SOME GEOMETRIC PROPERTIES OF A GENERALIZED CESÀRO
SEQUENCE SPACE
V. A. KHAN
Abstract. In this paper we define the generalized Cesàro sequence space Cesp (q) and exhibit some
geometrical properties of the space. The results herein proved are exhibit analogous to those by
Y. A. Cui [Southeast Asian Bull. Math. 24 (2000), 201–210] for the Cesàro sequence space Cesp .
1. Introduction
JJ J
I II
Go back
Let (X, k.k) be a real Banach space. By B(X) and S(X), we denote the closed unit ball and the
unit sphere of X, respectively. For any subset A of X, χA represents a characteristics function of
A.
A norm k.k is called uniformly convex (UC) (cf. [2]) if, for each ε > 0, there exists δ > 0 such
that, for x, y ∈ S(X), kx − yk > ε implies
1
(1.0.1)
2 (x + y) < 1 − δ.
A Banach space X is said to have the Banach-Saks property if every bounded sequence (xn ) in
X admits a subsequence {zn } such that the sequence {tk (z)} is convergent in norm in X (see [1]),
Full Screen
Close
Quit
Received August 10, 2005; revised November 02, 2009.
2000 Mathematics Subject Classification. Primary 46E30, 46E40, 46B20.
Key words and phrases. Banach-Saks property; ε-separated sequence; property β; Generalized Cesàro sequence
space.
where
tk (z) =
1
(z1 + z2 + · · · + zk ).
k
Every Banach space X with the Banach-Saks property is reflexive, but the converse is not true
(see [4, 5]). Kakutani [6] proved that any uniformly convex Banach space X has the Banach-Saks
property. Moreover, he also proved that if X is a reflexive Banach space and θ ∈ (0, 2) such that
for every sequence (xn ) in S(X) weakly convergent to zero, there exist n1 , n2 ∈ N satisfying the
Banach-Saks property.
For a sequence (xn ) ⊂ X, we define
A(xn ) = lim inf{kxi + xj k : i, j ≥ n, i 6= j}.
n→∞
In [6], the following new geometric constant connected with the packing constant (see [7]) and
with the Banach-Saks property was defined
C(X) = sup{A(xn ) : (xn ) is a weakly null sequence in S(X)}.
Recall that a sequence (xn ) is said to be an ε-separated sequence if, for some ε > 0,
JJ J
I II
sep(xn ) = inf{kxn − xm k : n 6= m} > ε.
Go back
Full Screen
Close
Quit
A Banach space X is said to satisfy property (β) if and only if, for every ε > 0, there exists δ > 0
such that, for each element x ∈ B(X) and each sequence (xn ) in B(X) with sep(xn ) ≥ ε, there is
an index k such that
x + xk for some k ∈ N.
2 | ≥ 1 − δ,
In this paper, we define the generalized Cesàro sequence space as follows: let p ∈ [1, ∞) and q
be a bounded sequence of positive real numbers such that
Qn =
n
X
qk ,
(n ∈ N),
k=0
Cesp (q) =


x = (x(i)) : kxk =

∞
X
n=1
n
1 X
|qi x(i)|
Qn i=1
!p !1/p


<∞ .

If qi = 1 for all i ∈ N, then Cesp (q) is reduced to Cesp (cf. [3, 8, 9]).
Lemma 1.1. Let x, y ∈ Cesp (q). Then for any ε > 0 and L > 0, ∃ δ > 0 such that
JJ J
I II
| kx + ykp − kxkp | < ε
Go back
Full Screen
whenever
Close
kxkp ≤ L
Quit
and
kykp ≤ δ.
Proof. For any fix ε > 0 and L > 0, take β = 2−p L−1 ε and δ = 2−p β p−1 ε. Then for any
x, y ∈ Cesp (q) with kxkp ≤ L and kykp ≤ δ, we have
∞
n
X
1 X
kx + yk =
(
|qi x(i) + qi y(i)|)p
Qn i=1
n=1
!p
∞
n X
qi y(i) 1 X =
(1 − β)qi x(i) + β(qi x(i) +
)
Qn i=1 β n=1
p
!p
n
n
1 X
1 X
qi y(i)
|
≤
(1 − β)
|qi x(i)| + β
|qi x(i) +
Qn i=1
Qn i=1
β
n=1
!p
!p
∞
n
∞
n
X
X
1 X
1 X
qi y(i)
≤ (1 − β)
|qi x(i)| + β
|qi x(i) +
|
Qn i=1
Qn i=1
β
n=1
n=1
∞
X
≤
∞
X
n
1 X
|qi x(i)|
Qn i=1
n=1
JJ J
I II
Go back
+
∞
βX
2 n=1
p
Full Screen
!p
p−1
≤ kxk + ε/2 + (2/β)
∞
X
n=1
Close
Quit
∞
βX
+
2 n=1
!
p
n
1 X 2qi y(i)
|
|
Qn i=1
β
!p
n
1 X
|2qi x(i)|
Qn i=1
n
1 X
|qi y(i)|
Qn i=1
!p
≤ kxkp + ε.
2. Main Results.
Theorem 2.1. The space Cesp (q) satisfies the property (β).
Proof. Let Cesp (q) not have property (β). Then there exists ε0 > 0 such that, for any δ ∈
1/p
(0, ε0 /(1 + 21+p )), there is a sequence (xn ) ⊂ S(Cesp (q)) with sep(xn ) > ε0 and an element
x0 ∈ S(Cesp (q)) such that
xn + x0 p
>1−δ
for any n ∈ N.
2
Fix δ ∈ (0, ε0 /(1 + 21+p )). We want to show that
(2.1.1)
JJ J
I II
lim sup
j→∞ k
∞
X
n=j+1
!p
n
21+p δ
1 X
|qi xk (i)|
≤ p
.
Qn i=1
2 −1
Otherwise, without loss of generality, we can assume that there exists a sequence (jk ) such that
jk → ∞ as k → ∞ and
!p
∞
n
X
1 X
21+p δ
(2.1.2)
|qi xk (i)|
> p
for every k ∈ N.
Qn i=1
2 −1
n=j +1
k
Go back
Full Screen
Close
Quit
Let δ > 0 be a real number corresponding to ε = δ and L = 1 in Lemma 1. By absolute
continuity of the norm of x0 , there exists a positive integer n1 such that
p
kx0 χn1 ,n1 +1,n1 +2,... k =
∞
X
n=n1 +1
!p
n
1 X
|qi x0 (i)|
< δ1 .
Qn i=1
Choose k so large that jk > n1 . By the Lemma 1 and (2.1.2), we have
1−δ <
∞
X
n=1
=
n1
X
n=1
!p
n 1 X qi xk (i) + qi x0 (i) Qn i=1 2
!p
n ∞
X
1 X qi xk (i)+qi x0 (i) +
Qn i=1 2
n=n +1
!p
n1
n
n
1 X
1X
1 X
|qi x0 (i)| +
|qi xk (i)|
Qn i=1
2 n=1 Qn i=1
!p
∞
n X
1 X qi xk (i) +
+δ
Qn i=1 2 n=n1 +1
!p
!p
n1
n
n
∞
1 X
1 1X
1 X
1 X
≤ +
|qi xk (i)| + p
|qi xk (i)| + δ
2 2 n=1 Qn i=1
2 n=n +1 Qn i=1
1
!p
!p
∞
n
∞
n
1 1X
1 X
2p − 1 X
1 X
≤ +
|qi xk (i)| −
|qi xk (i)| + δ
2 2 n=1 Qn i=1
2p n=n +1 Qn i=1
n1
1X
≤
2 n=1
JJ J
I II
1
!p
n 1 X qi xk (i) + qi x0 (i) Qn i=1 2
!p
1
Go back
Full Screen
Close
Quit
< 1 − 2δ + δ = 1 − δ.
That is a contradiction. Hence (2.1.1) must hold. Since
!p
!p
n1
n1
n
X
1 X
1 X
|qi xk (i)|
≤
|qi xk (i)|
≤ 1,
Qn1 i=1
Qn i=1
n=1
we have |qi xk (i)| ≤ Qn1 for k ∈ N and i = 1, 2, . . . , n1 . Hence there is a subsequence (zn ) of (xn )
and a sequence (an ) of real numbers such that
lim qi zk (i) = ai ,
k→∞
Therefore,
n1
X
n=1
n
1 X
|qi zk (i) − qi zm (i)|
Q n i=1
for ı = 1, 2, . . . , n1 .
!p
< δ for n,m sufficiently large.
Consequently,
kzk − zm kp
∞
n
X
1 X
|qi zk (i) − qi zm (i)|)p
(
Q
n
n=1
i=1
!p
!p
n1
n
∞
n
X
X
1 X
1 X
=
|qi zk (i) − qi zm (i)| +
|qi zk (i) − qi zm (i)|
Qn i=1
Qn i=1
n=n1 +1
n=1
!p
n1
n
X
1 X
≤
|qi zk (i) − qi zm (i)|
Qn i=1
n=1
!p
!p !
∞
n
∞
n
X
X
1 X
1 X
p
+2
(
|qi zk (i)| +
|qi zm (i)|
Qn i=1
Qn i=1
n=n +1
n=n +1
=
JJ J
I II
Go back
1
Full Screen
Close
Quit
≤δ+2
p+1
1
δ < ε0 ,
1/p
i.e., sep(xn ) ≤ sep(zn ) < ε0 . This is a contradiction. Therefore Cesp (q) must satisfy the property
(β).
Theorem 2.2. C(Cesp (q)) = 21/p .
Proof. Let
in
X
K = sup{A(un ) : un =
un (i)ei ∈ S(Cesp (q)),
i=in−1 +1
ω
0 = i0 < i1 < i2 < . . . , un →0, .
ω
Then C(Cesp (q)) ≥ K. Moreover, for any ε > 0, there is a sequence (xn ) ⊂ S(Cesp (q)) with xn →0
such that
A(xn ) + ε > C(Cesp (q)).
By the definition of A(xn ), there exists a subsequence (yn ) of (xn ) such that
kyn + ym k + 2ε > C(Cesp (q))
(2.2.1)
for any n, m ∈ N with m 6= n. Take v1 = y1 . Then, by the absolute continuity of the norm of y1 ,
there exists i1 ∈ N such that
∞
X
v1 (i)ei < ε.
i=i1 +1
JJ J
I II
Go back
Full Screen
Putting z1 =
i1
P
v1 (i)ei , we have
i=1
kz1 + ym k = ky1 + ym −
∞
X
v1 (i)ei k ≥ ky1 + ym k − ε
for any m > 1.
i=i1 +1
Close
Quit
Hence by (2.2.1), we have
kz1 + ym k + 3ε > C(Cesp (q))
for any m > 1.
Since yn (i) → 0 for i = 1, 2, · · · , there exists n2 ∈ N with n2 > n1 such that
i
1
X
yn (i)ei < ε
whenever n ≥ n2 .
i=1
Define v2 = yn2 . Then there is i2 > i1 such that
∞
X
v2 (i)ei < ε.
i=i2 +1
Taking z2 =
i2
P
v2 (i)ei , we obtain
i=i1 +1
kz1 + z2 k = ky1 −
∞
X
v1 (i)ei + yn2 −
i=i1 +1
i1
X
v2 (i)ei −
i=1
∞
X
v2 (i)ei
i=i2 +1
≥ ky1 + yn2 k − 3ε.
JJ J
I II
Hence and by (2.2.1), we immediately obtain
Go back
kz1 + z2 k + 5ε > C(Cesp (q)).
Full Screen
k−1
k−1
Suppose that increasing sequences (ij )k−1
j=1 , (nj )j=1 of natural numbers and a sequence (zj )j=1 of
elements of Cesp (q) are already defined and
Close
Quit
kzn + zm k + 6ε > C(Cesp (q))
for m, n ∈ {1, 2, . . . , k − 1}, m 6= n.
Since yn (i) → 0 for i = 1, 2, · · · , there exists a natural number nk > nk−1 such that
i
k−1
X
yn (i)ei < ε
i=1
provided n ≥ nk . Put v = ynk . Then there is ik > ik−1 such that
∞
X
vk (i)ei < ε.
i=ik +1
Defining zk =
ik
P
vk (i)ei , we obtain
ij−1
ik−1
∞
∞
X
X
X
X
kzj + zk k = y
−
v
(i)e
v
(i)e
−
v
(i)e
+
y
−
v
(i)e
−
k
i
j
i
j
i
nk
k
i
nj
i=1
ij+1
i=1
ik+1
i=ik−1 +1
≥ kynj + ynk k − 4ε
for j = 1, 2, . . . , k − 1.
Hence, by (2.2.1), we obtain
JJ J
I II
Go back
Full Screen
kzj + zk k + 6ε > C(Cesp (q))
for j = 1, 2, · · · , k − 1.
Using the induction principle, we can find a sequence (zn ) satisfying the following conditions:
P
(1) zn =
vn (i)ei , where 0 = i0 < i1 < i2 < . . . ;
i=in−1 +1
(2) kzn + zm k + 6ε > C(Cesp (q)) for m, n, ∈ N, m 6= n;
Close
(3) kzn k ≤ 1 for n = 1, 2, . . .;
Quit
ω
(4) zn →0 as n → ∞.
Define un = zn /kzn k for each n ∈ N. Then every un ∈ S(Cesp (q)) and
zn
zm ≥ kzn + zm k ≥ C(Cesp (q)) − 6ε
kun + um k = +
kzn k kzm k for any m, n ∈ N, m 6= n. By the arbitrariness of ε, we have C(Cesp (q)) = K.
Let ε > 0 be given. Take nε ∈ N such that
p
∞
X
a
< ε,
Qk
k=inε +1
where
a=
inε
X
|qi unε (i)|.
i=inε −1 +1
Hence for any m > nε , we have
!p
!p
k
∞
k
X
X
1
1 X
|qi unε (i)| +
(a +
|qi um (i)|
+ um k =
Qk i=1
Qk
i=inε −1 +1
i=1
k=im−1 +1
!p
!p
im−1
k
∞
k
X
X
1 X
1 X
≥
|qi unε (i)| +
(
|qi um (i)|
Qk i=1
Qk i=1
i=inε −1 +1
k=im−1 +1
!p ∞ !p
p
∞
k
∞
k
X
X
X
1 X
a
1 X
=
+
(
|qi um (i)|
|qi unε (i)| −
Qk i=1
Qk
Qk i=1
i=i
+1
im−1
kunε
JJ J
I II
Go back
Full Screen
Close
Quit
p
X
nε −1
> 1 − ε + 1 = 2 − ε,
k=im−1
k=im−1 +1
i. e. A(un ) ≥ (2 − ε)) ≥ (2 − ε)1/p .
On the other hand, for ε mentioned above, by Lemma 1.1, there exists δ > 0 such that
k|x + ykp − kxkp | < ε
whenever kxkp ≤ 1 and kykp < δ. Take nδ ∈ N such that
∞
X
(
k=inδ +1
a p
) < δ,
Qk
and
inδ
X
a=
|qi unδ (i)|.
i=inδ −1 +1
Hence for any m > nδ , we have
kunδ + um kp
im−1
X
=
i=inδ −1 +1
∞
X
≤
JJ J
I II
Go back
i=inδ −1 +1
!p
k
∞
X
1 X
|qi unδ (i)| +
Qk i=1
k=im−1 +1
!
p
k
∞
X
1 X
|qi unδ (i)| +
Qk i=1
= kunδ (i)kp +
k=im−1 +1
∞
X
k=im−1 +1
Full Screen
−
Close
Quit
∞
X
k1=im−1 +1
< 2 + ε,
k
X
k
X
1
(a +
|qi um (i)|)
Qk
i=1
!p
k
X
1
(a +
|qi um (i)|)
Qk
i=1
!p
!p
1
(a +
|qi um (i)|)
Qk
i=1
!p
k
1 X
|qi um (i)|
Qk i=1
+ kum kp
i. e. A((un )) ≤ (2 + ε)1/p .
Since ε was arbitrary, we obtain C(Cesp (q)) = 21/p .
Corollary 2.1. The space Cesp (q)) satisfies the Banach-Saks property .
Proof. The proof follows immediately from the above Theorem 1 of [3] and Theorem 2.2.
Acknowledgment. The author is grateful to Prof. Mursaleen and to the referee for their
valuable suggestions.
JJ J
I II
1. Banach S.and Saks S.; Sur la convergence forte dans les champs Lp , Studia Math. 2 (1930), 51–57.
2. Clarkson J. A., Uniformly convex spaces, Trans. Amer. Math. Soc. 40 (1936), 396–414.
3. Cui Y. A. and Meng C. and Pluciennik, R., Banach-Saks property and property (β) in Cesàro Sequence Space,
Southeast Asian Bull. Math. 24 (2000), 201–210.
4. Cui Y. A. and Hudzik H., On the Banach-Saks and weak Banach-Saks properties of some Banach sequence
spaces, Acta Sci. Math. (Szeged), 65 (1999), 179–187.
5. Diestel J., Sequences and series in Banach spaces, Graduate Texts in Math. 92, Springer- -Verlag, 1984.
6. Kakutani S., Weak convergence in Uniformly convex Banach spaces, Tohoku Math. J. Math. 45 (1938), 188–
193.
7. Kotmann C. A., Packing and reflexivity in Banach Spaces, Trans. Amer. Math. Soc. 150 (1970), 565–576.
8. Lee. P. Y., Cesàro sequence Spaces, Math cronical, New Zeland 13 (1984), 29–45.
9. Shue J. S., Cesàro sequence spaces, Tamkang J. Math. 1 (1970), 143–150.
Go back
Full Screen
Close
Quit
V. A. Khan, Department of Mathematics, A. M. U. Aligarh-202002, India,
e-mail: vakhan@math.com