Professional Documents
Culture Documents
v
x = (v
n
x
n
v
n1
x
n1
)
n=0
with x
1
= 0, where x = (x
n
) l
1
and v = (v
k
) is
either constant or strictly decreasing sequence of positive real numbers satisfying
certain conditions. In this paper we have obtained the results on spectrum and
point spectrum for the operator
v
over the sequence space l
1
. Further, the results
on continuous spectrum, residual spectrum and ne spectrum of the operator
v
on space l
1
are also derived
Keywords : Spectrum of an operator, Generalized dierence operator, Sequence
spaces.
2000 Mathematics Subject Classication : 47A10, 47B39, 46A45.
1 Introduction
Let v = (v
k
) be either constant or strictly decreasing sequence of positive real
numbers satisfying
lim
k
v
k
= L > 0 and (1.1)
sup
k
v
k
2L. (1.2)
We introduce the operator
v
on sequence space l
1
as follows;
v
: l
1
l
1
is dened by,
v
x =
v
(x
n
) = (v
n
x
n
v
n1
x
n1
)
n=0
with x
1
= 0, where x l
1
.
Copyright c 2010 by the Mathematical Association of Thailand. All rights
reserved.
222 Thai J. Math. 8(2) (2010)/ P.D. Srivastava and S. Kumar
It is easy to verify that the operator
v
can be represented by the matrix
v
=
_
_
_
_
_
v
0
0 0 . . .
v
0
v
1
0 . . .
0 v
1
v
2
. . .
.
.
.
.
.
.
.
.
.
.
.
.
_
_
_
_
_
.
The ne spectrum of the Cesaro operator on sequence space l
p
is studied by Gon-
zalez [5], where 1 < p < . The spectrum of the Cesaro operator on sequence
spaces bv
0
and bv is also investigated by Okutoyi [9] and Okutoyi [10], respectively.
Spectrum and ne spectrum of the dierence operator over sequence spaces l
1
and bv is determined by K. Kayaduman and H. Furkan [7]. The ne spectra of
the dierence operator over sequence space l
p
is determined by Akhmedov and
Basar [1], where 1 p < . Furthermore, the ne spectrum of the operator
B(r, s) on the sequence spaces l
1
and bv is examined by H. Furkan, H. Bilgic and
K. Kayaduman [3]. Recently, H. Bilgic and H. Furkan [2] studied the spectrum
and ne spectrum for the operaor B(r, s, t) over sequence spaces l
1
and bv.
In this paper we determine spectrum, point spectrum, continuous spectrum
and residual spectrum of the operator
v
on sequence space l
1
. The results of this
paper not only generalize the corresponding results of [7] but also give results for
some more operators.
2 Preliminaries and notation
Let X and Y be Banach spaces and T : X Y be a bounded linear operator.
The set of all bounded linear operators on X into itself is denoted by B(X). The
adjoint T
: X
of T is dened by
_
T
_
(x) = (Tx) for all X
and x X.
Clearly, T
.
Let X = {0} be a complex normed space and T : D(T) X be a linear
operator with domain D(T) X. With T, we associate the operator T
=
(T I), where is a complex number and I is the identity operator on D(T).
The inverse of T
and T
1
exists/ T
1
is bounded/ domain of T
1
is dense in X. We need
some denitions and known results which will be used in the sequel.
Denition 2.1. ([6], pp. 371) Let X = {0} be a complex normed space and
T : D(T) X be a linear operator with domain D(T) X. A regular value of T
is a complex number such that
(R1) T
1
exists,
(R2) T
1
is bounded,
(R3) T
1
exists and
satises (R3) but not (R2), i.e., range of T
is dense in X and T
1
is unbounded.
Residual spectrum
r
(T, X) is the set of all C such that T
1
exists but do
not satisfy (R3), i.e., domain of T
1
) and T
1
) = X,
(B) R(T
) = R(T
) = X,
(C) R(T
) = X,
and
(1) T
is injective and T
1
is continuous,
(2) T
is injective and T
1
is discontinuous,
(3) T
is not injective.
Remark 2.1. Combining (A), (B), (C) and (1), (2), (3); we get nine dierent
states. These are labeled by A
1
, A
2
, A
3
, B
1
, B
2
, B
3
, C
1
, C
2
and C
3
. We use
B
2
(T, X) means the operator T
B
2
, i.e., R(T
) = R(T
) = X and T
is
injective but T
1
A
1
or T
B
1
, then
belongs to the resolvent set (T, X) of T on X. The other classication gives rise
to the ne spectrum of T.
Denition 2.2. ([8], pp. 220-221) Let , be two nonempty subsets of the space
w of all real or complex sequences and A = (a
nk
) an innite matrix of complex
numbers a
nk
, where n, k N
0
= {0, 1, 2, }. For every x = (x
k
) and every
integer n we write
A
n
(x) =
k
a
nk
x
k
,
where the sum without limits is always taken from k = 0 to k = . The sequence
Ax = (A
n
(x)), if it exists, is called the transformation of x by the matrix A.
Innite matrix A (, ) if and only if Ax whenever x .
Lemma 2.1. ([11], pp. 126) The matrix A = (a
nk
) gives rise to a bounded linear
operator T B(l
1
) from l
1
to itself if and only if the supremum of l
1
norms of the
columns of A is bounded.
224 Thai J. Math. 8(2) (2010)/ P.D. Srivastava and S. Kumar
Lemma 2.2. ([4], pp. 59) T has a dense range if and only if T
is one to one,
where T
v
on sequence space l
1
In this section we obtain spectrum and point spectrum of the operator
v
on
sequence space l
1
. Throughout this paper, the sequence v = (v
k
) satisfy conditions
(1.1) and (1.2).
Theorem 3.1. The operator
v
: l
1
l
1
is a bounded linear operator and
(l1,l1)
= 2 sup
k
(v
k
).
Proof. Proof is simple. So we omit.
Theorem 3.2. The spectrum of
v
on sequence space l
1
is given by
(
v
, l
1
) =
_
C :
1
L
1
_
.
Proof. The proof of this theorem is divided into two parts. In the rst part, we
show that (
v
, l
1
)
_
C :
1
L
1
_
or equivalent to show that
C with
1
L
> 1 implies / (
v
, l
1
), i.e., (
v
, l
1
).
In the second part, we establish the reverse inequality, i.e.,
_
C :
1
L
1
_
(
v
, l
1
).
Let C with
1
L
n=0
|b
nk
| < . In order
to show sup
k
n=0
|b
nk
| < , rst we prove that the series
n=0
|b
nk
| is convergent for
each k N
0
.
Let S
k
=
n=0
|b
nk
|. Then the series
S
0
=
n=0
|b
no
|
=
1
v
0
n=1
v
0
v
1
v
n1
(v
0
)(v
1
) (v
n
)
(3.1)
is convergent because
lim
n
b
n+1,0
b
n0
= lim
n
v
n
v
n+1
=
1
1
L
< 1.
Similarly, we can show that the series S
k
=
n=0
|b
nk
| is convergent for any k =
1, 2, 3, .
Now we claim that sup
k
S
k
is nite. We have
S
k
=
1
|v
k
|
+
|v
k
|
|v
k
| |v
k+1
|
+ . (3.2)
Let = lim
k
v
k
v
k+1
L
L
, (3.3)
which shows that 0 < < 1 and gives
lim
k
1
v
k
= lim
k
_
v
k1
v
k
1
v
k1
_
=
L
. (3.4)
Taking limit both sides of equation (3.2) and using equations (3.3) and (3.4), we
get
lim
k
S
k
=
L
_
1
1
_
< .
Since (S
k
) is a sequence of positive real numbers and lim
k
S
k
< , so sup
k
S
k
< .
Thus,
(
v
I)
1
B(l
1
) for C with
1
L
> 1. (3.5)
226 Thai J. Math. 8(2) (2010)/ P.D. Srivastava and S. Kumar
Next, we show that domain of the operator (
v
I)
1
is dense in l
1
equivalent to
say that range of the operator (
v
I) is dense in l
1
, which follows immediately
as the operator (
v
I) is onto. Hence we have
(
v
, l
1
)
_
C :
1
L
1
_
. (3.6)
Conversely, it is required to show
_
C :
1
L
1
_
(
v
, l
1
). (3.7)
First we prove inclusion (3.7) under the assumption that = L as well as = v
k
for each k N
0
, i.e., one of the conditions of Denition 2.1 fails. Let C with
1
L
1. Clearly, (
v
I) is a triangle and hence (
v
I)
1
exists. So
condition (R1) is satised but condition (R2) fails as can be seen below:
Suppose C with
1
L
b
n+1,0
b
n0
= lim
n
v
n
v
n+1
=
1
1
L
> 1.
So sup
k
S
k
is unbounded. Hence
(
v
I)
1
/ B(l
1
) for C with
1
L
< 1. (3.8)
Next, we consider C with
1
L
n=0
1
v
n
is divergent due to the fact
that v
n
> 0 for all n and lim
n
1
v
n
=
1
L
= 0. Thus, sup
k
S
k
is unbounded. Hence
(
v
I)
1
/ B(l
1
) for C with
1
L
= 1. (3.9)
Finally, we prove the inclusion (3.7) under the assumption that = L as well as
= v
k
for all k N
0
. We have
(
v
v
k
I) x =
_
_
_
_
_
_
_
_
_
_
(v
0
v
k
)x
0
v
0
x
0
+ (v
1
v
k
)x
1
.
.
.
v
k1
x
k1
v
k
x
k
+ (v
k+1
v
k
)x
k+1
.
.
.
_
_
_
_
_
_
_
_
_
_
.
Fine spectrum of the generalized dierence operator
v
on sequence space l
1
227
Case(i): If (v
k
) is a constant sequence, say v
k
= L for all k N
0
, then
(
v
v
k
I) x = 0 x
0
= 0, x
1
= 0, x
2
= 0, .
This shows that the operator (
v
I) is one to one, but R(
v
I) is not
dense in l
1
. So condition (R3) fails. Hence L (
v
, l
1
).
Case(ii): If (v
k
) is strictly decreasing sequence, then for xed k,
(
v
v
k
I) x = 0
x
0
= 0, x
1
= 0, , x
k1
= 0, x
n+1
=
_
v
n
v
n+1
v
k
_
x
n
for all n k.
This shows that (
v
v
k
I) is not injective. So condition (R1) fails. Hence v
k
(
v
, l
1
) for all k N
0
.
Again, if = L, then |v
n
| < |v
n
| for each n, i.e.,
1
|v
n
|
<
1
|v
n
|
for each
n. Using this inequality and equation (3.1), the series S
0
>
n=0
1
v
n
is divergent due
to fact that v
n
> 0 for all n and lim
n
1
v
n
=
1
L
= 0. Thus, sup
k
S
k
is unbounded.
So condition (R2) fails. Hence
(
v
I)
1
/ B(l
1
) for = L. (3.10)
So L (
v
, l
1
). Thus, in this case also v
k
(
v
, l
1
) for all k N
0
and
L (
v
, l
1
). Hence we have
_
C :
1
L
1
_
(
v
, l
1
) . (3.11)
From inclusions (3.6) and (3.11), we get
(
v
, l
1
) =
_
C :
1
L
1
_
.
Theorem 3.3. Point spectrum of the operator
v
over l
1
is given by
p
(
v
, l
1
) =
_
_
_
, if (v
k
) is a constant sequence.
{v
0
, v
1
, v
2
, } , if (v
k
) is a strictly decreasing sequence.
Proof. The proof of this theorem is divided into two cases.
Case(i): Suppose (v
k
) is a constant sequence, say v
k
= L for all k N
0
. Consider
v
x = x for x = 0 = (0, 0, ) in l
1
, which gives
v
0
x
0
= x
0
v
0
x
0
+v
1
x
1
= x
1
v
1
x
1
+v
2
x
2
= x
2
.
.
.
v
k1
x
k1
+v
k
x
k
= x
k
.
.
.
_
_
(3.12)
228 Thai J. Math. 8(2) (2010)/ P.D. Srivastava and S. Kumar
Let x
t
be the rst non-zero entry of the sequence x = (x
n
), so we get Lx
t1
+
Lx
t
= x
t
, which implies = L and from the equation
Lx
t
+Lx
t+1
= x
t+1
,
we get x
t
= 0, which is a contradiction to our assumption. Therefore,
p
(
v
, l
1
) = .
Case(ii): Suppose (v
k
) is a strictly decreasing sequence. Consider
v
x = x for
x = 0 = (0, 0, ) in l
1
, which gives system of equations (3.12).
If = v
0
, then
x
k
=
_
v
k1
v
k
v
0
_
x
k1
for all k 1
=
_
v
k1
v
k2
v
0
(v
k
v
0
)(v
k1
v
0
) (v
1
v
0
)
_
x
0
for all k 1.
If we take x
0
= 0, then get non-zero solution of (
v
v
0
I) x = 0.
Similarly, if = v
k
for all k 1, then x
k1
= 0, x
k2
= 0, , x
0
= 0 and
x
n+1
=
_
v
n
v
n+1
v
k
_
x
n
for all n k
=
_
v
n
v
n1
v
k
(v
n+1
v
k
)(v
n
v
k
) (v
k+1
v
k
)
_
x
k
for all n k.
If we take x
k
= 0, then get non-zero solution of (
v
v
k
I) x = 0. Hence
p
(
v
, l
1
) = {v
0
, v
1
, v
2
, }.
4 Residual and continuous spectrum of the oper-
ator
v
on sequence space l
1
We need result of point spectrum of the operator
v
on l
1
for obtaining
residual and continuous spectrum. So rst we determine point spectrum of the
dual operator
v
of
v
on space l
1
.
Let T : l
1
l
1
be a bounded linear operator having matrix representation A
and the dual space of l
1
denoted by l
1
. Then the adjoint operator T
: l
1
l
1
is
dened by the transpose of the matrix A.
Theorem 4.1. Point spectrum of the operator
v
over l
1
is
p
(
v
, l
1
) =
_
C :
1
L
1
_
.
Fine spectrum of the generalized dierence operator
v
on sequence space l
1
229
Proof. Suppose
v
f = f for 0 = f l
= l
, where
v
=
_
_
_
_
_
v
0
v
0
0 . . .
0 v
1
v
1
. . .
0 0 v
2
. . .
.
.
.
.
.
.
.
.
.
.
.
.
_
_
_
_
_
and f =
_
_
_
_
_
f
0
f
1
f
2
.
.
.
_
_
_
_
_
.
This gives
f
k
=
_
(v
k1
)(v
k2
) (v
0
)
v
k1
v
k2
v
0
_
f
0
for all k 1.
Hence
|f
k
| =
(v
k1
)(v
k2
) (v
0
)
v
k1
v
k2
v
0
|f
0
| for all k 1. (4.1)
But
|v
k1
| (v
k1
L) +|L |
v
k1
v
k1
1
L
1.
Using equation (4.1), we get
|f
k
| |f
0
| for all k 1. So sup
k
|f
k
| < .
Hence
1
L
1 sup
k
|f
k
| < .
Converse follows from the fact that
sup
k
|f
k
| <
v
k1
v
k1
1 for all k m,
where m is a positive integer.
lim
k
v
k1
v
k1
1
L
1.
Hence
sup
k
|f
k
| <
1
L
1.
This means that f l
1
if and only if f
0
= 0 and
1
L
1. Hence
p
(
v
, l
1
) =
_
C :
1
L
1
_
.
230 Thai J. Math. 8(2) (2010)/ P.D. Srivastava and S. Kumar
Theorem 4.2. Residual spectrum
r
(
v
, l
1
) of the operator
v
over l
1
is
r
(
v
, l
1
) =
_
_
_
C :
1
L
1
_
, if (v
k
) is a constant sequence.
_
C :
1
L
1
_
\ {v
0
, v
1
, v
2
, }, if
(v
k
) is a strictly decreasing sequence.
Proof. The proof of this theorem is divided into two cases.
Case(i): Let (v
k
) be a constant sequence. For C with
1
L
1, the
operator (
v
I) is a triangle except = L and consequently, the operator
(
v
I) has an inverse. Further by Theorem 3.3, the operator (
v
I) is one
to one for = L and hence has an inverse.
But by Theorem 4.1, the operator (
v
I)
v
I is not one to one
for C with
1
L
r
(
v
, l
1
) =
_
C :
1
L
1
_
.
Case(ii): Let (v
k
) be a strictly decreasing sequence with lim
k
v
k
= L. For C
such that
1
L
1, the operator (
v
I) is a triangle except = v
k
for all k
N
0
and consequently, the operator (
v
I) has an inverse. Further by Theorem
3.3, the operator (
v
v
k
I) is not one to one and hence (
v
v
k
I)
1
does not
exists for all k N
0
.
On the basis of argument as given in case(i), it is easy to verify that the range
of the operator (
v
I) is not dense in l
1
. Thus,
r
(
v
, l
1
) =
_
C :
1
L
1
_
\ {v
0
, v
1
, v
2
, }.
Theorem 4.3. Continuous spectrum
c
(
v
, l
1
) of the operator
v
over l
1
is
c
(
v
, l
1
) = .
Proof. It is known that
p
(
v
, l
1
) ,
r
(
v
, l
1
) and
c
(
v
, l
1
) are pairwise disjoint
sets and union of these sets is (
v
, l
1
). But by Theorems 3.2, 3.3 and 4.2; we get
(
v
, l
1
) =
p
(
v
, l
1
)
r
(
v
, l
1
) .
Therefore,
c
(
v
, l
1
) = .
5 Fine spectrum of the operator
v
on sequence
space l
1
Theorem 5.1. If satises
1
L
> 1, then (
v
I) A
1
.
Fine spectrum of the generalized dierence operator
v
on sequence space l
1
231
Proof. It is required to show that the operator (
v
I) is bijective and has a
continuous inverse for C with
1
L
1
L
> 1 by statement
(3.5). Also the equation
(
v
I) x = y gives x = (
v
I)
1
y,
i.e., x
n
=
_
(
v
I)
1
y
_
n
, n N
0
.
Thus for every y l
1
, we can nd x l
1
such that
(
v
I)x = y, since (
v
I)
1
(l
1
, l
1
).
This shows that operator (
v
I) is onto and hence (
v
I) A
1
.
Theorem 5.2. Let (v
k
) be a constant sequence, say v
k
= L for all k N
0
. Then
L C
1
(
v
, l
1
).
Proof. We have
r
(
v
, l
1
) =
_
C :
1
L
1
_
.
Clearly, L
r
(
v
, l
1
). It is sucient to show that the operator (
v
LI)
1
is
continuous. By Lemma 2.3, it is enough to show that (
v
LI)
, we have to nd x = (x
n
) l
such that (
v
LI)
x = y.
Now (
v
LI)
x = y, i.e.,
Lx
1
= y
0
Lx
2
= y
1
.
.
.
Lx
i
= y
i1
.
.
.
Thus, Lx
n
= y
n1
for all n 1, which implies sup
n
|x
n
| < , since y l
and
L = 0. This shows that operator (
v
LI)
1
L
1.
232 Thai J. Math. 8(2) (2010)/ P.D. Srivastava and S. Kumar
Theorem 5.4. Let (v
k
) be a strictly decreasing sequence of positive real numbers
and
r
(
v
, l
1
). Then C
2
(
v
, l
1
).
Proof. It is sucient to show that the operator (
v
I)
1
is discontinuous for
r
(
v
, l
1
). The operator (
v
I)
1
is discontinuous by statements (3.8),
(3.9) and (3.10) for v
k
= C with
1
L
1.
References
[1] A.M. Akhmedov and F. Basar, On the ne spectra of the dierence operator
over the sequence space l
p
(1 p < ), Demonstratio Math., Vol. 39, No.
3 (2006) 585595.
[2] H. Bilgic and H. Furkan, On the ne spectrum of the operator B(r, s, t) over
the sequence spaces l
1
and bv, Mathematical and Comp. Modelling, Vol. 45
(2007) 883891.
[3] H. Furkan, H. Bilgic and K. Kayaduman, On the ne spectrum of the general-
ized dierence operator B(r, s) over the sequence spaces l
1
and bv, Hokkaido
Math. J., Vol. 35, No. 4 (2006) 893904.
[4] S. Goldberg, Unbounded linear operators, Dover Publications, Inc. New York,
1985.
[5] M. Gonzalez, The ne spectrum of the Cesaro operator in l
p
(1 < p < ),
Arch. Math., Vol. 44 (1985) 355358.
[6] E. Kreyszig, Introductory functional analysis with applications, John Wiley
and Sons Inc., New York, Chichester, Brisbane, Toronato, 1978.
[7] K. Kayaduman and H. Furkan, The ne spectra of the dierence operator
over the sequence spaces l
1
and bv, International Mathematical Forum, Vol.
1, No. 24 (2006) 11531160.
[8] I.J. Maddox, Elements of functional analysis, Cambridge University Press,
1988.
[9] J.I. Okutoyi, On the spectrum of C
1
as an operator on bv
0
, J. Austral. Math.
Soc. Ser. A, Vol. 48 (1990) 7986.
[10] J.T. Okutoyi, On the spectrum of C
1
as an operator on bv, Commun. Fac.
Sci. Univ. Ank. Ser. A
1
, Vol. 41 (1992) 197207.
[11] A. Wilansky, Summability through functional analysis, North-Holland Math-
ematics Studies, North-Holland, Amsterdam, Vol. 85, 1984.
(Received 3 December 2009)
Fine spectrum of the generalized dierence operator
v
on sequence space l
1
233
P.D. Srivastava
Department of Mathematics
Indian Institute of Technology, Kharagpur
Kharagpur - 721302, India
E-mail: pds@maths.iitkgp.ernet.in
Sudhanshu Kumar
Department of Mathematics
Indian Institute of Technology, Kharagpur
Kharagpur - 721302, India
E-mail: sudhanshu tomar@yahoo.com