INDUCTIVE LEARNING OF COMPLEX FUZZY RELATION

International Journal of Computer Science, Engineering and Information Technology (IJCSEIT), Vol.1, No.5, December 2011
DOI : 10.5121/ijcseit.2011.1503 29
INDUCTIVE LEARNING OF COMPLEX FUZZY
RELATION
S. K. Das1
, D.C. Panda2
, Nilambar Sethi3
and S. S. Gantayat3
1
Department of Computer Science, Berhampur University, Berhampur, Orissa
dr.dassusanta@yahoo.co.in
2
Department of Electronics & Communication Engineering
Jagannath Institute of Technology & Management, Paralakhemundi, Orissa
d_c_panda@yahoo.com
3
Department of Information Technology
Gandhi Institute of Engineering & Technology, Gunupur, Orissa
nilambar_sethi@rediffmail.com
sgantayat67@rediffmail.com
ABSTRACT
The objective of this paper to investigate the notion of complex fuzzy set in general view. In constraint to a
traditional fuzzy set, the membership function of the complex fuzzy set, the range from [0.1] extended to a
unit circle in the complex plane. In this article the comprehensive mathematical operations on the complex
fuzzy set are presented. The basic operation of complex fuzzy set such as union, intersection, complement
of complex fuzzy set and complex fuzzy relation are studied. Also vector aggregation and fuzzy relation
over the complex fuzzy set are discussed. Two novel operations of complement and projection for complex
fuzzy relation are introduced.
KEY WORD
Complex fuzzy set, complex fuzzy relation, complement of complex fuzzy relation, projection of complex
fuzzy relation.
1. INTRODUCTION
The notion of the fuzzy set was first introduced by Zadeh in a seminal paper in 1965 [7]. The
notion of the fuzzy set A on the universe of discourse U is the set of order pair {(x, µA(x)), x ε U}
with a membership function µA(x), taking the value on the interval [0,1]. Ramot et al [5],
extended the fuzzy set to complex fuzzy set with membership function, ( )siw x
sz r e=
where 1i = − , which ranges in the interval [0,1] to a unit circle. Ramot et al [4], also introduced
different fuzzy complex operations and relations, like union, intersection, complement etc., Still it
is necessary to determine the membership functions correctly, which will give the appropriate or
approximate result for real life applications. The membership function defined for the complex
fuzzy set ( )siw x
sz r e= , which compromise an amplitude term rs(x) and phase term ws. The
amplitude term retains the idea of “fuzziness” and phase term signifies declaration of complex

30
fuzzy set, for which the second dimension of membership is required. The complex fuzzy set
allows extension of fuzzy logic that is i to continue with one dimension gradeness of membership.
Xin Fu et al [8] defined the fuzzy complex membership function of the form z=a+ib, where x, y
are two fuzzy numbers with membership function µA(x), µB(x) respectively. If b does not exist, z
degenerates to a fuzzy number.
Figure 1.1. Fuzzy Complex Number
Xin Fu et al [8], also discussed a complex number in cartesian form where a= rscos(x) and
b=rssin(x), which are in polar forms defined in [4]. The fuzzy number is created by interpolating
complex number in the support of fuzzy set ([1], [2], [3][6]).
Complex fuzzy set is a unique framework over the advantage of traditional fuzzy set .The support
of complex fuzzy set is unrestricted, may include any kind of object such as number, name etc,
which is off course a complex number. The notion of T-norm and T-conorm are used through out
this paper [4].
1.1. Definition T-norm (Jang et al [4]).
Four of the most frequently used T-norm operators are
Minimum Tmin(a,b)=min(a,b)=a∧ b
Algebraic Product Tap(a,b)=ab.
Bounded product Tbp(a,b)=0∨ (a+b-1)
Drastic product Tdp(a,b)=
a, if b=1
b, if a=1
0, if a,b<1





With understanding that a and b are between 0 and 1.
1.2. Definition T-Conorm (Jang et al [4])
Corresponding to four T-norm operators in the previous definition we have the following four
T-conorm operators.
Maximum Smax(a,b)=max(a,b)=a∨ b
Algebraic Sum S(a,b)=a+b-ab
Bounded product S(a,b)=1 ∧ (a+b)

31
Drastic product S(a,b)=
a, if b=0
b, if a=0
1, if a,b>1





This paper is organized in the following order. The review of complex fuzzy set and operations
on it are discussed in section 2 and section 3 the complex fuzzy relation and operations are
mentioned on them. A novel idea to vector aggregation, complement of complex fuzzy relation
and projection of complex fuzzy relation are studied in section 4 followed by summary and
suggestion for future work are given in section 5.
2. COMPLEX FUZZY SET
2.1. Concept & Operations
Complex fuzzy set is the basis for complex fuzzy logic [7]. The formal definition of complex
fuzzy set was provided in [7] followed by a discussion and interpretation of the moral concept. In
addition, several set theoretic operations on complex fuzzy set were discussed with several
examples, which illustrate the potential applications of complex fuzzy set in information
processing.
2.2. Definition of Complex Fuzzy Set
A complex fuzzy set S, defined on a universe of discourse U, is characterized by a membership
function ( )xsµ , which can assign any element x∈U as a complex valued grade of membership in
S. The value of ( )xsµ lies in a unit circle in the complex plane and in the form ( )siw x
sz r e= ,
where 1−=i and rs ∈ [0,1].
The complex fuzzy set S, may be defined as the set of order pair given by
( ){ }, ( ) :sS x x x Uµ= ∈ (1)
2.3. Complex Fuzzy Union
2.3.1. Definition ([6,7]).
In the traditional fuzzy logic, union for two fuzzy set A and B on U denoted as A∪B is specified
by a function µ , where [ ] [ ] [ ]1,01,01,0: →×µ . (2)
The membership function µ (A∪B), may be defined as one of followings.
(i) Standard Union ( ) ( ) ( )max ,AUB A Bx x xµ µ µ=    (3)
(ii) Algebraic Sum ( ) ( ) ( ) ( ) ( ).AUB A B A Bx x x x xµ µ µ µ µ= + − (4)
(iii)Bound Sum ( ) ( ) ( )min 1,AUB A Bx x xµ µ µ= +   (5)

32
The fuzzy union must satisfy the definition 2.2 and is equivalent to the proposed T-Co norms [9].
2.3.2. Properties
The Complex fuzzy union is defined by a set of axioms. These axioms represent properties of
complex fuzzy union functions that must satisfy in order to intuitively acceptable values.
(a) Complex fuzzy union does not satisfy the closure property.
In the algebraic sum union function does not satisfy the closure property for complex fuzzy set,
which can be realized from the following example.
Example 2.3.1. ( ) ( ) ixx BA == µµ
( ) ( )
i
x
B
x
A
x
B
x
A
x
BA 21.
)()()(
+=−+=⇒ µµµµµ U
It shows that the membership function lies outside of the unit circle. i.e. 2 2
1 2+ = 5 >1.
(b) Complex Fuzzy union must be monotonic
Since complex number is not linearly order, monotonicity of the complex numbers is required.
Example 2.3.2. Suppose ( ) ( ), ,b d a b a dµ µ≤ ⇒ ≤ using definition2.3.1 (2)
Similarly the max and min operators used in (3) & (5) are not applicable to complex valued
grades of membership. Here we can apply traditional fuzzy definition of union by keeping same
approach to complex part and different approach to phase may be defined in the following way.
2.3.3. Definition ([6, 7]).
Suppose A and B be two complex fuzzy set on U with complex valued membership
function ( ) ( )xx Bn µµ , respectively. Then complex fuzzy union A∪B may be defined as
( ) ( ) ( ) ( )
( )
A Biw x
A B A Bx r x r x eµ ∪
∪ = ⊕   (7)
where ⊕ represents the T-Conorm and BAW ∪ is defined as follows.
a) (Sum) BABA WWW +=∪ (8)
b) (Max ) ( )BABA WWW ,max=∪ (9)
c) (Min) ( )BABA WWW −=∪ min (10)
and
d) (Winner takes all)
A A B
AUB
B B A
W if r r
W
W if r r
>
= 
≥
(11)
To illustrate the above, consider the following example.

33
Example 2.3.3: Suppose
i×0 i π iπ/2
1 e 0.4 e 0.8 e
, ,
-1 0 1
A
 
=  
 
and
i 3π/4 i 2π i π/5
0.2 e 0.3 e 1e
, ,
-1 0 1
B
 
=  
 
Then
0 π iπ i π/5
1e 0.4e 1e
, ,
-1 0 1
A B
 
∪ =  
 
Here the union operation is (1) of definition 2.3.1.
2.4. Fuzzy Complex Intersection
It should be noted that the derivation of fuzzy complex intersection closely related to complex
fuzzy union, which clear from the following definition.
2.4.1. Definition([6,7]). The intersection for two complex fuzzy set A and B on U defined as
( ) ( ) ( )
* A Bi w x
A BA B r x r x e ∩
∩ =    (12)
where * is denoted as T–norm.
Example 2.4.1. From the example 2.3.3, we can find the fuzzy intersection of the two complex
fuzzy sets A and B as follows.
i3π/4 i 2 π i5π/2
0.2e 0.3e 0.8 e
, ,
-1 0 1
A B
 
∩ =  
 
Note: The fuzzy union and intersection are not unique, As this depend on the type of functions
used in T-Conorm and T-norms.
2.5. Complex Fuzzy Complement
We known that the complement of a fuzzy set A on U with membership function ( )xAµ is
defined as, ( ( )xAµ−1 .) (13)
The similar concept can be applied to the complex fuzzy set. But it is not always true for some
cases. To illustrate this, consider the following example.
Example 2.5.1. Suppose the membership function of a complex fuzzy set is ( ) 1 i
s x e π
µ = .
That is, ( ) 1 i
s x e π
µ = [ ]ππ sincos1 i+= = −1
So the complement of ( ) 1 i
s x e π
µ = is ( ) 1 ( 1) 2s xµ = − − = , which is outside of the unit circle.
This is violating of the closure properties of Fuzzy complex number.
2.5.1. Definition ([6, 7]). Suppose A is a fuzzy set defined as U. The complex fuzzy
complement for the fuzzy set A is defined as ( )(1 )A xµ− , where
( ) ( )[ ])(sin)(cos xwxwxrx sssA +=µ (14)
This form of complex fuzzy complement does not reduce to its traditional fuzzy logic,
if ( ) 0=xws . That is,

34
when ( ) 0s xµ = , ( )( ) [ ]1 ( )s sC x r x iµ = − + which is again a complex number.
Hence we can write ( )( ) ( )( ) ( )
. sic wx
s sC x C r x eµ = (15)
Consider the function C: [0, 1] → [0, 1], if ( ) ( )( )s sx C xµ µ= then the equation (13) holds. So
we need ( ) ( )( )xCxC ss µµ =. to satisfy the equation (13). This can be achieved by adding π
to ( )sw x .
Hence the complement of ( )sw x will be ( )sw x π+ that is ( ) ( )( )s sC x C xµ µ π= + . (16)
3. COMPLEX FUZZY RELATION
In this section, the complex fuzzy relation is introduced. Here discussion is limited to the relation
between two complex fuzzy set.
3.1. Definition (Complex Fuzzy Relation) ([10]): Let A and B be two complex fuzzy sets
defined on U. Then
( ) ( ) ( )
( ) ( ) ( )
. a n d .A Biw x iw x
A A B Bx r x e x r x eµ µ= =
are their corresponding membership values. We can say A B≤ if and only if both amplitude and
phase of A and B are A B≤ that is ( )Ar x ≤ ( )Br x and ( )Aw x ≤ ( )Bw x .
A complex fuzzy relation R (U,V) is a complex fuzzy subset of the product space U x V.
The relation R(U,V) is characterized by a complex membership function
( ), , ,R x y x U y Vµ ∈ ∈ (17)
Where ( ) ( ) ( ) ( ){ }, ( , , , ) ,RR U V x y x y x y U Vµ= ∀ ∈ ×
Here the membership function ( ),R x yµ has the value within an unit circle in the complex plane
with the from ( ) ( ) ( )yxwi
R eyxryx ,
.,, =µ
where ( )yxwi
e ,
is a periodic function defined as ( ) ( ), , 2 , 0, 1, 2w x y w x y k kπ= + = ± ± (18)
3.2. Definition ([10])
Let X, Y and Z be three different Universes. Suppose A be a complex fuzzy relation defined from
X to Y and B be a complex fuzzy relation defined from Y to Z. Then composition of A and B
denoted on AoB is a complex fuzzy relation defined from X to Z which can be represented as
( ) ( ) ( )
( ) ( )( )
( ) ( )( )sup(inf , , ,
,
, , . sup inf , , , .
A B
yAoB
i w x y w y z
i w x z
AoB AoB A B
y
x z r x z e r x y r y z eµ ∀
∀
= = (19)

35
3.3. Definition (Complex Fuzzy Union Relation) ([10])
Let A and B be two complex fuzzy relations defined on U x V. The corresponding membership
functions of A and B are defined as
( ) ( ) ( ),
, , . Ai w x y
A Ax y r x y eµ = and ( ) ( ) ( ),
, , . Bi w x y
B Bx y r x y eµ = respectively.
The complex fuzzy union relation of A and B is defined as A∪B by membership function
( ),A B x yµ ∪ where
( ) ( ) ( ),
, , . AUBi w x y
A B AUBx y r x y eµ ∪ = ( ) ( )( ) ( ) ( )max ( , , , )
max , , , . A Bi w x y w x y
A Br x y r x y e= (20)
where ( ),A Br x y∪ is a relation valued function on [0,1] and ( ),A B x yi w
e ∪
is a periodic function
defined in (18)
3.4. Definition (Complex Fuzzy Intersection Relation) ([10])
Let A and B be two complex fuzzy sets defined on U x V with their corresponding membership
functions as
( ) ( ) ( )
( ) ( ) ( ), ,
, , and , , ,A Bi w x y i w x y
A A B Bx y r x y e x y r x y eµ µ= =
respectively. The complex fuzzy intersection relation of A and B is defined as A∩B by
membership function ( ),A B x yµ ∩
where ( ) ( ) ( ),
, , . A Bi w x y
A B A Bx y r x y eµ ∩
∩ ∩=
( ) ( )( ) ( ) ( )min( , , , )
min , , , . A Bi w x y w x y
A Br x y r x y e= (21)
Here ( )yxr BA ,∩ is a relation valued function on [0,1] and ( ),A B x yi w
e ∩
is a periodic function.
4. VECTOR AGGREGATION FOR COMPLEX FUZZY SET:
It allows the aggregation of complex fuzzy set which incorporate phase consideration.
4.1. Definition (Vector Aggregation) ([7])
Suppose C1, C2…… Cn be complex fuzzy sets on U. The vector aggregation of C1, C2…… Cn
may be defined by a function v as
{ } { }: : , 1 : , 1
n
v a a C a b b C b∈ ≤ → ∈ ≤ for all x U∈ , v is given by
( ) ( ) ( ) ( )( ) ( )1 2
1
, ................
n
A A A An i Ai
i
x v x x x w xµ µ µ µ µ
=
= = ∑ (22)
where { }
1
, 1 1
n
i i
c
w a a C a for all i and w
=
∈ ∈ ≤ =∑
Note: In Complex vector aggregation, the amplitude of the sum may reduce to individual
amplitude.

36
4.2. Definition (Complement of Complex Fuzzy Relation)
In traditional fuzzy set, the complement of a relation R denoted as R defined by
( ) ( ), 1 , , ,RR
x y x y x y U Vµ µ= − ∀ ∈ × (23)
Suppose the complex fuzzy relation R is defined as U V× with membership
function ( ) ( ) ( ),
, , . i w x y
R x y r x y eµ = . Then the complement of ( ),R x yµ is given by
( )( ) ( ) ( )[ ] ( ) ( )[ ]yxwyxriyxwyxryxC R ,sin.,1,cos.,1, −+−=µ 24)
And it should satisfy the periodic function defined in (16).
4.3. Definition (Projection of a Complex Fuzzy Relation)
Consider complex fuzzy relation R defined on U V×
where ( ) ( ){ }yxyxR R ,,, µ= . ,x U x V∀ ∈ ∀ ∈
i. The 1st
projection of R is defined as
( ) ( )
( )( ){ } ( )( )( ){ }1 1
, , ,max , ,R R
y
R x x y x x y x y U Vµ µ= = ∈ ×
( )
( ),max
,max , .
x y
y
i
y
x r x y e
µ
 
=   
 
(25)
ii. The 2nd
project is defined as
( ) ( )
( ){ } ( )
{ }max ,2 2
, , ,max ( , ) x
i x y
R
x
R x x y y r x y e
µ
µ= = (26)
iii. The total projection
( )
( ) ( ) ( ){ }max max , ,T
R
x y
R x y x y U Vµ= ∈ × (27)
In the above, max
y
means maximum with respect to y while x is to be considered as fixed
and max
x
maximum with respect to x while y is to be considered fixed. The projection
determines the maximal boundary of the complex fuzzy sets in a complex fuzzy relation,
Example 4.1. Consider the following complex fuzzy relation.
i1.2π i1.2π i1.2π i π
i1.2π i1.6π i1.6π i π
i1.2π i 2π i1.6π i π
i1.2π i1.6π i1.6π iπ
0.6e 0.6e 0.6e 0.5e
0.6e 0.8e 0.8e 0.5e
0.6e 1.0e 0.8e 0.5e
0.6e 0.8e 0.8e 0.5e
 
 
 
 
  
 

37
So by the definition 4.3, R(1)
=( ( )
Ti1.2π i1.6π i2π i1.6π
0.6e 0.8e 1.0e 0.8e ,
R(2)
= ( )
Ti1.2π i2π i1.6π iπ
0.6e 1.0e 0.8e 0.5e and R(T)
=
2
1.0 i
e π
. Which shows that the
maximum boundary of the given fuzzy complex relation is
2
1.0 i
e π
. So the projection determines
the maximal boundary of the complex fuzzy sets in a complex fuzzy relation,
5. CONCLUSION
The work presented in this paper is the novel frame work of complex fuzzy relation. In this paper
the various properties and operation of complex fuzzy set as well as complex fuzzy relation are
investigated. We also presented the complement of complex fuzzy relation as well as the
projection of complex fuzzy relation is presented. A further study is required to implement these
notions in real life applications such as knowledge representation and information retrieval in a
Complex plan corresponding of two variables which is not suitable for traditional fuzzy logic.
REFERENCE
[1] Buckley J.J (1987) “Fuzzy Complex Number” in Proceedings of ISFK, Gaungzhou, China, pp.597-
700.
[2] Buckley J.J (1989) “Fuzzy Complex Number”, Fuzzy Sets &Systems, vol33, pp 333-345.
[3} Buckley J.J (1991) “Fuzzy Complex Analysis I: Differentiation”, Fuzzy Sets &Systems, Vol 4l, pp.
269-284.
[4] Jang J.S.R, Sun C.T, Mizutani E, (2005) “Neuro-Fuzzy And Soft Computing”, PHI, New Delhi.
[5] Pedrycz and Gomide F, (2004) “Introduction To Fuzzy Set Analysis And Design”, PHI, New Delhi.
[6] Ramot Daniel, Menahem Friedman, Gideon Langholz and Abraham Kandel (2003) “Complex Fuzzy
Logic”, IEEE Transactions on Fuzzy Systems, Vol 11, No 4.
[7] Ramot Daniel, Menahem Friedman, Gideon Langholz and Abraham Kandel, (2002) “Complex Fuzzy
Set”, IEEE Transactions on Fuzzy Systems ,Vol 10, No 2.
[8] Xin Fu, Qiang Shen, (2009) “A Noval Framework Of Complex Fuzzy Number And Its Application
To Computational Modeling”, FUZZ –IEEE 2009, Korea, August 20-24.
[9] Zadeh L.A. (1965) “Fuzzy Sets”, Information & Control, Vol 8,pp 338-353.
[10] Zhang Guangquan, Tharam Singh Dillon, Kai-yung Cai Jun Ma and Jie Lu,(2010) “σ-Equality of
Complex Fuzzy Relations”, 24th
IEEE International Conference on Advance Information Networking
and Application.

38
Authors
Dr. S. K. Das is working as a Reader (Associate Professor) and Head of the
Department of Computer Science, Berhampur University, Berhampur, Orissa, India.
He has more than 17 research publications in National and International Journals and
conferences. He has supervised three Ph.D. scholars and more than 10 scholars are
under supervision. He is a life member of IEEE, ISTE, SGAT, CSI and OITS. He has
the research interests in the fields of Data Communication and Computer Networks,
Computer Security, Internet and Web Technologies, Database Management Systems,
Mobile Ad Hoc Networking and Applications.
Dr. D. C. Panda is working as an Assistant Professor in the Department of
Electronics and Communication Engineering, Jagannath Institute of Technology and
Management (JITM), Paralakhemundi, Orissa, India. He has more than 33 research
publications in National and International Journals and conferences. He attended
Several QIP of different Universities and Institutions of India. He is a reviewer of
IEEE Antenna and Wave Propagation Letters and Magazine. He published a book
“Electromagnetic Field Theory Simplified:- A New Approach for Beginners”, which
is under review. He is an exposure in Computational Electromagnetic Tools. He has
the research interests in the fields of Soft Computing Techniques, Computational Electromagnetic,
Microwave Engineering and MIMO Wireless Networks.
Mr. Nilambar Sethi is working as an Associate Professor in the Department of
Information Technology, Gandhi Institute of Engineering & Technology (GIET),
Gunupur, Orissa, India. He is under the supervision of Dr. S.K. Das and Dr. D. C.
Panda for his Doctoral degree in the field of Complex fuzzy sets and its applications.
He is a life member of ISTE. He has the research interests in the fields of Computer
Security, Algorithm Analysis and Fuzzy Logic.
Dr. S. S. Gantayat is working as a Professor in the Department of Information
Technology, Gandhi Institute of Engineering & Technology (GIET), Gunupur,
Orissa, India. He has more than 8 research publications in National and International
Journals and conferences. He is guiding one student for his doctoral degree in the
field of expert systems. He attended several workshops of different Universities and
Institutions of India. He is a reviewer of WSES and Information Sciences. He is
member and life member of IE(I), ISTE, OITS, IARCS, IACSIT, ISIAM, IAMT,
IMS and SSI. He has the research interests in the fields of Fuzzy Sets and Systems,
Rough Set Theory and Knowledge Discovery, List Theory and Applications, Soft Computing,
Cryptography and Discrete Mathematics.

INDUCTIVE LEARNING OF COMPLEX FUZZY RELATION

More Related Content

What's hot (15)

Similar to INDUCTIVE LEARNING OF COMPLEX FUZZY RELATION (20)

Recently uploaded (20)

INDUCTIVE LEARNING OF COMPLEX FUZZY RELATION