A STUDY OF FUZZY, SUPER

AND SUPER FUZZY MATRIX THEORY

Thesis submitted in partial fulfillment of the requirement for
The award of the degree of
Masters of Science
in

Mathematics and Computing

Submitted by

Nidhi Kalra
Roll no. 300803011

Under
the guidance of

Dr. M.A. Khan

JULY 2010

School of Mathematics and Computer Applications

Thapar University
Patiala-147004 (PUNJAB)
INDIA

PREFACE
Having paid my deep sense of gratitude, I feel myself deserve to present this thesis.
The work being presented in the thesis is devoted to A Study of Fuzzy, Super and Super
Fuzzy Matrix Theory. It consists of six chapters. In chapters, Tables and figures are
shown not as a sub-section but as different sections and in this way, in contents, list of
tables and figures are shown separately from chapters.
Starting from the first chapter on binary and fuzzy sets, which are further related to
fuzzy matrix theory. Then the determinant and adjoint theories of square fuzzy matrix are
included. Next, the concept of supermatrices is introduced and is further related to fuzzy
supermatrices. Last chapter is ended to recall the summary of the comparative study done
and to conclude the theories based on this work for further advanced study in this field
and its important applications in other related specific areas.
The main body of the thesis is then followed by a list of references which by no means
is a complete bibligraphy of the work, rather the work referred in this thesis has been
included in this literature.

CONTENTS
ACKNOWLEDGEMENT

PREFACE

II

CHAPTER 1 : BINARY AND FUZZY SET THEORY

1-10

1.1 : Introduction

1.2 : Examples

1.3 : Binary Set Theory

1.4 : Fuzzy Set Theory

CHAPTER 2 : FUZZY MATRIX THEORY

11-19

2.1 : Introduction

11

2.2 : Fuzzy Matrix Theory

11

2.3 : Operations on Fuzzy Matrices

14

CHAPTER 3 : THE DETERMINANT THEORY OF A SQUARE FUZZY

MATRIX

20-31

3.1 : Introduction

20

3.2 : The Determinant Theory of A Square Fuzzy Matrix

20

3.3 : Properties of Determinants of Square Fuzzy Matrices

23

CHAPTER 4 : THE ADJOINT THEORY OF A SQUARE FUZZY MATRIX 32-52

4.1 : Introduction

32

4.2 : The Adjoint Theory of A Square Fuzzy Matrix

32

4.3 : Properties of Adjoints of Square Fuzzy Matrices

34

CHAPTER 5 : SUPER AND SUPER FUZZY MATRIX THEORY

5.1 : Introduction

53-82
53

5.2 : Super Matrix Theory

53

5.3 : Super Fuzzy Matrix Theory

71

CHAPTER 6 : SUMMARY AND CONCLUSION

83-85

6.1 : Introduction

83

6.2 : Summary

83

6.3 : Conclusion

84

BIBLIOGRAPHY

86-87

LIST OF TABLES AND FIGURES

LIST OF TABLES
1.1: Intersection (Minimum)

1.2: Union (Maximum)

1.3: Complementation

5.1: An Index for the Orders of Submatrices

55

LIST OF FIGURES
5.1: A Diagrammatic Representation of Supermatrix

55

Chapter 1
BINARY AND FUZZY SET THEORY
1.1 Introduction
The purpose of present chapter is to introduce basic concepts of fuzzy sets for the
further

study

of

fuzzy

and

super

fuzzy matrices. In todays scientific and soci

otechnoligical fields, the classical approaches to decision making have been declining
while innovative concepts based upon the new mathematics such as fuzzy set theory, are
continually emerging.
Advances in science and technology have made our modern society very complex,
and with this, decision processes have become increasingly vague and hard to analyze.
The human brain possesses some special characteristics that enable it to learn and reason
in a vague and fuzzy environment. It has the ability to arrive at decisions based o
n
imprecise, qualitative data in contrast to formal mathematics and formal logic which
demand precise and quantitative data. Modern binary computers possess capacity but lack
human like ability. Undoubtedly, in many areas of cognition, human intelligence far
excels the computer intelligence of today, and the development of fuzzy concepts is a
step forward towards the development of tools capable of handling humanistic types of
problems.
We do have sufficient mathematical tools and computer-based technology for
analyzing and solving problems embedded in deterministic and uncertain (probabilistic)
environment. Here uncertainty may arise from the probabilistic behaviour of certain
physical phenomena in mechanistic systems. We know the important role that vagueness
and inexactitude play in human decision making, but we did not know until 1965 how the
vagueness arising from subjectivity which is inherent in human thought processes can be
modelled and analyzed.
In 1965, Professor Loffi A. Zadeh [15] laid the foundation of fuzzy set theory. In
effect, fuzzy set theory is a body of concepts and techniques that gave a form

of
mathematical precision to human cognitive processes that in many ways are imprecise

9and ambiguous by the standards of classical mathematics. Today, these concepts are
gaining a growing acceptability among

mathematicians,

engineers,

scientists and

philosophers. Since its inception, research in the fuzzy set field has faced an increasing
exponential growth. This fuzzy field has blossomed into a many faceted field of inquiry,
drawing on and contributing to a wide spectrum of areas ranging from pure mathematics
to human cognition, preception and judgement. Its influence in science, engineering and
social sciences has been felt already and is certain to grow in the decade to come.
1.2 [7] Examples
Imprecision in characterizing the possible presence or absence of properties in a
concept leads to a fuzzy set. Suppose we consider the character recognition problem of a
faded typescript. We are aware of the character set available on the typewriter. Suppose a
compound symbol such as O appears on the typescript. It is easily identifiable as being
made up of characters L and O. We would then say the compound symbol i
s
associated with the subset {L,O} of the character set on the typewriter.
Suppose instead a compound symbol such as Q appears. On a first glance we think
we recognize a Q, but on a closer examination we find it as superimposed Q, G,
O and possible C. We are not sure as to the presence or absence of C in Q. Thus,
if this symbol has to be represented as a subset of the symbols on the typewriter, then we
either assign a doubtful label to C or decide that may be it is present. How do we
represent a partial presence?
Such examples may be multiplied. How is one to measure the beauty of a snow-clad
mountain? The marks assigned to the student in a course represent the degree to which
his/her teacher measures his/her ability or performance in that subject. These marks may
range from 0 to 100% or graded from A to F. A fuzzy concept of his/her overall
performance is obtained by his/her mark list for the term. Thus, instead of treatin
g
presence or absence in a binary manner, as in ordinary set theory, one deals with the
presence in a graded manner, from 0 to 1.

Before we define the theory of fuzzy sets, let us first define the notion of the theory of
binary sets starting from the definition of a set. A set is a collection of objects that are

well specified and possess some common properties. These objects may represent some
abstract concept, or may be a collection of some physical properties. It can be finite or
infinite, enumerable or non-enumerable.
1.3 [5] Binary Set Theory
The representation of a binary set can be made in many ways, the most usual form is,
For a finite set,
E = {a1,a2,,an},
and for an infinite enumerable set,
E = {a1,a2,,an,}.
In a set E, an element ai is called the member of the set.
In the case of an infinite non-enumerable set, real numbers are usually represented by a
symbol like R (real numbers), R

(non-negative real numbers), R (positive real

numbers), N (natural numbers), Z (integers) etc.

Given a set E,
E = {a,b,c,d,e,f}.
A subset is a collection of some members from a set. Then the subset A of E is
A = {a,d,e}.
Symbolically, we write a subset as
A E.
By definition, E is also a subset of itself and we can write E E. A strict inclusion i
s
represented by when a set cannot be a subset of itself.
A set E and its subset A can also be represented in the following way

The numbers 0 and 1 define the membership of each element of the subset, where 1
means the element belongs to the subset and 0 means the element does not belong to the
subset.
If x A , we denote the corresponding function as

x E : A (x) {0,1}
The function

A(x) is called the 'characteristic function' or 'membership function'. We

see that a subset is the function of E on {0,1}.

Let us define the 'empty subset' as

x E : ( x) 0
and for the set E itself

x E : E ( x) 1
E

The set of all subsets is called the 'Power Set' and is written as P(E) or {0,1} .
For example, let E = {a,b,c}
Then power set is defined as
P(E) = { ,{a},{b},{c},{a,b},{a,c},{b,c},{a,b,c}}
1.3.1 [7] Binary Set Operations
In the theory of sets, the main operators are defined as follow:

Intersection:

A B (X

(X )

(X ),

Union:

A B (X )
Complementation:

where

(X )

(X ),

A ( X ) 1 A ( X),

(and) means 'minimum' and

(or) means 'maximum'. These operators are

shown below in a tabular form.

Intersection (Minimum)
Table 1.1 [7]

Union (Maximum)
Table 1.2 [7]

Complementation
Table 1.3 [7]

We shall now illustrate these operators by means of an example.

Let us define a finite referential set E as
E = {a,b,c,d,e,f},
and the subsets A and B as

and

The intersection, union and complementation on these subsets are given by

and

1.3.2 [6] Algebraic Properties of Sets

For any subset A,B,C from referential set E, the following properties hold:
(i)

Commutativity:
AB = BA
AB = BA.

(ii)

Associativity:
(AB)C = A(BC),
(AB)C = A(BC).

(iii)

Idempotence:
AA = A,
AA = A.

(iv)

Distributivity:
A(BC) = (AB)(AC),
A(BC) = (AB)(AC).

(v)

Exclusion:
AA = .

(vi)

Non-contradiction:
AA = E,
A = ,
A = A,
AE = A,

AE = E.
(vii)

Involution:

() = A
(viii)

De-Morgans theorems:
A B

= A B,

A B A B.
Then set E is called a 'Boolean lattice' or a 'Distributive and Complemented lattice'.
In the next section, we will review the fuzzy set theory.
1.4 [3] Fuzzy Set Theory
Consider a referential set E with x as its element. The 'Characterstic function' or the
'membership function' of x is

x E : A (x) [0,1]
where [0,1] is the segment or closed interval from 0 to 1, called fuzzy unit interval. Then
subset A of E given by {(x,

A (x)); x E, A (x) [0,1]} , is called a fuzzy set.

Thus a fuzzy set has a membership function with not only values of 0 (does not belong
to) or 1 (belong to), but any number over the interval 0 and 1,
For example; 0.3,0.43,0.99, .
To be more non-technical a fuzzy set can be defined mathematically by assigning to
each possible individual in the universe of discourse a value representing its grade of
membership in the fuzzy set. For Zadeh [15] introduced a theory whose objects fuzzy sets
are set with boundaries that are not precise. The membership in a fuzzy set is not a matter
of affirmation or denial but rather a matter of a degree. The significance of Zadeh
s
contribution, Fuzzy Sets as a Basis for Theory of Possibility [15], was that

it

challenged not only probability theory as a sole agent for uncertainty; but the ver
y
foundations upon which the probability theory is based, Aristotelian two-valued logic.
For when A is a fuzzy set and x is a relevant object, the proposition x is a member of A, is
not necessarily either true or false as required by two valued

logic, but it may be true

only to some degree the degree to which x is actually a member of A. It is most common,
but not required to express degrees of membership in the fuzzy sets as well as degrees of
truth of the associated propositions by numbers in the closed unit interval [0,1]. The
extreme values in this interval 0 and 1, then represent respectively, the total denial and
affirmation of the membership in a given fuzzy set as well as falsity or truth of th
e
associated proposition.
The capability of fuzzy sets to express gradual transitions to membership to nonmembership and vice-versa has a broad utility. This not only helps in the representation
of the measurement of uncertainties but also gives a meaningful representation of vague
concepts in a simple natural language.
For example; a worker wants to find the moods of his master, he cannot say cent
percent in mood or 0 percent of mood or depending on the person who is going to meet
the boss he can say some 20% in mood or 50% in mood or 1% in mood or 98% in mood.
So even 98% in mood or 50% in mood the worker can meet with some confidence. If 1%
in mood the worker may not meet his boss. 20% in mood may fear while meating him.
Thus we see however that this definition eventually leads us to accept all degrees of
mood of his boss as in mood no matter how gloomy the boss mood is! In order to resolve
this paradox the term mood may introduce vagueness by allowing some sort of gradual
transition from degrees of good mood that are considered to be in mood and those that are
not. This is infant precisely the basic concept of the fuzzy set, a concept that is bot
h
simple and intuitively pleasing and that form is essence, a generalization of the classical
or crisp set. The crisp set is defined in such a way as to dichotomize the individuals in
some given universe of discourse into two groups: members (those that certainly belong
to the set) and non-members (those that certainly do not). Since full membership and full

non-membership in a fuzzy set can still be indicated by the values of 1 and

0,
respectively, the function can be generalized such that the values assigned to the elements

of the universal set fall within a specified range and indicate the membership grade of
these elements in the set in the example. Larger values denote higher degrees of se
t
membership. Such a function is called a membership function and the set defined by it a
fuzzy set. For example:
Let E = {a,b,c,d,e,f} and a subset given by

Then A represents a fuzzy set.

1.4.1 [6] Fuzzy Set operations
The intersection (minimum), union (maximum) and complementation operators
defined earlier can be used also for fuzzy sets. We will show these operators by means of
an example.
Let

and

The intersection of A and B is defined as

The union of A and B is defined as

The complements of A and B are given by

and

1.4.2 [6] Algebraic Properties of Fuzzy Sets

The algebraic properties of fuzzy subsets are the same as for ordinary subsets, except for
the following relations:
AA = (not necessarily) and AA = E (not necessarily).
For example, taking A from the above example we have

and

Chapter 2
FUZZY MATRIX THEORY
2.1 Introduction
The purpose of present chapter is to introduce the basic concept of fuzzy matrix
theory. Morever, we shall consider the operations defined on these matrices for further
treatment of determinant and adjoint theory of square fuzzy matrices. First of all, we shall
give fuzzy matrix theory and some operations defined on fuzzy matrices. However the
book of Paul Horst, Matrix Algebra for Social Scientists, [2] would be a boon to social
scientists who wish to make use of matrix theory in their analysis.
2.2 Fuzzy Matrix Theory
A fuzzy matrix is a matrix which has its elements from [0,1], called fuzzy unit interval.

Definition 2.2.1 [3]: Consider a matrix A= aij

mn

where
d

[0,1], 1 i m an

. Then A is fuzzy matrix.

2.2.1 [3] Types of Fuzzy Matrices

(a)

(i) Fuzzy Rectangular Matrix:

aij mn
Let
(mn) where is
i rectangular
m andmatrix.
1
3 1fuzzy
For A=
example:
a 2[0,1],

Then A is a fuzzy rectangular matrix.

0

0.2

0.1 0.5 0.3

(ii) Fuzzy Square Matrix

a11

a12...aij...a1n

a21
an1

a22...a2j ...a
2 n

an 2

... anj

nn

Let A= ai1

ai 2...aij

...ain

Let A = a1

... an ,

[0,1]; j=1,2,,n.

... a
[0,1], 1 i , j n.

where

Then A is a fuzzy square matrix.

(iii) Fuzzy Row Matrix:
Let A=

Then A is called a 1 n fuzzy row matrix or fuzzy row vector.

For example: 0.3

0.7 0.05

(iv) Fuzzy Column Matrix:

a1
a
2

1 is a 14 fuzzy row matrix.

am

where

[0,1]; i=1,2,,m.

Then A is called a m 1 fuzzy column matrix.

For example:
0.5

is a 3 1 fuzzy column matrix or fuzzy column vector.

0.4

(v) Fuzzy Diagonal Matrix:

A fuzzy square matrix A= aij

nn

is said to be fuzzy diagonal matrix if

=0

[0,1], 1 i , j n.

when i j ,where

0
0

0 0.3
0.4 0

is a fuzzy diagonal matrix of order 3.

For example:
0

0.9

This diagonal matrix is also denoted by 0.4,0.3,0.9 .

(vi) Fuzzy Scalar Matrix:

A fuzzy diagonal matrix is said to be fuzzy scalar matrix, if all its diagonal entries
are equal.

aij nn is said to be a fuzzy scalar matrix

Thus, a fuzzy square matrix A=

For example: 0.3 and

are
fuzzy scalar matrices of order 1 and 2

if

a ij 0 w h e n i

where

[0,1], 1 i , j n.

j
a ij w h e n i
j
0.4 0
0 0.4

respectively.
(b)

Usual identity matrix and zero matrix are fuzzy matrices as their entries are from
the fuzzy crisp set {0,1}.

(c)

If the entries in upper triangular matrix and lower triangular matrix are from fuzzy
unit interval [0,1], then these matrices are said to be fuzzy upper triangular and
fuzzy lower triangular matrices respectively.

2.2.2 Equality of Fuzzy Matrices

Two Fuzzy matrices of the same type are said to be equal iff their elements in th
e
corresponding positions are equal.

2.3 [3] Operations on Fuzzy Matrices

2.3.1 Operations of Maximum and Minimum
We shall define the following three operations on fuzzy matrices:
(a)

maximum of matrices

(b)

minimum of a matrix by a scalar

(c)

max min of matrices

(a)

Operation I Maximum of matrices

If two fuzzy matrices are of the same type, then they are said to be conformable for
addition. But the question arises that when we add two fuzzy matrices, then the
resultant matrix is not a fuzzy matrix. So in case of fuzzy matrices of same type,

max operation is defined.

Definition 2.3.1: Let A = aij

and B = bij

mn

mn

be two fuzzy matrices.

Then their sum, denoted by A+B, is defined as

i.e., aij

ij

A+B = max {A,B}

max(aij

mn

For example: Let A =

0.4 0.3,
0

0.3

ij

mn

b0.1

; for 1 0.9
i m ,1 j n

and B =

0.9

0.6

0.8

0.4

Then

0.7 1

0.5

0.2

0.6

0.7
0.5

0
0.5
0

0.6

0.8

0.5

max(0, 0.6)
max(0.4, 0.9)

max(0.3, 0.7)
max(0.3, 0.5)

A+B = max {A,B} =

max(1, 0.7)

max(0.8,1)

max(0.9, 0)
max(0.1, 0.5)

0.6

0.7

max(0.4, 0)

0.9 0.5 0.5

= max(0.5, 0.6)

max(0.2, 0.8)

max(0.6, 0.5)

0.9

0.4

0.6

0.8

0.6

In a similar way, we can define the difference of two fuzzy matrices of same type as the
max operation.

Thus, in case of fuzzy matrices of same type, AB = max {A,B} = A+B.

(b)

Operation II minimum of a matrix by a scalar

kA = A k = ka ij m in( k , aij )

mn

be any fuzzy matrix and k F, where F = [0,1] is a

Definition 2.3.2: Let A= a ij

fuzzy unit interval. Then scalar multiple

of A by k, denoted by kA or Ak is given by

; aij [0,1] ,1 i m ,1 j n.

m in(0.3, 0.2)

m n

m n

Thus kA or Ak is the matrix obtained when each entry of A is multiplied by k.

0.5
For example: 0.3 0 .4
0 .2

0.8

1
0 .6

= m in(0.3, 0.4) m in(0 .3, 0.5)

m in(0 .3, 0.8)

m in(0.3,1)
m in(0.3, 0.6)

0.3
= 0 .3
0.3

0 .2

(c)

0.3
0.3

Operation III max min of matrices

If we wish to find the product AB of two fuzzy matrices A and B where A and B
are compatible under multiplication i.e., number of columns of A = number of rows
of B ; still we may not have the product AB to be a fuzzy matrix. So in case o

fuzzy matrices compatible under multiplication , max min operation is defined.

Definition 2.3.3 : Let A =

aij

B = b jk n p be two fuzzy matrices.

mn
and

Then their product, denoted by AB, is defined to be the fuzzy matrix cik
where

= n

j
i

a ij b

jk

= max{min (

m p ,

); 1 i m , 1 k p} for j = 1,2,,

0.4 0.7

0.4

Remark: If the fuzzy product AB is defind then BA may not be befined.

34
0.3

0.2

For example: Let A =

0.5

0.3

0.7

and B =
0.6 0.3 1 0.1

Then the product AB is 0.4

defined
is given
by AB =
0 and0.6

23

11 12

c24

Since no. of columns of A = no. of rows in B

where

c21

c22 c23

c11 = max{min(0.3,0),min(0.2,0.4),min(0.5,0.6)}
= max{0,0.2,0.5}
= 0.5
c12 = max{min(0.3,0.3),min(0.2,0.7),min(0.5,0.3)}
= max{0.3,0.2,0.3}
= 0.3

0.5

c13

24

c13 = max
{min(0.3,0.7),mi
n(0.2,0.4),min(0
.5,1)}
= max{
0.3,0.2,0.5
= 0.5
c14 = m
ax{min(0.3,0.5),
min(0.2,1),min(
0.5,0.1)}

c14

= max{0.3,0.2,0.1}
= 0.3
c21 = max{min(0.4,0),min(0,0.4),min(0.6,0.6)}
= max{0,0,0.6}
= 0.6
c22 = max{min(0.4,0.3),min(0,0.7),min(0.6,0.3)}
= max{0.3,0,0.3}
= 0.3
c23 = max{min(0.4,0.7),min(0,0.4),min(0.6,1)}
= max{0.4,0,0.6}
= 0.6
c24 = max{min(0.4,0.5),min(0,1),min(0.6,0.1)}
= max{0.4,0,0.1}
= 0.4
0.5
Thus AB =
0.6

0.3
0.3

0.5
0.6

0.3

0.4

But since, no. of columns of B no. of rows of A.

Thus the product BA is not defined.
Using the max-min fuction, we can find the positive integral powers of a square fuzzy
matrix.

2.3.2 Transpose of fuzzy matrix

Definition 2.3.4: Let A = aij

mn

be any fuzzy matrx. Then the transpose of A, denoted

by Aor At or AT , is n m fuzzy matrix obtained from A by interchanging its rows and

columns.

i.e., A= bij

nm

[0,1]; for 1 i n and 1

where

m.

Remarks: 1. The transpose of a fuzzy row matrix is a fuzzy column matrix and viceversa.
For example: Let A = 0.6
2. The product AA and AA of two fuzy matrices are always defined.

31

0.2

be a fuzzy column matrix.

0.7

Now A.A = 0.2 0.6 0.7 0.6

Then A= 0.2 0.6 0.713 is a fuzzy row matrix.

Observe that the products AA and AA of two fuzzy matrices A and Aare defined.
0.2

0.7

0.2 0.6
and A. A = max{0.2,0.6,0.7}


= 0.7

0.7

Then A.A is a singleton fuzzy matrix.

0.2

0.6
0.7

min(0.2, 0.2)

min(0.2, 0.6)

min(0.2, 0.7)

min(0.7, 0.6)

min(0.7, 0.7)

min(0.7, 0.2)

= min(0.6, 0.2) min(0.6, 0.6)

min(0.6, 0.7)

= 0.2
0.2 0.6
0.2

0.2

0.6

0.6
0.2

0.7

Then A. A is a symmetric fuzzy matrix as the matrix obtained by interchanging its

rows and columns is the matrix itself and the elements of the matrix belong to fuzzy unit
interval [0,1].

Chapter 3
THE DETERMINANT THEORY OF A SQUARE FUZZY MATRIX

3.1 Introduction
The purpose of this chapter is to introduce and investigate some properties of the
determinant of a square fuzzy matrix. Kim [11] defined the determinant of a square fuzzy
matrix and contributed with very research work [8,9,10,11], a lot to the study
of
determinant theory of square fuzzy matrices.The properties of a square fuzzy matrix are
somewhere analogous to the crisp case for determinant of a square matrix. We shall
introduce the determinant theory of a square fuzzy matrix and their properties and
investigate the elementary properties of determinant theory for fuzzy matrices in fuzzy
matrix semiring Mn (F) [13].
3.2 The Determinant Theory of a Square Fuzzy Matrix
This section lays down the foundation for the determinant theory of a square fuzzy
matrix where the elements takes its values from the unit interval [0,1].
Definition 3.2.1 [11]: The determinant A of an n n fuzzy matrix A is defined a
s
follows:

a1 (1) a2 ( 2) ... an (n )

Sn

a
i

Sn

(i )

i 1

where Sn denotes the symmetric group of all permutations of the indices (1,2,,n). We
may use det(A) instead of A .We may call det(A) the permanent of A.

Preliminaries:

1. It may be noted with care that the non-square fuzzy matrices do not have determinants.
2. The elements of the determinant of a fuzzy matrix takes its values from the unit

interval [0,1].
3. Here multiplication and addition takes respectively the meanings of min. and max.
operations as defined usually in fuzzy matrices.
(i)

The determinant of a 11 fuzzy matrix [a] is denoted by a and is defined as a.

a
(ii)

The determinant of a 22 fuzzy matrix

and is defined

a1

c d

b is denoted by a b

as ad+bc or max {min{a,d},min{b,c}}.

(iii) The determinant of a 33 fuzzy matrix

a1

a3

a2
b2

c2

c1

a2

a3

b2

b3

is denoted by b1

c1

c3

c2

c3

and is defined as

b3 a2 b1 b3 a3 b1 b2

a1 b2
c2

c3

)+

i.e.,

or

max{min(

max{min(
=

c1 c3
(

c1 c2
+

)+

, ) , min( , )} +

max{min(

)
, ) , min( , )} +

, ) , min( , )}
+

, where = max{min( ,

) , min( ,

)}

= min( ,
= max{min(

) + min( ,
,

= max{min( ,

) , min( ,

)}

= max{min( ,

) , min( ,

)}

) + min( ,

),min( ,

),min( ,

)
)}

We can expand the determinant along any other row or column as the same value of

the determinant can be obtained by expanding along any row or column. It can be easily
verified that expanding along any row or column gives the value of the determinant as
+
+
+
+
+
Notation 3.2.1: The determinant of a square fuzzy matrix A of order n is defined as
follows:
n

det(A) = |A|
=

aijA

, i {1, 2,..., n} [Expanding along ith row]

ij

j 1

ij

ij

, j {1,2,, n} [Expanding along jth column]

i1

where A is the determinant of the fuzzy matrix of order (n1) obtained from a square
fuzzy matrix A of order n by deleting (striking out) row i and column j.
Explanation: Consider a 33 matrix
A= a21 22 23
a a1131 a3212 a13

33

= a22
a32

a23 , A
a33

=a21
a31 33

a
a23

,A

a
a

a
a

= a21
a31

a22
a32

Then the value of the det(A) is given by

A a11A11 a12A12 a13A13.

We can expand the determinant along any other row or column.
The value of the determinant does not depend on along which row or column it
is
expanded.
Thus, we have:
= a11A11 a12A12 a13A13 [Expanding by Ist row]

= a21A21 a22A22 a23A23 [Expanding by IInd row]

= a31A31 a32A32 a33A33 [Expanding by IIIrd row]

= a11A 11 a 21A 21 a 31A 31 [Expanding by Ist column]

= a12A12 a22A22 a32A32 [Expanding by IInd column]

= a13A13 a23A23 a33A33 [Expanding by IIIrd column]
Remark: It may be noted with care that the value of the determinant of a square fuzzy
matrix A is one among the elements of A.
Thus 0 det(A) 1.

A =a square
0.6 0.2fuzzy
0.9matrix
Example 3.2.1: For

0.5

0.3

0.8

0.7

0.4

0.0

We calculate the determinant |A| as follows:

|A| =

= 0.5 0.2 0.9 + 0.3 0.6 0.9 + 0.8 0.6 0.2

0.7

0.4

0.0

0.4

0.0

0.7

= 0.5 (0.2 + 0.7) + 0.3 (0.4 + 0.0) + 0.8(0.6 + 0.0)

= 0.5(0.7) + 0.3(0.4) + 0.8(0.6)
= 0.5 + 0.3 + 0.6
= 0.6 .
It can be easily verified that the same value of the deterninant can be obtained by
expanding along any other row or column.

3.3 Properties of Determinants of Square Fuzzy Matrices

The following properties of determinants are true for determinants of any order.
However we shall show these for determinants of order 3. These properties are used in
order to simplify the determinant before expanding it. Some properties of determinants of

square fuzzy matrices are analogues to the properties of determinants of square matrices

while some other properties differ a lot and a few properties when considered in fuzzy
matrix semiring Mn (F) give rise to some new problems and sorting out these problems
establishes some new results in fuzzy matrix semiring Mn (F). Before we proceed, it is
worth to note that whatever operation or result is true for rows, is also true for columns.

Property 3.3.1 The value of the determinant remains unaltered by interchanging its rows
and columns.
Proof:
Let
a1 a2 a3
= b 1 b2 b 3
c1 2 3
c

Expanding along first row, we get

b3 a2 b1 b3 a3 b1 b2

= a1 b2

c2
=

c3

c1 c3

)+

c1 c2
+

)+

Interchanging the rows and columns of , then the new determinant is

a1

b1

c1

= a2 b2

c2

a3

c3

b3

Exapanding along first column, we get:

= a1b2

c2 a2 b1

b3 c3

b1

c1 b3
c1

c3 b2

c2

 a3
=

=.
Thus we have =

)+

)+

Remark: Interchange of the rows and columns does not change the value of determinant,
i.e., if A is a square fuzzy matrix, then det(A) = det(A') where A' denotes the transpose of
the square fuzzy matrix A.
Property 3.3.2 The value of the determinant remains unaltered interchanging its any two
2
rows (or columns).

a1 a
Proof:

a3

Let = b1 b2
c1

b3

c2

c3

Expanding along the first row, we get:

= a b2

b3 a2 b1 b3 a3 b1 b2

c3

c2
=

c1 c3

)+

c1 c2
+

)+

Interchanging the first and second rows, then the new determinant is
b1

b2
a2
= a1
c2
c1

b3
a3
c3

Expanding along second row, we get:

=

b2

b3 a2 b1 b3 a3 b1 b2

a1

c2

c3

c1 c3
)+

c1 c2
+

)+

=
Thus we have =
Property 3.3.3 If all rows (or columns) of a determinant are identical then its values is
minimum element along all elements of the determinant.

a
Proof:

Let =

b
a

c
b

c
c

= a(bc+bc) + b(ac+ac) + c(ab+ab)

= a(bc) + b(ac) + c(ab)
= abc + abc + abc
= abc
= min{a,b,c}
= minimum value among all elements of the determinant
Definition 3.3.1 [12]: An m n fuzzy matrix A is said to be constant if =
) for all i, j,k i.e., its rows (or columns) are equal to each-other.
=

(or

Note: Thus property 3.3.3 can be stated as:

The determinant of a constant square fuzzy matrix is its minimum element.
Property 3.3.4 If all the elements of a row (or column) are equal to a (say) and all other
rows (or columns) have elements a then the value of the determinant =
a (which is the minimum element among all elements).
Proof: We know that each term in the determinant contains a factor of each row (or
column) and hence each term contains a factor of that row (or column) in which al
l
elements are equal to a (say) and all other elements in that term of other rows (
or
columns) a so that each term in the determinant of a square fuzzy matrix is equal to a
and consequently

= a.

For example:

Let

= b1 b2

c1 c

b3
c3

= a(

)+

= a(

(
+

)+
+

each element

=a

min element =
Another statement of property 3.3.4[11]

Let

Ak ak1,ak2, akn

be the kth row of A. If

aki ai 1, 2,, n

and a pq a p, q 1, 2,, n
Then det(A) = a.
Corollary [12]: The value of the determinant containing a zero row (or column) is zero.
Proof: We know that each term in the determinant contains a factor of each row (or
column) and hence contains a factor of zero row (or column) so that each term in th
e
determinant of a square fuzzy matrix is equal to zero (min. element) and consequently
2

= 0.

a1 a
Example: Let

= b1 b2
0

a3
b3
0

(0+0) + (0+0) +

(0) +

= 0+0+0
=0

(0) +

(0)

(0+0)

Note: In a square fuzzy matrix, as all the elements of a zero row (or column) are equal to
0 and all other elements 0. Hence by property 3.3.4, the value of the determinant = 0.

Property 3.3.5 The determinant of

a diagonal matrix is the product of

its diagonal

elements.
Proof: Let A =

a ij

be a diagonal matrix i.e., aij = 0 for i j

Take a term t of A ,

t a1 (1) a2 ( 2) ... an ( n)
Let

(1) 1, then a1 (1) = 0 and thus t = 0.

This means that each term is zero if (1) 1.

Let now

(1) = 1 but

(2) 2, then a2(2) = 0 and thus t = 0.

This mean each term is zero if

(1) 1 or (2) 2.

a11 22way, we
ann can see athat
ii
However in the similar
each term for which
or
(2) 2 or (n) n must be zero

(1) 1

Consequently |A| =

a
i 1

= Product of its diagonal elements.

Corollary 1 The determinant of a scalar matrix is the non-zero (diagonal) element.
Proof:

Let be the determinant of a scalar matrix with each of its non-zero

diagonal element equal to k (say).
Since every scalar matrix is a diagonal matrix.

 By property 3.3.5, = product of its diagonal elements

= k.k. k
= k.
Corollary 2 The determinant of an identity matrix is unity.

Proof:

Since an identity matrix is a diagonal matrix.

By property 3.3.5, |I| = products of its diagonal elements
= 1.1. 1
=1

Corollary 3 The determinant of the matrix obtained from an identity matrix by interchanging its any two rows (or columns) is unity.
Proof:

By property 3.3.2, the value of determinant remains unaltered by interchanging its any two rows (or columns).
E ij I n

where Eij is the matrix obtained from identity matrix I

interchanging its rows (or columns).

i.e., E ij = 1.
Remarks:

(by corollary 2)

1. det(InA) = det(A) = det(AIn)

2. det(EijA) = det(A) = det(AEij)

Property 3.3.6 [12] The determinant of a triangular matrix is given by the product of its
diagonal elements.
Proof: Suppose = a ij is triangular from below.
i.e.,

= 0 for i < j

Take a term t of ,

t a 1
Let

(1 )

a2

(2)

...

a n

(n)

(1) 1, so that 1 < (1) and so a1 (1) = 0 and thus t = 0.

This means that each term is zero if

Let now

(1) 1.

(1) =1 but (2) 2 , then 2 < (2) and so

and thus t = 0.

a 2

(2 )

by

(1) 1 or

This means that each term is zero if

(2) 2.

However, in a similar manner we can see that each term must be zero if
or

(2) 2 or (n) n.

a11

22

ann

ii

Consequently =

(1) 1,

a
i1

= product of its diagonal elements.

Similar is the case with the determinant triangular from above.
Property 3.3.7 [12] If each element of a row (or column) of a determinant is multiplied
by a constant k F where F = [0,1] is the fuzzy unit interval of the
real line i.eS k [0,1] then its value gets multiplied by k .

Proof:

By definition of determinant,

a i

(i)

i 1

a a

1 (1)

2 (2) ....

i (i)....

n (n)

Sn

Multiply by k, the elements of the ith row (say) then the new
determinant is
'

1 (1)

Sn

2)

2 (

...kai (i)...an (n)

a
Sn

a
i

Sn

(i )

i1

=k

k
(n)

a
1

(1)

...ai

2 (2)

...an

(i)

Remark: In case of matrices kA is the matrix obtained when each entry of A is muliplied
by k F = [0,1].
Thus if A is a square matrix of order n, then
| A| =

|A| = k |A| i.e., det(kA) = det(A) = k det(A).

Chapter 4
THE ADJOINT THEORY OF A SQUARE FUZZY MATRIX
4.1 Introduction
In this chapter the adjoint theory of a square fuzzy matrix will be studied.The adjoint
of a square fuzzy matrix is defined by Thomason [13] and Kim[11]. We state a formula
for the adjoint matrix of a square fuzzy matrix and this formula shall be used anywhere in
this chapter. Then we shall establish the relationship between the adjoints of two fuzzy
matrices. Also we shall find the relationship between the adjoints of two fuzzy matrices
corresponding the relationship between the fuzzy matrices. For a square fuzzy matrix
satisfying some property, we shall verify the same property for its adjoint. In this chapter
we define the symmetric, reflexive, transitive, circular and idempotent fuzzy matrices and
show that some properties of a square fuzzy matrix such as symmetry, reflexivity,
transitivity, circularity and idempotence are carried over to the adjoint matrix and shall
understand the same with the help of illustrations. For a given square fuzzy matrix A,
through the adjoint matrix adj(A), we shall construct a transitive fuzzy matrix A (adj A).
Along with its illustration, we shall prove it before. We establish some results including
that A(adjA)|A|I and (adjA)A |A| I where |A| denotes the determinant of a square
fuzzy matrix A and adjA denotes the adjoint matrix of a square fuzzy matrix A.
4.2 The Adjoint Theory of a Square Fuzzy Matrix
Let us first define the notion of adjoint of a sqaure fuzzy matrix.
Definition4.2.1 [11]: The adjoint matrix B = [bij] of a square fuzzy matrix A = [aij] of
order n, is a square fuzzy matrix of same order n, denoted by adjA, is defined
as
1)

= A ji ; where

A ji is the determinant of the square fuzzy matrix of order (n

obtained from a square fuzzy matrix A of order n by deleting row j and column i and B =
[bij] = adjA.
Remark: Note that

A ji

can be obtained from A by replacing the element of A by 1

and all other row j factors ajk , k i by 0.

Example 4.2.1: For a Square fuzzy matrix

0.4
0.5

0 0.3

A = 0.2 0.4
1

0.7

0.3

We find adjA as follows:

b11 = A11 = 0.4 0.5 = 0.4+0.3 = 0.4
0.3

0.7

b12 = A21 = 0.3 0.4

0.3

= 0.3+0.3 = 0.3

0.7

b13 = A31 = 0.3 0.4 = 0.3+0.4 = 0.4

0.2 0.5
b21 = A12 = 0.4 0.5

b22 = A22 = 1

0
b23 = A32 =

0
b32 = A23 = 0.2

0.7 = 0.2+0.5 = 0.5

0.4
0.7

= 0+0.4 = 0.4

0.4 = 0+0.2 = 0.2

0.5

b31 = A13 = 0.2 0.4 = 0.2+0.4 = 0.4

0
b33 = A33 =
1

0.2

0.3

0 .4 = 0+0.4 = 0.4
0 .7
0.3
0.4 = 0+0.2 = 0.2

 0.4

0.3

0.4

0.4

0.2

0.4

adjA = [bij] = 0.5 0.4

0.2

4.3 Properties of Adjoints of Square Fuzzy Matrices

Notation 4.3.1[12]: We also can rewrite the element bij of adj A as

a t

bij

(t)

S n j ni t
n

where

= {1,2,,n}\{j}
= {1,2,,n}\{i}

and

S n j ni

is the set of all permutations of set

over the set

Proposition 4.3.1 [12]: Comparison of the adjoints of two fuzzy matrices

For n n fuzzy matrices A and B,

A B adj A adj B
adj A + adj B adj (A+B)

(i)
(ii)

Proof: Let A = a ij and B = bij

where i, j {1,2,,n}.

(i) since A B

( )

i, j {1.2,,n}

( )

for every t j,

a
t

(t

S
nj

ni n

(t) i

t (t

tn j

ni

adjA adjB
(ii) since A , B A + B

[ A + B = max {A, B}]

adjA , adjB adj(A+B)

[A B adjA adjB]

adjA + adjB adj(A+B)

Proposition 4.3.2 [12]: The adjoint of the transpose of a matrix is the transpose of the
adjoint of the matrix. i.e., for a square fuzzy matrix A of order n, adj A = ( A).

Proof: Let A= aij be a n n fuzzy matrix.

= adj A and C= cij

Let B = bij
Then bij =

a t

= adj A.

(t)

Snj S j (t i)t
n n

nj

and cij

ni

Hence adj A = (

, which is element b
t (t )
a
ni

ji

A), which proves the assertion.

Proposition 4.3.3 [12]:

Let A be a n n fuzzy matrix. Then

(i)
A(adj A) | |In
| |In
(ii)
(adj A)A
where In is a unit matrix of order n.
Proof (i): Let A = aij
(

Suppose B = bij

nn

nn

the jth column of B = bij

, then ith row of A is given by

= adjA

Then by definition of adjA,

nn

= adjA is given by

,,
(

).
,

,,

)=

Let C = cij

nn

j1

, A j 2 , ..., A
jn

= A(adj A)

Then (i,j)th element of C = cij

nn

= A(adj A) is given by

cij =

Ajk
ik
a

k1
n

and where cii

ik

A ik A .

a
|A| In = ... ... ... ... 0
k 1

Thus C = A(adj A) |A| In, where

...0

......

...

...

...

00

...

,,

).

(ii)
Let A = aij

nn

Then jth column of A is (

Let B = bij

nn

= adj A.

Then by definition of adj A,

The ith row of B = bij

nn

= adj A is ( ,

Let C = cij

nn

,,

)=

1i

, A 2 i , ..., A
ni

= (adjA)A

Then (i,j)th element of C = cij

j

nn

= (adjA)A is ci

ki

ak

k 1
n

a
where
k 1

Thus C = (adjA) A |A| In where In is a unit matrix of order n.

cii

ki ki

A.

0 and

Remark: Any diagonal element of the fuzzy matrix A(adj A) is |A| and non-diagonal
element 0.
Proposition 4.3.4 [12]: Let A be a square fuzzy matrix, then the following properties
hold:
(i)
(ii)

If A has a zero row then (adjA)A = O (the zero matrix)

If A has a zero column then A(adjA) = O (the zero matrix)

Proof (i): Let C = cij = (adjA)A, then c ij

Aki a kj

If the ith row of A is zero, then

So that Ak i k j = 0 i , j.

= 0 for every k = i

and for k i , A contains a zero row and so

A k i = 0 for k i

(by corollary of property 3.3.4 of determinants)

Hence (adjA)A = O (the zero matrix).

(ii) Let C = cij

= A(adjA),then =

a ik

jk

If the jth column of A is zero, then

= 0 for every k = j

and for k j, A contains a zero column and so A jk = 0 for k j

(by corollary of property 3.3.4 of determinants)
so that

ik

jk

0 i, j

Hence A(adjA) = O (the zero matrix).

Proposition 4.3.5 [12]: Let A be an n n constant fuzzy matrix, then
) is constant.

(i)

(ii)

C = A(adjA) is constant and

= | | , which is the least element in A.

Proof: Let A be an n n constant fuzzy matrix where its all rows are equal to each-other

i.e.,
(i)

i ,j

Let B = (adj A), then

bij

t (t

and

bik

Sn j ni tn j

at ( t )

S nk ni tnk

Since the numbers (t) of columns cannot be changed in the two expansions of
and
as A is constant and so
=
i,j,k .
In order that,
(ii)

i, j, k, we must have (

Since A is constant i.e.,

i, j, k .

Then A = A i,j,k and so

Aik

= A jk

i,j,k .

Let C = cij

= A(adjA), the

cij

ik

A jk

k 1 k1

= |A| i ,j

Thus C = A(adjA)a is constant.

n

Now

i (i)

Sn

i1

1 (1)

Sn

) is constant.

a2 (2) ...

an (n)

ik

Aik

= a1

(1)

a 2 ( 2 ) ...a n ( n ) for any Sn

( A is constant i. e. ,

Taking
|A| =

, the identity permutation i.e.,

= , ,

(i) = i i,we get

; which is the least element in a constant fuzzy matrix A .

Definition 4.3.1 [14]: Let A be a square fuzzy matrix of order n , then following hold:
(i)

A is said to be reflexive fuzzy matrix iff A In i.e., iff all diagonal elements in
fuzzy matrix A are unity i.e., iff = 1 i .

(ii)

A is said to be symmetric iff A = A i.e., iff the square fuzzy matrix A remains
unaltered by interchanging its rows and columns i.e., iff

(iii)

i,j {1,2,,n}.

A is said to be transitive iff A A i.e., iff the square fuzzy matrix A multiplied
by itself gives the elements less than or equal to the corresponding elements of the
square fuzzy matrix A. i.e., iff

for every k =1,2,,n.

A square fuzzy matrix is similarity (equivalence

symmetric and transitive.

relation) iff it is reflexive,

Let us understand the same with the help of an example.

Example 4.3.1: Let A be a square fuzzy matrix of order 3.
(i)

Consider
a square fuzzy matrix
A = 0.3 1 0.4

0.2

0.9

0.3

Since all the diagonal elements in square fuzzy matrix are unity, then A is a

reflexive
A = 0.4fuzzy
0.6 matrix.
0.1

(ii)

Consider
a square fuzzy matrix
0.3

0.4

0.5

Then

0.5

0.1

 0.3

0.4

0.5

0.5

0.1

A = 0.4 0.6

0.1

=A.

Thus A is a symmetric fuzzy matrix.

(iii)

Consider a square fuzzy matrix

A = 0.6
0.5 0.7
0.6

0.6

Then

0.6
0.5

0.7

0.6

A = 0.5 0.6

0.5

0.6

0.7

0.6

0.6

0.7

= 0.5 0.5

0.5

0.6

0.5

0.6 0.7

0.6

0.6
A

0.6
0.5

0.7

0.6

0.6

0.6

0.6

0.6

0.6

0.6

A= 1 1 0
as the element in A are the corresponding
elements in A. Thus A is a transitive

fuzzy matrix. Next, consider the square fuzzy matrix.

1

All diagonal elements in A equal to 1 implies that A is reflxive. Further, observe that
A = A so A is symmetric. Also A =A leads to the conclusion that A is transitive. Hence
A is similarity.
Let us see how the properties of a square fuzzy matrix are carried over to its adjoint.
Theorem 4.3.1[12]: Let A be a square fuzzy matrix of order n, Then we have the
following properties:
(i) If A is reflexive, then adj A is reflexive.
(ii) If A is symmetric , then adj A is symmetric.
(iii) If A is transitive, then adj A is transitive.

Proof: (i) Since A = a ij

Let B = bij

= adj A

is reflexive, then

=1 i

Then

bij

Snjn

tn

(t)

a t

S n tn
i j

and so bii

(t)

Taking only the identity permutation

i.e;

)(

1i

)(

(t) = t; we get

=1 )

=1 i

and so

Hence adj A is reflexive.

(ii)

Since A is symmetric, then

Sn j n

Let B = bij

Then bij

i , j.

tn j

= adjA

a
t

(t

a
t

(t )

[ aij a ji i, j ]

Sni n j tni

)
i

=
.
Hence adj A is symmetric.

(iii)

Since A is transitive, then

Let B = bij
Let D = A
as follows:

i,j .

= adjA.
, we can determine the elements of D in terms of the elements of A

dhk =

ahk

if h i, k j,

a(h1)k

if h i, k j,

ah(k1)

if h i, k j,

a(h1)(k1)

if h i, k j.

where A denotes the (n1) (n1) fuzzy matrix obtained from A by deleting
ith row and jth column.
Now we show that A A
Let R = A

,C=A

A for every t {1,2,,n}.

, F = A and W = A

n 1

Now w

rik c k j

k 1

n1

ik kj

ij

if i s, k t, j u,

a f

aik
k 1
u,

ij

k( j1)

ai( j1) fij if i s, k t, j

n1

a a

i(k1)

(k1) j

ij

ij

f if i s, k t, j u,

k1
n1

ai(k1)
u,
a

(k1)( j1)

ai( j1) fij if i s,k t, j

k1

a(i1)k
u,n1

kj

a(i1) j

k1
a(i1)(k1)

u,

(k1) j

ij

f if i s, k t, j

a(i1) j ijf if i s, k t, j

n1

a(i1)(k1)
a
u,

(k1)( j1)

ai1( j1) fij if i s,k t, j

k1

n1a(i1)k
u.
k1

n1

k( j1)

ai1( j1) = fij if i s, k t, j

a
k1

n1

a
k1

Thus

in every case and therefore A A

A for every t {1,2,..,n}

Since we know that |

| | || |; we have

This means btsbut bus i.e.,

Ast Atu Ast Atu Asu .

but bts bus for every t {1,2,..,n}

Hence B = adjA is transitive.

Corollary [12]: If a square fuzzy matrix is similarity then adj A is also similarity.
Example 4.3.2: Consider
a square
fuzzy matrix A
be
(i)
Let A = 0.1 1
a reflexive fuzzy matrix, then
0

adjA = 0.1
0.4

0.3

1
0.1
0.5
1

0.3

0.3

0.4

(ii)

0.5

be a symmetric fuzzy matrix, then

Let A all
= the
Since
0 diagonal
1
0.1elements
in adjA are unity, then adj A is a reflexive fuzzy

matrix.

adjA = 0.1 0.6

0.2

0.1

0.6

0.1

0.9

0.6
0.9

0.1

0.6

0.1

0.2

0.6

since(

A) = adjA , then adjA is a symmetric fuzzy matrix.

0.6

0.7

0.6


0.6

0.6

0.6

0.6

0.6

0.7

0.6

0.6

0.6

Now

(iii)

Let A = 0.5 0.6

0.5 be transitive fuzzy matrix,then

adjA = 0.5 0.6

0.5

A) = 0.6
0.5 0.6
0.6

0.5
0.6 0.5
0.6 0.6
0.6 0.50.6

0.6
= 0.5

0.6

0.7
0.6
0.6

0.6

0.6

0.6

0.6

0.6

0.5

0.6

0.7

0.6

adjA, then adjA is a transitive fuzzy matrix.

A= 1 1 0
Next, consider the similarity fuzzy matrix

Then A is reflexive,
symmetric and transitive.
adjA = 1 1 0

Now

Then adjA is reflexive, symmetric and transitive.

Hence adjA is similarity fuzzy matrix.
Definition 4.3.2 [12]: A square fuzzy matrix A of order n is called circular fuzzy matrix
iff (A ) A or more explicity,

for every k = 1, 2, , n.

Theorem 4.3.2 [12]: If a square fuzzy matrix A of order n is circular, then adjA is
circular.

Proof: Since A = aij

is circular then

Let B = bij = adjA

i,j

Let D = A

, we can determine the elements of D in terms of the elements of A as

follow:

ahk
a
(h 1)k

if h i, k j,

if h i, k j,
dhk = a(h 1)( k 1) if h i, k j.
ah( k 1)

if h i, k j,

where A denotes the (n1)(n1) fuzzy matrix of order (n1) obtained from A
by deleting ith row and jth column.
Now we show that A

A A for every t {1,2,,n}

Let R = A ,C = rAjk ,kiF = A and W = A

n1

ajk
u, k k 1

Now

ki

ij

ij

an1
( j1)k
u, n1

f if i s,k t, j

a ( j1) fij if i s, k t, j

ki

k1

k1

j(k 1) (k 1)i

ij

ij

if i s, k t, j u,

n1

k1

( j1)(k 1)

(k 1)i

i( j1)

fij if i s, k t, j u,

n1

k 1jk k (i1)

(i1) j

ij

if i s, k t , j u,

n1

aj(k1)
k 1
u,
n1

(k1)( j1)

(i1) j

f if i s,k t, j

ij

c
a
a

a a f aa f

aa

 n1
a( j1)(k1)a
u,

ai1( j1) = fij if i s, k t, j

(k1)( j1)

k1

( j1)k k (i1)

n1

ai1( j1) fij if i s, k t, j u.

a
k 1

Thus wij fij in every case and therefore Ast Atu Aus for every t {1,2,..,n}
Since we know that |

| | || |; we have A st A tu A st A tu A us .

This means btsbut bsu i.e.,

butbts bsu for every t {1,2,..,n}

Hence B = adjA is circular.

Theorem 4.3.3 [12]: To construct a transitive fuzzy matrix from a given fuzzy
matrix through adjoint matrix.

For any n n fuzzy matrix A, the fuzzy matrix A(adjA) is transitive.

Proof: Let C = cij

= A(adjA).

Then cij =

Ajk

= aif

k1

and

1
ast

1is csj

c
s1

nn

A if

l
jt

for some f {1,2,,n}

a il
1

sl

AA
n

ih

Ash

a
s 1

= aih Agh

agu A ju

su

A ju

, for some h , u {1,2,,n}

 aih A ju

Ajf = ci j

aif

Thus

ci j

and so (A adjA)2 A (adjA).

= 0.3is 0.6
0.4
Hence AA(adjA)
transitive.

Example 4.3.3: For a square fuzzy matrix

0.7 0.4
0.8
Then adjA = 0.5
0.4 0.8

0.9

0.2

Now A (adjA) = 0.3 0.6

0.6 0.7

0.6

0.5

0.6
=

0.7

0.4

0.4

0.4

0.8

0.5

0.7

0.8 0.6

0.9

0.2

0.6

0.7 0.5

0.4

0.6 0.4

0.6

0.7 0.6

0.7

0.6

1 0.6

0.7

0.5

which is transitive fuzzy matrix as

0.6

0.7

0.5 0.6

0.7

0.7

0.6

0.7

0.6 0.6

0.6

0.6

0.5

0.6

0.6

0.6

(A(adjA)) = 0.4 0.6

0.4

0.4

= 0.4 0.6

0.4

0.5

0.6

0.6

0.4

 A (adjA)
Hence for any square fuzzy matrix A, A(adjA) is transitive.
Definition 4.3.3 [15]: An n n fuzzy matrix A is called idempotent fuzzy matrix iff
A2 = A.
Let us understand by an example to convert reflexive fuzzy matrix into idempotent
fuzzy matrix by taking its adjoint matrix:Example 4.3.4:
For a reflxive fuzzy matrix
A = 0.3 1 0.4

0.1

0.2

0.5

0.6

adjA = 0.4
Then

1
0.2

1
0.4

0.2

0.5 0.6
(adjA) = 0.4 1

1

0.4 0.4

which is idempotent fuzzy matrix as

0.2

0.2

0.2

=
0.4

0.4

0.2

0.4

= adjA

0.2

0.2

0.5

0.6

1 0.5

0.6

1 0.5

0.6

Hence idempotent fuzzy matrix is formed by taking adjoint of reflexive fuzzy matrix.
Remark[13]: For a reflexive fuzzy matrix A of order n , adjA = Ac where Ac
is
idempotent and c n1.

Example 4.3.5: For a 3 3 reflexive fuzzy matrix

0.1

0.2

0.5

0.6

A = 0.3

0.4

Then

0.2
1

0.5

0.6

adjA = 0.4

0.2
0.4 is idempotent fuzzy matrix

Now
1 0.1
A = 0.3
1

0.2 0.3
1
0.4

0.6
1

1 0.5
0.4

0.4
0.2

0.5
= 0.4

1
0.1

0.6

1
0.2

0.2

0.5

0.6

= adjA
Then A2 is idempotent fuzzy matrix.
Theorem4.3.4 [12]: Let A be a n n

reflexive fuzzy matrix. If A is idempotent then

adjA = A is idempotent.
Proof: We know that for a reflexive fuzzy matrix A of order n, adjA =
where

is idempotent.

But we have also that A is idempotent and so

=A.

Thus adjA = A.
Since A is idempotent and hence adjA is idempotent.

(c n 1)

Theorem 4.3.5 [12]: Let A be an n n

following properties:
(i)
(ii)
(iii)
(iv)
Proof:

adj = ( A) = adjA
adj (adjA) = adjA
adjA A
A(adjA) = (adjA)A = adjA.

reflexive fuzzy matrix. Then we have the

(i)

Since A is reflexive =and thus

A
Then adj

c 2

is also reflexive.
where

= adjA is idempotent.

adjA2 = (adjA)2
Since adjA is idempotent and so (adjA)2 = adjA

2
Hence adj
= (adjA) = adjA

(ii)

(iii)

Since A is reflexive and then adjA is reflexive.

Also for a reflexive fuzzy matrix A, adjA is idempotent.
Hence adj (adjA) = adjA.
(by theorem 4.3.4)
Since A is reflexive, then aii =1 i
Let B = bij = adjA
Then

bij

t (t )

Sn j ni tn j

Taking the permutation of set nj over the set ni such that

(h) = h ,

(i) = j , h

i
i.e., the permutation
1 2 3
1 +1
1 2 3
1 +1
11
22
33
(j-1)(j-1)
(j+1)(j+1)
Thus a a a a
a
ann is a term of bij so that
bij a11 a22 a33a(j-1)(j-1)a(j+1)(j+1) ann = aij
(
=1 )
Hence adjA A.
(iv)

Since A = aij is reflexive , then

=1 i

Then ij ij ik jk
A
Let Bc= b a= adjA

aii Aji Aji bij aii 1i)

k1
n

Let C = cij
and ij

= A(adjA) and D = dij = (adjA)A.

ki

akj Aji ajj Aji bij ajj 1j)

k1
n

Thus we have A(adjA) adjA and (adjA)A adjA.

But adjA = (adjA) (adjA)
[by (i)]

(
(

 A(adjA) [by (iii) and using the result A B AC BC]

So that A(adjA) = adjA
Also adjA = (adjA)(adjA) (adjA) A
[by (i),(iii) and using the result A B CA BA]
So that (adjA)A= adjA
Hence A(adjA) = adjA = (adjA)A.
Example 4.3.6: For a reflexive fuzzy matrix

A = 0.3
0.7

0.5

0.8

1
0.6

0.4
1

1 0.5
We have A = 0.3 1

0.7

adjA = 0.4

0.6

0.8 1
0.4 0.3

0.5
1

1 0.7

0.6

1
0.6

0.4

0.8 = 0.4
1 0.6
0.4
1

0.7

0.8
0.4

0.6

0.8

adjA = 0.4
0.7

1
0.6

1 0.6
A(adjA) = 0.3 1

0.4 A , (adjA) = 0.4

1

0.8
0.4 0.4

0.7

0.6

0.5

0.8

0.7

0.6

0.4

0.4

=1 0.6 0.8
0.4 1 0.4

0.7

0.6

1 0.7

0.6

0.5

= adj(adjA)

0.8

0.6

0.6

0.7

0.6

0.6

0.8

1
0.6

0.8

0.8

0.8

0.7

0.6

1 0.7

0.6

0.7

0.6

It is clear that this example satisfies all the statements of the above proposition.
Hence adjA2 = (adjA)2 = adjA = adj(adjA) = A(adjA) = (adjA)A and (adjA) A.

(adjA)A = 0.4

0.4

0.3

0.4

= 0.4

0.4

Chapter 5
SUPER AND SUPER FUZZY MATRIX THEORY
5.1 Introduction
Since now, we are well familiar with basic and fuzzy matrix theory. In this chapter,
our purpose is to compare super and super fuzzy matrix theory. Firstly, we shall study
super matrices and some operations defined on super matrices. Further in this chapter, we
shall study fuzzy supermatrices and some operations defined on fuzzy supermatrices.
However, the concept of supermatrix for social scientists was first introduced by Paul
Horst, Matrix Algebra for Social Scientist [15]. Let us first study the concept o
f
supermatrices.
5.2 Supermatrix Thoery
The general rectangular or square array of numbers are known as matrices whose
elements are just an ordinary number or a letter that stands for a number. In other words,
the elements of a simple matrix are scalars of scalar quantities. Now we define the notion
of a supermatrix.
Definition5.2.1 [4] A matrix whose elements are themselves matrices with elements that
can be either scalars or other matrices is called a supermatrix.
In general the kind of supermatrices we shall deal, the matrix elements which have
any scalar for their elements. Thus a supermatrix is a general rectangular or square array
of matrices.
Let us understand how to construct a supermatrix from the given matrices with the
help of an example. Also we shall understand the same with the help of figure and
a
table.
Example 5.2.1 Consider the following matrices

a11 = 2

0 , a12 = 2

0 , a13 = 1

2 5

73

a21 =

4
6

aa41
31

, a22

96

3 , a33 = 3
3
2

0
6

5
1
=
7

, a32 = 6
3

3 4 , a42 =
=

= 1
1 35 3
5
6
2

6 5 3
3 6 1

, a33 =

1
6
4

4 1 7
, a43 =1 3 1 6
8 5 03

55

Here aij denotes a matrix

and anot
a scalar of a matrix (1 i 4,1 j 3).
a
a22
23

Let
a11

a12 a13

a = 21

3 2
6 1
a31 a32 a33

a41

a42

a43

We can write out the supermatrix a in terms of the original (natural) matrix elements i.e.,

0 1
32
3

7 4

33

3 6

3 4

a=

6 5 3 6

0 5

Here the elements are divided vertically and horizontally by thin lines. If the lines
were not used, the matrix a would be read as a simple matrix.
A Diagrammatic representation of supermatrix a showing within submatrices aij along
with their orders, is given by the following figure:

Figure 5.1 [4] A Diagrammatic representation of supermatrix a

Remark: All submatrices within a given row must have the same number of rows.
Likewise all submatrices within a given column must have the same number of columns.
Now let us show the orders of submatrices aij (1 i 4, 1 j 3) within supermatrix
a, in the
tabular form a
1
2
3
Column
(j)

s below:
Row (i)
1

13

12

14

33

32

34

23

22

24

23

22

24

Table 5.1 An index for the orders of submatrices aij (1 i 4 , 1 j 3)

The Order of a Supermatrix [4] is defined

in the same

way

as that

of a simple

matrix. The height of a supermatrix is the number of rows of submatrices in it and the
width of a supermatrix is the number of columns of submatrices in it. Thus we have for
this supermatrix 4 rows and 3 columns and so order of supermatrix is 4 3. However
the order of the corresponding simple matrix (considering the matrix without horizontal
and vertical thin lines) is 8 9.
Observations:
1.

The order of a supermatrix tells us nothing about the orders of the submatrices which
by general rectangular or square array construct a supermatrix.
2. The order of a supermatrix tells us nothing about the order of the simple matrix which
on dividing the elements by horizontal and vertical thin lines construct a supermatrix.
This process of dividing the elements of a simple matrix by horizontal and vertical
thin lines to construct a supermatrix is called partition [4].

Note: Different supermatrices constructed from a simple matrix by different partitions are
all equal. This we shall discuss later on while dealing with type II supervectors.
From a given simple matrix, different supermatrices either of same order or of
different orders can be constructed, each time partitioning between different rows and
columns in any way that happens to suit our purpose.
Example 5.2.2 Consider a 5 5 simple matrix
78

22 7 4 0 7

A= 3

25

This matrix can be partitioned by drawing thin lines in a number of ways to obtain

different supermatrices.
Let us partition between the columns one and two and three and four. Also let us

partition between the rows one and two and three and four.
Then the supermatrix is obtained is
2
5

7
8

as = 3

2
2

4
7

7
0

which is symmetrically partitioned.

By a symmetric partition [4] of a matrix, we mean that the rows and columns are
partitioned in exactly the same way. According to this rule of symmetric partition, only
square simple matrix can be symmetrically partitioned. Now we proceed on to recall the
notion of symmetric partition of a symmetric simple matrix [4].
Example 5.2.3 Consider a fourth order symmetric matrix and partition it between the
second and third rows and also between the second and third columns. Then the second
order supermatrix obtained is
6 9 6 5
a=

4 6

2 5

a11

=
a21

a11

a12

a22

'

 a12

a12

a 22

The general expression for a symmetrically partitioned symmetric matrix is given by

a22 23
2n
a11 a12 a13 ... a1n
a = a13 a23 a33 ... a3n
'

a2' n a3' n
nn
a
...
a
a21
'

...

'

a1n
The simple rule about the matrix elements of a symmetrically partitioned symmetric
simple matrix are
1.
2.
e

The diagonal submatrices

of the supermatrix are all symmetric matrices i.e., = aii .

D
=
The non-diagonal submatrices are symmetric about symmetric about the diagonal. In
'
' elements below the diagonal are the transposes of th
other words the matrix

corresponding elements above the diagonal and vice-versa.

i.e.

= aji and

= aij for i j.

The general experssion for a symmetrically partitioned simple diagonal matrix [4]
is given by
D11 0 ... 0
'

D12 ...

0... D nn

0 denote the matrices with zero as all entries and 0 only represents the order
is
reversed or transformed. We can write simply,

D1

...

D= 0

D2 ...0

Dn

...

which is reffered to as the super diagonal matrix.

Example 5.2.4 Consider a simple partitioned symmetrically diagonal matrix

00 0 0 0 0
=
6 0 0

4 0
0 5
D= 0 0

D1

0 t 0

0 D
0
2

0
3

The Identity Matrix I =

I
r

0 1 0
0 0the 1orders of the identity matrices Is , It and Ir respectively, is a
where s, t and Ir =denote
0 0 0 =

super identity matrix [4].

Example 5.2.5 Consider a super identity matrix

0 0

0
I4

0
1

0 0

0
I2

0 1

0 0

0 1

Now we define the notion of partial triangular matrix as a supermatrix [4].

where T is a upper triangular submatrix.

Then u is a partial upper triangular matrix partitioned as a supermatrix [4].
Example 5.2.7 Let

6
2

0
0
0

7
0

l= 8 5

T
a'

where T is a lower triangular submatrix.

Then l is av partial
lower triangular matrix partitioned as a supermatrix [4].

Now
we
define
the
notion of supervectors of type I and type II.
Let v =
A simple vector is a vector each of whose elements is a scalar. In this way a supervector
is a vector each of whose elements is either a scalar or a vector.
Definition 5.2.2 [4] Type I Column supervector
v1
2

Let v =
Example 5.2.8
4

vn

where each vi (1 i n) is a column subvector of the column vector v. Then v is a type I

column supervector.
6

v1

v2

3
6

where v1 , v2 are column subvectors.

Then v is a type I column supervector.
Definition 5.2.3 [4] Type I row supervector

'

(1 i n) is a row subvector of the row vector v.

Then v is a type I row super vector.

'

where

Then

and

are row subvector of the row vector .

is a type I row supervector.

Next we recall the definition of type II supervectors.

Let
each

= v1 2 ' v

vn'

Example 5.2.9 Let

= 2
'v
2

where

5 1 5

5 = v1

Definition 5.2.4 [4] Type II column and row supervector

The general m n supermatrix is given by

aa11
21

aa
22

a=

a1n

... a 2 n

a1 a11 a12 ...

a1n

a m1 a m 2...

amn

If
1

a12 a21 a22...

a2n

..
a=

am am11

am2

...

amn

Then

a
a
a
1 , a2 ,..., an
a
a

a2

am

is defined to be the type II column supervector. Similarly if

a11 a12 a1n

1

21 22 2n

am1 am2

then a a 1 a 2

...

amn

an

is defined to be the type II row supervector.

Clearly

a1

1
m

am

shows the equality of supermatrices [4].

a a2

...

Remark: Different supervectors (supermatrices) constructed by different partitions from

a simple matrix are all equal. In other words, two supermatrices are equal iff their
corresponding simple forms are equal.
Now we define the notion of transpose of a supermatrix [4].
The transpose of the general m n supermatrix a, denoted by is given by
'

'

a11

'

'

am1

'

a21...

'

'

'

'

a12 a22...

'

am2

a1n

a2n...

amn

Then is a n m supermatrix obtained by taking the transpose of each element i.e.,

the submatrices of a.
'
'
'
'
Next we will find the transpose of supermatrix constructed by symmetrically
'
partitioned symmetric simple matrix.
The transpose of ' a symmetrically
partitioned
n n symmetric simple matrix denotedby
'
''
is given by

aa11' 1n a(a'2n12)'

'

a(a3n13 ......

ann(a 1n )'

)'

'

a12

'

'

a22

(a23)

a' a13 a23

'

...

a2n

a33... (a3n )'

(1) The diagonal submatrices are all symmetric matrices and so are unaltered b
y
transposition i.e.,

= .

(2) Recall also the transpose of a transpose is the original matrix i.e., (aij )' aij
Thus the transpose of supermatrix constructed by symmetrically partitioned
symmetric simple matrix a, denoted by

is given by

'

a22 a2313 ... 2n a1n

a11 a12

'
a a13 a23 a33 ... a3n a
'

...
'
a nn
a2' n a 3...
n
a12
'

'

...
a

a1n
Hence the transpose of supermatrix constructed
symmetric simple matrix is the supermatrix itself.

by symmetrically

partitioned

Similarly the transpose of symmetrically partitioned diagonal matrix and identity

matrix are the original supermatrices itself.
Also it can be easily verified that

The transpose of a partial triangular supermatrix is the partial lower triangular

supermatrix and vice versa.

The transpose of type I column supervecrtor is a type I row supervector and vice
versa.

The transpose of type II column supervector is a type II row supervector and vice

va2
vb2
versa.

va
and vb
be two column supervectors of type I.

Now we proceed on to define the notion of minor and major product of supervectors.
Firstlly we recall the definition of minor and major product of type I supervectors.
Definition 5.2.5[4] Minor product of type I supervectors:
Suppose

 vb1

va1

The minor product is defined as

v
a

v
b

vavb v

vb1
va'

v
b
n
vb2

...

va

'

= vavb va bv
v

... a vb

n
vb

'

'

1 v1a

'

nvb n

Example 5.2.10 Consider two type I supervectors

v a
3 , va
1 2

va

,v

v a v a and vb vb
2

vb

where

and

, vb

4
2

1
0

0 2
,v

23

1
'

'

'

'
1

'

vavb va va

vb

va vb

vb

= va vb va

v va vb

3 1 0 + 4 1 0 3
1

= 2

2
0

+ 2 1

= (2+3) + (4+12) + (4+1)

= 5+163
It can be easily verified that vavb vv
b a .
= 18
'

'

Example 5.2.11 We just recall if

v=

is a column vector and

v = 1

, the transpose of v is a row vector then we have

4
4

2
2
3is
+a 4column
vx == 1 2 +
vector , then

= 1+9+16
= 26

Thus if
x1
x
2

xn

x
x1

= x1

x2

...

xn

=
+

xn

Also x1 x2

...xn

= x1 x2 ... xn

x .

i.e., vx .1 1.vx

va 1 x1
x

vb1
v
b2

vb
v 2
and 1 1 ... 1 = x1 x2 ... xn

xn

va
'
'
'

v 2 . vib
b
bv

v
a Major product of type I supervectors
Definition 5.2.6[4]
'

'

'
a b

va2

Suppose

and vb =

am

be any two supervectors of type I

The major product is defined as

va1

v...

 va vb

'
b

va

'

va vb
1 1
=

'

v ...

va1

'
b
2

2'

va b v

v ... va b'v
1

'

1'

... v2a

n'
b

vamvb1

vamvb2

va = va2 and vb =
Example 5.2.12 Consider two type I supervectors

va

vb1

va

where va = [3] , va =
=
3
v
b2

, v a 2

and vb = 1 , vb =
= [2] 1 2 v b3

, 1v b =

vb4

3
5
2
1

0
6

, vb

 va
Now va

''

'

'

'

'
b

= va2

'
a

'

v
'
'
b1vb2 vb vb

va

va b
v

va vb

va

= va vb va vb

va vb

v v

v v

va vb

'
v vb

v v
a

3 1

'

va vb

 13 55 11 00
3

=
1

0 2 5 1

31 22 11 31 33

1

2 2 1

0
0

1

2

11 00 66 3 1 22

1

3 1 0 6 2 2

5
15
=
15 5

3
5

0 6

3 9 3

1 3

1
10 2 0

'

4 2 6 2 0

18

18

'

06
12

2
4

0 0 0 0 0 0

0 0

It can be easily verified that (va v b )' vbva .

... a2n
b
b
... b
a=
Next we recall the definition of minor and major product of type II supervectors.
Definition 5.2.7 [4] Minor and Major Product of Type II Supervectors
Consider two general supermatrices

a11

a12

...

a1n

am1

am2 ...

amn

 b11
and b =

21 22 2 s

...

b12

b1s

br1

br 2

...

b1

a1

a1

a2

...

an

a2

b1

bs

b 2 ...

b2

and b =

brs

am

=
n

b1

The minor product of type II supervectors is given by

Then a =

ab = a a2

an b 1

...

= a b
1 a b
2 ... a br s

ms

b2

ab =

bs

a1b

b b 2 b 1 ...

2 1

n 1

1 2

a1 b

...

a11bs

...

a2 b

am1 b s

and the major product of type

II supervectors is given by
1 s

a1
am
b
1

a b ...
1
m

a2
1

am

a2b
=

a2b

Example 5.2.13 The minor product of type II supervectors

a1 a2
1 3

a= 2

and b b2

b1

=
1 4

is given by
1

b2
1

ab =

1
= a b
1 a b
2

1
1

 1 7 +
2
3 6
1


=
4
1

= 2 14 + 12 0

= 14
1

14
7

1 3 a1
b

9
0

a = 2 6 a12 and b =
1

a3

b2

And the major product of type II supervector

1

ab = 1a12 4 b b2
1
a
is given by

a11b a11b2
a1
= a12b a12b2

2
ab ab

3
= 2 6

1
1

1
2

7
0

= 14

14

It is easily verified that the product of two matrices a and b as a minor product of type II
supervectors coincides with the major product.
Next, Let us discuss the concept of fuzzy supermatrices.
5.3 Super Fuzzy Matrix Theory
We are well femiliar with fuzzy matrices whose elements belong to [0,1]. In this
section of our dealing, we for the first time introduce the notion of fuzzy supermatrices
and operations on them. Throughout this section we consider matrices with entries only
from the fuzzy interval [0,1]. Thus all matrices in this section unless we make a specific
partition or mention of them will be fuzzy matrices. Now we define the notion of a fuzzy
supermatrix
Definition 5.3.1[4] Let us consider a fuzzy matrix

A
A 11

A
... A
A 12...A 1n

A=

21

22

2n

A m1

Am2

...

A mn

where Aij (1 i m and 1 j n ) are fuzzy submatrices of A; with entries from [0,1]
such that number of rows in fuzzy submatrices Ai1 , Ai2, ,Ain for each i=1,2,,m are
equal and similarly number of columns in fuzzy submatrices A1j , A2j, ,Amj for each
j=1,2,,n are equal.
Then A is a general super fuzzy matrix or a general fuzzy supermatrix.
Now we define the notion of transpose of a super fuzzy matrix [4].

The transpose of the fuzzy matrix A, denoted by A is given by

'

'

'

A 12 A 22 ... A
= A
11
A 21
...

A ' 1n

'
2n

...

'
m2

mn

m1

'

where

0.4

are transpose of Aij (1 i m and 1 j n ) .

0.1 0.6

Example 5.3.1 Consider a super fuzzy matrix

0.4

0.6

0.5

0.7

0.7
0.5

0.6
0.4

0.2 0
0.9
0.6 0.7 0.6

0.4

0.4

0.3

1
0.3
0.3

0.4

A = 0.5 0.2
= A 21 A 22
0.7 0.1

0.7 0.9
A 23
0.4 0.4

0.4

0.1

0.3

0.1

0.6 0.7

0.3

0.4

0.6

0.8

0.9

0.8

0.1

0.3

0.4

0.6

0.3

0.3

0.4

0.6

0.1

0.4

0.2

0.3

A 11

0.5

A 12

A 13

A 32

A 33

A 31

0.6

0.6

0.2

0.4

0.4

41

42

43

where Aij are fuzzy submatrices of A with entries from [0,1] such that Ai1 , Ai2,Ai3
have same number of rows for i =1,2,3,4 and similarily A1j , A2j,A3j,A4j have sam

e
number of columns for j = 1, 2, 3.
The transpose of the fuzzy supermatrix A denoted by A is given by

'

A 11

'

A= A 12

'
'

23

21

'

13

'

'

'

'

22
'

'

A 31

A 33

A 33

43

41

'

42

10

0.4
0.1

1
0.4 0.6

0.6
0.5

0.5
0.2

0.7 0.8
0. 0.9 1 0

0.7

0.7

0.4

0.8

0.5

0.3

0.7 0.5

0.4

= 0.6 0.4

0.9

0.4 0.4

0.4
0.1

0.1
0.3

0.6

0.4

0.6

0.6

0.2 0.6 0.3

0.1

0.6

0.4 0.1 0.2

0 0.7

0.3

0.7

0.6

0.4

0.9 0.6 0.3

1 0.3 0.4

0
0.4 0.3 0.2 0.4
0.3
0.6
1 0.3
0.4

Next we defne the notion of fuzzy super row matrix, fuzzy super column matrix and
their transpose.
Definition 5.3.2 [4] Fuzzy Super Row Vector (Fuzzy Super Row Matrix)
A

Let A = (m > 1)
Let A= [A1 A2 An] or simply [A1 A2 . An] (n>1) where each Ai , i=1,2,,n. is
a
fuzzy row vector (fuzzy row matrix). Then A is called as the fuzzy super row vector or
fuzzy super row matrix.
Definition 5.3.3 [4] Fuzzy Super Column Vector (Fuzzy Super Column Matrix)
A1
2

Am

where each Ai (i=1, 2, , m) is a fuzzy column matrix (fuzzy column vector).

Then A is called the super fuzzy column matrix or super fuzzy column vector.
Now we define the notion of transpose of super fuzzy row matrix and super fuzzy
column matrix [4].

The transpose of super fuzzy row matrix A , denoted by A is a super fuzzy column
matrix is given by
A1

A =

'

A2

An

'

'

where each A is fuzzy column matrix.

'
'
'
Similarly the transpose of super
fuzzy
column
matrix A, denoted by A is a super fuzzy

row matrix given by A = A1

A 2 ...A m

where each A is a fuzzy row matrix.

Example 5.3.2 Let A = [A1 A2 A3 A4] be a super fuzzy row vector where
A1 = [ 0 0.2 0.3], A2 = [0.4 0.2 0 0.3], A3 = [1 0] , A4 = [0.2 0.3 0.2 ]
i.e., A = [0 0.2 0.3 0.4 0.2 0 0.3 1 0 0.2 0.3 0.2]
'
The transpose of super fuzzy row vector A is given by

A = A' 2
A
A1

'

, A = 0 .2 , A = 1

= 0.2
3

where A

A '4

0.2
0

0.3

0.2
and A 0.3
=

0.2

0 .4

0

0 .3

Thus A =
0

0
0.3

0.4

0.2

0
.3

0.2

0.3

0.2

0.3

,
A1

Example
5.3.3 Let
A = A2

be a super fuzzy column vector, where A1 =

A3

1

0.6

A2 =

, A3 = 0.3

0.3

0.4

0.5

0.7

0.3

0.6

0.3

i.e., A = 0.5

0.3
A = A1A 2 A 3'
0.4

0.7
0

The transpose of super fuzzy column vector A is given by

'

'

where A = 1
Thus A= 1

0.3

0.3

0.6 , A = 0.3

0.6 0.3

0.5 0.3

0.4

0.5 , A = 0.3

0.4

0.7 0

0.7 0

is a super fuzzy row matrix.

Next we define the two types of products of these fuzzy super column and ro
w
matrices i.e. A . A and where A . A is a fuzzy super row matrix.
Firstly, Let us define A . A where A is a fuzzy super row matrix.
Definition 5.3.4 Let A= [A1

A2 . An] be a super fuzzy super row matrix i.e., each Ai

is a 1ti fuzzy row submatrix of A , i=1,2,,n and A be the transpose of A.

A1

max{A. A } = ma
A 1
x

. A

An

= max {A1. A , A2. A , , An. A },

which is known as the minor product of two fuzzy supermatrices.

'

A
'2

'

i.e., max{A. A }= max {max min (ai , ai' ) , max min (ai , ai' ) ,, max min (ai ,
ai' ) }
1
1
2
2
n
n
1 < i1,
2

'

< t1 , 1 < i < t2 ,, 1 < in

A1 = a1 , a2 , ...,, at

a
1

A2 =

,...,at2

< tn where

and A =

,
at'
a1
'

'

a2

and A =

at

'

'

An =

a1, a2 ,..., ant and A =

aa1 '
t

'

a2
by minimum of (ai , ) and the sum of the elements ai ++ ai replaced by the

,a

maximum of ( ai , , ai ).

....

a1
'

a2

Thus max{A. A } = a [0,1].

This fuzzy supermatrix operation is the usual operation with product of ai , replaced
1

Thus given any fuzzy super row matrix A, its transpose A is the fuzzy super column
matrix such that the product A. A , defined as the product of two fuzzy super row matrix
(vector) and fuzzy super column matrix, is always an element from the fuzzy interval
[0,1].
Example 5.3.4 Let A = A1 A2 be a super fuzzy row matrix.
where A1 = 0.1 0.4

0 , A2 = 0.3

A'
A =

is a fuzzy super column matrix, where

'

A2


A = 0.1
0.4

and A = 0.3

Now A. A = max min{A , A }

(This is a notational convenience of the product)
= max min

A1

'
A

A 2 ,

A2

0.1 0.4 0
= max {A1 . A ., A2. A }

0.1

= max

0.4 , 0.3 0.

= max {min (0.1,0.4),min(0.4,0.4),min(0,0)},max{min(0.3,0.3)}

= max(0.1,0.4,0),max(0.3)
= 0.4,0.3
Then max{ A.A } = max {0.4,0.1}
= 0.4
The way in which max {A.A }= max {max min{A , A }} is defined as peculiar which
may be defined as the super pseudo product of the fuzzy supermatrix A and the transpose
of the fuzzy supermatrix A .
Now how is A A defined where A is a n 1 fuzzy super column vector and A is a
n 1 fuzzy super row vector.
As we have defined A.A as the super pseduo product, now we define A .A as a super
fuzzy n n matrix.
a1
Definition 5.3.5 [4] Let A = (a1 ,a2 ,,an) = (A1 A2 ... An ) be a super fuzzy row matrix
and A be the transpose of A
min{A ,A} =
min

'

'

a2 a1

a2

'

... an an

 min(a1' , a1 ) min(a1' , a2 )...min(a1, an )'

'
2

min{A ,A} =

'

, 1 ) min( , 2 ) ...2 min(a a

a .

' '
min(
, )min( ,
)
11
)n n n na aa a a

'
a 2amin( , )na

...' min(
,a

Now min{A ,A} is a n n square fuzzy matrix. It is partitioned as per the divison of
rows and columns of A and A respectively. It is important to note min{A ,A} is
a
symmetric matrix about the diagonal.

min{A ,A} = min(a1 , a1) min(a1 , a2 ) ...

min(a2 , a1) min(a2 , a2 ) ...

where

min(a1 , a n )

min(a2 , an )

,a

= ai for the elements remain as it is while transposing the elements. Clearly

since min(a2,a1) = min (a1,a2) , we get min{A ,A} matrix to be a symmetric matrix.
Further

if

A = ( A1 A2

An) with number of elements in Ai (i=1,2,,n) is ti

(1 < ti < n), min{A ,A} is a super fuzzy matrix with i i fuzzy submatrices i = 1,2,,n.
and min {A ,A}is a n n fuzzy matrix.
We illustrate this by the following example.
Example 5.3.5 Let A = 0.2 0.3 0 0.3 0.6 0.4 0.7 0.7 0.5 = [A1 A2

A3]

be a fuzzy super row matrix with

A1 = 0.2 ' 10.3
A 0 , A2 = 0.3 0.6 and A3 = 0.4 0.7 0.7 0.5

0.7 ,
and A, = '2A
; where A =0.30.3

0.4

0
' and A =
0.2
A =
0.7

A
3

0.6

0.5

 0.3
0.2

0.2 0.3 0 0.3 0.6 0.4 0.7 0.7 0.5

0.6 ,
we find min{A ,A} = mi 0

n
0.7

0.3
0.5

min(0.3, 0.5)

0.2)
min(0.3, 0.3) min(0.3,0.7)
min(0.3,
0.4

0.7

min(0.2, 0.2) min(0.2, 0.3) min(0.2, 0.7) min(0.2, 0.5)

0.2

min{A ,A}=

0.2 0.3 0 0.3 0.3 0.3 0.3 0.3 0.3

min{A ,A}= min(0.7,

min(0.7,
0.2 0.3 0.2)
0 0.3
0.6 0.3)
0.4 0.6 min(0.7,
0.6 0.50.7) min(0.7, 0.5)
min(0.5, 0.2)

min(0.5, 0.3)

min(0.5, 0.7)

min(0.5, 0.5)

Here min {A ,A} is a 9 9 fuzzy supermatrix partitioned between 3rd and 4th row , 5th
and 6th row and similarly between 3rd and 4th column, 5th and 6th column.

0.2

0.2

0.2

0.2

0.2

0.2

0.2

0.2

0.3

0.3

0.3

0.3

0.3

0.3

0.3

0 0 0

0.2

0.3 0 0.3 0.4 0.4

0.4

0.4

0.4

0.2

0.3 0 0.3 0.6 0.4 0.7

0.7

0.5

0.3 0 0.3 0.6 0.4 0.7 0.7 0.5

0.3 0 0.3 0.5 0.
0.5 0.5 0.5
4

0.2
0.2

B11

B13

B12

'

B
'

'

13

23

B
33

= B12

22

B
23

where
B11 min{A1' , A
1}
B12 min{A1 , A2}
'

B13 min{A1 , A3}

'
'
B12' min{A
'
2 , A1} min{min(A1 , A2
)}
B22 min{A2 , A
2}
'
B23 min{A2 , A
3}
B13 min{A3' , A
1 } min{min(A
1
,3 )A}
B23 min{A3, A2} min{min(A2 , A3 )'}
B33 min{A3' , A
3}
'

'

'

'

'

'

'

Thus A .A is a 9 9 super fuzzy matrix. Further A .A under the super pseduo product
is a symmetric fuzzy square supermatrix i.e., A .A is a symmetric fuzzy supermatrix.
Next we define the notion of square fuzzy symmetric supermatrix but before, firstly let us
define the notion of square fuzzy supermatrix.
Definition 5.3.6 [4]: Square Fuzzy Supermatrix
Let A is n n square fuzzy matrix. Then A is called a square fuzzy supermatrix o
r
super fuzzy square matrix, if A can be partitioned arbitrarily between the columns i1 and
i1+1 , i2 and i2 and i2+1, ,ir and ir+1 and similarly between the rows i1 and i1+1 , i2 and i2
and i2+1, ,ir and
0.3 ir+1 (r+1< n) We illustrate this by a following example:
Example 5.3.6: 0.5
Let A be a fuzzy square matrix
A=
0.2

0.4
0.5

1
0.1

0
0.2

0.6

0.3

0.4

0.3

0.4

0.3

0.9

0.8

0.3
0.1

0.3

0.4
0

0.1
0.3

0.4
0

0.5
0.6

0.3

0.4

 0.3

0.5

0.9

0.6

The 6 6 square fuzzy matrix has been partitioned between the columns 2nd and 3rd,
3rd and 4th and also partitioned between the rows 2nd and 3rd, 3rd and 4th . Thus A is a
square fuzzy supermatrix. Now having defined a square fuzzy supermatrix, we next
proceed on to define the notion of symmetric square fuzzy supermatrix.

Definition 5.3.7 [4]: Square Fuzzy Symmetric Supermatrix

Let A be a fuzzy super square matrix or fuzzy square supermatrix. Then A is called as
a symmetyric fuzzy super square matrix or a fuzzy symmetric super square matrix or a
symmetric square super fuzzy matrix , if
A11

A 12...A1n

A 21

A 22

A 2 n

...

A=

A m1

Am2

...

A mn

is such that Aii are square fuzzy matrices and each of these fuzzy square matrices are
symmetric square matrices and diagonal submatrices Aii are symmetric square fuzzy
matrices i.e., A = Aii and non diagonal fuzzy submatrices are symmetric about th
e

diagonal i.e., A = Aji for0.31 i, j n.

We illustrate this by the following example:
0.3 0 .4
0 .6

0 .5
0 .4 0.5

0.5
0.2

0 .3
0.5

0
0.2

0.3
0.4

0
0.1

0.1

0.2

0.2

0.6

0.5

0.3
0.7

0.6

0.2

A = 0 .4 0.7

0.1

0 .3

0.2

0.4

0.3

0.4

0 .3

0 .2
1
A

0 .2 0 .3 0.1

0.4

0.4

0.1

0.2

0.7

0 .3

0.6

0 .7

0.9

0.4

0.5 0 .6

0.5

0 .1

0.2

0.2

0.1

0.6

0 .1

0.4

0.8

0.6

0.2

0.3

0.2

0.4

0 .2

A11

0 .3
0

0 .3
0

A12

21 22 23

A13

0.4

0 .3

0 .5

0.4

0.8

0.5

Clearly A is a square fuzzy supermatrix. For it is 9 9 fuzzy matrix partitione

d
between the columns 3rd and 4th , 5th and 6th and also partitioned the rows 3rd and 4th ,
5th and 6th . Further the diagonal fuzzy matrices A11, A22,,A33 are all symmetric fuzzy
matrices and non-diagonal fuzzy matrices are symmetric about the diagonal i.e.,
A = A21 , A = A31 , A = A32.
Thus A is symmetric fuzzy supermatrix.

A31

32

33

Also we have shown that if A is a fuzzy super row vector and A , the transpose of A is
a fuzzy super column vector then A A under the super pseduo product is a symmetric
fuzzy square supermatrix i.e., A A is symmetric fuzzy supermatrix.
Thus we can say if one wants to construct fuzzy symmetric supermatrices then one
can take a fuzzy super row vector A and find the pseudo product of A and A. Then A A
will always be a symmetric fuzzy supermatrix.

Chapter 6
SUMMARY AND CONCLUSION
6.1 Introduction
George Cantor (1845-1918), a renowned German mathematician, was the first to start
the formulation of set theory. The first axiomatic study was attempted by G. Frege in
1879. But in 1902, this theory received a setback. Bertrand Russel, an Engli
sh
mathematician and Philosopher, showed that Freges system was inconsistent by devising
a contradiction, now called Russels paradox. In 1908, Zermelo proposed a theory of sets
which was later improved by Fraenkel and Skolem. Most parts of present d
ay
mathematics are based on Zermelos set theory.
Considering the important role that vagueness and inexactitude play in human
decision making, Professor Loffi A. Zadeh [15] through a seminal paper in 1965, laid the
foundation of fuzzy set theory which gave a form of mathematical precision to human
cognitive processes that in many ways are imprecise and ambiguous by the standards of
classical mathematics.
The fuzzy set theory is further related to fuzzy matrix in the same way as the set theory
is related to matrix theory. Kim [11] defined the determinant of a square fuzzy matrix.
The adjoint of a square matrix theory. The adjoint of a square fuzzy matrix is defined by
Thomason [13] and Kim [11].
The concept of supermatrix for social scientists was first introduced by Paul Horst,
Matrix Algebra for social scientists [2]. This concept was then observed in the light of
fuzzy matrices to introduce fuzzy supermatrices.
6.2 Summary
This section recalls the summary of the work done in the previous five chapters a
s
follow:

Chapter 1 (Binary And Fuzzy Set Theory) sets the platform for matrix theory by
introducing set

theory. The aspects of binary set theory involving set operations and

algebraic properties of sets are explained which in essence, from the prerequisites for the
similar understanding of fuzzy sets. Some examples have also been shared to understand
fuzzy set theory.
Chapter 2 (Fuzzy Matrix Theory) consists of fuzzy matrix theory involving the types
and equality of matrices. Moreover, operations on these matrices including transpose of
these matrices are explained.
Chapter 3 (The Determinant Theory of a Square Fuzzy Matrix) contains the
determinant theory of square fuzzy matrix along with their properties including another
statement, illustration, corollary parts and remarks, wherever observed. Moreover, some
theorems for fuzzy matrices is considered.
Chapter 4 (The Adjoint Theory of a Square Fuzzy Matrix) deals with the adjoint
theory of square fuzzy matrix with their properties including corollary parts and remarks,
wherever observed. Moreover, some special properties of square fuzzy matrices such as
symmetry, reflexivity, transitivity, circularity and idempotence are dealt with and carried
over to the adjoint matrix, which can be easily understand with the help of give
n
examples.
Chapter 5 (Super and Super Fuzzy Matrix Theory) concerns with supermatrix and
fuzzy supermatrix theory. Moreover operations defined on these matrices including
transpose and then two types of products are also explained. Most of the chapter is tried
to develop and explain with the help of examples for the sake of simplification.
6.3 Conclusion
The work being presented in the thesis is devoted to The Study of Fuzzy, Super and
Super Fuzzy Matrix Theory. The study being comparative shows vividly the difference
between fuzzy, super and super fuzzy matrices. Moreover operations defined on fuzzy,
super and super fuzzy matrices are involved which shows the comparative working of
these matrix theories. This comparative study also includes the determinant and adjoint

theory of square fuzzy matrix. Only those operations on super fuzzy matrices ar
e
provided which are essential for developing super fuzzy multi expert models.
The Comparative study done in this illustration is complete in itself but not complete
in all regards. Beyond the thesis work, we can describe simple
fuzzy

fuzzy matrix model,

cognitive maps, bidirectional associative memories model, fuuzy associative

memories model and comparatively new super fuzzy relational maps model, new super
fuzzy bidirectional and associative memories models can be introduced which will be
highly useful to social scientists who wish to work in future with multi expert models to
analyze their problems. All in all, this will certainly be a boon not only to social scientists
but also to mathematicians, engineers, doctors, researchers and students. Its influence is
certain to grow more in the decade to come.

BIBLIOGRAPHY
1.

Dubois D. and Prade H., Fuzzy Sets and Systems: Theory and Appliccation,
Academic Press, New York (1986).

2.

Horst P., Matrix Alegbra for Social Scientists, Holt, Rinehart and Winston, inc.
(1963).

3.

Kandasamy W.B.V., Smarandache F. and llanthenral K., Elementary Fuzzy

Matrix Theory and Fuzzy Models for Social Scientists, Los Angeles (2007).

4.

Kandasamy W.B.V., Smarandache F. and Amal K., Super Fuzzy Matrices and
Super Fuzzy Models for Social Scientists, Ann Arbor (2008).

5.

Kaufmann A., Introduction to the Theory of Fuzzy Subsets, Vol. 1, Academic

press, New York (1975).

6.

Kaufmann A. and Gupta M.M., Introduction to Fuzzy Arithmetic Theory and

Applications, Van Nostrand Reinhold, New York (1988).

7.

Kaufmann A. and Gupta M.M., Fuzzy Mathematical Models in Engineering and

Management Science, North-Holland (1988).

8.

Kim J.B., Idempotents and Inverses in Fuzzy Matrices, Malaysian Math 6(2)
(1983) 5761.

9.

Kim J.B., Inverses of Boolean Matrices, Bull. Inst. Math. Acod. Sinica 12(2)
(1984) 125128.

10.

Kim J.B., Determinant Theory for Fuzzy and Boolean Matrices, Congressus
Numerantium, Utilitas Mathematica Pub. (1978) 273276.

11.

Kim J.B., Baartmans A. and Sahadin N.S., Determinant Theory for Fuzzy
Matrices, Fuzzy Sets and Systems 29 (1989) 349356.

12.

Ragab M.Z. and Emam E.G., The Determinant and Adjoint of a Square Fuzzy
matrix, Fuzzy Sets and Systems 61 (1994) 297307.

13.

Thomason M.G., Convergence of Power of Fuzzy matrix, J. of Math. Anal.

Appl. 57 (1977) 476486.

14.

Xin L.J., Controllable Fuzzy Matrices, Fuzzy Sets and Systems, 45 (1992),
313319.

15.
16.

Zadeh L.A., Fuzzy Sets as a Basis for Theory of Possibility, Fuzzy Sets and
Systems, Vol. 1, No. 1 (1978) 328.
Zimmermann H.J., Fuzzy Set Theory and Its Applications, Kluwer Nijhoff
Publishing, Dordrecht (1985).