Since UI T I = I T IU holds. Hence T is hyponormal. Hence Quasi normal Hyponormal. Next we show that Quasi normal Subnormal. Define [T] = T*T â TT*. Then ...
OPERATOR THEORY LECTURE NOTES UNIT IV FOR M.SC., MATHEMATICS INVITATION TO LINEAR OPERATORS FROM MATRICES TO BOUNDED LINEAR OPERATORS ON A HILBERT SAPCE BY TAKAYUKI FURUTA Dr. D. SENTHILKUMAR DEPARTMENT OF MATHEMATICS GOVT. ARTS COLLEGE
RELATION AMONG SEVERAL CLASSES OF NON-NORMAL OPERATORS:
2.6.1 PARANORMAL OPERATORS: Definition:
An operator T on a Hilbert space H is said to be a paranormal operator if for any unit vector x H. Definition: An operator T on a Hilbert space H is said to be a subnormal operator if T has a normal extension N, that is, there exists a normal operator N on a larger Hilbert space K H such that Tx = Nx for all x H.
Therom 1:
If T is a paranormal operator, then the following inequalities hold for any x H. +1 x
4x
3x
2x
‖ x‖ -----------------(1) ‖ x‖ ‖ 3x ‖ ‖ 2x‖ ‖ x‖ ‖x ‖ If T is an invertible paranormal operator then the following inequalities hold for any vector x H.
…….
…….
-----------------(2) ‖x ‖ ‖ −1 x ‖
−1 x
‖ −2 x ‖
−2 x
‖ −3 x ‖
……
− +1 x
‖ − x‖
……..
1 ‖ −1 ‖
2.6.2 IMPLICATION RELATION AMONG SEVERAL CLASSES OF NON NORMAL OPERATORS:
To prove: Quasi normal Subnormal. For this first we show that Quasi normal Hyponormal. Let T be quasi normal and T = U ITI be the polar decomposition of T. We know that, “ T is quasi normal UI T I = I T IU”. Let T*T - TT* = (UI T I)* (UI T I) - (UI T I) (UI T I)* = I T IU* UI T I - UI T I I T IU* = - UI T I I T IU* = I T I (I - UU*) I T I T*T - TT* 0.
Since UI T I = I T IU holds. Hence T is hyponormal. Hence Quasi normal Hyponormal. Next we show that Quasi normal Subnormal. Define [T] = T*T – TT*. Then, [T] 0. Since T is quasi normal, [T] T = T*TT – TT*T = T*TT – T*TT [T]T = 0. So that, T*[T] = 0 by taking adjoint [T]T = 0. T= T* = 0 for all natural number n. Since the square root ‘S’ of [T] is approximated uniformly by polynomials of [T] without constant terms, so that ST = T*S.
[since T is quasi normal and [T] = T*T – TT*, ST = T*S]. Therefore T is subnormal. Hence Quasi normal Subnormal Next we prove that Subnormal Hyponormal. Let T be subnormal then there is a normal extension N.
2.7. CHARACTERIZATION OF CONVEXOID OPERATORS AND RELATED EXAMPLES 2.7.1.CHARCTERIZATION OF CONVEXOID OPERATORS
Definition:
An operator T is said to be a convexoid operator if (T) = co (T) where (T) means the convex hull of the spectrum (T) of T. Definition: An operator T is said to be a condition ‖
‖ =
the spectrum (T) of T.
for all
operator
(T) where (T) means
. Definition: An operator T is said to be a transoloid operator if T - is normaloid for any c. Lemma 1: If x is any bounded closed set in the complex plane, then (i)
co x = {the intersection of all circles which is contain the set X} =
.
(ii) co x = {the intersection of all closed half planes which contain the set X} =
.
. Theorem 1: An operator T is convexoid if and only if T all complex number .
is spectraloid for
Proof: Taking X = (T) and since X = (T) is convex. Since,
, then the following (1) and (2) hold
(T) =
--------(1)
=
--------(2)
co From (1) and (2),
The proof follows from the relation co
-
= co
- ).
Theorem 2: −1
An operator t is convexoid iff ‖ all
‖
1 ( ,
(T))
for
co
We prove this theorem, we have to prove the following (5) lemma. Proof: The resolvent of any operator T satisfies a first order growth with respect to (T) by ( iii) of theorem For any operator T, the following (i),(ii) and (iii) hold (i)
co (iii)
(iv) ‖
, (ii)
1 ( ,
(T)) −1
‖
w(T) −1
‖ 1 ( ,
( ))
‖
for all
‖ T‖ 1 ( ,
( ))
(T).
We have only to show if part since the opposite implication follows easily from this fact. The condition (*) is equivalent to the following. ‖
x‖
co
and ‖x‖ = 1.
So that, 2
– Re(Tx, x)
+
2
2
2
Taking = | | to ∞, we obtain
−( + )
Re(Tx, x)
and dividing by |
Re s
| and transferring |
for ‖x‖ = 1.
|
This implies (T) co inclusion always holds.
by (ii) of lemma 1 and the opposite
So we obtain (T) = co
.
Theorem 3: Let T be a hyponormal operator,then the following properties hold. (i) (ii) (iii) (iv)
is also hyponormal for any c. is a transaloid operator. is also a hyponormal operator if exists. is a condition operator.
Proof: (i) To prove:
is also hyponormal for any
Since T is hyponormal
T*T
c.
TT*
Consider, ∗
(
= (T*T -
)-( T - T* +
= T*T - TT*
∗
) 2
) – (TT* -
T - T* +
0.
Therefore, is also a hyponormal for any
c.
2
)
(ii)To prove:
is a transaloid operator.
is hyponormal for any c, so that is normaloid c. Since a hyponormal operator is normaloid.
for any
Hence
is a transaloid operator.
(iii)To prove: Since T*T
−1
is also a hyponormal operator if −1
TT*, we have
*
−1 −1
This is equivalent to, I
*
−1
−1∗ −1
Hence
−1
−1∗ −1
−1∗
is also a hyponormal operator if
−1
exists.
exists. I.
(iii)To prove: is a condition
operator.
We know that, II
x‖
=r
‖ ‖ =r holds by (i) and (iii) and by the result that a hyponormal operators is normaloid. Therefore, ‖
‖ =
Hence is a condition Hence proved.
operator.
Theorem 4: The following properties hold. (i) (ii)
If T is a transaloid operator,then T is a convexoid operator. If T is a condition 1 operator, then T is a convexoid operator.
Proof: (i)To prove: If T is a transaloid operator,then T is a convexoid operator. Let T be a transaloid operator. (i.e)
is normaloid for any
c.
Then, is spectraloid for any
c.
Since a normaloid operator is spectraloid by theorem 3(2.6.1)
So that, T is convexoid by theorem 2 Hence If T is a transaloid operator,then T is a convexoid operator. (ii)To prove: If T is a condition operator. Let T be a condition
1
1
operator, then T is a convexoid
operator.
Then, we have ‖
−1
‖ =
1 ( ,
1 (T ))
( ,
(T))
for any
Because the inequality follows by the definition of condition operator. Hence T is a convexoid operator. Hence proved.
co
1
Corollary 5: (i) (ii)
Every hyponormal operator is convexoid. Every normal operator is convexoid.
Proof: (i)To prove: Every hyponormal operator is convexoid. If T is hyponormal, then the transloid by (ii) of theorem 3. So that T is convexoid by theorem 4. Hence every hyponormal operator is convexoid. (ii)To prove: Every normal operator is convexoid. We know that, Every normal operator is a hyponormal operator and also every hyponormal operator is convexoid. Therefore, Every normal operator is convexoid. Hence proved.
Theorem 6: An operator T is convexoid iff for any 0 2 where (s) denotes co
Re ( .
)=
(Re(
))
Proof: We know that, Every self adjoint operator is convexoid, so that the condition ( - ) means Re (
)=
(Re(
=
(Re(
= Re Re ( Therefore, Then,
( )=
))
(
) = Re (
)) ) ( )) for any
with 0
( )
The reverse is obvious.
Hence proved.
2 .
Theorem
′
: ′
Any operator T is convexoid iff co (Re( )) for any with 0
co Re (
)=
2 .
Proof: We know that, co ( ) = co (Re(
)) =
( ) (Re(
))
= Re ( (
))
= Re ( co (Re(
)) = co(Re
( )) (
)) for any 0
2 .
Theorem 7: If an operator T is convexoid such that both ( ) and (Re( )) are connected, then Re ( ) = (Re( )) hold. Proof: We know that, “T is convexoid then ′ [by theorem 6]. (i.e) is Therefore,
′
′
at
condition is true”. = 0 is true.
Re ( ) and (Re( )) are both closed convex sets.
Since connected sets coinsides with a convex set in the real line, ′ then the eliminate co in . Re (
)=
Hence Re ( ) =
(Re( (Re( )).
Hence proved.
))
Corollary 8: Let T be any convexoid operator. If [ 0 , interval containing Re ( ),then 0 , 0 (Re( )).
0]
is the smallest
Proof: Suppose T is convexoid operator and [ 0 , 0 ] is the smallest interval containing Re ( ( )). By theorem 6, we have Re ( ( )) = Therefore,
0,
(Re( )). 0
.
Hence proved. Corollary 9: Suppose T is a convexoid operator and ( ) is connected. If [ , ] is the smallest interval containing (Re( )), then (Re( )) [ , ] Re ( ( )).
Proof: Suppose T is a convexoid operator and ( ) is connected. If the smallest interval [ , ]
(Re( )) then Re ( ( ))
[ , ]
[by theorem 6’] By using corollary 8, we have
(Re( ))
[ , ]
Re ( ( ))
Hence proved. 2.7.2 Examples: Example 1: An example of non convexoid, non-pranormal and normaloid operator.
Let T =
where M =
Solution: 2
Here
2
‖
2
= T.T =
2
x‖
2
2
does not hold.
Thus T is non paranormal operator. Hence proved.