Damascus University Journal for BASIC SCIENCES Vol.

24, No 1, 2008

A Finite BCI-algebra of KL-Product

Ahmad Habil
Department of Mathematics Faculty of Sciences-University of Damascus-Syria
Received 09/05/2007 Accepted 15/07/2008

ABSTRACT
A non-empty set X with a binary operation * and a distinguished element 0 is called a BCI-algebra if the following axioms are satisfied: I) ((x y ) (x z )) (z y ) 0 II) (x (x

y )) y 0 III) x x 0 IV) x y 0, y x 0 for every x , y ,z X .

Let be X a finite BCI- algebra, it is known that X is of KL- product if and only if the following condition is satisfied:

(a e ) (0 e ) a ;

a X , e L (X )

We present a necessary and sufficient condition for BCI-algebra X to be of KL- product, this condition is pure numerical, that is the number of elements of the row which is opposite to the zero element in the Cayley table of the operation * divides the number of elements in each row of the mentioned table. Key words: BCI- algebra, KL -product.

Habil A Finite BCI-algebra of KL-Product

- KL

BCI

2007/ 05/ 09 2008/ 07/ 15

BCII)

X :

((x y ) (x z )) (z y ) 0

II)

(x (x y )) y

III) x x 0 IV) x y :

0, y x

x
- KL

y
.X X

x , y ,z
BCIX

(a e ) (0 e ) a ;
KLBCIX

a X , e L (X )

KL- BCI-

10

Damascus University Journal for BASIC SCIENCES Vol. 24, No 1, 2008

Introduction
The notion of BCK- algebras was proposed by Y. Iami and K. Iseki in 1966. In the same year K. Iseki [3] introduced the notion of BCI -algebra, which is a generalization of BCK- algebra. After that, many mathematical papers have been published investigating some algebraic properties of BCK/BCI-algebras and their relationship with other universal structures including lattices and Boolean algebras. 1) - Basic definitions and results Definition (1): A non-empty set X with a binary operation * and a distinguished element 0 is called a BCI - algebra if the following axioms are satisfied for every x , y , z X :

((x II) (x III) x IV) x

I)

y ) (x z )) (z y ) 0 (x y )) y 0 x 0 y 0, y x 0 x y

Definition (2): A BCI -algebra X is called BCK-algebra if it satisfied: V) 0 x 0 ; x X A partial ordering relation can be defined on BCI -algebra X for some x , y X by: x y if and only if x y 0 Remark: the axioms in definition (1) can be rewriting by using the symbol in the following simpler way: I ) (x II ) III )

y ) (x z ) z y

x (x y ) y
x y

x x IV ) x y , y x

And the axiom V in definition (2):

11

Habil A Finite BCI-algebra of KL-Product

V) 0 x ; x

Definition (3):[5] An element a in BCI -algebra X is called an atom if and only if : x a 0 x a ( x X ). Definition (4): [2] A BCI - algebra X is called p-semisimple when the following condition is satisfied: 0 (0 x ) x ( x X ) J. Meng and X. L. Xin [6] introduced the notion of KL-product BCI-algebras. Definition (5): Let X be BCI -algebra. If there exist BCK- algebra Y and p-semisimple BCI -algebra Z such that X Y Z , then X is called BCI -algebra of KL-product. Lemma (1): [7] An element a in BCI - algebra X is an atom if and only if x (x a) a ( x X ) Notation: We shall denote the subset of all atoms in BCI -algebra X by L(X) Lemma (2): [7] Let X be BCI -algebra then L (X ) 0 X . Theorem (1): [7] A BCI -algebra X is of KL- product if and only if it satisfies the condition:

(a e ) (0 e ) a ;

a X , e L (X )

Proposition (1): [4], [1] In any BCI -algebra X the following Proprieties hold for every x , y , z X : (1) x 0 (2) x (x

(x y )) x y (3) (x y ) z (x z ) y (4) (x y ) (z y ) x z
a
2) - Main Results Definition (6): Let X be BCI -algebra, then for any element X the subset T is defined by:
a

T a {x X :a (a x ) x }

12

Damascus University Journal for BASIC SCIENCES Vol. 24, No 1, 2008

Lemma (3): Let X be BCI -algebra, and (1) L (X ) T

0

a X

, then:

(2)

0, a T a

Proof: (1)(i) x T 0 (ii)

0 (0 x ) x

x 0 X , so T 0

0 X

x 0 X y X ;x 0 y 0 (0 x ) 0 (0 (0 y )) 0 y x
the other inclusion, 0 X (2)(i) a (a 0) a a 0 (ii)

x T 0 , So we have

T 0 and by lemma (2) L (X ) T 0 .

0 Ta

a (a a) a 0 a

a Ta

Proposition (2): Let X be BCI -algebra, then for any element

a X we have the following Proprieties: (1) T a X {a x : x X } a

(2) T (3) T (4) T (5) T
a
a x
0
a

a L (X ) T 0 ( a X )
Ta ( x X )
Ta
X Ta
Tx Ta

(6) x T a (7) T
0

T a if a is an atom.

Proof: (1) (i) y T a

y a (a y )

a X , so we have:

Ta

a X

(ii) on the other hand: y

a X

x X ;y a x

13

Habil A Finite BCI-algebra of KL-Product

a (a y ) a (a (a x )) a x
This implies

y Ta

a X

T a . So, T a a X .
X}

(2) T a {(a x ) a : x a

{(a a) x : x
(3) Let y T a then: y

X}

{0 x : x X } 0 X
x

L (X ) T 0

(a x ) ((a x ) y )

(a x ) ((a y ) x ) But by the proposition (1): a (a y )

y Therefore we have: a (a y ) y
Ta
Let y
a

y Ta .
T0 Ta

(4) By putting x=a in the preceding property we have:

Ta

(5) (i) Suppose that z T a

x X ;z a x .

X then by (3):

Tz
So

T (a
z y

x) y

Ta

Ta .
Ta

T a and T a X

(ii) Now (6)

Ta a X

Ta X

x Ta

u X ; x a u.
u

We have by (3): T a (7) (i) a is an atom And by (4)

T a , therefore T x

T a.

a L (X ) T 0

Ta

T0.
Ta

T0

Ta

, hence we have: T 0

14

Damascus University Journal for BASIC SCIENCES Vol. 24, No 1, 2008

(ii) If T 0

T a then a T 0

0 (0 a) a , so by lemma (2)

a L (X ) . Definition (7): Let X be a BCI- algebra, then for any element a X the subset S is defined by: a

S a {x X : x (x a) a}.
Proposition (3): Let X be a BCI- algebra, then for any elements a,b X we have the following Proprieties: (1) a S
a

(2) (3)

x Sa
Sa Sb

Sx

Sa

Tb T a

(4) S a

S a x for any x X

(5) S X , if a is an atom in X. a (6) S X . 0 (7) If a is not an atom then (X \ S ) X a Proof: (1) We have a (a a) a 0 a (2) Let

X \Sa

a Sa

x S a and

let

y S x then

x (x a ) a

and

y ( y x ) x which imply by proposition (1) a ( y ( y x )) (( y ( y x )) a) ( y ( y x )) (( y a) ( y x )) y ( y a) So a y ( y a) and clearly y ( y a) a So y ( y a) a y S a.

15

Habil A Finite BCI-algebra of KL-Product

(3) Suppose that S a

S b and x T b then b (b x ) x

b Sx
Tb T a

Sb

Sx

Sa

S x , So a S x

x T a Concluding
Sa Sb

In a similar way we can prove that (4) We know that

Tb T a

T a x T a therefore S a S a x . (5) We know that x (x a ) a ( x X ) if a is an atom by

lemma (1). So, S X if a is an atom. a (6) We have

0 0 0 0 X

0 is an atom

S0 X .

(7) If a is not an atom then S a Now let Then S x so, x y

x
y

X. X \ S a and y X , and suppose that x y S a

S a , but S x
X \Sa

Sx

x S a which is not true,

a

X \ S a which implies (X \ S a ) X

(X \ S a ) X

X \ S a , but of course: Because for any x X \ S we have:

x x 0 and finally (X \ S ) X X \ S . a a As a consequence of the preceding proposition we can write the following: Corollary (1): In any BCI -algebra X, the following properties are equivalents: (1) S S a b (2) T T
a b

b Sa Ta (4) a S Tb b
(3) Notation: We define a relation x y on BCI - algebra X by: T T .
x y

16

Damascus University Journal for BASIC SCIENCES Vol. 24, No 1, 2008

It is clear that is an equivalence relation, and by corollary (1) the equivalence class of an element a X is S T .
a a

Proposition (4): In any BCI - algebra X the function :S T S T ; x a x is well defined and injective,
a a 0 0

a X

Proof: If x S T then: a a

(x ) a x

(x (x a)) x
(x x ) (x a )

0 (x a)
So
a

(x ) 0 X
a

T0 X

T0 S0 T0

Therefore If

is well defined.

x 1, x 2 S a T a such that a (x 1) a (x 2 ) then: a x1 a x 2 Now x 1 a (a x 1) a (a x 2 ) x 2

is injective. a Theorem (2): In any finite BCI- algebra X the following conditions are equivalents: (1) (a e ) (0 e ) a ( a X and e L (X )) . (2) :S T S T ; x a x is surjective , a X .
a a a 0 0

Therefore

(3) Card (0*X) divides Card (I) ( I X : I X I) Proof: (1) (2) By proposition (4) the function is well defined and injective. a Also is surjective. a For e S T X T T 0 X L (X )
0 0 0 0

17

Habil A Finite BCI-algebra of KL-Product

e Is an atom Let

a (a e ) e .
x S a because

x a e , then clearly x T a , and

x (x a) (a e ) ((a e ) a) (a e ) ((a a) e ) (a e ) (0 e )
a
Finally So, for
a

x Sa Ta .
e S0 T 0 we have found

x S a T a such that

(x ) e so a is surjective, which prove (2).

(2) (1) Let be a X , e L (X ) S T , since 0 0

is surjective there

exists some x

S a T a such that

So, a x e But x T so x = a * e, also x S implies: a a

(x ) e a (a x ) a e .
a

a (a e ) ((a e ) a) (a e ) ((a a) e ) (a e ) (0 e )
Which prove (1). (2) (3) let I be any subset of X such that I X I , in this case it is easy to see that I is the union of disjoint subsets of the form S T ,
a a

because if a I then:

Ta a X
Sa Ta
so subsets S
a

I X
(S a

, but S T a a where

T a which implies that

C I,
and the

I.
I
a C

Ta)

T a ;a C are disjoints, this is possible because they are

18

Damascus University Journal for BASIC SCIENCES Vol. 24, No 1, 2008

equivalence classes of the equivalence relation have: Card(I ) Card(S a T a )

a C

defined above, so we

a have: Card(S a

Since

is surjective by (2) and injective by proposition (4) we

Card(I )

T a ) Card(S 0 T 0 ) and: T0) C Card(S 0 C Card(0 X )

= CardC Card(0 X )
Which implies that: Card (0 * X) divides Card (I). (3) (2): If Card (0 * X) divides Card (I) ( I X : I X For any a X we have: T X T
a a

I ) then:

(S a (X \ S a )) T a (S a T a ) ((X \ S a ) T a )
This implies CardT a Or Card(S a

Card(S a T a ) Card((X \ S a ) T a )

T a ) CardT a Card((X \ S a ) T a )

We have tow cases: (i)If a is not an atom then: (X And T X a Hence ((X \ S ) T ) X a a This implies: Card(0*X) [ Card(T ) a

\Sa) X

(X \ S a ) (Proposition 3)

T a (Proposition 2)
(X \ S a ) T a By (3) we have:

Card (0 * X) Card(T ) , Card (0 * X) Card((X \ S ) T ) a a a

Card((X \ S a ) T a ) ]

Card (0* X) Card(S T ) a a However, clearly S T 0 X , which implies: 0 0

Card(S 0 T 0 ) Card(S a T a )

Card(S 0 T 0 )

Card(S a T a )

19

Habil A Finite BCI-algebra of KL-Product

By proposition (4) the function inequality: Card(S a

is injective, so we got the

T a ) Card(S 0 T 0 ) concluding that:

is surjective, because it a is an injective function of finite subsets which have the same cardinality. (ii) If a is an atom, we know that S X S by Proposition (3), a 0 by Proposition (2). T T

Card(S a T a ) Card(S 0 T 0 )

Therefore, we have: S a

Ta S0 T0

similarly, as in (i),

is surjective. So is surjective a X , which prove (2). a Which conclude the proof of the theorem. By the preceding theorem and by theorem (1) we have: Corollary (1): A finite BCI - algebra X is of KL- product iff: Card (0 * X) divides Card (I) ( I X : I X I ). Remark: let I be a subset of finite BCI- algebra X such that I X I then we can see that if x I then T x X I , now
x

using the fact that

x T x by (2) in lemma (3) we can write:

1 2

I
x I

T x Or
j

I Ta Ta

..... T a in
s

such

way

that

Ta

T a whenever i

j , in this case we shall say that I is properly

written. Theorem (3): Let be X a finite BCI - algebra, then the following conditions are equivalents: (1) Card (0 * X) divides Card (I) ( I X : I X I) (2) Card (0 * X) divides Card (0 * X) ( a X ) Proof: (1) (2) Let I T a where a X .

20

Damascus University Journal for BASIC SCIENCES Vol. 24, No 1, 2008

By the property (5) of the proposition (2) the subset I satisfies the condition I X I , so, by (1) Card (0 * X) divides Card (I), but I T a X , which is clear by the proposition (2), and condition
a

(2) is proved. (2) (1) First we define the subsets:

X n {x

X :Card(T x ) n } ,

for n any natural number, then we have a sequence of natural numbers: n n n .... n .... where we suppose that for
0 1 2 k

any number

nk

there exist at least one element x

X such that:

Card(T x ) nk and if we have n k

element

m n k 1 then there is no

X such that Card(T x ) m .

k

It is clear by propriety (4) of the proposition (2) that Card(T ) n . 0 0 We will prove (1) by induction on k the index of the numbers n . Now, the statement of induction for k-1 is: If I is any subset of Xsuch that: I X I and Card(T x ) nk 1; x I then: Card (0 * X) divides Card (I). Now for k 0 , Suppose that I is a subset of X such that I X I and Card(T ) n ; x I , but T T by proposition (2), so
x 0

Card(T 0 ) Card(T x ) n0 then it is clear that card(T x ) n0 , because the fact mentioned above that Card(T ) n , also we have: T Tx , 0 0 0

I however I
x I

T x therefore, in this case I T 0 .

So, Card(T ) divides Card (I), or Card (0 * X) divides Card (I), and 0 the statement of induction is true for k=0. Now, we prove the statement for k: Suppose that I is any subset of Xsuch that I X

I and

21

Habil A Finite BCI-algebra of KL-Product

Card(T x ) nk ; x I , We shall prove that Card (0 * X) divides

Card (I), provided that the assertion is true for k-1. As we have seen above in the remark, the subset I can be properly Written in the form: I T T ..... T .
a1 a2 as

By set theory we have:

Card(I )
i

Card(T ) ai

i j

Card(T T ) ai a j

i j k

Card(T T T ) ... ai a j ak

Now it is clear that the first sum is divisible by Card(0 X ) , because this number divides Card(T ), i by (2).
ai

For the other sums, the subsets like T a

Ta , Ta
j

Ta

Ta ,
k

will be denoted by J. We shall prove the following for the subset

J Ta

Ta

,
j

which we can do with the remaining subsets in the same way. It is clear by property (5) in the proposition (2) that the subset J Satisfies the condition: J X J and can be written as the union of subsets of the form T , just as it was remarked above so:
x

J Ta

Ta

j
bt

Tb

Tb
1)

..... T b , (Properly Written), here we

r
t

have: Card(T ) n (k some 1

,(1 t r ) , because if Card(T b ) nk , for

T a ) Card(T a ) nk
j j i

t
t t

r then we have:
i i

nk =Card(T b ) Card(T a nk =Card(T b ) Card(T a

T a ) Card(T a ) nk
j
i

So, we have: Card(T a

T a ) Card(T a ) Card(T a ) which implies

j i j

Ta

Ta

Ta
in

Ta
the

but this is impossible because I is properly

j

Written

form

I Ta Ta
1

..... T a

so

22

Damascus University Journal for BASIC SCIENCES Vol. 24, No 1, 2008

Card(T b ) n(k
t

1)

,(1 t r ) , from this we have by using property (6)

in proposition (2) that:

Card(T x ) n(k

1)

; x J And by the statement of induction:

i

Card(0 X ) divides Card(J ) Card(T a

Ta )
j

In the same way we proceed for the remaining subsets in the Other sums, so Card (0 * X) divides each of these sums, hence: Card (0 * X) divides Card (I). Which prove (1). So the theorem is proved. By the preceding theorem and corollary (1) of the theorem (2) we have: Corollary (1): A finite BCI- algebra X is of KL- product iff Card (0 * X) divides Card(a X ) ( a X ) . In other words a finite BCI -algebra X is of KL- product iff the number of atoms in X divides the number of elements of any row in the Cayley table of the binary operation*. Example: The last corollary gives us a simple method for examining if a BCI -algebra is of KL- product or not. Here we have these tow examples: 0 1 2 3 4 5 0 1 2 3 4 5 0 0 0 0 3 3 3 0 0 0 0 3 3 3 1 1 0 0 4 3 3 1 1 0 0 3 3 3 2 2 2 0 5 5 3 2 2 2 0 5 5 3 3 3 3 3 0 0 0 3 3 3 3 0 0 0 4 4 3 3 1 0 0 4 4 3 3 1 0 0 5 5 5 3 2 2 0 5 5 5 3 2 2 0 Where the first table represents a BCI- algebra of KL- product but the second does not. Because the number of atoms in the first and the second BCI-algebras is 2 (the number of the elements of the first row in the tow tables), it is clear that the condition of the precedent corollary is satisfied in the first table but not in the second.

23

Habil A Finite BCI-algebra of KL-Product

REFERENCES
[1] Dudek, W. A. (1988). On the axioms system for BCI-algebras, Prace Nauk. Wsp, Czestochows, Mathematyka 2. [2] Hoo, C. S. (1990). Closed ideals and p-semisimple BCI-algebras, Math. Japonica. 35, 1103-1112. [3] Iseki, K. (1966) An algebra related with a prepositional czlculus, Pros. Japan. Acad. 42, 26-29. [4] Iseki, K. (1980). On BCI-algebras, Math. Sem. Notes 8, 125-130. [5] Meng, J. and Xin, X. L. (1992). Characterizations of atoms in BCIalgebras, Math. Japonica, 37, 359-361. [6] Meng, J. and Xin, X. L. (1993). A problem in BCI-algebras, Math. Japonica, 38, 723-725. [7] Wei, S. M., Bai, G. Q., Meng, J. and Wang, Y. Q. (1999). Quasi- Implicative BCI-algebras, Math. Japonica 50, No. 2, 227- 233.

24