\alpha(\epsilon-\alpha(e^{||A||}+e^{1})|B||)>0$.

(3)

Here we should note an interesting characterization of chaotic order recently obtained by Yamazaki and Yanagida [22], which is associated with Kantorovich inequality and

con-sequently Specht’s ratio,

see

[23]. 3. Proof.

We begin with the Putnam inequality for $\mathrm{p}$-hyponormal operators; we turn it into the

following lemma viathe L\"owner-Heinz inequality:

Lemma 1.

If

$T$ is a$q$-hyponormal operator, then$T$

satisfies

(6)

for

all $0<p\leq q$

.

Conse-quently (4) holds

for

any $q- h\uparrow Jponomlal$ operators $T$.

Thesecond halfis

ensur.ed.by

the fact that$\mathrm{u}-\lim_{tarrow 0}\frac{A^{t}-1}{t}=\log$ $A$for a positive invertible operator $A$.

Thus Lemma 1 suggests us to find

a

family $\{T_{q}; q>0\}$ of$q$-hyponormal operators such

that $||T_{q}-T||arrow 0$

as

$qarrow \mathrm{O}$ for a given $\log$-hyponormal operator $T$

.

In this situation, the Aluthge transform completely responds to

our

demand. As a matter offact, Tanahashi prepared the following result in [20]:

Lemma 2.

If

$T$ is an invertible $log$-hyponormal operator with the polar decomposition$T=$ $U|T|$, then the Aluthge

transform

$\tilde{T}=|T|^{q}U|T|^{1-q}$ is $q$-hyponormal

for

$0<q \leq\frac{1}{2}$

.

To prove Theorem $\mathrm{T}$, we take $T_{q}=|T|^{q}U|T|^{1-q}$ for $0<q< \frac{1}{2}$. Since $\sigma(T_{q})=\sigma(T)$ for

all $q$, it is

compiete.

We finally give a short proof to Lemma 2 via Theorem A Putting $p=2q$ and

$r=1-q$

in (1), we have

$(T_{q}^{*}\tau_{q})1-q=(|T|^{1-q}U^{*}|T|2qU|T|^{1q}-)1-q$

$=U^{*}(U|T|^{1-}qU^{*}|T|^{2q}U|\tau|^{1-q}U*)^{1}-qU$

$=U^{*}(|\tau*|1-q|\tau|2q|\tau^{*}|^{1-}q)^{1q}-U$

$\geq U^{*}|T^{*}|^{2(-}1q)U$

$=|T|^{2(}1-q)$,

so that $(T_{q}^{*}Tq)2q\geq|T|^{2q}$ viathe L\"owner-Heinz inequality by $1-q\geq q$.

On

the other hand,

we have also

$(T_{q}T_{q}^{*})^{q}=(|T|^{q}U|T|^{2}(1-q)U*|T|^{q})^{q}$

$=(|T|^{q}|\tau*|^{2(}1-q)|\tau|^{q})^{q}$

$\leq|T|^{2q}$

.

Therefore it follows that

$(\tau_{q}^{*}T_{q})^{q}\geq|T|^{2q}\geq(T_{q}T_{q}^{*})^{q}$, as desired.

Remark. (1) (5) is obtained by Putnam [17] for $p=1$, Xia [21] for $\frac{1}{2}\leq p<1$ and

Cho-Itoh [3] for $0<p< \frac{1}{2}$

.

(2) The proof ofLemma 2 is available to show Tanahashi’s result [20; Theorem 4]. Acknowledgement. The authors would like to express their thanks to Prof. K.Tana-hashi for his fascinating talk in the meeting of the Japan Mathematical Society at March 26, 1998. He also thanks to Prof. E.Kamei for his critical discussion.

(4)

REFERENCES

1. A.Aluthge, On$p$-hyponormal operatorsfor$0<p<1$, Integr. Equ. Oper. Theory, 13 (1990), 307-315.

2. A.Aluthge, Some generalized theorems on $p$-hyponormal operators, Integr. Equ. Oper. Theory, 24

(1996), 497-501.

3. M.Cho and M.Itoh, Putnam’s inequalityfor$p$-hyponomal operators, Proe. Amer. Math. Soc., (19).

4. $\mathrm{B}.\mathrm{P}$.Duggal, On the spectrum of$p$-hyponormal operators, Acta Sci. Math. (Szeged), 63 (1997),623-637.

5. M.Fujii, Furuta’s inequality and itsmean theoretic approach, J. Operator theory, 23 (1990), 67-72.

6. M.Fujii,T.Furuta andE.Kamei, Operatorfunctions associated with Furuta’sinequality, Linear Alg. and

its Appl., 179 (1993), 161-169.

7. M.Fujii, J.-F. Jiang and E.Kamei, Characterization of chaotic order and its application to Furuta

inequaity, Proc. Amer. Math. Soc., 125 (1997), 3655-3658.

8. M.Fujii, J.-F. Jiang and E.Kamei, Characterization ofchaotic order and its applications to $\Pi_{4ru}ta’ s$

type operator inequalities, Linear Multilinear Alg., 43 (1998), 339-349.

9. M.Fujii,J.-F.JiangandE.Kamei, A characterizationoforders definedby$A^{\delta}\geq B^{\delta}$ via Fbruta inequaity,

Math. Japon., 45 (1997), 519-525.

10. M.Fujii, J.-F. Jiang, E.Karnei andK.Tanahashi, A characterization ofchaotic order and a problem, J.

lneq. Appl., to appear.

11. M.Fujii and E.Kamei, Furuta’s inequalityfor the chaotic order, I and II, Math. Japon., 36 (1991), 603-606 and 717-722.

12. T.Furuta, $A\geq B\geq 0$ assures$(B^{r}A^{\mathcal{P}}B^{\Gamma})1/q\geq B^{(}p+2r)/q$for$r\geq 0,$ $p\geq 0,$$q\geq 1$ with $(1+2r)q\geq p+2r$,

Proc. Arner. Math. Soc., 101 (1987), 85-88.

13. T.Furuta, Elementary proofofan order preserving inequality, Proc. Japan Aead., 65 (1989), 126.

14. T.Furuta, Furtherextensions ofAluthgetransformation on$p$-hyponormal operators,Integr. Equ.Oper.

Theory, 29 (1997), 122-125.

15. T.Huruya, A note on$p$-hyponormal operators, Proc. Amer. Math. Soc., 125 (1997), 3617-3624. 16. E.Kamei, A satellite to $F\alpha ru$ta’s inequality, Math. Japon., 33 (1988), 883-886.

17. $\mathrm{C}.\mathrm{R}$.Putnam, An inequalityfor the area ofhyponormal spectra, Math. Z., 116 (1970), 323-330.

18. K.Tanahashi, Best possibility ofthe Furuta inequality, Proc. Amer. Math. Soe., 124 (1996), 141-146. 19. K.Tanahashi, Putnam’s inequalityforlog-h?/ponormaloperators, prep$7\dot{\mathrm{Y}}nt$.

20. K.Tanahashi, On $log$-hyponormal $operat_{\mathit{0}}rsr$ prepnnt.

21. D.Xia, Spectral Theory ofHyponormal Operators, Birkh\"auserVerlag, Basel, 1983.

22. T.Yamazaki and M.Yanagida, Characterizations ofchaotic order associatedwith Kantorovich

inequal-$ity$, Sci, Math., to appear.

23. $\mathrm{J}.\mathrm{I}$.Fujii, T.Furuta, T.Yamazakiand M.Yanagida, Simplified proofofCharacterizations ofchaotic order

via Specht’s ratio, Sci. Math., to appear.

DEPARTMENT OF MATHEMATICS, OSAKA KYOIKU UNIVERSITY, KASHIWARA, OSAKA 582-8582, JAPAN

参照

関連したドキュメント

Recently Verma [21] introduced an iterative scheme characterized as an auxiliary variational inequality type of algorithm and applied to the approximation-solvability of a class

YANAGIDA, Powers of class wA(s, t) operators associated with gen- eralized Aluthge transformation, J. YUAN, Extensions of the results on powers of p- hyponormal and

Furuta, Log majorization via an order preserving operator inequality, Linear Algebra Appl.. Furuta, Operator functions on chaotic order involving order preserving operator

Zaslavski, Generic existence of solutions of minimization problems with an increas- ing cost function, to appear in Nonlinear

In [10, 12], it was established the generic existence of solutions of problem (1.2) for certain classes of increasing lower semicontinuous functions f.. Note that the

The following result is useful in providing the best quadrature rule in the class for approximating the integral of a function f : [a, b] → R whose first derivative is

We present a new reversed version of a generalized sharp H¨older’s inequality which is due to Wu and then give a new refinement of H¨older’s inequality.. Moreover, the obtained

Theorem 5 was the first result that really showed that Gorenstein liaison is a theory about divisors on arithmetically Cohen-Macaulay schemes, just as Hartshorne [50] had shown that