Contents lists available atSciVerse ScienceDirect
Mathematical and Computer Modelling
journal homepage:www.elsevier.com/locate/mcm
Some I-convergent lambda-summable difference sequence spaces of fuzzy real numbers defined by a sequence of Orlicz functions
Bipan Hazarika
a,∗, Ekrem Savas
baDepartment of Mathematics, Rajiv Gandhi University, Rono Hills, Doimukh-791112, Arunachal Pradesh, India
bDepartment of Mathematics, Istanbul Ticaret University, Uskudar-Istanbul, Turkey
a r t i c l e i n f o
Article history:
Received 13 June 2011 Accepted 24 July 2011
Keywords:
Ideal I-convergent Orlicz function Fuzzy number Difference space
a b s t r a c t
In this paper we introduce certain new sequence spaces of fuzzy numbers defined by I- convergence using sequences of Orlicz functions and a difference operator of order m. We study some basic topological and algebraic properties of these spaces. Also we investigate the relations related to these spaces.
© 2011 Elsevier Ltd. All rights reserved.
1. Introduction
The concept of fuzzy sets was initially introduced by Zadeh [1]. It has a wide range of applications in almost all branches of study, in particular in science, where mathematics is used. Now the notion of fuzziness is used by many researchers in cybernetics, artificial intelligence, expert systems and fuzzy control, pattern recognition, operations research, decision making, image analysis, projectiles, probability theory, agriculture, weather forecasting, etc. The fuzziness of all the subjects of mathematical sciences has been investigated. It has attracted many workers on sequence spaces and summability theory to introduce different types of fuzzy sequence spaces and to study their different properties. Our studies are based on the linear spaces of sequences of fuzzy numbers which are very important for higher level studies in quantum mechanics, particle physics and statistical mechanics, etc. Different classes of sequences of fuzzy real numbers have been discussed by Nanda [2], Nuray and Savas [3], Matloka [4], Altinok et al. [5], Colak et al. [6] and many others.
The notion of I-convergence was initially introduced by Kostyrko et al. [7]. Later on, it was further investigated from the sequence space point of view and linked with the summability theory by ˘Salàt et al. [8,9], Tripathy and Hazarika [10,11], Kumar and Kumar [12], Savas [13,14] and many other authors.
Let X be a non-empty set, then a family of sets I
⊂
2X(the class of all subsets of X ) is called an ideal if and only if for each A,
B∈
I, we have A∪
B∈
I and for each A∈
I and each B⊂
A,
we have B∈
I. A non-empty family of sets F⊂
2X is a filter on X if and only ifΦ̸∈
F , for each A,
B∈
F , we have A∩
B∈
F and for each A∈
F and each A⊂
B, we have B∈
F . An ideal I is called a non-trivial ideal if I̸=
Φand X̸∈
I. Clearly I⊂
2Xis a non-trivial ideal if and only if F=
F(
I) = {
X−
A:
A∈
I}
is a filter on X . A non-trivial ideal I⊂
2X is called admissible if and only if{{
x} :
x∈
X} ⊂
I. A non-trivial ideal I is maximal if there cannot exist any non-trivial ideal J̸=
I containing I as a subset. Further details on ideals of 2Xcan be found in [7].Let
w
F be the set of all sequences of fuzzy numbers. The operator∆n: w
F→ w
Fis defined by(
∆0X)
k=
Xk; (
∆1X)
k=
∆Xk
=
Xk−
Xk+1; (
∆nX)
k=
∆nXk=
∆n−1Xk−
∆n−1Xk+1, (
n≥
2)
for all n∈
N. The generalized difference has the following∗Corresponding author. Tel.: +91 360 2278512; fax: +91 360 2277881.
E-mail addresses:bh_rgu@yahoo.co.in(B. Hazarika),ekremsavas@yahoo.com(E. Savas).
0895-7177/$ – see front matter©2011 Elsevier Ltd. All rights reserved.
doi:10.1016/j.mcm.2011.07.026
B. Hazarika, E. Savas / Mathematical and Computer Modelling 54 (2011) 2986–2998 2987
binomial expression for n
≥
0,∆nxk
=
n
−
ν=0
nν (−
1)
νxk+ν.
(1)Recall in [15] that an Orlicz function M is a continuous, convex, nondecreasing function defined for x
>
0 such that M(
0) =
0 and M(
x) >
0. If the convexity of the Orlicz function is replaced by M(
x+
y) ≤
M(
x) +
M(
y)
then this function is called the modulus function and was characterized by Ruckle [16]. An Orlicz function M is said to satisfy∆2-condition for all values of u, if there exists K>
0 such that M(
2u) ≤
KM(
u),
u≥
0.Lemma 1.1. Let M be an Orlicz function which satisfies the∆2- condition and let 0
< δ <
1. Then for each t≥ δ
, we have M(
t) <
Kδ
−1M(
2)
for some constant K>
0.Two Orlicz functions M1and M2are said to be equivalent if there exist positive constants
α, β
and x0such that M1(α) ≤
M2(
x) ≤
M1(β)
for all x with 0
≤
x<
x0.Lindenstrauss and Tzafriri [17] studied some Orlicz type sequence spaces defined as follows:
ℓ
M=
(
xk) ∈ w :
∞
−
k=1
M
|
xk| ρ
< ∞,
for someρ >
0 .
The space
ℓ
Mwith the norm‖
x‖ =
inf ρ >
0:
∞
−
k=1
M
|
xk| ρ
≤
1
becomes a Banach space which is called an Orlicz sequence space. The space
ℓ
Mis closely related to the spaceℓ
pwhich is an Orlicz sequence space with M(
t) = |
t|
p, for 1≤
p< ∞
.The following well-known inequality will be used throughout the article. Let p
= (
pk)
be any sequence of positive real numbers with 0≤
pk≤
supkpk=
G,
H=
max{
1,
2G−1}
then|
ak+
bk|
pk≤
H(|
ak|
pk+ |
bk|
pk)
for all k
∈
N and ak,
bk∈
C. Also|
a|
pk≤
max{
1, |
a|
G}
for all a∈
C.In a later stage different classes of Orlicz sequence spaces were introduced and studied by Parashar and Choudhary [18], Tripathy and Sarma [19], Mursaleen and Basarir [20] and many others.
Throughout the article
w
F denotes the class of all fuzzy real-valued sequence spaces. Also N and R denote the set of positive integers and set of real numbers respectively. The zero sequence is denoted byθ
.In this paper we study some new sequence spaces of fuzzy numbers defined by I-convergence using the sequence of Orlicz functions and the difference operator. We establish an inclusion relation between the sequence spaces
w
λI(F)[
M,
∆m,
p] , w
Iλ(F)[
M,
∆m,
p]
0, w
λF[
M,
∆m,
p]
∞andw
λI(F)[
M,
∆m,
p]
∞where p= (
pk)
denote the sequence of positive real numbers for all n∈
N, a non-decreasing sequenceλ = (λ
n)
of positive real numbers such thatλ
n+1≤ λ
n+
1, λ
1=
1, λ
n→ ∞ (
n→ ∞ )
and M= (
Mk)
are sequences of Orlicz functions. In addition, we study some algebraic and topological properties of these spaces.2. Definitions and notations
We now give a brief introduction about the sequences of fuzzy real numbers. Let D denote the set of all closed and bounded intervals X
= [
x1,
x2]
on the real line R. For X,
Y∈
D, we define X≤
Y if and only if x1≤
y1and x2≤
y2,d
(
X,
Y) =
max{|
x1−
y1| , |
x2−
y2|} ,
where X= [
x1,
x2]
and Y= [
y1,
y2] .
Then it can be easily seen that d defines a metric on D and
(
D,
d)
is a complete metric space (see [4]). Also the relation‘‘
≤
’’ is a partial order on D. A fuzzy number X is a fuzzy subset of the real line R i.e. a mapping X:
R→
J(=[
0,
1] )
associating each real number t with its grade of membership X(
t)
. A fuzzy number X is convex if X(
t) ≥
X(
s) ∧
X(
r) =
min{
X(
s),
X(
r)}
, where s<
t<
r. If there exists t0∈
R such that X(
t0) =
1, then the fuzzy number X is called normal. A fuzzy number X is said to be upper semicontinuous if for eachϵ >
0,
X−1([
0,
a+ ϵ))
for all a∈ [
0,
1]
is open in the usual topology in R. Let R(
J)
denote the set of all fuzzy numbers which are upper semicontinuous and have compact support, i.e. if X∈
R(
J)
the for anyα ∈ [
0,
1] , [
X]
αis compact, where[
X]
α= {
t∈
R:
X(
t) ≥ α,
ifα ∈ [
0,
1]} ,
[
X]
0=
closure of({
t∈
R:
X(
t) > α,
ifα =
0} ).
The set R of real numbers can be embedded in R
(
J)
if we define r∈
R(
J)
by r(
t) =
1,
if t=
r;
0,
if t̸=
r.
The additive identity and multiplicative identity of R
(
J)
are defined by 0 and 1 respectively.The arithmetic operations on R
(
J)
are defined as follows:(
X⊕
Y)(
t) =
sup{
X(
s) ∧
Y(
t−
s)},
t∈
R, (
X⊖
Y)(
t) =
sup{
X(
s) ∧
Y(
s−
t)},
t∈
R, (
X⊗
Y)(
t) =
sup
X(
s) ∧
Y
t s
,
t∈
R,
X Y
(
t) =
sup{
X(
st) ∧
Y(
s)},
t∈
R.
Let X
,
Y∈
R(
J)
and theα
-level sets be[
X]
α= [
xα1,
xα2] , [
Y]
α= [
yα1,
yα2] , α ∈ [
0,
1]
. Then the above operations can be defined in terms ofα
-level sets as follows:[
X⊕
Y]
α= [
xα1+
yα1,
xα2+
yα2] , [
X⊖
Y]
α= [
xα1−
yα1,
xα2−
yα2] , [
X⊗
Y]
α=
[
mini∈{1,2}xαiyαi
,
maxi∈{1,2}xαiyαi
]
,
[
X−1]
α= [ (
xα2)
−1, (
xα1)
−1] ,
xαi>
0,
for each 0< α ≤
1.
For r∈
R and X∈
R(
J)
, the product rX is defined as follows:rX
(
t) =
X(
r−1t),
if r̸=
0;
0,
if r=
0.
The absolute value,
|
X|
of X∈
R(
J)
is defined by (for details see [21])|
X| (
t) =
max{
X(
t),
X(−
t)},
if t≥
0;
0
,
if t<
0.
Define a mappingd
¯ :
R(
J) ×
R(
J) →
R+∪ {
0}
by d¯ (
X,
Y) =
sup0≤α≤1
d
([
X]
α, [
Y]
α).
It is known that
(
R(
J), ¯
d)
is a complete metric space (for details see [21]).A metric on R
(
J)
us said to be translation invariant if d¯ (
X+
Z,
Y+
Z) = ¯
d(
X,
Y),
for X,
Y,
Z∈
R(
J).
A sequence X
= (
Xk)
of fuzzy numbers is said to converge to a fuzzy number X0if for everyϵ >
0, there exists a positive integer n0such thatd¯ (
Xk,
X0) < ϵ
for all n≥
n0.A sequence X
= (
Xk)
of fuzzy numbers is said to be bounded if the set{
Xk:
k∈
N}
of fuzzy numbers is bounded.A sequence X
= (
Xk)
of fuzzy numbers is said to be I-convergent to a fuzzy number X0if for eachϵ >
0 such that A= {
k∈
N: ¯
d(
Xk,
X0) ≥ ϵ} ∈
I.
The fuzzy number X0is called the I-limit of the sequence
(
Xk)
of fuzzy numbers and we write I−
lim Xk=
X0. A sequence X= (
Xk)
of fuzzy numbers is said to be I-bounded if there exists M>
0 such that{
k∈
N: ¯
d(
Xk, ¯
0) >
M} ∈
I.
Let EF denote the sequence space of fuzzy numbers.
A sequence space EFis said to be solid (or normal) if
(
Yk) ∈
EFwhenever(
Xk) ∈
EFand|
Yk| ≤ |
Xk|
for all k∈
N.A sequence space EFis said to be symmetric if
(
Xk) ∈
EFimplies(
Xπ(k)) ∈
EFwhereπ
is a permutation of N.A sequence space EFis said to be monotone if EFcontains the canonical pre-images of all its step spaces.
Example 2.1. If we take I
=
If= {
A⊆
N:
A is a finite subset}
. Then If is a non-trivial admissible ideal of N and the corresponding convergence coincides with the usual convergence.Example 2.2. If we take I
=
Iδ= {
A⊆
N: δ(
A) =
0}
whereδ(
A)
denotes the asymptotic density of the set A,
then Iδis a non-trivial admissible ideal of N and the corresponding convergence coincides with the statistical convergence.Lemma 2.1. A sequence space EFis normal implies that EF is monotone. (For the crisp set case, one may refer to Kamthan and Gupta [22, page 53]).
Lemma 2.2 (Kostyrko et al. [7, Lemma 5
.
1]). If I⊂
2Nis a maximal ideal, then for each A⊂
N we have either A∈
I or N−
A∈
I.B. Hazarika, E. Savas / Mathematical and Computer Modelling 54 (2011) 2986–2998 2989
3. Some new sequence spaces of fuzzy numbers
In this section using the sequence of Orlicz functions and difference operator∆mdefined by I-convergence; we introduce the following new sequence spaces and examine some properties of the resulting sequence spaces. Let I be an admissible ideal of N and let p
= (
pk)
be a sequence of positive real numbers for all k∈
N, and a nondecreasing sequenceλ = (λ
n)
of positive real numbers such thatλ
n+1≤ λ
n+
1, λ
1=
1, λ
n→ ∞ (
n→ ∞ )
. Let M= (
Mk)
be a sequence Orlicz functions and X= (
Xk)
be a sequence of fuzzy numbers, we define the following sequence spaces as:w
λI(F)[
M,
∆m,
p] =
(
Xk) ∈ w
F: ∀ ε >
0,
n
∈
N:
1λ
n−
k∈In
[
Mk
d¯ (
∆mXk,
X0) ρ
]
pk≥ ε
∈
I,
for some
ρ >
0 and X0∈
R(
J)
,
w
λI(F)[
M,
∆m,
p]
0=
(
Xk) ∈ w
F: ∀ ε >
0,
n
∈
N:
1λ
n−
k∈In
Mk ¯
d
∆mXk
, ¯
0 ρ
pk≥ ε
∈
I,
for some
ρ >
0 ,
w
λF[
M,
∆m,
p]
∞=
(
Xk) ∈ w
F:
supn
1
λ
n−
k∈In
Mk ¯
d
∆mXk
, ¯
0 ρ
pk< ∞,
for someρ >
0 ,
andw
λI(F)[
M,
∆m,
p]
∞=
(
Xk) ∈ w
F: ∃
K>
0 s.t.
n
∈
N:
1λ
n−
k∈In
Mk ¯
d
∆mXk
, ¯
0 ρ
pk≥
K
∈
I,
for some
ρ >
0 ,
where In= [
n− λ
n+
1,
n]
.From the definitions it is clear that
w
λI(F)[
M,
∆m,
p]
0⊂ w
λI(F)[
M,
∆m,
p] ⊂ w
Fλ[
M,
∆m,
p]
∞.
Also that the inclusions are strict follows from the following example.Example 3.1. Let M
(
x) =
x2for all x∈ [
0, ∞),
pk=
1 for all k∈
N. Consider the sequence(
Xk)
of fuzzy numbers to be defined as follows:For k
=
2i,
i=
1,
2,
3, . . .
.Xk
(
t) =
4k
+
1,
if−
k4
≤
t≤
0;
−
4k
+
1,
if 0<
t≤
k 4;
0
,
otherwiseand Xk
(
t) = ¯
0, otherwise.It is easy to prove that the sequences
(
Xk)
and(
∆mXk)
are bounded but these are not I-convergent.Some classes are obtained by specializing m
,
M, (λ
n),
I and(
pk)
. Here are some examples:(i) If m
=
0, then we obtainw
λI(F)[
M,
p] =
(
Xk) ∈ w
F: ∀ ε >
0,
n
∈
N:
1λ
n−
k∈In
[
Mk
d¯ (
Xk,
X0) ρ
]
pk≥ ε
∈
I,
for some
ρ >
0 and X0∈
R(
J)
,
w
λI(F)[
M,
p]
0=
(
Xk) ∈ w
F: ∀ ε >
0,
n
∈
N:
1λ
n−
k∈In
Mk ¯
d
Xk, ¯
0
ρ
pk≥ ε
∈
I,
for someρ >
0 ,
w
λF[
M,
p]
∞=
(
Xk) ∈ w
F:
supn
1
λ
n−
k∈In
Mk ¯
d
Xk, ¯
0
ρ
pk< ∞,
for someρ >
0
and
w
Iλ(F)[
M,
p]
∞=
(
Xk) ∈ w
F: ∃
K>
0 s.t.
n
∈
N:
1λ
n−
k∈In
Mk ¯
d
Xk, ¯
0
ρ
pk≥
K
∈
I,
for someρ >
0 .
(ii) If Mk
(
x) =
x for all k∈
N, then we obtainw
Iλ(F)[
∆m,
p] =
(
Xk) ∈ w
F: ∀ ε >
0,
n
∈
N:
1λ
n−
k∈In
d¯
∆mXk
,
X0
pk≥ ε
∈
I,
for some X0∈
R(
J)
,
w
Iλ(F)[
∆m,
p]
0=
(
Xk) ∈ w
F: ∀ ε >
0,
n
∈
N:
1λ
n−
k∈In
d¯
∆mXk
,
X0
pk≥ ε
∈
I ,
w
Fλ[
∆m,
p]
∞=
(
Xk) ∈ w
F:
supn
1
λ
n−
k∈In
d¯
∆mXk
,
X0
pk< ∞
and
w
Iλ(F)[
∆m,
p]
∞=
(
Xk) ∈ w
F: ∃
K>
0 s.t.
n
∈
N:
1λ
n−
k∈In
d¯
∆mXk
,
X0
pk≥
K
∈
I .
(iii) If
(λ
n) = (
1,
2,
3, . . .)
, then we obtainw
I(F)[
M,
∆m,
p] =
(
Xk) ∈ w
F: ∀ ε >
0,
n
∈
N:
1 nn
−
k=1
[
Mk
d¯ (
∆mXk,
X0) ρ
]
pk≥ ε
∈
I,
for some
ρ >
0 and X0∈
R(
J)
,
w
I(F)[
M,
∆m,
p]
0=
(
Xk) ∈ w
F: ∀ ε >
0,
n
∈
N:
1 nn
−
k=1
Mk ¯
d
∆mXk
, ¯
0 ρ
pk≥ ε
∈
I,
for some
ρ >
0 ,
w
F[
M,
∆m,
p]
∞=
(
Xk) ∈ w
F:
supn
1 n
n
−
k=1
Mk ¯
d
∆mXk
, ¯
0 ρ
pk< ∞,
for someρ >
0
and
w
I(F)[
M,
∆m,
p]
∞=
(
Xk) ∈ w
F: ∃
K>
0 s.t.
n
∈
N:
1 nn
−
k=1
Mk ¯
d
∆mXk
, ¯
0 ρ
pk≥
K
∈
I,
for some
ρ >
0 .
(iv) If p
= (
pk) = (
1,
1,
1, . . .)
, then we obtainw
Iλ(F)[
M,
∆m] =
(
Xk) ∈ w
F: ∀ ε >
0,
n
∈
N:
1λ
n−
k∈In
[
Mk
d¯ (
∆mXk,
X0) ρ
]
≥ ε
∈
I,
for some
ρ >
0 and X0∈
R(
J)
,
w
Iλ(F)[
M,
∆m]
0=
(
Xk) ∈ w
F: ∀ ε >
0,
n
∈
N:
1λ
n−
k∈In
Mk ¯
d
∆mXk
, ¯
0 ρ
≥ ε
∈
I,
for some
ρ >
0
,
B. Hazarika, E. Savas / Mathematical and Computer Modelling 54 (2011) 2986–2998 2991
w
λF[
M,
∆m]
∞=
(
Xk) ∈ w
F:
supn
1
λ
n−
k∈In
Mk ¯
d
∆mXk
, ¯
0 ρ
< ∞,
for someρ >
0
and
w
λI(F)[
M,
∆m]
∞=
(
Xk) ∈ w
F: ∃
K>
0 s.t.
n
∈
N:
1λ
n−
k∈In
Mk ¯
d
∆mXk
, ¯
0 ρ
≥
K
∈
I,
for some
ρ >
0 .
(v) If I
=
If, then we obtainw
λF[
M,
∆m,
p] =
(
Xk) ∈ w
F:
limn→∞
1
λ
n−
k∈In
[
Mk
d¯ (
∆mXk,
X0) ρ
]
pk=
0,
for someρ >
0 and X0∈
R(
J)
,
w
λF[
M,
∆m,
p]
0=
(
Xk) ∈ w
F:
limn→∞
1
λ
n−
k∈In
Mk ¯
d
∆mXk
, ¯
0 ρ
pk=
0,
for someρ >
0 ,
w
λF[
M,
∆m,
p]
∞=
(
Xk) ∈ w
F:
supn
1
λ
n−
k∈In
Mk ¯
d
∆mXk
, ¯
0 ρ
pk< ∞,
for someρ >
0 .
If X
= (
Xk) ∈ w
Fλ[
M,
∆m,
p]
then we say that X= (
Xk)
is strongly∆mλ(p)-Cesàro convergent with respect to the sequence of Orlicz functions M.4. Main results
In this section we examine the basic topological and algebraic properties of the new sequence spaces and obtain the inclusion relation related to these spaces.
Theorem 4.1. Let
(
pk)
be a bounded sequence. Then the sequence spacesw
Iλ(F)[
M,
∆m,
p] , w
λI(F)[
M,
∆m,
p]
0, andw
λI(F)[
M,
∆m,
p]
∞are linear spaces.Proof. We will prove the result for the space
w
λI(F)[
M,
∆m,
p]
0only and the others can be proved in a similar way.Let X
= (
Xk)
and Y= (
Yk)
be two elements inw
λI(F)[
M,
∆m,
p]
0. Then there existρ
1>
0 andρ
2>
0 such that Aε2
=
n
∈
N:
1λ
n−
k∈In
[
Mk
d¯ (
∆mXk, ¯
0) ρ
1]
pk≥ ε
2
∈
I andBε
2
=
n
∈
N:
1λ
n−
k∈In
[
Mk
d¯ (
∆mYk, ¯
0) ρ
2]
pk≥ ε
2
∈
I.
Let
α, β
be two scalars. By the continuity of the function M= (
Mk)
the following inequality holds:1
λ
n−
k∈In
[
Mk
d¯ (
∆m(α
Xk+ β
Yk, ¯
0))
| α|ρ
1+ | β|ρ
2]
pk≤
D1λ
n−
k∈In
[ | α|
| α|ρ
1+ | β|ρ
2Mk
d¯ (
∆mXk, ¯
0) ρ
1]
pk+
D1λ
n−
k∈In
[ | β|
| α|ρ
1+ | β|ρ
2Mk
d¯ (
∆mYk, ¯
0) ρ
2]
pk≤
DK 1λ
n−
k∈In
[
Mk
d¯ (
∆mXk, ¯
0) ρ
1]
pk+
DK 1λ
n−
k∈In
[
Mk
d¯ (
∆mYk, ¯
0) ρ
2]
pk,
where K
=
max
1,
|α|
|α|ρ1+|β|ρ2
G,
|α|
|β|ρ1+|β|ρ2
G
.From the above relation we obtain the following:
n
∈
N:
1λ
n−
k∈In
[
Mk
d¯ (
∆m(α
Xk+ β
Yk, ¯
0))
| α|ρ
1+ | β|ρ
2]
pk≥ ε
⊆
n
∈
N:
DK 1λ
n−
k∈In
[
Mk
d¯ (
∆mXk, ¯
0) ρ
1]
pk≥ ε
2
∪
n
∈
N:
DK 1λ
m−
k∈In
[
Mk
d¯ (
∆mYk, ¯
0) ρ
2]
pk≥ ε
2
∈
I.
This completes the proof.Remark 4.1. It is easy to verify that the space
w
Fλ[
M,
∆m,
p]
∞is a linear space.Theorem 4.2. The spaces
w
λI(F)[
M,
∆m,
p] , w
Iλ(F)[
M,
∆m,
p]
0, andw
Iλ(F)[
M,
∆m,
p]
∞w
λF[
M,
∆m,
p]
∞ are paranormed spaces (not totally paranormed) with the paranorm g∆defined byg∆
(
X) = −
k∈In
d
¯ (
Xk, ¯
0) +
inf
ρ
pnH:
supk
[
Mk
d¯ (
∆mXk, ¯
0) ρ
]
pk≤
1,
for someρ >
0,
n=
1,
2,
3, . . .
,
where H
=
max{
1,
supkpk}
.Proof. Clearly g∆
(−
X) =
g∆(
X)
and g∆(θ) =
0. Let X= (
Xk)
and Y= (
Yk)
be two elements inw
Iλ(F)[
M,
∆m,
p]
0. Then forρ >
0 we putA1
=
ρ >
0:
supk
Mk
d¯ (
∆mXk, ¯
0) ρ
≤
1
and A2
=
ρ >
0:
supk
Mk
d¯ (
∆mYk, ¯
0) ρ
≤
1 .
Let
ρ
1∈
A1andρ
2∈
A2. Ifρ = ρ
1+ ρ
2then we obtain the following Mk
d¯ (
∆m(
Xk+
Yk), ¯
0) ρ
≤ ρ
1ρ
1+ ρ
2Mk
d¯ (
∆mXk, ¯
0) ρ
1 + ρ
2ρ
1+ ρ
2Mk
d¯ (
∆mYk, ¯
0) ρ
2 .
Thus we havesup
k
[
Mk
d¯ (
∆m(
Xk+
Yk), ¯
0) ρ
]
pk≤
1 andg∆
(
X+
Y) = −
k∈In
d
¯ (
Xk+
Yk, ¯
0) +
inf(ρ
1+ ρ
2)
pnH: ρ
1∈
A1, ρ
2∈
A2
≤ −
k∈In
d
¯ (
Xk, ¯
0) +
inf
ρ
1pnH: ρ
1∈
A1
+ −
k∈In
d
¯ (
Yk, ¯
0) +
inf
ρ
2pnH: ρ
2∈
A2
=
g∆(
X) +
g∆(
Y).
Let tm
→
t where tm,
t∈
C and let g∆(
Xm−
X) →
0 as m→ ∞
. To prove that g∆(
tmXm−
tX) →
0 as m→ ∞
, we putA3
=
ρ
m>
0:
supk
[
Mk
d¯ (
∆mXkm, ¯
0) ρ
m]
pk≤
1
and
A4
=
ρ
m/´>
0:
supk