In this paper, we investigate the relationships between fuzzy relations and Alexandrov
L
-topologies in complete residuated lattices. Moreover, we give their examples.
1. INTRODUCTION
Pawlak
[9
,
10]
introduced rough set theory as a formal tool to deal with impre- cision and uncertainty in data analysis. Hájek
[3]
introduced a complete residuated lattice which is an algebraic structure for many valued logic. Radzikowska
[11]
de- veloped fuzzy rough sets in complete residuated lattice. Bělohlávek
[1]
investigated information systems and decision rules in complete residuated lattices. Lai
[7
,
8]
in- troduced Alexandrov
L
-topologies induced by fuzzy rough sets. Algebraic structures of fuzzy rough sets are developed in many directions
[1
-
13]
.
In this paper, we investigate the relationships between fuzzy relations and Alexan- drov
L
-topologies in complete residuated lattices. Moreover, we give their examples.
2. PRELIMINARIES
- Definition 2.1([1,3]). An algebra (L,∧,∨,⊙, →⊥,⊤) is called acomplete residuated latticeif it satisfies the following conditions:
-
(L1)L= (L, ≤ , ∨,∧,⊥,⊤) is a complete lattice with the greatest element ⊤ and the least element ⊥;
-
(L2) (L,⊙,⊤) is a commutative monoid;
-
(L3)x⊙y≤ziffx≤y→zforx, y, z∈L.
In this paper, we assume (
L
,∧,∨,⊙, →, *⊥,⊤) is a complete residuated lattice with the law of negation;i.e.
x** = x
. For
α
∈
L
,
A
,⊤
x
∈
LX
, (
α
→
A
)(
x
) =
α
→
A
(
x
), (
α
⊙
A
)(
x
) =
α
⊙
A
(
x
) and ⊤
x
(
x
) = ⊤,⊤
x
(
x
) = ⊥, otherwise.
- Definition 2.2([1,7]). LetXbe a set. A functionR : X×X → Lis called afuzzy relation. A fuzzy relationRis called afuzzy preorderif satisfies (R1) and (R2).
-
(R1)reflexiveifR(x, x) = ⊤ for allx∈X,
-
(R2)transitiveifR(x, y) ⊙R(y, z) ≤R(x, z), for allx, y, z∈X.
-
We denote
- Lemma 2.3([1,3]).Let(L,∨,∧,⊙, →, *⊥,⊤)be a complete residuated lattice with a negation *. For each x,y,z,xi,yi∈L,the following properties hold.
-
(1)If y ≤ z, then x⊙y ≤ x⊙z.
-
(2)If y ≤ z, then x → y ≤ x → z and z → x ≤ y → x.
-
(3)and
-
(4)and
-
(5) (x → y) ⊙x ≤ y and(y→ z) ⊙ (x →y) ≤ (x → z).
-
(6) (x⊙y) →z = x→ (y → z) =y→ (x → z)and(x⊙y)* =x→y*.
-
(7)x* →y* =y→x and(x→y)* =x⊙y*.
- Definition 2.4([5-7]). A subsetτ⊂LXis called anAlexandrov topologyif it satisfies satisfies the following conditions.
-
(T1) ⊥X, ⊤X∈τwhere ⊤X(x) = ⊤ and ⊥X(x) = ⊥ forx∈X
-
(T2) If
-
(T3)α⊙A∈τfor allα∈LandA∈τ.
-
(T4)α → A∈τfor allα∈LandA∈τ.
- Definition 2.5([7]). LetR∈LX×Xbe a fuzzy relation. A setA∈LXis calledextensionalifA(x) ⊙R(x, y) ≤A(y) for allx,y∈X.
3. FUZZY RELATIONS AND ALEXANDROVL-TOPOLOGIES
- Theorem 3.1.Let R∈LX×Xand R−1∈LX×Xwith R−1(x, y) =R(x, y).
-
(1)τ is an Alexandrov topology on X iff τ*= {A*∈LX|A∈τ}is an Alexandrov topology on X.
-
(2)τR= {A∈LX|A(x) ⊙R(x, y) ≤A(y),x,y∈X}is an Alexandrov topology on X. Moreover,
-
(3)Ifis the smallest fuzzy preorder such that R≤then
where Rr
(
x, y
) = Δ∨
R
(
x, y
)
and
Δ (
x, y
) = ⊤
if x = y and
Δ(
x, y
) = ⊥
if x
≠
y
.
Moreover,
-
(4)wherefor each y∈X.
-
(5)wherefor each x∈X.
-
(6)wherefor each y∈X.
-
(7)wherefor each x∈X.
-
(8)Moreover,CτR(A) ∈τR.
-
(9)Moreover,IτR(A) ∈τR.
-
(10)A∈τRiff A = CτR(A) =IτR(A).
-
(11)CτR(A) = (IτR−1(A*))*for all A∈LX.
Proof.
(1) Let
A
* ∈
τ
* for A ∈
τ
. Since
α
⊙
A
* = (
α →A
)* and
α → A
* = (
α
⊙
A
)*,
τ
* is an Alexandrov topology on
X
.
(2) (T1) Since ⊤
X
(
x
) ⊙
R
(
x, y
) ≤ ⊤
X
(
y
) = ⊤ and ⊥
X
(
x
) ⊙
R
(
x, y
) = ⊥ = ⊥
X
(
y
), Then⊥
X
,⊤
X
∈
τR
.
(T2) For
Ai
∈
τR
for each
i
∈ Γ, since
Similarly,
(T3) For
A
∈
τR
,
α
⊙
A
∈
τR
.
(T4) For
A
∈
τR
, by Lemma 2.3(5), since
α
⊙ (
α → A
(
x
)) ⊙
R
(
x, y
) ≤
A
(
x
) ⊙
R
(
x, y
) ≤
A
(
y
), (
α → A
(
x
)) ⊙
R
(
x, y
) ≤
α → A
(
y
). Then
α → A
∈
τR
. Moreover
A
∈
τR iff A
* ∈
τ
R−1
from:
-
A(x) ⊙R(x, y) ≤A(y) iffR(x, y) →A* ≥A*(y)
-
iffA*(y) ⊙R(x, y) ≤A*(x) iffA*(y) ⊙R−1(y, x) ≤A*(x).
(3) Define
Then
RτR
is a fuzzy preorder. Since
B
∈
τR
and
B
(
x
) ⊙
R
(
x, y
) ≤
B
(
y
), then
R
(
x, y
) ≤
B
(
x
) →
B
(
y
). Hence
R
(x, y
) ≤
RτR
. If
P
is a fuzzy preorder with
R ≤ P
, for
Pw
(
x
) =
P
(
w, x
), then
Pw
(
x
) ⊙
R
(
x, y
) ≤
Pw
(
x
) ⊙
P
(
x, y
) ≤
Pw
(
y
). Hence
Pw
∈
τR
. Thus
RτR
(
x, y
) =
Thus,
Since
Rr
(
x, y
) = Δ∨
R
(
x
,
y
), we have (
Rr
)
n
(
x
,
x
) = ⊤ for each
n
∈
N
. So
Since
then
Hence
is a fuzzy preorder. If
R ≤ P
and
P
is fuzzy preorder, then
Rr ≤ P
and (
Rr
)
n
≤
Pn = P
, thus,
Hence
(4) Put
and
Since
A
∈
τR
,
Hence
. Since
Thus,
A
∈
τR
.
Let
A
∈
τ
. Since
Thus,
A
∈
τR
.
Let
A
∈
τ
. Then
Put
A
(
x
) =
ax
. Then
Let
Then
Thus,
D
∈
τ
. Hence
τR = τ
=
τ
1
.
(5) Put
and
Since
A
∈
τR
,
RτR
(
x, y
) →
A
(
y
) =
→
A
(
y
) ≥ (
A
(
x
) →
A
(
y
)) →
A
(
y
) ≥
A
(
x
). Hence
Since
,
Thus,
A
∈
η
.
Let
A
∈
η
. Since
Thus,
R
(
x, y
) →
A
(
y
) ≥
A
(
x
) iff
A
(
x
) ⊙
R
(
x, y
) ≤
A
(
y
). So,
A
∈
τR
.
Let
A
∈
η
Then
Put
A
(
y
) =
by
. Then
Let
Then
Thus,
A
∈
η
. Hence
τR = η
=
η
1
.
(6) It follows from
iff
(7) It follows from
iff
(8) Put
Then
B
∈
τR
from:
If
A ≤ E
and
E
∈
τR
, then
B ≤ E
from:
Hence
CτR = B
.
(9) Let
from
If
E ≤ A
and
E
∈
τR
, then
E ≤ B
from:
Hence
IτR = B
.
(11)
Theorem 3.2.
Let RX and RY be fuzzy relations and f
:
X → Y a map with RX
(
x
,
y
) ≤
RY
(
f
(
x
),
f
(
y
))
for all x
,
y
∈
X. Then the following equivalent conditions hold.
-
(1)f−1(B) ∈τRXfor all B∈τRY.
-
(2)for all
-
(3)RτRX(x,y) ≤RτRY(f(x),f(y))for all x,y∈X.
-
(4)for all x,y∈X.
-
(5)f(CτRX(A)) ≤CτRY(f(A))for all A∈LX.
-
(6)for all A∈LX.
-
(7)CτRX(f−1(B)) ≤f−1(CτRX(B))for all B∈LY.
-
(8)for all B∈LY.
-
(9)f−1(IτRX(B)) ≤IτRY(f−1(B))for all B∈LY.
-
(10)for all B∈LY.
Proof.
(1) For all
B
∈
τRY
,
f
−1
(
B
) ∈
τRX
from:
-
f−1(B)(x) ⊙RX(x,y) ≤B(f(x)) ⊙RY(f(x),f(y)) ≤B(f(y)) =f−1(B)(y).
(1) ⇔ (2) It follows from (1) and Theorem 3.1(2).
(1) ⇒ (3)
(1) ⇒ (5)
(3) ⇒ (5)
(5) ⇒ (7) By (5), put
A
=
f
−1
(
B
). Since
f
(
CτRX
(
f
−1
(
B
))) ≤
CτRY
(
f
(
f
−1
(
B
))) ≤
CτRY
(
B
), we have
CτRX
(
f
−1
(
B
)) ≤
f
−1
(
CτRX
(
B
)).
(7) ⇒ (1) For all
B
∈
τRY
,
CτY
(
B
) =
B
. Since
CτRX
(
f
−1
(B)) ≤
f
−1
(
CτRX
(
B
)) =
f
−1
(
B
),
f
−1
(
B
) ∈
τRX
.
(1) ⇒ (9)
(9) ⇒ (1) For all
B
∈
τRY
,
IτY
(
B
) =
B
. Since
IτRX
(
f
−1
(
B
)) ≥
f
−1
(
IτRX
(
B
)) =
f
−1
(
B
),
f
−1
(
B
) ∈
τRX
.
Other cases are similarly proved.
Example 3.3.
Let (
L
= [0, 1],⊙,→,* ) be a complete residuated lattice with the law of double negation defined by
-
x⊙y= (x + y− 1) ∨ 0,x → y= (1 −x + y) ∧ 1,x* = 1 −x.
Let
X
= {
a
,
b
,
c
},
Y
= {
x
,
y
,
z
} be sets and
f
:
X → Y
as follows:
(1) Define
RX
∈
LX×X
,
RY
∈
LY ×Y
as follows
Then
RX
(
a
,
b
) ≤
RY
(
f
(
a
),
f
(
b
)) for all
a
,
b
∈
X
.
For
n
≥ 2,
and
as follows:
Then
Moreover,
Then
RτRX
(
a, b
) ≤
RτRX
(
f
(
a
),
f
(
b
)) for all
a
,
b
∈
X
.
(2)
where
ai
∈
L
and
For
For
where
bi
∈
L
and
For
For
(3)
where
ai
∈
L
and
where
bi
∈
L
and
(4) For
A
= (0.2. 0.8, 0.6) ∈
LX
,
Bělohlávek R.
2002
Fuzzy Relational Systems
Kluwer Academic Publishers
New York
Hájek P.
1998
Metamathematices of Fuzzy Logic
Kluwer Academic Publishers
Dordrecht
Kim Y.C.
2014
Alexandrov L-topologies
International Journal of Pure and Applied Mathematics
93
(2)
165 -
179
Kim Y.C.
2014
Alexandrov L-topologies and L-join meet approximation operators
International Journal of Pure and Applied Mathematics
91
(1)
113 -
129
Lai H.
,
Zhang D.
2009
Concept lattices of fuzzy contexts: Formal concept analysis vs. rough set theory
Int. J. Approx. Reasoning
50
695 -
707
DOI : 10.1016/j.ijar.2008.12.002
Pawlak Z.
1984
Rough probability
Bull. Pol. Acad. Sci. Math.
32
607 -
615
She Y.H.
,
Wang G.J.
2009
An axiomatic approach of fuzzy rough sets based on residuated lattices
Computers and Mathematics with Applications
58
189 -
201
DOI : 10.1016/j.camwa.2009.03.100
Ma Zhen Ming
,
Hu Bao Qing
2013
Topological and lattice structures of L-fuzzy rough set determined by lower and upper sets
Information Sciences
218
194 -
204
DOI : 10.1016/j.ins.2012.06.029