Engineering and Technology Journal
Vol. 37, Part B, No. 01, 2019
DOI: http://dx.doi.org/10.30684/etj.37.1B.5
Jehad R. Kider
Department of Mathematics
and Computer Applications,
School of Applied Sciences,
University of Technology.
jehadkider@gmail.com
Noor A. Kadhum
Department of Mathematics
and Computer Applications,
School of Applied Sciences,
University of Technology.
Properties of Fuzzy Closed Linear Operator
Abstract- In this paper we recall the definition of fuzzy norm of a fuzzy bounded
linear operator and the fuzzy convergence of sequence of fuzzy bounded linear
operators in order to prove the uniform fuzzy bounded theorem and fuzzy open
mapping theorem. The definition of fuzzy closed linear operators on fuzzy
normed spaces is introduced in order to prove the fuzzy closed graph theorem.
Keywords- Fuzzy bounded linear operator, fuzzy norm of a fuzzy bounded
operator, Fuzzy open mapping, Fuzzy closed graph, Fuzzy closed operator.
Received on: 12/03/2018
Accepted on: 13/12/2018
Published online: 25/04/2019
How to cite this article: J.R. Kider and N.A. Kadhum, “Properties of Fuzzy Closed Linear Operator,”
Engineering and Technology Journal, Vol. 37, Part B, No. 1, pp. 25-31, 2019.
1. Introduction
Zadeh in 1965[1] was the first author who find the
theory of fuzzy set. When Katsaras in 1984 [2]
studying the notion of fuzzy topological vector
spaces he was the first researcher who find the
notion of fuzzy norm on a linear vector space. A
fuzzy metric space was found by Kaleva and
Seikkala in 1984 [3]. The type of fuzzy norm on a
linear space was found by Felbin in 1992 [4] where
the corresponding fuzzy metric is of Kaleva and
Seikkala type. Another type of fuzzy metric space
was found by Kramosil and Michalek in [5]. The
type of fuzzy norm on a linear space was found by
Cheng and Mordeson in 1994 [6] so that the
corresponding fuzzy metric is of Kramosil and
Michalek type. A finite dimensional fuzzy normed
linear spaces was studied by Bag and Samanta in
2003 [7]. Some results on fuzzy complete fuzzy
normed spaces was studied by Saadati and
Vaezpour in 2005 [8]. Fuzzy bounded linear
operators on a fuzzy normed space was studied by
Bag and Samanta in 2005 [9]. The fixed point
theorems on fuzzy normed linear spaces of Cheng
and Mordeson type was proved by Bag and
Samanta in 2006, 2007 [10], [11]. The fuzzy
normed linear space and its fuzzy topological
structure of Cheng and Mordeson type was studied
by Sadeqi and Kia in 2009 [12]. Properties of
fuzzy continuous mapping on a fuzzy normed
linear spaces of Cheng and Mordeson type was
studied by Nadaban in 2015 [13]. The definition of
the fuzzy norm of a fuzzy bounded linear operator
was introduced by Kider and Kadhum in 2017
[14]. Fuzzy functional analysis is developed by the
concepts of fuzzy norm and a large number of
Copyright © 2019 by UOT, IRAQ
researches by different authors have been
published for reference please see [ 16, 17, 18,
19,20, 21, 22, 23].The structure of this paper is as
follows:
In section two we recall the definition of fuzzy
normed space duo to Cheng and Mordeson [6] and
we recall some basic definitions and properties of
this space that we will need it later in this paper. In
section three the definition of three types of fuzzy
convergence sequence of operators is recalled in
order to prove the uniform fuzzy bounded
theorem. Also the fuzzy open mapping theorem is
proved and the fuzzy closed graph theorem is
proved after defining the fuzzy closed linear
operator.
2. Properties of Fuzzy normed space
In this section we recall basic properties of fuzzy
normed space
Definition 2.1: [1]
Suppose that U is any set, a fuzzy set 𝐴̃ in U is
equipped with a membership function, μ𝐴̃ (u): U→
̃ is represented by 𝐴̃ ={(u,μ ̃ (𝑢)):
[0,1]. Then 𝐴
𝐴
uU, 0 ≤ μ𝐴̃ (𝑢) ≤ 1}.
Definition 2.2: [25]
Let ⨂ ∶ [0, 1] × [0, 1 ] → [0, 1] be a binary
operation then ⨂ is called a continuous t -norm
(or triangular norm) if for all α , β , γ , δ ∈
[0, 1] it has the following properties
(1)α⨂β = β⨂α,
(2)α⨂1 = α,
(3)(α ⨂ β)⨂γ = α ⨂(β⨂γ)
(4)If α ≤ β and γ ≤ δ then α⨂γ ≤ β⨂δ
Remark 2.3: [22]
(1)If α > 𝛽 then there is 𝛾 such that α⨂γ ≥ β
52
Vol. 37, Part B. No. 01, 2019
Engineering and Technology Journal
(2)There
is
𝛿
such
that
δ⨂δ ≥
σ where α , β , γ , δ, σ ∈ [0,1]
Definition 2.4: [8]
The triple (V, L, ⨂ ) is said to be a fuzzy normed
space if V is a vector space over the field 𝔽, ⨂ is
a t-norm and L: V × [0 , ∞ ) →[0,1] is a fuzzy set
has the following properties for all a, b∈ V and
α , β > 0.
1-L(a , α) > 0
2-L( a , α) = 1 ⇔ a = 0
α
3-L(ca , α) = L (a, | | ) for all c ≠ 0 ∈ 𝔽
c
4-L(a , α)⨂L( b , β ) ≤ L (a + b , α + β)
5-L(a , . ): [0, ∞) → [0,1] is continuous function
of 𝛼.
6- lim L(a, α) = 1
α→∞
Remark 2.5: [24]
Assume that (V, L, ⨂) is a fuzzy normed space and
let a ∈ V , t > 0 , 0 < 𝑞 < 1. If
L(a, t) > (1 − q) then there is s with 0 < s <
𝑡 such that L(a, s) > (1 − q).
Definition 2.6: [6]
Suppose that (V, L, ⨂) is a fuzzy normed space.
Put
FB(a, p, t) = {b ∈ V ∶ L (a − b, t) > (1 − p)}
FB[a, p, t] = {b ∈ V ∶ L (a − b, t) ≥ (1 − p)}
Then FB(a, p, t) and FB[a, p, t] is called open and
closed fuzzy ball with the center a ∈Vand radius
p, with p > 0.
Lemma 2.7: [8]
Suppose that (V, L, ⨂) is a fuzzy normed space
then L(x − y, t) = L(y − x, t) for all x, y ∈ V and
t>0
Definition 2.8: [6]
Assume that (V, L, ⨂) is a fuzzy normed space. W
⊆V is called fuzzy bounded if we can find t >
0 and 0 < 𝑞 < 1 such that L(w, t) > (1 − q) for
each w ∈ W.
Definition 2.9: [8]
A sequence (vn ) in a fuzzy normed space (V, L, ⨂)
is called converges to v ∈ V if for each q >
0 and t > 0 we can find N with L[v n − v, t] >
(1 − q) for all n ≥ N. Or in other word lim vn =
n→∞
v or simply represented by vn →v, v is known the
limit of (vn ) or lim L[vn − v, t] = 1.
n→∞
Definition 2.10: [8]
A sequence (v n ) in a fuzzy normed space
(V, L, ⨂) is said to be a Cauchy sequence if for all
0 < 𝑞 < 1 , 𝑡 > 0 there is a number N with
L[v m − vn ,t] > (1 − q) for all
m , n ≥ N.
Definition 2.11: [4]
Suppose that (V, L, ⨂) is a fuzzy normed space and
let W be a subset of V. Then the closure of W is
̅ or CL(W) and which is W
̅=
written by W
⋂{W⊆B: B is closed in V}.
Lemma 2.12: [14]
Assume that (V, L, ⨂) is a fuzzy normed space and
̅ if
suppose that W is a subset of V. Then w ∈ W
and only if there is a sequence (wn) in W with
(wn ) converges to w.
Definition 2.13: [14]
Suppose that (V, L, ⨂) is a fuzzy normed space and
̅=
W ⊆ V. Then W is called dense in V when W
V.
Theorem 2.14: [14]
Suppose that (V, L, ⨂) is a fuzzy normed space and
assume that W is a subset of V. Then W is dense
in V if and only if for every u ∈ V there is w ∈ W
such that
L[u − w , t] > (1 − ε) for some 0 < 𝜀 < 1 and
t > 0.
Definition 2.15: [10]
A fuzzy normed space (V, L, ⨂) is said to be
complete if every Cauchy sequence in V
converges to a point in V.
Definition 2.16: [8]
Suppose that (V, LV , ⨂) and (W, LW , ⨀) are two
fuzzy normed spaces .The operator S: V → W is
said to be fuzzy continuous at 𝐯𝟎 ∈ V if for all
t > 0 and for all 0 < 𝛼 < 1 there is
s[depends on t , α and v 0 ] and there is β [depends
on t, α and v 0 ] with, LV [v − v0 ,s] > (1 − p) we
have LW [S(v) − S(v 0 ), t] > (1 − α) for all v ∈V.
Definition 2.17: [14]
Let (V, LV , ⨂) and (U, LU ,⨀) be two fuzzy
normed spaces. The operator T:D(T)→ U is said to
be fuzzy bounded if there exists r, 0 < r < 1 such
that
LU (Tv , t) ≥ (1 − r) ⨂LV (v,t) for each v ∈ D(T)
⊆ V and t > 0 where D(T) is the domain of T.
Theorem 2.18: [14]
Suppose that (V, LV , ⨂) and (U, LU ,⨀) are two
fuzzy normed spaces. The operator S: D(S) →U is
fuzzy bounded if and only if S(A) is fuzzy bounded
for every fuzzy bounded subset A of D(S).
Put FB(V,U) ={S:V→U, S is a fuzzy bounded
operator} when (V, LV , ⨂) and (U, LU , ⨀) are two
fuzzy normed spaces [14].
Lemma 2.19: [14]
Let (V, LV , ⨂) be a fuzzy normed space. If A and
B are fuzzy bounded subset of V then A+B and 𝛼A
are fuzzy bounded for any 𝛼 ≠ 0 ∈ 𝔽.
Theorem 2.20: [14]
Suppose that (V, LV , ⨂) and (U, LU ,⨀) are two
fuzzy normed spaces. Define L(T, t) =
𝐢𝐧𝐟𝐱∈𝐃(𝐓)𝐋𝐘(𝐓𝐱,𝐭) for all T ∈ FB(V, U) and t >
0 then (FB(V, U), L,∗) is fuzzy normed space.
Theorem 2.21: [14]
Suppose that (V, LV , ⨂) and (U, LU ,⨀) are two
fuzzy normed spaces with S: D(S) →U is a linear
62
Vol. 37, Part B. No. 01, 2019
Engineering and Technology Journal
operator where D(S) ⊆ V. Then S is fuzzy
bounded if and only if S is fuzzy continuous.
Corollary 2.22: [14]
Let (V, LV , ⨂) and (U, LU , ⨀) be two fuzzy
normed spaces. Assume that T: D(T) →U is a
linear operator where D(T) ⊆V. Then T is a fuzzy
continuous if T is a fuzzy continuous at x ∈ D(T).
Theorem 2.23: [14]
Let (V, LV , ⨂) and (U, LU , ⨀) be two fuzzy
normed spaces. If U is fuzzy complete then
FB(V, U) is fuzzy complete.
Definition 2.24: [14]
A linear functional f from a fuzzy normed space
(V, LV , ⨂) into the fuzzy normed space (F, LF , ⨀)
is said to be fuzzy bounded if there exists r, 0 <
𝑟 < 1 such that LF [f(x), t] ≥ (1 − r)⨂LX [x, t]
for all x ∈ D(f) and t > 0. Furthermore, the fuzzy
norm of f is L(f, t) = inf LF (f (x),t) and
LF (f(x), t) ≥ L(f, t)⨂LX (x, t).
Definition 2.25: [14]
Suppose that (V, LV , ⨂) is a fuzzy normed space.
Then the vector space FB(V,𝔽 ) ={f:V→ 𝔽, f is
fuzzy bounded linear functional } with a fuzzy
norm defined by
𝐋 (𝐟, 𝐭) = 𝐢𝐧𝐟 𝐋𝐅 (𝐟(𝐱), 𝐭) form a fuzzy normed
space which is called the fuzzy dual space of V.
3. Fuzzy Convergence of Sequence of
Operators and Functional
Definition 3.1: [24]
A sequence (vn ) in a fuzzy normed space
(V, LV , ⨂) is said to weakly fuzzy convergent if
we can find v ∈ V with every h ∈ FB(V,ℝ)
lim h(v n) = h(v ). This is written vn →w v the
n→∞
element v is said to be the weak limit of (v n) and
(vn ) is said to be fuzzy converges weakly to v.
Theorem 3.2: [24]
Suppose that (vn ) is in the fuzzy normed
space(V, LV , ⨂).
1.
If
vn →v
then
vn →wv.
w
2.If vn → v and dimension of V is finite then
vn →v.
Definition3.3: [24]
Suppose that (V, LV , ⨂) and (U, LU , ⨀) are two
fuzzy normed spaces. A sequence (Tn ) operators
Tn ∈ FB(V, U) is said to be
1. Uniform operator fuzzy convergent
if there is T: V → U with L[Tn –T, t] → 1 for any
t > 0 and n≥N.
2. Strong operator fuzzy convergent
if there is T: V → U with LU [Tnv– Tv,t] → 1 for
every t > 0, v ∈ V and n ≥ N.
3. Weak operator fuzzy convergent if there is
T: V → U with LR [f (Tnv) − f (Tv ), t] for every t >
0, f ∈ FB(U, R) and n≥N.
Respectively, T is called uniform strong and weak
operator limit of (Tn )
Definition3.5: [24]
Let (V, LV , ⨂) be a fuzzy normed space. A
sequence (hn ) of functional hn ∈ FB(V,ℝ) is
called
1) Strong fuzzy converges if there is h ∈
FB(V, ℝ) with L[hn − h, t] → 1 for all t > 0 this
written hn → h
2) Weak fuzzy converges if there is h ∈FB(V,R)
with hn (v) → h(v) for every v ∈ V written by
lim hn (v) = h(v).
n→∞
Definition 3.6:
A subset E of a fuzzy normed space (V, L, ⨂) is
said to be
1- Rare (or no where dense ) in V if its
closure Ē has no interior points
2-Meager (or of the first category) in V if E is the
union of countable many sets each of which is rare
in V
3-Nonmeager (or of the second category) in V if
E is not meager in V
Theorem [Fuzzy Baire’s theorem] 3.7:
Let (V, LV , ⨂) be a complete fuzzy normed space
such that V ≠ ∅ , then it is non meager in itself.
Proof:
We must prove that when V ≠ ∅ is complete and
V=∪∞
k=1 Vk ( Vk closed) then there is one
Vk contains open subset which is not equal to ∅.
Let V be meager in itself that is
V=∪∞
k=1 Mk ….(1)
such that every Mk is rare in V, we will find a
Cauchy sequence (pk ) whose limit is p is in
no Mk this contradict the equation (1) by
̅̅̅̅1 does not
assumption M1 is rare in V so that M
contains a nonempty open set but V does. It
̅̅̅̅c = V − M
̅̅̅̅1 ≠ ∅ and
follows ̅̅̅̅
M1 ≠ V. Thus M
1
c
̅̅̅̅
open. Now choose a point q1 in M1 and an open
fuzzy ball about it, say
̅̅̅̅c . By assumption M2 is
B1 = FB(q1 , ε1 , t) ⊂ M
1
̅̅̅̅2 does not contains a nonempty
rare in V so that M
ε
open set hence FB (q1 , 1 , t) ⊈ ̅̅̅̅
M2 this implies
2
̅̅̅̅c ∩ FB (q1 , ε 1 , t) is not empty and open, let
that M
2
2
̅̅̅̅2 ∩ FB(q1 , ε1 , t)
B2 = FB(q2 , ε2 , t) ⊆ M
by
induction we obtain a sequence of open fuzzy balls
Bk = FB(qk , εk , t), such that
ε
Bk ∩ Mk = ∅ and Bk+1 ⊂ FB (qk , k , t) ⊂ Bk .
2
Let εk < 2−k and (1 − εk ) > 1 − 2−k the
sequence (qm ) of the centers is a Cauchy and
converges say qm ⟶ q ∈ V because V is
complete as we assume. Now for all m and n
with n > 𝑚 with Bn ⊂ FB (qm ,
LV (qm − q, t)
εm
2
, t) so that
62
Vol. 37, Part B. No. 01, 2019
Engineering and Technology Journal
t
t
≥ LV ( qm − qn , ) ⨂LV (qn − q, )
2
2
ε
t
≥ (1 − m) ⨂LV (qn − q, )
= (1 −
2
𝜀𝑚
2
) ⨂1 = (1 −
𝜀𝑚
2
2
)
as n → ∞
c , so
Hence q ∈ BM for every m since Bm ⊂ ̅̅̅̅̅
Mm
∞
q ∉ Mm for any m, so that q ∉∪m=1 Mm = V.
This contradicts q ∈ V.
We need the following theorem in the next results
Uniform fuzzy Bounded Theorem 3.8:
Let (Tn ) be sequence in FB(V, U) where (V, LV , ⨂)
is a complete fuzzy normed space and (U, LU ,⨀)
is a fuzzy normed space such that (LU [Tn v, t]) is
fuzzy bounded for every v ∈ V with LU [Tnv, t] ≥
(1 − cv ) ….(2) n=1, 2,… where 0 < cv < 1.
Then (L[Tn, t]) is fuzzy bounded so there is 0 <
𝑐 < 1 with
L[Tn , t] ≥ (1 − c)….(3) n ∈ ℕ.
Proof:
For m ∈ (0, 1) let Vm={ v∈V: LU [Tnv,t] ≥
(1 − m) } for all v ∈ ̅
Vm there is a sequence
(vj ) in Vm converges to v. This means that for
every fixed 𝑛 we have LU [Tnv j ,t] ≥ (1 − m) and
obtain LU [Tnv,t] ≥ (1 − m) since Tn is fuzzy
continuous. Hence v ∈ Vm and Vm is closed now
by (2) each v ∈ V belongs to some Vm. Hence V=
⋃∞
𝑚=1 Vm . Since V is complete at this point Fuzzy
Baire’s theorem implies that some Vm contains an
open fuzzy ball say,
B0 = FB(v0 , r, t) ⊂ Vm0…(4)
Let v ∈ V be nonzero, let z = v0 + v with
r
that LV [z − v 0 , t] >
LV [z − v0 ,t] > (1 − ) so
2
(1 − 𝑟) which implies 𝑧 ∈ B0. Now by (4) and
from definition of Vm0 we have
t
LU [Tnz, ] ≥ (1 − m0 ) for all n. Also
2
t
LU [Tn v 0, ] ≥ (1 − m0 ) since v0 ∈ B0 .
2
Now for all v ∈ V
LU [Tn v, t] = LU [Tn(z − v 0 ), t]
t
t
≥ LU [Tn z, ] ⨀LU [Tnv 0 , ]
2
2
≥ (1 − m0 )⨀(1 − m0 ).
Hence for all n,
L[Tn , t] = inf LU [Tnv,t] ≥ (1 − c) where for
some (1 − m0 )⨀(1 − m0 ) ≥ (1 − c) for some
0 < 𝑐 < 1.
The following lemma is needed later
Lemma 3.9:
Let Tn ∈ FB(V, U) where (V, LV , ⨂) is a complete
fuzzy normed space and (U, LU ,⨀) is a fuzzy
normed space if (Tn ) is a strong operator fuzzy
convergent with limit T then T ∈ FB(V, U).
Proof
Suppose that T is linear since Tn v → Tv for every
v ∈ V and Tn is linear. The sequence (Tn v) is fuzzy
bounded for any v ∈ V. Now since V is complete
so (L[Tn ,t]) is fuzzy bounded by uniform fuzzy
bounded Theorem 3.8, say L[Tn ,t] ≥ (1 − c ) for
all n and for some 0 < 𝑐 < 1. From this it follows
that LU [Tnv, t] ≥ L[T, t] ⨂LV [v,t]. This implies
LU [Tv, t] ≥ (1 − c)⨂LV [v, t] hence T ∈ FB(V,
U).
Definition 3.10:
Let (V, LV , ⨂) be a fuzzy normed space then a
subset D of V is called a fuzzy total set if
̅̅̅̅̅̅̅̅
spanD =V.
Theorem 3.11:
A sequence (Tn ) ∈ FB(V, U) where (V, LV , ⨂) and
(U, LU , ⨀) are complete fuzzy normed spaces is
strong operator fuzzy convergent if and only if
(i) (L[Tn ,t]) is fuzzy bounded sequence.
(ii) (Tn d) is a Cauchy sequence in U for any d ∈ E
where E is a fuzzy total subset of V.
Proof:
Since Tn v → Tv for every v ∈ V then (𝑖) is
satisfied from Theorem 3.8 since V is complete,
also (ii) is satisfied.
Conversely suppose that (i) and (ii) holds let
t
L [Tn , ] ≥ (1 − c) for all n and t >0
3
0 < 𝑐 < 1. We consider any v ∈ V and prove that
(Tn v) converges strongly in U. Let 0 < 𝜀 < 1
and t > 0 be given since spanE is dense in V we
t
have y ∈ spanE with LV [v − y , ] > (1 − 𝜀) for
3
some 0 < 𝜀 < 1 the sequence (Tn y) is Cauchy by
(ii) hence for given 0 < 𝑟 < 1 and t > 0 there is a
t
number N such that LU [Tn y –Tm y, ] > (1 − r).
3
Now
for
n,
m
≥N
we
get
LU [Tn v –Tm v, t]
≥
t
t
LU [Tnv –Tny, ] ⨀LU [Tn y – Tm y, ]
3
3
t
⨀LU [Tmy –Tmv, ]
3
t
t
t
≥ L [Tn , ] ⨂LV [v − y, ] ⨀LU [Tn y –Tm y, ]
3
3
3
t
t
⨀L [Tm , ] ⨂LV [v − y, ]
3
3
> (1 − c)⨂(1 − ε)⨀(1 − r)⨀(1 − c)⨂(1 − ε).
Now we can find 0 < (1 − 𝑝) < 1 so that
(1 − c)⨂(1 − ε)⨀(1 − r)⨀(1 − c)⨂(1 − ε) >
(1 − p) .Therefore
LU [Tnv − Tm v,t] > (1 − 𝑝) for every n, m ≥N.
Hence (Tnv ) is Cauchy in U but U is complete
(Tn v) convergence in U. Since v ∈ V was arbitrary
this proves strong operator fuzzy convergence of
(Tn ).
62
Engineering and Technology Journal
Definition 3.12:
Suppose that (V, LV , ⨂) and (U, LU ,⨀) are two
fuzzy normed spaces. The operator T: D(T) → U
where D(T) ⊆ V is said to be a fuzzy open if for
any open set E in D(T)
T(E) is open set in U.
The following lemma is the key of proving the next
main results
Lemma 3.13:
Suppose that (V, LV , ⨂)and (U, LU , ⨀) are two
complete fuzzy normed spaces and let T: V → U be
a fuzzy bounded linear surjective operator. Then
T(B0 ) the image of the open fuzzy ball B0 =
FB(0,1, t) ⊂ V contains an open fuzzy ball a bout
0 ∈ U.
Proof:
First step we will prove that the closure of the
image of the open fuzzy ball B1 =
1
FB (0, , t) contains an open fuzzy ball B ∗. For a
2
subset E of V we know that αE = {v ∈ V: v = 𝛼z,
z ∈ E} where 𝛼 ≠ 0 ∈ 𝐹 and E + w ={v ∈ V: v =
z + w, z ∈ E}. We consider the fuzzy open ball
1
Bk = FB (0, k , t). Hence V= ⋃∞
k=1 Bk since T is
2
surjective and linear U= T(V) = T(⋃∞
k=1 Bk )
∞
∞
̅̅̅̅̅̅̅̅
=⋃ k=1 T(Bk )= ⋃ k=1 T(Bk ) since U is complete it
is nonmeager in itself by Theorem 3.7. Hence we
concluded that ̅̅̅̅̅̅̅̅
T(B1 ) must contain some open
̅̅̅̅̅̅̅̅
(B1 ). It
fuzzy ball, say B ∗ = FB(y 0 ,ε, t) ⊂ T
follows that
B ∗ −y0 = FB(0, ε, t) ⊂ ̅̅̅̅̅̅̅̅
T(B1 ) − y 0 … (5).
In the second step we will show that
̅̅̅̅̅̅̅̅
T(B
n ) contains an open fuzzy ball Dn about 0 ∈ U
where Bn = FB(0, 2−n , t) ⊂V. To prove ̅̅̅̅̅̅̅̅
T(B1 ) −
̅̅̅̅̅̅̅̅
(B0 ) let y ∈ ̅̅̅̅̅̅̅̅
y0 ⊂ T
T(B1 ) − y0 . Then y + y0 ∈
̅̅̅̅̅̅̅̅
T(B1 ) we know that y 0 ∈ ̅̅̅̅̅̅̅̅
T(B1 ). Then by lemma
2.12 there are un = Twn ∈ T(B1 ) such that un →
y + y0 , vn = Tzn ∈ T(B1 ) such
that vn → y0
since wn ,zn ∈ B2 it follows that
t
t
LV [wn − zn ,t] ≥ LV [wn, ] ⨂LV [zn, ]
2
2
1
1
≥ (1 − ) ⨂ (1 − )
2
2
so that wn − zn ∈ B0 . From
T(wn − zn ) = Twn − Tzn = un − vn → y
̅̅̅̅̅̅̅̅
̅̅̅̅̅̅̅̅
we say that y ∈ ̅̅̅̅̅̅̅̅
T(B0 ). So T(B
1 ) − y 0 ⊂ T(B0 ).
Thus from (5) we get B ∗ − y 0 = FB(0, ε, t) ⊂
̅̅̅̅̅̅̅̅
T(B0 ) from this we obtain
ε
̅̅̅̅̅̅̅̅
(Bn )…(6)
Vn = FB (0, n , t) ⊂ T
2
−n
where Bn = FB(0, 2 , t) ⊂ V with B0 ⊃ B1 ⊃
̅̅̅̅̅̅̅̅
(Bn ) ⊃ ⋯
T(B1 ) ⊃ ⋯ ⊃ T
B2 ⊃ ⋯ so ̅̅̅̅̅̅̅̅
T(B0 ) ⊃ ̅̅̅̅̅̅̅̅
̅̅̅̅̅̅̅̅
̅̅̅̅̅̅̅̅
that is T(Bn ) ⊂ T(B0 ). In the final step we prove
that T(B0 ) contains an open fuzzy ball a bout 0 ∈
ε
U that is we must prove that V1 = FB (0, , t) ⊂
2
T(B0 ) let y ∈ V1 from (6) with n = 1 we
Vol. 37, Part B. No. 01, 2019
T(B1 ) there must be
have V1 ⊂ ̅̅̅̅̅̅̅̅
T(B1 ) hence y ∈ ̅̅̅̅̅̅̅̅
ε
v ∈ T(B1 ) such that LU [y − v, t] > (1 − ).
4
Now v ∈ T(B1 ) implies v = T(x1 ) for some x1 ∈
ϵ
B1 . Hence LU [(y − Tx1 ), t] > (1 − ) from this
4
and (6) with n = 2 we see that y − Tx1 ∈ V2 ⊂
̅̅̅̅̅̅̅̅
(B2 ) . As before there is x2 ∈ B2 such that
T
ϵ
LU [(y − Tx1 ) − Tx2 , t] > (1 − ) hence
8
y − Tx1 − Tx2 ∈ V3 ⊂ ̅̅̅̅̅̅̅̅
T(B3 ) and so on. In the
nth step we can find xn ∈ Bn such that
ε
LU [y − ∑nk=1 Txk , t] > (1 − n+1) (n =
2
1,2,3, … ) …(7)
Let zn = ∑ni=1 xi since xk ∈ Bk we have
1
LV [xk , t] > (1 − k). For m>n
2
LV [zn − zm ,t] >
1
1
1
(1 − m+1 ) ⨂ (1 − m+2 ) ⨂ … ⨂ (1 − n )
2
2
2
1
1
1
Put (1 − m+1) ⨂ (1 − m+2) ⨂ … ⨂ (1 − n ) >
2
2
2
(1 − r) for some 0 < 𝑟 < 1 hence
LV [zn − zm , t] > (1 − r) for all n, m > N for
some positive number N. Thus (zn ) is Cauchy
so zn → x because V is complete. Also x ∈ B0
since x = ∑∞
k=1 xk and
B0 ⊃ B1 ⊃ B2 ⊃ ⋯
and
T
is
fuzzy
continuous, Tzn → Tx and (7( show that T(x)= y.
Hence y ∈ T(B0 ).
The following result is the main result
Theorem [fuzzy open mapping theorem] 3.14:
Suppose that (V, LV , ⨂) and (U, LU ,⨀) are two
complete fuzzy normed spaces and let S: V → U be
a surjective fuzzy bounded linear operator then S
is a fuzzy open mapping. Hence if S is bijective
then S −1 is fuzzy continuous hence fuzzy bounded.
Proof:
We will prove that if E⊂V is open then S(E) is
open in U that is we will show that for every u=
S(v) ∈ S(E) the set S(E) contains an open fuzzy
ball about u= S(v) ∈ S(E). Let u= S(v) ∈ S(E )
but E is open so it contains an open fuzzy ball with
center v. Hence E−v contain an open fuzzy ball
with center 0 let the radius of the fuzzy ball be r,
0 < 𝑟 < 1. Since FB(0, 1, t) ⊃ FB(0, r, t) for each
0 < 𝑟 < 1. Hence E−v contains the open fuzzy
ball FB(0, 1, 𝑡). Now Lemma 3.13 implies that
S(E − v) = S(E) − S(v ) contains an open fuzzy
ball a bout 0 and so dose S (E) − S(v). Thus S(E)
contains an open fuzzy ball about u= S(v ). Since
u ∈ S(E) was arbitrary S(E) is open in U.
Now if S −1 :U → V exists it is continuous since S
is fuzzy open. Because S −1 is linear then by
Theorem 2.21 it is fuzzy bounded.
62
Vol. 37, Part B. No. 01, 2019
Engineering and Technology Journal
Definition 3.15:
Suppose that (V, LV , ⨂ ) and (U, LU , ⨂) are two
fuzzy normed spaces and let S: D(S) →U be a
linear operator with D(S) is a subset of V then S is
said to be a fuzzy closed linear operator if its
graph G(S) = {(v, u): v ∈ D(S), u = S(v)} is
closed in the fuzzy normed space (V × U, L, ⨂)
where L[(v, u), t] = LV [v, t]⨂LU [u, t]
Theorem [fuzzy closed graph theorem] 3.16:
Suppose that (V, LV , ⨂) and (U, LU ,⨂) are
complete fuzzy normed spaces and let S: D(S) →U
be a closed linear operator with D(S) is a subset of
V. Then S is fuzzy bounded if D(S) is closed in V.
Proof:
First (V × U, L,∗) is complete fuzzy normed space
[24]. By assumption G(S) is closed in V ×
U and D(S) is closed in V. Hence G(S) and D(S)
are complete [24]. Now define P: G(S) → D(S) by
P [(v, S(v))] =v then P is linear and P is fuzzy
bounded since
L[P (v,S (v)),t] = LV [v, t] ≥ LV [v, t]⨂LU [Sv,t]
= L[(v, S(v)), t]
and P is bijective so P −1 exists where P −1 :D(S) →
G(S) defined by P −1 [v] = (v,S (v )) since G(S)
and D(S) are complete we can apply Theorem 3.14
and see P −1 is fuzzy bounded, say
L[(v,S(v))] ≥ (1 − r) ⨂LV [v,t] for some
0 < 𝑟 < 1 and for all v ∈ D(S).
Hence S is fuzzy bounded because
LU [Sv, t] ≥ LU [Sv, t]⨂LV [v,t] = L[(v, S(v )), t]
≥ (1 − r)⨂LV [v,t] for all v ∈ D(S).
We need the following lemma in the next result.
Lemma 3.17:
Suppose that (V, LV , ⨂ ) and (U, LU , ⨂) are two
fuzzy normed spaces and let S: D(S) →U be a
linear operator with D(S) is a subset of V then S is
fuzzy closed if and only if when vn → v where
vn ∈ D(S) and Svn → u then
v ∈ D(S)
and S(v ) = u where u ∈ U.
Proof:
G(S) is closed if and only if z = (v, u) ∈ ̅̅̅̅̅̅
G(S)
implies z ∈ G(S), Now from Theorem 3.16 we see
that z ∈ ̅̅̅̅̅̅
G(S) if and only if there are zn =
(vn, S(v n )) ∈ G(S) such that zn → z hence vn →
v, Svn → u and z = (v,u) ∈ G(S) if and only if
v ∈ D(S) and u = S (v).
Theorem 3.18:
Suppose that (V, LV , ⨂ ) and (U, LU , ⨂) are two
fuzzy normed spaces and let S: D(S) →U be fuzzy
bounded linear operator with D(S) is a subset of V
then:
1-If D(S) is closed in V then S is closed
2-If S is closed and U is complete then D(S) is
closed in V.
Proof:
1-Suppose that (vn ) ∈ D(S) and v n → v such
̅̅̅̅̅̅
(S) = D(S) since D(S) is
that Sv n → Sv then v ∈ D
closed and Sv n → Sv since S is fuzzy continuous.
Hence S is fuzzy closed by Lemma 3.17.
̅̅̅̅̅̅
(S) there is (vn ) in D(S) with vn → v
2-Let v ∈ D
since S is fuzzy bounded
LU [Sv n − Svm, t] = LU [S(v n − v m ),t]
≥ L[S, t]⨂LV [vn − vm ,t].
This show that (Svn ) is Cauchy so Svn → u ∈ U
since U is complete because S is fuzzy closed v ∈
D(S) and Sv = u. Hence D(S) is closed because
̅̅̅̅̅̅
(S) was arbitrary.
v∈D
4.Conclusion
The main goal of this paper is to prove the fuzzy
open mapping theorem and to introduce the
definition of fuzzy closed linear operators on fuzzy
normed spaces in order to prove the fuzzy closed
graph theorem and to investigate some basic
properties of this type of operators. We have tried
here to translate the results in functional analysis
to fuzzy context and we succeeded in this
situations.
References
[1] L. Zadeh, “Fuzzy sets,” Inf. Control, 8, 338-353,
1965.
[2] A. Katsaras, “Fuzzy topological vector spaces II,”
Fuzzy sets and Systems, 12, 143-154, 1984.
[3] O. Kaleva and S. Seikkala, “On fuzzy metric
spaces,” Fuzzy sets and systems, 121, 215-229, 1984.
[4] C. Felbin, “Finite dimensional fuzzy normed linear
spaces,” Fuzzy sets and Systems, 48, 239-248, 1992.
[5] O. Kramosil and J. Michalek, “Fuzzy metrics and
statistical metric spaces,” Kybernetika, 11, 326-334,
1975.
[6] S. Cheng and J. Mordeson, “Fuzzy linear operators
and fuzzy normed linear spaces,” Ball. Cal. Math. Soc.
86, 429 – 436, 1994.
[7] T. Bag and S. Samanta, “Finite dimensional fuzzy
normed linear spaces,” J. Fuzzy Math. 11, 3, 687-705,
2003.
[8] R. Saadati and M. Vaezpour, “Some results on fuzzy
Banach spaces,” J. Appl. Math. and Computing, 171,
475-484, 2005.
[9] T. Bag and S. Samanta, “Fuzzy bounded linear
operators,” Fuzzy sets and Systems, 151, 3, 513-547,
2005.
[10] T. Bag and S. Samanta, “Fixed point theorems on
Fuzzy normed spaces,” Inf. sci.176, 2910-2931, 2006.
[11] T. Bag and S. Samanta, “Some fixed point
theorems in fuzzy normed linear spaces,” Information
sciences 177, 3271-3289, 2007.
03
Engineering and Technology Journal
Vol. 37, Part B. No. 01, 2019
[12] I. Sadeqi and F. Kia, “Fuzzy normed linear space
and its topological structure,” Chaos Solitions and
Fractals, 40, 5, 2576-2589, 2009.
[13] S. Nadaban, “Fuzzy continuous mapping in fuzzy
normed linear spaces,” I. J. Computer Communications
and Control, Vol 10, 6, 834-842, 2015.
[14] J.R. Kider and N.A. Kadhum, “Properties of Fuzzy
Norm of fuzzy Bounded Operators,” Iraqi Journal of
Science, Vol. 58, No.3A, pp: 1237-1281, 2017.
[15] H. Si, H. Cao and P. Yang, “Continuity in an
intuitionistic fuzzy normed Space,” Seventh, I.
Conference on fuzzy systems and knowledge
Discovery, 144-148., 2010.
[16] W. Congxin and M. Ming, “Continuity and
bounded mappings between fuzzy normed spaces,”
Fuzzy Math, 1, 13-24, 1993.
[17] M. Goudarzi and S. Vaezpour, “Best simultaneous
approximation in fuzzy normed spaces,” Iranian J. of
Fuzzy Systems, 7, 87-69, 2010.
[18] R. Jameel, “On some results of analysis in a
standard fuzzy normed spaces ,” M.Sc. Thesis,
University of Technology, Iraq, 2014.
[19] J.R. Kider, “New fuzzy normed spaces ,” J.
Baghdad Sci., 9, 559-564, 2012.
[20] D. Oregan and R. Saadati, “L-Random and fuzzy
normed spaces and classical theory,” CUBO
Mathematical J. 2, 71-81, 2010.
[21] J. Xiao and X. Zhu, “Fuzzy normed spaces of
operators and it is completeness ,” Fuzzy sets and
Systems, 133, 437-452, 2004.
[22] A. George and P. Veeramani, “On some results in
fuzzy metric spaces,” Fuzzy sets and Systems, 64, 395399, 1994.
[23] I. Golet, “On generalized fuzzy normed spaces and
coincidence theorems,” Fuzzy sets and Systems, 161,
1138-1144, 2010.
[24] N. A. Kadhum, “On fuzzy norm of a fuzzy bounded
operators on fuzzy normed spaces,” M.Sc. Thesis
University of Technology, Iraq, 2017.
[25] A. Amini and R. Saadati, “Some Properties of
continuous t-norms and s-norms,” Int. J. Pure Appl.
Math, 16, 157-164, 2004.
03