Asma Bektaş / J Zhejiang Univ SCIENCE A 2006 7(12):2093-2096 2093

Journal of Zhejiang University SCIENCE A

ISSN 1009-3095 (Print); ISSN 1862-1775 (Online)
www.zju.edu.cn/jzus; www.springerlink.com
E-mail: [email protected]

On some difference sequence spaces defined by a sequence

of Orlicz functions

ASMA BEKTAŞ Çiğdem

(Department of Mathematics, Firat University, Elazig 23119, Turkey)
E-mail: [email protected]; [email protected]
Received Oct. 20, 2005; revision accepted July 15, 2006

Abstract: The idea of difference sequence spaces was introduced in (Kızmaz, 1981) and this concept was generalized in (Et and
Çolak, 1995). In this paper we define some difference sequence spaces by a sequence of Orlicz functions and establish some
inclusion relations.

Key words: Difference sequence, Orlicz function, Sequence of Orlicz functions, Strongly almost convergent
doi:10.1631/jzus.2006.A2093 Document code: A CLC number: O34

INTRODUCTION  1 n 
[cˆ, p]0 =  x =( xk ):lim ∑ | xk + s | pk =0, uniformly in s  ,
n n
 k =1 
Let w be the set of all sequences of real or com-
 1 n 
plex numbers and R∞, c and c0 be the sequence spaces [cˆ, p]∞ =  x = ( xk ):sup ∑ xk + s
pk
< ∞.
of bounded, convergent and null sequences x=(xk),  s , n n k =1 
respectively.
A sequence x∈R∞ is said to be almost convergent Kızmaz (1981) defined the sequence spaces
(Lorentz, 1948) if all Banach limits of x coincide.
X(∆)={x=(xk):(∆xk)∈X}
Lorentz (1948) proved that
for X=R∞, c or c0, where ∆x=(∆xk)=(xk−xk+1).
 1 n 
cˆ =  x = ( xk ) : lim ∑ xk + s exists, uniformly in s  . After Et and Çolak (1995) generalized the above
n n
 k =1  sequence spaces to the sequence spaces

Maddox (1967; 1978) has defined x to be X(∆m)={x=(xk):(∆mxk)∈X}

strongly almost convergent to a number L if
for X=R∞, c or c0, where m∈ù, ∆0x=(xk), ∆x=(xk−xk+1),
1 n ∆mx=(∆mxk)=(∆m−1xk−∆m−1xk+1) and so that
lim
n
∑ xk + s − L = 0, uniformly in s.
n k =1
m
m
∆ m xk = ∑ ( −1)v   xk + v .
Let p=(pk) be a sequence of strictly positive real v =0 v
numbers. Nanda (1984) defined
An Orlicz function is a function M: [0,∞ ) →
 1 n
 [0,∞ ) which is continuous, non-decreasing and con-
[cˆ, p]=  x=(xk ):lim ∑| xk + s − L | pk =0, uniformly in s  ,
n n
 k =1  vex with M(0)=0, M(x)>0 for x>0 and M(x)→∞ as
2094 Asma Bektaş / J Zhejiang Univ SCIENCE A 2006 7(12):2093-2096

x→∞. c0 ( M, ∆) = { x = ( xk ) : ( M k (| ∆xk | / ρ )) → 0, k → ∞} .
It is well known that if M is a convex function
and M(0)=0, then M(λx)≤λM(x) for all λ with 0<λ<1.
Lindenstrauss and Tzafriri (1971) used the idea SOME NEW SEQUENCE SPACES DEFINED BY
of Orlicz function to construct the sequence space A SEQUENCE OF ORLICZ FUNCTIONS

 ∞

RM =  x ∈ w : ∑ M (| xk | / ρ ) < ∞, for some ρ >0 . Definition 1 Let M=(Mk) be a sequence of Orlicz
 k =1  functions and p=(pk) be any sequence of strictly
positive real numbers. We define the following se-
The space RM with the norm
quence sets

 ∞
 [cˆ, M, p ](∆ m )
|| x ||= inf  ρ > 0 : ∑ M (| xk | / ρ ) ≤ 1
 k =1   1 n
=  x = ( xk ) : lim ∑ [ M k (| ∆ m xk + s − L | /ρ )] pk = 0,
n n
becomes a Banach space which is called an Orlicz  k =1

sequence space. The space RM is closely related to the 

uniformly in s, for some ρ >0 and L >0 ,
space Rp which is an Orlicz sequence space with 
M(x)=xp for 1≤p<∞. [cˆ, M, p ]0 (∆ m )
Let M be an Orlicz function and p=(pk) be any
 1 n
sequence of strictly positive real numbers. Güngör =  x = ( xk ) : lim ∑ [ M (| ∆ m xk + s | / ρ )] pk = 0,
n n
and Et (2003) defined the following sequence spaces  k =1


uniformly in s, for some ρ >0  ,
[cˆ, M , p](∆ m ) 
 1 n [cˆ, M, p ]∞ (∆ m )
=  x = ( xk ) : lim ∑ [ M (| ∆ m xk + s − L | / ρ )] pk = 0,
n n
 k =1
 1 n
=  x = ( xk ) : sup ∑ [ M (| ∆ m xk + s | / ρ )] pk < ∞ ,
  s , n n k =1
uniformly in s, for some ρ >0 and L >0 ,
 
for some ρ >0  .
[cˆ, M , p]0 (∆ m ) 
 1 n
=  x = ( xk ) : lim ∑ [ M (| ∆ m xk + s | / ρ )] pk = 0,
 n n
k =1 If Mk(x)=x for every k, then [cˆ, M, p](∆m)= [cˆ,
 p](∆m), [cˆ, M, p]0(∆m)= [cˆ, p]0(∆m) and [cˆ, M, p]∞(∆m)
uniformly in s, for some ρ >0  ,
 = [cˆ, p]∞(∆m). We denote [cˆ, M, p](∆m), [cˆ, M, p]0(∆m)
[cˆ, M , p]∞ (∆ m ) and [cˆ, M, p]∞(∆m) by [cˆ, M](∆m), [cˆ, M]0(∆m) and
 1 n [cˆ, M]∞(∆m), respectively, when pk=1 for all k.
=  x = ( xk ) : sup ∑ [ M (| ∆ m xk + s | / ρ )] pk < ∞ ,
 s , n n k =1 Theorem 1 Let M=(Mk) be a sequence of Orlicz
 functions. Then the following statements are equiva-
for some ρ > 0  , lent:

(i) [cˆ, p]∞(∆m)⊆ [cˆ, M, p]∞(∆m);
Let M=(Mk) be a sequence of Orlicz functions. (ii) [cˆ, p]0(∆m)⊆ [cˆ, M, p]∞(∆m);
Mursaleen et al.(2001) defined the following se- 1 n
quence spaces (iii) sup
n
∑
n k =1
[ M k (t / ρ )] pk < ∞ (t, ρ>0).

{ }
Proof (i)⇒(ii) is obvious, since [cˆ, p]0(∆m)⊆ [cˆ,
R∞ (M,∆) = x = ( xk ) : sup( M k (| ∆xk | / ρ )) < ∞ ,
k p]∞( ∆m).
 Asma Bektaş / J Zhejiang Univ SCIENCE A 2006 7(12):2093-2096 2095

(ii)⇒(iii). Let [cˆ, p]0(∆m)⊆ [cˆ, M, p]∞(∆m). Sup- 1 n

pose that (iii) is not satisfied. Then for some t, ρ>0
inf
n
∑[M k (t / ρ )] pk = 0 (t, ρ>0).
n k =1
(3)

1 n We can choose an index sequence (ni) such that

sup
n
∑ [M k (t / ρ )] pk = ∞,
n k =1
ni
1
and therefore there is sequence (ni) of positive inte- ni
∑[M
k =1
k (i / ρ )] pk < i −1 , i = 1, 2, ...
gers such that
Define the sequence x=(xk) by
ni
1
ni
∑[M k (i −1 / ρ )] pk > i, i = 1, 2, ... (1) i, 1 ≤ k ≤ ni , i = 1, 2, ...;
k =1 xk = 
0, k > ni .
Define x=(xk) by
Thus by Eq.(3), x∈ [cˆ, M, p]0(∆m) but x∉ [cˆ, p]∞(∆m)
i , 1 ≤ k ≤ ni , i = 1, 2, ...;
−1
which contradicts (II). Hence (III) must hold.
xk = 
0, k > ni . (III)⇒(I). Let (III) hold and x∈ [cˆ, M, p]0(∆m),
i.e.,
Then x∈ [cˆ, p]0(∆m), but by Eq.(1), x∉ [cˆ, M, p]∞(∆m)
which contradicts (ii). Hence (iii) must hold. 1 n
(iii)⇒(i). Let (iii) be satisfied and x∈ [cˆ, p]∞(∆m). lim
n
∑[ M k (| ∆ m xk + s | / ρ )] pk = 0, uniformly in s. (4)
n k =1
Suppose that x∉ [cˆ, M, p]∞(∆m). Then
Suppose that x∉ [cˆ, p]0(∆m). Then for some
1 n
sup
s ,n
∑ [ M k (| ∆ m xk + s | / ρ )] pk = ∞.
n k =1
(2) number ε0>0 and index n0, we have

|∆mxk+s|≥ε0, for some s≥s′ and 1≤k≤n0.

Let t=|∆mxk+s| for each k and fixed s, then by
Eq.(2) Therefore

1 n [ M k (ε 0 / ρ )] pk ≤ [ M k (| ∆ m xk + s | / ρ )] pk
sup
n
∑ [M k (t / ρ )] pk = ∞,
n k =1 and consequently by Eq.(4)
1 n
which contradicts (iii). Hence (i) must hold. lim ∑ [ M k (ε 0 / ρ )] pk = 0,
n n
k =1
Theorem 2 Let 1≤pk≤ sup pk <∞. Then the following
k

statements are equivalent for a sequence of Orlicz which contradicts (III). Hence
functions M=(Mk):
(I) [cˆ, M, p]0(∆m)⊆ [cˆ, p]0(∆m); [cˆ, M, p]0(∆m)⊆ [cˆ, p]0(∆m).
(II) [cˆ, M, p]0(∆m)⊆ [cˆ, p]∞(∆m);
Theorem 3 Let 1≤pk≤ sup pk <∞. The inclusion [cˆ,
1 n
∑[M k (t / ρ )] pk > 0 (t, ρ>0).
k
(II) inf
n n k =1 M, p]∞(∆m)⊆ [cˆ, p]0(∆m) hold if
Proof (I)⇒(II) is obvious.
(II)⇒(III). Let [cˆ, M, p]0(∆m)⊆ [cˆ, p]∞(∆m). 1 n
Suppose that (III) does not hold. Then
lim
n
∑
n k =1
[ M k (t / ρ )] pk = ∞ (t, ρ>0). (5)
2096 Asma Bektaş / J Zhejiang Univ SCIENCE A 2006 7(12):2093-2096

Proof Let [cˆ, M, p]∞(∆m)⊆ [cˆ, p]0(∆m). Suppose that Eq.(8) does not hold. Then for some t0>0,
Eq.(5) does not satisfied. Therefore there is a number
t0>0 and an index sequence (ni) such that 1 n
lim
n
∑
n k =1
[ M k (t0 / ρ )] pk = L ≠ 0. (9)
ni
1
ni
∑[M k (t0 / ρ )] pk ≤ N < ∞, i=1, 2, … (6)
k =1 k −m
 m + k − v − 1
Define x=(xk) by xk = t0 ∑ (−1)m  
v=0  k −v 
Define the sequence x=(xk) by
m
for k=1, 2, … Thus x∉ [c, M, p]0(∆ ) by Eq.(9), but
ˆ
t0 , 1 ≤ k ≤ ni , i = 1, 2, ...; x∈ [cˆ, p]∞(∆m). Hence Eq.(8) must hold.
xk = 
0, k > ni . Conversely, suppose that Eq.(8) hold and x∈
[cˆ, p]∞(∆m). Then for every k and s
Thus by Eq.(6), x∈ [cˆ, M, p]∞(∆m), but x∉ [cˆ, p]0(∆m).
Hence Eq.(5) must hold. |∆mxk+s|≤N<∞.
Conversely, let Eq.(5) be satisfied. If x∈ [cˆ, M,
p]∞(∆m), then for each s and n Therefore

1 n [ M k (| ∆ m xk + s | / ρ )] pk ≤ [ M k ( N / ρ )] pk ,
∑
n k =1
[ M k (| ∆ m xk + s | / ρ )] pk ≤ N < ∞. (7)
and
1 n
Suppose that x ∉ [cˆ, p]0(∆m). Then for some
lim
n
∑ [ M k (| ∆ m xk + s | / ρ )] pk
n k =1
number ε0>0 there is a number s0 and index n0 1 n
≤ lim
n
∑ [ M k ( N / ρ )]pk = 0
n k =1
m
|∆ xk+s|≥ε0, for s≥s0.
by Eq.(8). Hence x∈ [cˆ, M, p]0(∆m).
Therefore
References
[ M k (ε 0 / ρ )] pk ≤ [ M k (| ∆ m xk + s | / ρ )] pk , Et, M., Çolak, R., 1995. On some generalized difference se-
quence spaces. Soochow J. of Math., 21(4):377-386.
and hence for each k and s we get Güngör, M., Et, M., 2003. ∆r-strongly almost summable
sequences defined by Orlicz functions. Indian J. Pure
Appl. Math., 34(8):1141-1151.
1 n
∑
n k =1
[ M k (ε 0 / ρ )] pk ≤ N < ∞, Kızmaz, H., 1981. On certain sequence spaces. Canad. Math.
Bull., 24(2):169-176.
Lindenstrauss, J., Tzafriri, L., 1971. On Orlicz sequence
spaces. Israel J. Math., 10:379-390.
for some N>0, by Eq.(7) which contradicts Eq.(5).
Lorentz, G.G., 1948. A contribution to the theory of divergent
Hence [cˆ, M, p]∞(∆m)⊆ [cˆ, p]0(∆m). sequences. Acta Mathematica, 80(1):167-190.
Theorem 4 Let 1≤pk≤ sup pk <∞. Then the inclusion [doi:10.1007/ BF02393648]
k Maddox, I.J., 1967. Spaces of strongly summable sequences.
[cˆ, p]∞(∆m)⊆ [cˆ, M, p]0(∆m) hold if Quart. J. Math., 18:345-355.
Maddox, I.J., 1978. A new type of convergence. Math. Proc.
Camb. Phil. Soc., 83:61-64.
1 n Mursaleen, Mushir, Qamaruddin, 2001. Some difference se-
lim
n
∑[M k (t0 / ρ )] pk = 0 (t, ρ>0).
n k =1
(8) quence spaces defined by Orlicz functions. Matimyas
Matematica, 24(1):42-47.
Nanda, S., 1984. Strongly almost convergent sequences. Bull.
Proof Let [cˆ, p]∞(∆m)⊆ [cˆ, M, p]0(∆m). Suppose that Cal. Math. Soc., 76:236-240.