generalized anti fuzzy bi-ideals in ordered semigroups
TRANSCRIPT
ISSN 1995-0802, Lobachevskii Journal of Mathematics, 2010, Vol. 31, No. 1, pp. 65–76. c© Pleiades Publishing, Ltd., 2010.
Generalized Anti Fuzzy Bi-Ideals in Ordered Semigroups
Young Bae Jun1*, Asghar Khan2**, Muhammad Shabir3***, and Seok Zun Song4
(Submitted by M.M. Arslanov)1Department of Mathematics Education Gyeongsang National University Chinju 660-701, Korea
2Department of Mathematics COMSATS Institute of IT Abbottabad, Pakistan3Department of Mathematics Quaid-i-Azam University Islamabad, Pakistan
4Department of Mathematics Cheju University, 690-756, KoreaReceived November 12, 2009
Abstract—Using the notion of anti fuzzy points and its besideness to and non-quasi-coincidentwith a fuzzy set, new concepts of an anti fuzzy bi-ideals in ordered semigroups are introduced andtheir inter-relations and related properties are investigated.
DOI: 10.1134/S1995080210010105
Key words and phrases: besides to, non-quasi coincidence with, (α, β)-anti fuzzy bi-ideal.
1. INTRODUCTION
The concept of a fuzzy set was first initiated by Zadeh [8]. Since then it has become a vigorous areaof research in engineering, medical science, social science, physics, statistics, graph theory, etc. In thispaper, we introduce the concept of an anti fuzzy bi-ideal of ordered semigroups by using the notionof anti fuzzy points and besideness to and non-quasi-coincidence with a fuzzy set, and investigatetheir inter-relations and related properties.
2. PRELIMINARIES
An ordered semigroup is a structure (S, ·,≤) satisfying the following conditions:(OS1) (S, ·) is a semigroups,(OS2) (S,≤) is a poset,(OS3) (∀a, b, x ∈ S)(a ≤ b =⇒ xa ≤ xb and ax ≤ bx).Let (S, ·,≤) be an ordered semigroup. For A ⊆ S, we denote
(A] := {t ∈ S|t ≤ h for some h ∈ A}.
For A,B ⊆ S, we denote, AB := {ab|a ∈ A, b ∈ B}. Let A,B ⊆ S. Then A ⊆ (A], (A](B] ⊆ (AB],and ((A]] = (A] [5].
Let S be an ordered semigroup and ∅ �= G ⊆ S. Then G is called a subsemigroup of S if G2 ⊆ G.A subsemigroup G of an ordered semigroup S is called a bi-ideal of S if (1) GSG ⊆ G and (2)(∀x, y ∈ S)(∀y ∈ G) (x ≤ y =⇒ x ∈ G) [5].
Definition 2.1. A fuzzy subset A of S is called an anti fuzzy bi-ideal of S if it satisfies:(i) (∀x, y ∈ S)(x ≤ y =⇒ A(x) ≤ A(y)),(ii) (∀x, y ∈ S)(A(xy) ≤max{A(x),A(y)}),
*E-mail: [email protected]**E-mail: [email protected]
***E-mail: [email protected]
65
66 JUN et al.
(iii) (∀a, x, y ∈ S)(A(xay) ≤max{A(x),A(y)}).Proposition 2.2. Let (S, ·,≤) be an ordered semigroup and A is a fuzzy subset of S. Then A is
an anti fuzzy bi-ideal of S if and only if the complement Ac of A is a fuzzy bi-ideal of S.Proof. =⇒. Suppose that A is an anti fuzzy bi-ideal of S. Let x, y ∈ S be such that x ≤ y. Then
Ac(x) = 1 −A(x) ≥ 1 −A(y) = Ac(y).
For a, x, y ∈ S, we have
Ac(xy) = 1 −A(xy) ≥ 1 − max{A(x),A(y)} = min{1 −A(x), 1 −A(y)} = min{Ac(x),Ac(y)},
Ac(xay) = 1 −A(xay) ≥ 1 − max{A(x),A(y)} = min{1 −A(x), 1 −A(y)} = min{Ac(x),Ac(y)}.
Hence Ac is a fuzzy bi-ideal of S.⇐=. Assume that Ac is a fuzzy bi-ideal of S. Let x, y ∈ S be such that x ≤ y. Then
A(x) = 1 −Ac(x) ≤ 1 −Ac(y) = A(y).
For a, x, y ∈ S we have
A(xy) = 1 −Ac(xy) ≤ 1 − min{Ac(x),Ac(y)} = max{1 −Ac(y), 1 −Ac(y)} = max{A(y),A(y)},
A(xay) = 1 −Ac(xay) ≤ 1 − min{Ac(x),Ac(y)} = max{1 −Ac(y), 1 −Ac(y)}= max{A(y),A(y)}.
Hence A is an anti fuzzy bi-ideal of S. �Given a fuzzy subset A of a set S and for every t ∈ [0, 1], the subsets
C(A; t) := {x ∈ S|A(x) ≤ t} and O(A; t) := {x ∈ S|A(x) < t}are called the closed t-cut and open t-cut of A, respectively.
Theorem 2.3. A fuzzy subset A of an ordered semigroup S is an anti fuzzy bi-ideal of S if andonly if C(A; t) is a bi-ideal of S.
Proof. Suppose that A is an anti fuzzy bi-ideal of S. Let x, y ∈ S be such that x ≤ y. If y ∈ C(A; t)then A(y) ≤ t. Since x ≤ y and A is an anti fuzzy bi-ideal of S, we have A(x) ≤ A(y), and henceA(x) ≤ t =⇒ x ∈ C(A; t). Let x, y ∈ C(A; t) and a ∈ S. Then A(x) ≤ t, and A(y) ≤ t. By (ii) and (iii) ofDefinition 2.1, we have
A(xy) ≤ max{A(x),A(y)} ≤ t, A(xay) ≤ max{A(x),A(y)} ≤ t.
Hence xy ∈ C(A; t) and xay ∈ C(A; t). Thus C(A; t) is a bi-ideal of S.Conversely, assume that C(A; t) is a bi-ideal of S. Let x, y ∈ S be such that x ≤ y. If A(y) <
A(x) then there exists t ∈ (0, 1) such that A(y) ≤ t < A(x). Then y ∈ C(A; t) but x /∈ C(A; t). Thisimpossible and hence A(y) ≤ A(x) for all x, y ∈ S with x ≤ y. Suppose that (ii) is not valid. Then
max{A(x),A(y)} ≤ t < A(xy),
for some x, y ∈ S and t ∈ (0, 1). Then x, y ∈ C(A; t) but xy /∈ C(A; t). This is a contradiction. HenceA(xy) ≤ max{A(x),A(y)} for all x, y ∈ S. Suppose that (iii) is false. Then
max{A(x),A(y)} ≤ t < A(xay),
for some a, x, y ∈ S and t ∈ (0, 1). Then x, y ∈ C(A; t) but xay /∈ C(A; t). This is a contradiction.Therefore A(xay) ≤ max{A(x),A(y)} for all a, x, y ∈ S. �
Example 2.4. Consider the set S = {a, b, c, d, e} with the following multiplication table andorder relation “≤”
(a, e), (b, d), (b, e), (c, e), (d, e)}≤:= {(a, a), (b, b), (c, c), (d, d), (e, e), (a, c), (a, d).
LOBACHEVSKII JOURNAL OF MATHEMATICS Vol. 31 No. 1 2010
GENERALIZED ANTI FUZZY Bi-IDEALS 67
· a b c d e
a a d a d d
b a b a d d
c a d c d e
d a d a d d
e a d c d e
Then S is an ordered semigroup (see [5]) and {a}, {a, b, d} and {a, b, c, d} are bi-ideals of S. Define afuzzy subset A : S −→ [0, 1] by
A(a) = 0.2, A(b) = 0.4, A(c) = 0.5, A(d) = 0.3, A(e) = 0.6.
Then
C(A; t) =
⎧⎪⎪⎪⎪⎪⎪⎨
⎪⎪⎪⎪⎪⎪⎩
S if t ∈ [0.6, 1){a, b, c, d} if t ∈ [0.5, 0.6){a, b, d} if t ∈ [0.4, 0.5){a} if t ∈ [0.2, 0.3)∅ if t ∈ [0, 0.2).
Then C(A; t) is a bi-ideal and by Theorem 2.3, A is an anti fuzzy bi-ideal of S.A fuzzy subset A of a set S of the form
A(y) =
{t ∈ [0, 1) if y = x,
1 otherwise,
is called an anti fuzzy point with support x and value t and is denoted byt
x. A fuzzy subset A of S is
said to be non-unit if there exists x ∈ S such that A(x) < 1.
For an anti fuzzy pointt
xand a fuzzy subset A in a set S, Jun and Song [4] gave meaning to the
symbolt
xαA, where α ∈ {[∈], [q], [∈] ∨ [q], [∈] ∧ [q]}.
To say thatt
x[∈]A(resp.
t
x[q]A) means that A(x) ≤ t(resp. A(x) + t < 1), and in this case,
t
xis said
to be beside to (resp. be non-quasi-coincident with) a fuzzy subset A. To say thatt
x[∈] ∨ [q]A (resp.
t
x[∈] ∧ [q]A) means that
t
x[∈]A or
t
x[q]A (resp.
t
x[∈]A and
t
x[q]A). To say that
t
xαA means that
t
xαA
does not hold.
3. GENERALIZED ANTI FUZZY BI-IDEALS
In what follows let S denote an ordered semigroup unless otherwise specified.Definition 3.1. A fuzzy subset A of S is called an ([∈], [∈])-fuzzy bi-ideal of S if it satisfies:
(i) (∀x, y ∈ S)(∀t ∈ [0, 1))(
x ≤ y,t
y[∈]A =⇒ t
x[∈]A
)
,
(ii) (∀x, y ∈ S)(∀t1, t2 ∈ [0, 1))(
t1x
[∈]A,t2y
[∈]A =⇒max{t1, t2}xy
[∈]A)
,
(iii) (∀x, y, a ∈ S)(∀t1, t2 ∈ [0, 1))(
t1x
[∈]A,t2y
[∈]A =⇒max{t1, t2}xay
[∈]A)
.
LOBACHEVSKII JOURNAL OF MATHEMATICS Vol. 31 No. 1 2010
68 JUN et al.
Theorem 3.2. A fuzzy subset A of S is an ([∈], [∈])-fuzzy bi-ideal of S if and only if it satisfies:(i) (∀x, y ∈ S)(x ≤ y =⇒ A(x) ≤ A(y)),(ii) (∀x, y ∈ S)(A(xy) ≤max{A(x),A(y)}),(iii) (∀x, y, a ∈ S)(A(xay) ≤max{A(x),A(y)}).Proof. Assume that A satisfies the conditions (i), (ii) and (iii). Let x, y ∈ S and t ∈ [0, 1) be such that
x ≤ y andt
y[∈]A. Using (i), we have A(x) ≤ A(y) ≤ t, and so
t
x[∈]A. Let a, x, y ∈ S and t1, t2 ∈ [0, 1)
be such thatt1x
[∈]A andt2y
[∈]A. Then A(x) ≤ t1 and A(y) ≤ t2, which implies from (ii) and (iii) that
A(xy) ≤ max{A(x),A(y)} ≤ max{t1, t2},A(xay) ≤ max{A(x),A(y)} ≤ max{t1, t2}.
Hencemax{t1, t2}
xy[∈]A and
max{t1, t2}xay
[∈]A.
Conversely, assume that a fuzzy subset A of S is an ([∈], [∈])-fuzzy bi-ideal of S. Let x, y ∈ S be
such that x ≤ y. If A(y) < A(x), then there exists t ∈ (0, 1) such that A(y) ≤ t < A(x). Thust
y[∈]A,
butt
x[∈]A. This is impossible, and therefore A(x) ≤ A(y) for all x, y ∈ S with x ≤ y. Suppose that (ii)
is not valid. Thenmax{A(x),A(y)} ≤ t < A(xy)
for some x, y ∈ S and t ∈ (0, 1). It follows thatt
x[∈]A, and
t
y[∈]A, but
t
xy[∈]A. This is a contradicition.
Hence A(xy) ≤ max{A(x),A(y)} for all x, y ∈ S. Finally assume that (iii) is false. Then
max{A(x),A(y)} ≤ t < A(xay),
for every a, x, y ∈ S and t ∈ (0, 1). Hencet
x[∈]A, and
t
y[∈]A, but
t
xay[∈]A. This is a contradiction.
Therefore A(xay) ≤ max{A(x),A(y)} for all a, x, y ∈ S. �Remark 3.3. From Theorem 3.2, it follows that every anti fuzzy bi-ideal of S is an ([∈], [∈])-
fuzzy bi-ideal of S. But the converse is not true.Example 3.4. Consider the ordered semigroup given in Example 2.4, and define a fuzzy subset
A as follows:
A(a) = 0.2, A(b) = 0.4, A(c) = 0.8, A(d) = 0.6, A(e) = 0.5.
Then A is an ([∈], [∈])-fuzzy bi-ideal of S. But A is not an anti fuzzy bi-ideal of S, because
0.6 = A(d) = A(ab) ≤ max{A(a),A(b)} = max{0.2, 0.4} = 0.4.
Definition 3.5. A fuzzy subset A of S is called an ([∈], [∈] ∨ [q])-fuzzy bi-ideal of S if it satisfies:
(i) (∀x, y ∈ S)(∀t ∈ [0, 1))(
x ≤ y,t
y[∈]A =⇒ t
x[∈] ∨ [q]A
)
,
(ii) (∀x, y ∈ S)(∀t1, t2 ∈ [0, 1))(
t1y
[∈]A,t2x
[∈]A =⇒ max{t1, t2}xy
[∈] ∨ [q]A)
,
(iii) (∀a, x, y ∈ S)(∀t1, t2 ∈ [0, 1))(
t1y
[∈]A,t2x
[∈]A =⇒ max{t1, t2}xay
[∈] ∨ [q]A)
.
Theorem 3.6. A fuzzy subset A of S is an ([∈], [∈] ∨ [q])-fuzzy bi-ideal of S if and only if itsatisfies:
(i) (∀x, y ∈ S)(x ≤ y =⇒ A(x) ≤ max{A(y), 0.5}),(ii) (∀x, y ∈ S)(A(xy) ≤ max{A(x),A(y), 0.5}),(iii) (∀a, x, y ∈ S)(A(xay) ≤ max{A(x),A(y), 0.5}).
LOBACHEVSKII JOURNAL OF MATHEMATICS Vol. 31 No. 1 2010
GENERALIZED ANTI FUZZY Bi-IDEALS 69
Proof. Suppose that A satisfies the conditions (i), (ii) and (iii). Let x, y ∈ S, x ≤ y and t ∈ [0, 1) be
such thatt
y[∈]A. Then A(y) ≤ t, by using (i), we have
A(x) ≤ max{A(y), 0.5} ≤ max{t, 0.5}.
Thus A(x) ≤ t or A(x) ≤ 0.5, according to t > 0.5 or t ≤ 0.5. Thust
x[∈] ∨ [q]A. Let a, x, y ∈ S and
t, r ∈ [0, 1) be such thatt
x[∈]A and
r
y[∈]A. Then A(x) ≤ t and A(y) ≤ r and by using (ii) we have
A(xy) ≤ max{A(x),A(y), 0.5} ≥ max{t, r, 0.5},
If max{t, r} < 0.5 then A(xy) ≤ 0.5 and so A(xy) + max{t, r} < 0.5 + 0.5 = 1, and we havemax{t, r}
xy[q]A. If max{t, r} ≥ 0.5 then A(xy) ≤ max{A(x),A(y)} and so
max{t, r}xy
[∈]A and hence
max{t, r}xy
[∈] ∨ [q]A. Let a, x, y ∈ S and t, r ∈ [0, 1) be such thatt
x[∈]A and
r
y[∈]A. Then A(x) ≤ t
and A(y) ≤ r and by using (iii) we have
A(xay) ≤ max{A(x),A(y), 0.5} ≥ max{t, r, 0.5}.
If max{t, r} < 0.5 then A(xay) ≤ 0.5 and so A(xay) + max{t, r} < 0.5 + 0.5 = 1, and we havemax{t, r}
xay[q]A. If max{t, r} ≥ 0.5 then A(xay) ≤ max{A(x),A(y)} and so
max{t, r}xay
[∈]A and hence
max{t, r}xay
[∈] ∨ [q]A.
Conversely, let x, y ∈ S and x ≤ y. We consider the following cases:a) A(y) > 0.5,b) A(y) ≤ 0.5.Case a): Let x, y ∈ S and x ≤ y. Assume that max{A(y), 0.5} < A(x), which implies that A(y) <
A(x). Choose t such that A(y) ≤ t < A(x). Thent
y[∈]A but
t
x[∈]A and so
t
x[∈] ∨ [q]A. This is a
contradiction.Case b): Let x, y ∈ S and x ≤ y. Assume that max{A(y), 0.5} < A(x). Then A(x) > 0.5 and so
0.5y
[∈]A but0.5x
[∈]A and so0.5x
[∈] ∨ [q]A. This is a contradiction. Hence A(x) ≤ max{A(y), 0.5} for
all x, y ∈ S with x ≤ y.Let x, y ∈ S and we consider the following cases:a) max{A(x),A(y)} ≤ 0.5,b) max{A(x),A(y)} > 0.5.
Case a): Let x, y ∈ S be such that max{A(x),A(y), 0.5} < A(xy). Then A(xy) > 0.5,0.5x
[∈]A and
0.5y
[∈]A but0.5xy
[∈] ∨ [q]A. This is a contradiction.
Case b): Let x, y ∈ S be such that max{A(x),A(y), 0.5} < A(xy). Then max{A(x),A(y)} <
A(xy). Choose r ∈ [0, 1) such that max{A(x),A(y), 0.5} ≤ r < A(xy). Thenr
x[∈]A and
r
y[∈]A but
r
xy[∈]A and so
r
xy[∈] ∨ [q]A. This is a contradiciton. Hence A(xy) ≤ max{A(x),A(y), 0.5} for all
x, y ∈ S.Let a, x, y ∈ S and we consider the following cases:a) max{A(x),A(y)} ≤ 0.5,b) max{A(x),A(y)} > 0.5.
LOBACHEVSKII JOURNAL OF MATHEMATICS Vol. 31 No. 1 2010
70 JUN et al.
Case a): Let a, x, y ∈ S be such that max{A(x),A(y), 0.5} < A(xay). Then A(xay) > 0.5,0.5x
[∈]A
and0.5y
[∈]A but0.5xay
[∈]A and so0.5xay
[∈] ∨ [q]A. This is a contradiction.
Case b): Let a, x, y ∈ S be such that max{A(x),A(y), 0.5} < A(xay). Then max{A(x),A(y)} <
A(xay). Choose s ∈ [0, 1) such that max{A(x),A(y), 0.5} ≤ s < A(xay). Thenr
x[∈]A and
r
y[∈]A but
r
xay[∈]A and
r
xay[∈] ∨ [q]A. This is a contradiciton. Hence A(xay) ≤ max{A(x),A(y), 0.5} for all
a, x, y ∈ S. �Remark 3.7. From Theorem 3.6, and Remark 3.3, it follows that every anti fuzzy bi-ideal of S
is an ([∈], [∈] ∨ [q])-fuzzy bi-ideal of S. But the converse is not true.Example 3.8. Consider the ordered semigroup given in Example 2.4, and define a fuzzy subset
A as follows:
A(a) = 0.2, A(b) = 0.4, A(c) = 0.8, A(d) = 0.6, A(e) = 0.5.
Then A is an ([∈], [∈] ∨ [q])-fuzzy bi-ideal of S. But A is not an anti fuzzy bi-ideal of S, because
0.6 = A(d) = A(ab) ≤ max{A(a),A(b)} = max{0.2, 0.4} = 0.4.
In the following Theorem we provide a condition for an ([∈], [∈] ∨ [q])-fuzzy bi-ideal to an ([∈], [∈])-fuzzy bi-ideal.
Theorem 3.9. Let A be an ([∈], [∈] ∨ [q])-fuzzy bi-ideal of S such that A(x) > 0.5 for all x ∈ S.Then A is an ([∈], [∈])-fuzzy bi-ideal of S.
Proof. Let x, y ∈ S, x ≤ y and t ∈ [0, 1) be such thatt
y[∈]A. Then A(y) ≤ t and by (i) of Theorem
3.6, it follows that
A(x) ≤ max{A(y), 0.5} ≤ A(y) ≤ t,
and sot
x[∈]A. Let a, x, y ∈ S and r, s ∈ [0, 1) be such that
r
x[∈]A. and
s
x[∈]A. Then A(x) ≤ r and
A(y) ≤ s. By (ii) and (iii) of Theorem 3.6, it follows that
A(xy) ≤ max{A(x),A(y), 0.5} ≤ max{r, s},A(xay) ≤ max{A(x),A(y), 0.5} ≤ max{r, s}.
Hencemax{r, s}
xy[∈]A and
max{r, s}xay
[∈]A and therefore A is an ([∈], [∈])-fuzzy bi-ideal of S. �
For any fuzzy subset A of S and t ∈ [0, 1), we denote
Q(A; t) :={
x ∈ S
∣∣∣∣t
x[q]A
}
and [A; t] :={
x ∈ S
∣∣∣∣t
x[∈] ∨ [q]A
}
.
Obviously [A; t] = C(A; t) ∪ Q(A; t).In the following Theorem we provide another characterization of ([∈], [∈] ∨ [q])-fuzzy bi-ideal of S by
using the set [A; t].Theorem 3.9. A fuzzy subset A of S is an ([∈], [∈] ∨ [q])-fuzzy bi-ideal of S if and only if [A; t]
is a bi-ideal of S for all t ∈ [0, 1).We call [A; t] an ([∈] ∨ [q])-level bi-ideal of A.Proof. =⇒. Suppose that A is an ([∈], [∈]∨ [q])-fuzzy bi-ideal of S. Let x, y ∈ S, x ≤ y and t ∈ [0, 1)
be such that y ∈ [A; t]. Thent
y[∈]∨ [q]A and so A(y) ≤ t or A(y) + t < 1. Since A is an ([∈], [∈]∨ [q])-
fuzzy bi-ideal of S and x ≤ y we have
A(x) ≤ max{A(y), 0.5},then we have the following cases:
LOBACHEVSKII JOURNAL OF MATHEMATICS Vol. 31 No. 1 2010
GENERALIZED ANTI FUZZY Bi-IDEALS 71
Case 1 A(y) ≤ t. If t ≤ 0.5, then A(x) ≤ 0.5 and so A(x) + t < 0.5 + 0.5 = 1. Hencet
x[q]A If t > 0.5
then A(x) ≤ max{A(y), 0.5} ≤ t and sot
x[∈]A.
Case 2 A(y) + t < 1. If t ≤ 0.5 then
A(x) ≤ max{A(y), 0.5} ≤ max{1 − t, 0.5} = 1 − t,
and so A(x) < 1 − t =⇒ A(x) + t < 1. Hencet
x[q]A. If t > 0.5 then
A(x) ≤ max{A(y), 0.5} ≤ max{1 − t, 0.5} = 0.5 ≤ t,
and sot
x[∈]A. Thus in both cases, we have
t
x[∈] ∨ [q]A and hence x ∈ [A; t]. Let x, y ∈ S and t ∈ [0, 1)
be such that x, y ∈ [A; t]. Thent
x[∈] ∨ [q]A and
t
y[∈] ∨ [q]A. Hence A(x) ≤ t or A(x) + t < 1 and
A(y) ≤ t or A(y) + t < 1. Since A is an ([∈], [∈] ∨ [q])-fuzzy bi-ideal of S, we have
A(xy) ≤ max{A(x),A(y), 0.5}.
Then we have the following cases:Case 1 Let A(x) < t and A(y) < t. If t ≤ 0.5. Then
A(xy) ≤ max{A(x),A(y), 0.5} = 0.5
and hencet
xy[q]A. If t > 0.5. Then
A(xy) ≤ max{A(x),A(y), 0.5} ≤ t
and sot
xy[∈]A. Hence
t
xy[∈] ∨ [q]A.
Case 2 Let A(x) ≤ t and A(y) + t < 1. If t ≤ 0.5, then
A(xy) ≤ max{A(x),A(y), 0.5} = max{A(y), 0.5}max{1 − t, 0.5} = 1 − t,
so A(xy) + t < 1 and hencet
xy[q]A. If t > 0.5. Then
A(xy) ≤ max{A(x),A(y), 0.5} ≤ max{t, 1 − t, 0.5} = t,
and hencet
xy[∈]A. Thus
t
xy[∈] ∨ [q]A.
Case 3 Let A(x) + t < 1 and A(y) ≤ t. If t ≥ 0.5, then
A(xy) ≤ max{A(x),A(y), 0.5} ≤ max{A(x), 0.5} = max{1 − t, 0.5} = 1 − t,
so A(xy) + t < 1 and hencet
xy[q]A. If t < 0.5, then
A(xy) ≤ max{A(x),A(y), 0.5} ≤ max{1 − t, t, 0.5} = t,
and so A(xy) ≤ t =⇒ t
xy[∈]A. Hence
t
xy[∈] ∨ [q]A.
Case 4 Let A(x) + t < 1 and A(y) + t < 1. If t ≤ 0.5, then
A(xy) ≤ max{A(x),A(y), 0.5} ≤ max{1 − t, 0.5} = 1 − t,
and so A(xy) + t < 1, hencet
xy[q]A. If t > 0.5 then
A(xy) ≤ max{A(x),A(y), 0.5} ≤ max{1 − t, 0.5} = 0.5 ≤ t,
hencet
xy[∈]A. Thus
t
xy[∈] ∨ [q]A. Therefore in any case we have
t
xy[∈] ∨ [q]A and hence xy ∈ [A; t].
LOBACHEVSKII JOURNAL OF MATHEMATICS Vol. 31 No. 1 2010
72 JUN et al.
Now, let a, x, y ∈ S and t ∈ [0, 1) be such that x, y ∈ [A; t]. Thent
x[∈]∨ [q]A and
t
y[∈]∨ [q]A. Hence
A(x) ≤ t or A(x) + t < 1 and A(y) ≤ t or A(y) + t < 1. Since A is an ([∈], [∈] ∨ [q])-fuzzy bi-ideal ofS, we have
A(xay) ≤ max{A(x),A(y), 0.5}.
Then we have the following cases:
Case 1 Let A(x) < t and A(y) < t. If t ≤ 0.5. Then
A(xay) ≤ max{A(x),A(y), 0.5} = 0.5
and hencet
xay[q]A. If t > 0.5. Then
A(xay) ≤ max{A(x),A(y), 0.5} ≤ t
and sot
xay[∈]A. Hence
t
xay[∈] ∨ [q]A.
Case 2 Let A(x) ≤ t and A(y) + t < 1. If t ≤ 0.5, then
A(xay) ≤ max{A(x),A(y), 0.5} = max{A(y), 0.5}max{1 − t, 0.5} = 1 − t,
so A(xay) + t < 1 and hencet
xay[q]A. If t > 0.5. Then
A(xay) ≤ max{A(x),A(y), 0.5} ≤ max{t, 1 − t, 0.5} = t,
and hencet
xay[∈]A. Thus
t
xay[∈] ∨ [q]A.
Case 3 Let A(x) + t < 1 and A(y) ≤ t. If t ≥ 0.5, then
A(xay) ≤ max{A(x),A(y), 0.5} ≤ max{A(x), 0.5} = max{1 − t, 0.5} = 1 − t,
so A(xay) + t < 1 and hencet
xay[q]A. If t < 0.5, then
A(xay) ≤ max{A(x),A(y), 0.5} ≤ max{1 − t, t, 0.5} = t,
and so A(xay) ≤ t =⇒ t
xay[∈]A. Hence
t
xay[∈] ∨ [q]A.
Case 4 Let A(x) + t < 1 and A(y) + t < 1. If t ≤ 0.5, then
A(xay) ≤ max{A(x),A(y), 0.5} ≤ max{1 − t, 0.5} = 1 − t,
and so A(xay) + t < 1, hencet
xay[q]A. If t > 0.5 then
A(xay) ≤ max{A(x),A(y), 0.5} ≤ max{1 − t, 0.5} = 0.5 ≤ t,
hencet
xay[∈]A. Thus
t
xay[∈] ∨ [q]A. Therefore in any case we have
t
xy[∈] ∨ [q]A and hence xay ∈
[A; t]. Thus [A; t] is a bi-ideal of S.
Conversely, assume that A is a fuzzy subset of S and t ∈ [0, 1) be such that [A; t] is a bi-ideal of S. Ifpossible, let max{A(y), 0.5} ≤ t < A(x) for some t ∈ [0, 1). Then y ∈ C(A; t) ⊆ [A; t]. Since x ≤ y and[A; t] is a bi-ideal of S, we have x ∈ [A; t]. Then A(x) ≤ t of A(x) + t < 1. This is a contradition. Thus
A(x) ≤ max{A(y), 0.5} for all x, y ∈ S with x ≤ y.
Let x, y ∈ S and t ∈ [0, 1) be such that
max{A(x),A(y), 0.5} ≤ t < A(xy),
LOBACHEVSKII JOURNAL OF MATHEMATICS Vol. 31 No. 1 2010
GENERALIZED ANTI FUZZY Bi-IDEALS 73
then x, y ∈ C(A; t) ⊆ [A; t]. Since [A; t] is a bi-ideal of S, we have xy ∈ [A; t]. Then A(xy) ≤ t orA(xy) + t < 1. This is a contradiction. Hence A(xy) ≤ max{A(x),A(y), 0.5} for all x, y ∈ S. Finally,let a, x, y ∈ S and t ∈ [0, 1) be such that
max{A(x),A(y), 0.5} ≤ t < A(xay),
then x, y ∈ C(A; t) ⊆ [A; t]. Since [A; t] is a bi-ideal of S, we have xay ∈ [A; t]. Then A(xay) ≤ tor A(xay) + t < 1. This is a contradiction. Hence A(xay) ≤ max{A(x),A(y), 0.5} for all x, y ∈ S.Therefore A is an ([∈], [∈] ∨ [q])-fuzzy bi-ideal of S. �
Theorem 3.10. Every ([∈]∨ [q], [∈]∨ [q])-fuzzy bi-ideal is an ([∈], [∈]∨ [q])-fuzzy bi-ideal of S.Proof. Let A be an ([∈] ∨ [q], [∈] ∨ [q])-fuzzy bi-ideal of S. Let x, y ∈ S with x ≤ y and t ∈ [0, 1)
be such thatt
y[∈]A. Then
t
y[∈] ∨ [q], [∈] ∨ [q]A. Since A is an ([∈] ∨ [q], [∈] ∨ [q])-fuzzy bi-ideal of S,
and x ≤ y we havet
x[∈] ∨ [q]A. Let x, y ∈ S and r, s ∈ [0, 1) be such that
t
x[∈]A, and
t
y[∈]A. Then
r
x[∈] ∨ [q]A, and
s
y[∈] ∨ [q]A and hence
max{r, s}x
[∈] ∨ [q]A. Finally, let a, x, y ∈ S and r, s ∈ [0, 1) be
such thatt
x[∈]A, and
t
y[∈]A. Then
t
x[∈] ∨ [q]A, and
t
y[∈] ∨ [q]A and we have
max{r, s}xay
[∈] ∨ [q]A.
Therefore A is an ([∈], [∈] ∨ [q])-fuzzy bi-ideal of S. �The converse of Theorem 3.10 is not true in general. Consider the ordered semigroup given in
Example 2.4, and define a fuzzy subset A : S −→ [0, 1] by
A(x) = 0.2, A(b) = 0.4, A(c) = 0.5, A(d) = 0.3, A(e) = 0.6.
Then A is an ([∈], [∈] ∨ [q])-fuzzy bi-ideal of S but A is not an ([∈] ∨ [q], [∈] ∨ [q])-fuzzy bi-ideal,
because0.22a
[∈]A and0.42
b[∈]A but
max{0.22, 0.42}ab
=0.42d
[∈]A.
Theorem 3.11. Every ([∈], [∈])-fuzzy bi-ideal is an ([∈], [∈] ∨ [q])-fuzzy bi-ideal of S.Proof. Straightforward. �Theorem 3.12. A fuzzy subset A of S is an ([∈], [∈] ∨ [q])-fuzzy bi-ideal of S if and only if
C(A; t) is a bi-ideal of S for all t ∈ [0.5, 1).Proof. Suppose that A is an ([∈], [∈] ∨ [q])-fuzzy bi-ideal of S. Let x, y ∈ S with x ≤ y and t ∈
[0.5, 1) be such that y ∈ C(A; t). Then A(y) ≤ t. Since A is an ([∈], [∈] ∨ [q])-fuzzy bi-ideal we have,
A(x) ≤ max{A(y), 0.5} ≤ t,
and so x ∈ C(A; t). Let x, y ∈ C(A; t). Then A(x) ≤ t and A(y) ≤ t. By hypothesis,
A(xy) ≤ max{A(x),A(y), 0.5} ≤ t,
and so xy ∈ C(A; t). Finally, let a, x, y ∈ S and t ∈ [0.5, 1) be such that x, y ∈ C(A; t). It follows from(iii) of Theorem that
A(xay) ≤ max{A(x),A(y), 0.5} ≤ t,
hence xay ∈ C(A; t). Therefore C(A; t) is bi-ideal of S.Conversely, assume that C(A; t) is a bi-ideal of S for all t ∈ [0.5, 1). Let x, y ∈ S with x ≤ y. If
possible, let there exists t ∈ [0.5, 1) such that
max{A(y), 0.5} ≤ t < A(x),
then A(y) ≤ t and we have y ∈ C(A; t). Since x ≤ y, we have x ∈ C(A; t). This is a contradiction. HenceA(x) ≤ max{A(y), 0.5} for all x, y ∈ S with x ≤ y. Let x, y ∈ S and t ∈ [0.5, 1) be such that
max{A(x),A(y), 0.5} ≤ t < A(xy),
then x, y ∈ C(A; t). Since C(A; t) is a bi-ideal we have xy ∈ C(A; t). Then A(xy) ≤ t. This is a contra-diction. Hence A(xy) ≤ max{A(x),A(y), 0.5} for all x, y ∈ S. Finally, let a, x, y ∈ S and t ∈ [0.5, 1) besuch that
max{A(x),A(y), 0.5} ≤ t < A(xay),
LOBACHEVSKII JOURNAL OF MATHEMATICS Vol. 31 No. 1 2010
74 JUN et al.
then x, y ∈ C(A; t) and hence xay ∈ C(A; t). Then A(xay) ≤ t. This is a contradiction.Hence A(xay) ≤ max{A(x),A(y), 0.5} for all a, x, y ∈ S. Thus A is an ([∈], [∈] ∨ [q])-fuzzybi-ideal of S. �
The following example shows that there exists an ([∈], [∈]∨ [q])-fuzzy bi-ideal of S such that Q(A; t)is not a bi-ideal of S for some t ∈ [0, 1).
Example 3.13. Consider the ordered semigroup S = {a, b, c, d, e} as given in Example 2.4, anddefine a fuzzy subset A by:
A(a) = 0.2, A(b) = 0.3, A(c) = 0.8, A(e) = 0.5, A(d) = 0.6.
Then Q(A; t) = {a, b, d, e} for all 0.5 < t ≤ 0.6. Since0.3a
[∈]A and0.6b
[∈]A. But
max{0.3, 0.6}ab
=0.6d
[q]A.
Definition 3.14. A fuzzy subset A of S is called an ([q], [q])-fuzzy bi-ideal of S if it satisfies:
(i) (∀x, y ∈ S)(t ∈ [0, 1))(x ≤ y,t
y[q]A =⇒ t
x[q]A),
(ii) (∀x, y ∈ S)(r, s ∈ [0, 1))(r
x[q]A,
s
y[q]A =⇒ max{r, s}
xy[q]A),
(iii) (∀a, x, y ∈ S)(r, s ∈ [0, 1))(r
x[q]A,
s
y[q]A =⇒ max{r, s}
xay[q]A).
Theorem 3.17. Let B be a bi-ideal of S and let A be a fuzzy subset of S such that(i) (∀x ∈ S\B)(A(x) = 1),(ii) (∀x ∈ B)(A(x) ≤ 0.5).Then A is both a ([q], [∈] ∨ [q])-fuzzy bi-ideal and an ([∈], [∈] ∨ [q])-fuzzy bi-ideal of S.
Proof. Let x, y ∈ S, x ≤ y and t ∈ [0, 1) be such thatt
y[q]A. Then y ∈ B. Since x ≤ y ∈ B, then
x ∈ B. If t ≥ 0.5 then A(x) ≤ 0.5 ≤ t and hencet
x[∈]A. If t < 0.5 then A(x) + t < 0.5 + 0.5 = 1 and
sot
x[q]A. Hence
t
x[∈]∨ [q]A for all x, y ∈ S with x ≤ y. Let x, y ∈ S and r, s ∈ [0, 1) be such that
t
x[q]A
andt
y[q]A. Then x, y ∈ B and hence xy ∈ B. If max{r, s} ≥ 0.5 then A(xy) ≤ 0.5 ≤ max{r, s} and so
max{r, s}xy
[∈]A. If max{r, s} < 0.5 then A(xy) + max{r, s} < 0.5 + 0.5 = 1, and hencemax{r, s}
xy[q]A.
Thusmax{r, s}
xy[∈] ∨ [q]A. Let a, x, y ∈ S and r, s ∈ [0, 1) be such that
t
x[q]A and
t
y[q]A. Then
x, y ∈ B and hence xay ∈ B. If max{r, s} ≥ 0.5 then A(xay) ≤ 0.5 ≤ max{r, s} and somax{r, s}
xay[∈]
A. If max{r, s} < 0.5 then A(xay) + max{r, s} < 0.5 + 0.5 = 1, and hencemax{r, s}
xay[q]A. Thus
max{r, s}xay
[∈] ∨ [q]A. Therefore A is an ([q], [∈] ∨ [q])-fuzzy bi-ideal of S.
Let x, y ∈ S, x ≤ y and t ∈ [0, 1) be such thatt
y[∈]A. ThenA(y) ≤ t and so y ∈ B. Since x ≤ y ∈ B,
then x ∈ B. If t ≥ 0.5 then A(x) ≤ 0.5 ≤ t and hencet
x[∈]A. If t < 0.5 then A(x) + t < 0.5 + 0.5 = 1
and sot
x[q]A. Hence
t
x[∈] ∨ [q]A for all x, y ∈ S with x ≤ y. Let x, y ∈ S and r, s ∈ [0, 1) be such that
t
x[∈]A and
t
y[∈]A. Then A(x) ∈ t and A(y) ≤ t and so x, y ∈ B and hence xy ∈ B. If max{r, s} ≥ 0.5
LOBACHEVSKII JOURNAL OF MATHEMATICS Vol. 31 No. 1 2010
GENERALIZED ANTI FUZZY Bi-IDEALS 75
then A(xy) ≤ 0.5 ≤ max{r, s} and somax{r, s}
xy[∈]A. If max{r, s} < 0.5 then A(xy) + max{r, s} <
0.5 + 0.5 = 1, and hencemax{r, s}
xy[q]A. Thus
max{r, s}xy
[∈] ∨ [q]A. Let a, x, y ∈ S and r, s ∈ [0, 1)
be such thatt
x[∈]A and
t
y[∈]A. Then A(x) ∈ t and A(y) ≤ t and so x, y ∈ B and hence xay ∈ B.
If max{r, s} ≥ 0.5 then A(xay) ≤ 0.5 ≤ max{r, s} and somax{r, s}
xay[∈]A. If max{r, s} < 0.5 then
A(xay) + max{r, s} < 0.5 + 0.5 = 1, and hencemax{r, s}
xay[q]A. Thus
max{r, s}xay
[∈] ∨ [q]A. Therefore
A is an ([q], [∈] ∨ [q])-fuzzy bi-ideal of S. �Note that every ([∈], [∈] ∨ [q])-fuzzy bi-ideal of S may not be ([q], [∈] ∨ [q])-fuzzy bi-ideal of S.Example 3.17. Consider an ordered semigroup S = {a, b, c, d, e} as given in Example 2.4, and
define an fuzzy subset A by:
A(a) = 0.2, A(b) = 0.4, A(c) = 0.5, A(d) = 0.3, A(e) = 0.6.
Then A is an ([∈], [∈] ∨ [q])-fuzzy bi-ideal. But A is not a ([q], [∈] ∨ [q])-fuzzy bi-ideal of S. Since0.11a
[q]A and0.21
b[q]A. But
max{0.11, 0.21}ab
=0.21d
[∈]A.
Proposition 3.18. If {Ai}i∈I is a family of ([∈], [∈] ∨ [q])-fuzzy bi-ideals of S. Then⋂
i∈I
Ai is an
([∈], [∈] ∨ [q])-fuzzy bi-ideal of S, where⋂
i∈I
Ai(x) =∧
i∈I
Ai(x) for all x ∈ S.
Proof. Let x, y ∈ S be such that x ≤ y. Then
⋂
i∈I
Ai(x) =∧
i∈I
Ai(x) ≤∧
i∈I
{Ai(y) ∨ 0.5} =
{∧
i∈I
Ai(y) ∨ 0.5
}
=
{(⋂
i∈I
Ai
)
(y) ∨ 0.5
}
.
Let a, x, y ∈ S. Then
⋂
i∈I
Ai(xy) =∧
i∈I
Ai(xy) ≤∧
i∈I
{Ai(x) ∨ Ai(y) ∨ 0.5} =
{∧
i∈I
Ai(x) ∨∧
i∈I
Ai(y)
}
∨ 0.5
=
{(⋂
i∈I
Ai
)
(x) ∨(
⋂
i∈I
Ai
)
(y) ∨ 0.5
}
,
and⋂
i∈I
Ai(xay) =∧
i∈I
Ai(xay) ≤∧
i∈I
{Ai(x) ∨ Ai(y) ∨ 0.5} =
{∧
i∈I
Ai(x) ∨∧
i∈I
Ai(y)
}
∨ 0.5
=
{(⋂
i∈I
Ai
)
(x) ∨(
⋂
i∈I
Ai
)
(y) ∨ 0.5
}
.
Thus⋂
i∈I
Ai is an ([∈], [∈] ∨ [q])-fuzzy bi-ideal of S. �
Note that if {Ai}i∈I is a family of ([∈], [∈] ∨ [q])-fuzzy bi-ideals of S then⋃
i∈I
Ai may not be an
([∈], [∈] ∨ [q])-fuzzy bi-ideals of S.Example 3.19. Let S = {a, b, c, d, e} be an ordered semigroup, which is given in Example 2.4, and
let A1 and A2 be a fuzzy subset of S which are defined as follows:
A1(a) = 0.3, A1(b) = 0.5, A1(d) = 0.6, A1(c) = A1(e) = 0.7
LOBACHEVSKII JOURNAL OF MATHEMATICS Vol. 31 No. 1 2010
76 JUN et al.
A2(a) = 0.5, A2(b) = A2(c) = A2(d) = 0.8, A2(e) = 0.75.
Then
C(A1; t) :=
{S if 0.7 ≤ t < 1{a, b, d} if 0.6 ≤ t < 0.7
C(A2; t) :=
{S if 0.8 ≤ t < 1{a} if 0.5 ≤ t < 0.8.
The union of A1 and A2, A1 ∪ A2 is given by:
A1 ∪ A2(a) = 0.5, A1 ∪A2(b) = A1 ∪ A2(c) = A1 ∪ A2(d) = 0.8, A1 ∪A2(a) = 0.75.
Then
C(A1 ∪ A2; t) :=
{S if 0.8 ≤ t < 1{a, e} if 0.5 ≤ t < 0.8.
Since {a, e} is not a bi-ideal of S, it follows from Theorem 3.6 that A1 ∪ A2 is not an ([∈], [∈] ∨ [q])-fuzzy bi-ideal of S.
ACKNOWLEDGEMENT
We would like to thank for financial support during this work under CIIT research grant program.
REFERENCES1. S. K. Bhakat and P. Das, On the definition of a fuzzy subgroup, Fuzzy Sets and Systems 51, 235 (1992).2. S. K. Bhakat and P. Das, (∈,∈ ∨q)-fuzzy subgroups, Fuzzy Sets and Systems 80, 359 (1996).3. A. Borumand Saeid and Y. B. Jun, Redefined fuzzy subalgebras of BCK/BCI-algebras, Iranian J. Fuzzy
Systems (to appear).4. Y. B. Jun and S. Z. Song, Generalized anti fuzzy subalgebras, East Asian Math. J. 22 (2), 195 (2006).5. A. Khan, Y. B. Jun and M. Shabir, Ordered semigroups characterized by their (∈,∈ ∨q)-fuzzy bi-ideals,
Bull. Malays. Math. Sci. Soc. (2) 32 (3), 391 (2009).6. P. M. Pu and Y. M. Liu, Fuzzy topology I, Neighbourhood structure of a fuzzy point and Moore-Smith
convergence, J. Math. Anal. Appl. 76, 571 (1980).7. X. Yuan, C. Zhang and Y. Ren, Generalized fuzzy groups and many-valued implications, Fuzzy Sets and
Systems 138, 205 (2003).8. L. A. Zadeh, Fuzzy sets, Inform. and Control 8, 338 (1965).
LOBACHEVSKII JOURNAL OF MATHEMATICS Vol. 31 No. 1 2010