close

Вход

Log in using OpenID

embedDownload
MATEMATIQKI VESNIK
originalni nauqni rad
research paper
66, 3 (2014), 317–322
September 2014
NEW VERSION OF PROPERTY (az)
Hassan Zariouh
Abstract. In this paper we define new spectral properties (h), (gh), (ah) and (gah) as a
continuation of [H. Zariouh, Property (gz) for bounded linear operators, Mat. Vesnik, 65 (2013),
94–103], which are variants to the properties (z), (gz) and (az) introduced in the mentioned
paper. The purpose of the paper is to study the relationship between these properties and other
Weyl-type theorems.
1. Introduction
This paper is a continuation of [9] where we introduced and studied the spectral
properties (z) and (az). We investigate the spectral properties (h), (gh), (ah) and
(gah) (see Definitions 2.1 and 2.8) for bounded linear operators as new versions
of properties (z) and (az). The results obtained are summarized in the diagram
presented at the end of this paper. For further definitions and symbols we refer the
reader to [9], and we also refer to [3, 4, 5, 6, 10] for more details. In addition, we
have the following usual notations that will be needed later.
E(T ): eigenvalues of T that are isolated in the spectrum σ(T ),
E 0 (T ): eigenvalues of T of finite multiplicity that are isolated in σ(T ),
Ea (T ): eigenvalues of T that are isolated in the approximate point spectrum σa (T ),
Ea0 (T ): eigenvalues of T of finite multiplicity that are isolated in σa (T ),
Π(T ): poles of T ,
Π0 (T ): poles of T of finite rank,
Πa (T ): left poles of T ,
Π0a (T ): left poles of T of finite rank,
σBW (T ): B-Weyl spectrum of T ,
σW (T ): Weyl spectrum of T ,
σSBF − (T ): essential semi-B-Fredholm spectrum of T ,
+
σSF − (T ): Weyl essential approximate point spectrum of T ,
+
∆a (T ) = σa (T ) \ σSF − (T ),
+
2010 Mathematics Subject Classification: 47A53, 47A10, 47A11
Keywords and phrases: Property (h), property (ah), Weyl-type theorems.
317
318
H. Zariouh
∆ga (T ) = σa (T ) \ σSBF − (T ),
+
∆+ (T ) = σ(T ) \ σSF − (T ),
+
∆g+ (T ) = σ(T ) \ σSBF − (T ).
+
The inclusion of the following table [2, 5, 8, 9] which contains the meaning of
some properties is motivated by having overview of the subject.
(z)
σ(T ) \ σSF − (T ) = Ea0 (T )
(gz)
σ(T ) \ σSBF − (T ) = Ea (T )
(az)
σ(T ) \ σSF − (T ) = Π0a (T )
(gaz)
σ(T ) \ σSBF − (T ) = Πa (T )
(w)
σa (T ) \ σSF − (T ) = E 0 (T )
(gw)
σa (T ) \ σSBF − (T ) = E(T )
(b)
σa (T ) \ σSF − (T ) = Π0 (T )
(gb)
σa (T ) \ σSBF − (T ) = Π(T )
+
+
+
+
+
+
+
+
2. Main results
Definition 2.1. A bounded linear operator T ∈ L(X) is said to satisfy
property (h) if ∆+ (T ) = E 0 (T ) and is said to satisfy property (gh) if ∆g+ (T ) =
E(T ).
Example 2.2. Let L be the unilateral left shift operator defined on `2 (N). It
is well known that σ(L) = D(0, 1) is the closed unit disc in C, σSF − (L) = D(0, 1)
+
is the closed unit disk in C, E(L) = E 0 (L) = ∅. Moreover, σSBF − (L) = D(0, 1).
+
Hence the properties (h) and (gh) are satisfied by L.
Proposition 2.3. Let T ∈ L(X). Then T satisfies property (h) ⇐⇒ T
satisfies property (w) and σ(T ) = σa (T ).
Proof. Assume that T satisfies property (h). If λ ∈ ∆a (T ), then λ ∈ ∆+ (T ).
Therefore λ ∈ E 0 (T ) and ∆a (T ) ⊂ E 0 (T ). Now if λ ∈ E 0 (T ), then λ ∈ σa (T ) and
since T satisfies property (h), it follows that λ ∈
/ σSF − (T ). Thus λ ∈ ∆a (T ). Hence
+
∆a (T ) = E 0 (T ), i.e. T satisfies property (w). We then have σa (T ) \ σSF − (T ) =
+
E 0 (T ) and σ(T )\σSF − (T ) = E 0 (T ). This implies that σa (T ) = isoσa (T )∪σSF − (T )
+
+
and σ(T ) = isoσa (T ) ∪ σSF − (T ). So σ(T ) = σa (T ). Conversely, assume that T
+
satisfies property (w) and σ(T ) = σa (T ). Then ∆a (T ) = E 0 (T ) and σ(T ) = σa (T ).
So ∆+ (T ) = E 0 (T ), i.e. T satisfies property (h).
The next proposition gives a generalization of Proposition 2.3 to the context
of B-Fredholm theory.
Proposition 2.4. Let T ∈ L(X). Then T satisfies property (gh) ⇐⇒ T
satisfies property (gw) and σ(T ) = σa (T ).
Proof. Assume that T satisfies property (gh). If λ ∈ ∆ga (T ), then λ ∈ ∆g+ (T ).
Thus λ ∈ E(T ) and ∆ga (T ) ⊂ E(T ). Now if λ ∈ E(T ), then λ ∈ σa (T ) and since
T satisfies property (gh), λ ∈
/ σSBF − (T ) and so λ ∈ ∆ga (T ). Hence ∆ga (T ) =
+
New version of property (az)
319
E(T ) and T satisfies property (gw). We then have σa (T ) \ σSBF − (T ) = E(T )
+
and σ(T ) \ σSBF − (T ) = E(T ). Thus σa (T ) = isoσa (T ) ∪ σSBF − (T ) and σ(T ) =
+
+
isoσa (T ) ∪ σSBF − (T ) which entails that σ(T ) = σa (T ). Conversely, assume that T
+
satisfies property (gw) and σ(T ) = σa (T ). Then ∆ga (T ) = E(T ) and σ(T ) = σa (T ).
So ∆g+ (T ) = E(T ) and T satisfies property (gh).
Remark 2.5. The equality σa (T ) = σ(T ) which establishes a link between
property (gh) [resp. property (h)] and property (gw) [resp. property (w)] is satisfied
if T ∗ has the SVEP at every µ ∈
/ σSF − (T ), which in turn implies that properties
+
(w) and (h) are equivalent, and properties (gh) and (gw) are equivalent. Of course,
in this case it is easily seen that σW (T ) = σSF − (T ). Thus if λ ∈
/ σa (T ), then
+
T − λI is injective and λ ∈
/ σW (T ). Therefore T − λI is surjective, which implies
that λ ∈
/ σ(T ). Hence σ(T ) = σa (T ).
From Proposition 2.3 and Proposition 2.4, if T ∈ L(X) satisfies property (h)
[resp. property (gh)] then it satisfies property (w) [resp. property (gw)]. However,
the converses do not hold in general as seen by the following example.
Example 2.6. Let T = R ⊕ S defined on the Banach space `2 (N) ⊕ `2 (N),
where R is the unilateral right shift operator defined on `2 (N) and S is defined
on `2 (N) by S(x1 , x2 , x3 , . . . ) = ( 12 x2 , 13 x3 , . . . ). Then σ(T ) = D(0, 1) which is
the closed unit disc in C, σa (T ) = C(0, 1) ∪ {0} where C(0, 1) is the unit circle
of C, σSF − (T ) = C(0, 1) ∪ {0}, E 0 (T ) = ∅. Therefore ∆a (T ) = E 0 (T ) and T
+
satisfies property (w). Moreover, we have σSBF − (T ) = C(0, 1) ∪ {0} and E(T ) = ∅.
+
So ∆ga (T ) = E(T ) and T satisfies property (gw). But T does not satisfy either
property (h) nor property (gh), since ∆+ (T ) 6= E 0 (T ) and ∆g+ (T ) 6= E(T ). Here
σa (T ) 6= σ(T ).
Theorem 2.7. Let T ∈ L(X). Then the following statements hold.
(i) T satisfies property (gh) ⇐⇒ T satisfies property (gz).
(ii) T satisfies property (h) ⇐⇒ T satisfies property (z).
Proof. (i) If T satisfies (gh), that is ∆g+ (T ) = E(T ), then from Proposition
2.4, we have σ(T ) = σa (T ). So E(T ) = Ea (T ). Therefore ∆g+ (T ) = Ea (T ) and
T satisfies property (gz). Conversely, if T satisfies property (gz) that is ∆g+ (T ) =
Ea (T ), then from [9, Theorem 2.4] we have σ(T ) = σa (T ). So E(T ) = Ea (T ).
Thus ∆g+ (T ) = E(T ), i.e. T satisfies property (gh).
(ii) Follows directly by Proposition 2.3 and [9, Theorem 2.4].
Definition 2.8. A bounded linear operator T ∈ L(X) is said to satisfy
property (ah) if ∆+ (T ) = Π0 (T ) and is said to satisfy property (gah) if ∆g+ (T ) =
Π(T ).
Example 2.9. Recall that the Volterra operator V on L2 ([0, 1]) is defined
Rt
by V (f )(t) = 0 f (s) ds, for f ∈ L2 [0, 1]. It is well known that σ(V ) = {0},
320
H. Zariouh
σSF − (T ) = {0}, Π0 (V ) = Π(V ) = ∅ and σSBF − (V ) = {0}. Hence the Volterra
+
+
operator is an example satisfying the two properties (ah) and (gah).
Theorem 2.10. Let T ∈ L(X). Then T satisfies property (gah) ⇐⇒ T
satisfies property (ah).
Proof. Assume that T satisfies property (gah). If λ ∈ ∆+ (T ), then λ ∈
∆g+ (T ) = Π(T ). Since T − λI is an upper semi-Fredholm operator, then α(T − λI)
is finite. Thus λ ∈ Π0 (T ) and ∆+ (T ) ⊂ Π0 (T ). As we always have that ∆+ (T ) ⊃
Π0 (T ), then ∆+ (T = Π0 (T ) and T satisfies property (ah). Conversely, assume that
T satisfies property (ah). Let λ ∈ ∆g+ (T ) be arbitrary. We can assume without
loss of generality that λ = 0. Then T is an upper semi-B-Fredholm operator with
ind(T ) ≤ 0 and in particular is an operator of topological uniform descent [4]. From
the punctured neighborhood theorem for semi-B-Fredholm operators [4, Corollary
3.2], there exists ² > 0 such that if 0 < |µ| < ², then T − µI is an upper semiFredholm operator with ind(T − µI) = ind(T ). Let |µ| < ² and µ ∈
/ σ(T ). Then
a(T − µI) = δ(T − µI) = 0. The second possibility is that µ ∈ σ(T ), then
µ ∈ σ(T ) \ σSF − (T ) = Π0 (T ), since T satisfies property (ah). Thus a(T − µI) =
+
δ(T − µI) < ∞. In the two cases, we have a(T − µI) = δ(T − µI) < ∞. By [7,
Corollary 4.8] we then deduce that a(T ) = δ(T ) < ∞ and since 0 ∈ σ(T ) then
0 ∈ Π(T ). Hence ∆g+ (T ) ⊂ Π(T ), and since the opposite inclusion holds for every
operator, it then follows that property (gah) is satisfied by T .
Proposition 2.11. Let T ∈ L(X). Then T satisfies property (gah) ⇐⇒ T
satisfies property (gb) and σ(T ) = σa (T ).
Proof. Assume that T satisfies property (gah). If λ ∈ ∆ga (T ), then λ ∈ ∆g+ (T ).
Therefore λ ∈ Π(T ) and ∆ga (T ) ⊂ Π(T ). As ∆ga (T ) ⊃ Π(T ), then ∆ga (T ) = Π(T ),
i.e. T satisfies property (gb). We then have ∆g+ (T ) = Π(T ), ∆ga (T ) = Π(T ). Hence
σa (T ) = isoσa (T ) ∪ σSBF − (T ) and σ(T ) = isoσa (T ) ∪ σSBF − (T ). So σ(T ) =
+
+
σa (T ). Conversely, assume that T satisfies property (gb) and σ(T ) = σa (T ). Then
∆ga (T ) = Π(T ) and σ(T ) = σa (T ). So ∆g+ (T ) = Π(T ), i.e. T satisfies property
(gah).
Corollary 2.12. Let T ∈ L(X). Then T satisfies property (ah) ⇐⇒ T
satisfies property (b) and σ(T ) = σa (T ).
Proof. Assume that T satisfies property (ah), then from Theorem 2.10, T satisfies property (gah), which implies from Proposition 2.11 that T satisfies property
(gb) and σ(T ) = σa (T ). By [5, Theorem 2.3], T satisfies property (b). Conversely,
assume that T satisfies property (b) and σ(T ) = σa (T ). Then ∆a (T ) = Π0 (T ) and
σ(T ) = σa (T ). So ∆+ (T ) = Π0 (T ), i.e. T satisfies property (ah).
Remark 2.13. 1) From Corollary 2.12, if T ∈ L(X) satisfies (ah) then it
satisfies property (b). But the converse is not true in general: for example, let
T = 0 ⊕ R be defined on the Banach space `2 (N) ⊕ `2 (N), where R is the unilateral
right shift operator. Then σa (T ) = C(0, 1) ∪ {0}, σ(T ) = D(0, 1), Π0 (T ) = ∅,
New version of property (az)
321
σSF − (T ) = C(0, 1) ∪ {0}. So ∆a (T ) = Π0 (T ), i.e. T satisfies property (b). But it
+
does not satisfy property (ah), since ∆+ (T ) 6= Π0 (T ). Here σa (T ) 6= σ(T ).
2) From Proposition 2.11, if T ∈ L(X) satisfies property (gah) then it satisfies
property (gb). However, the converse does not hold in general. For this, if we
consider the operator T = R⊕S defined as in Example 2.6 then T satisfies property
(gb), because ∆ga (T ) = Π(T ) = ∅. But it does not satisfy property (gah), because
∆g+ (T ) 6= Π(T ). Here σa (T ) 6= σ(T ).
Theorem 2.14. Let T ∈ L(X). Then T satisfies property (ah) ⇐⇒ T satisfies
property (az).
Proof. Assume that T satisfies property (ah), that is ∆+ (T ) = Π0 (T ). As the
inclusion Π0 (T ) ⊂ Π0a (T ) is always true then ∆+ (T ) ⊂ Π0a (T ). If λ ∈ Π0a (T ), then
T − λI is upper semi-Fredholm operator with ind(T − λI) ≤ 0 and λ ∈ σ(T ), so
that λ ∈ ∆+ (T ). Hence ∆+ (T ) = Π0a (T ), i.e. T satisfies property (az). Conversely,
assume that T satisfies property (az), that is ∆+ (T ) = Π0a (T ). Let λ ∈ ∆+ (T ),
then λ ∈ isoσa (T ) and this implies from [9, Theorem 3.2] that λ ∈ isoσ(T ). Since
T −λI is an upper semi-Fredholm of negative index, from [1, Theorem 3.77] we have
λ ∈ Π0 (T ). Thus ∆+ (T ) ⊂ Π0 (T ). As ∆+ (T ) ⊃ Π0 (T ) holds for every operator
then ∆+ (T ) = Π0 (T ), i.e. T satisfies property (ah).
The next corollary gives a similar result to Theorem 2.14 in the case of property
(gah).
Corollary 2.15. Let T ∈ L(X). Then T satisfies property (gah) ⇐⇒ T
satisfies property (gaz).
Proof. Assume that T satisfies property (gah), then T satisfies property (ah).
By Theorem 2.14, T satisfies property (az). This is equivalent from [9, Corollary
3.5] to saying that T satisfies property (gaz). Conversely, assume that T satisfies
property (gaz), then T satisfies property (az). This implies that T satisfies property
(ah). Hence T satisfies property (gah).
322
H. Zariouh
As a conclusion, we give a summary of the results obtained in this paper. In
the previous diagram, arrows signify implications between the properties studied
in this paper and other Weyl type theorems. The numbers near the arrows are
references to the results in the present paper (numbers without brackets) or to the
bibliography therein (numbers in square brackets).
REFERENCES
[1] P. Aiena, Fredholm and Local Spectral Theory, with Application to Multipliers, Kluwer Academic, 2004.
[2] M. Amouch, M. Berkani, On the property (gw), Mediterr. J. Math. 5, (2008), 371–378.
[3] M. Berkani, On a class of quasi-Fredholm operators, Integr. Equ. and Oper. Theory 34
(1999), 244–249.
[4] M. Berkani, M. Sarih, On semi B-Fredholm operators, Glasgow Math. J. 43 (2001), 457–465.
[5] M. Berkani, H. Zariouh, Extended Weyl type theorems, Math. Bohemica 134 (2009), 369–378.
[6] M. Berkani, H. Zariouh, New extended Weyl type theorems, Mat. Vesnik 62 (2010), 145–154.
[7] S. Grabiner, Uniform ascent and descent of bounded operators, J. Math. Soc. Japan 34
(1982), 317–337.
[8] V. Rakoˇ
cevi´
c, On a class of operators, Mat. Vesnik. 37 (1985), 423–426.
[9] H. Zariouh, Property (gz) for bounded linear operators, Mat. Vesnik, 65 (2013), 94–103.
[10] H. Zariouh, Zguitti, Variations on Browder’s theorem, Acta Math. Univ. Comenianae 81
(2012), 255–264.
(received 28.01.2013; available online 01.08.2013)
Centre r´
egional des m´
etiers de l’´
education et de la formation, B.P 458, Oujda, Maroc et Equipe
de la Th´
eorie des Op´
erateurs, Universit´
e Mohammed Premier, Facult´
e des Sciences d’Oujda,
D´
epartement de Math´
ematiques et Informatique, Morocco
E-mail: h.zariouh@yahoo.fr
1/--pages
Пожаловаться на содержимое документа