2 RELATIONS

2.1. INTRODUCTION

In the
previous chapter, we have discussed various
setsfrom given sets. We now operations on sets to generato.
products of sets which will helpdiscuss
one more
us in
property of sets which is known as eat
understanding the concept of relations. rtesianw
2.2. ORDERED PAR

Let A amd B be
where any two sets. Then by
xe A. ye B. an ordered pair of elements, we mean a
For pair (t
exanmple, the ordered
pairs (1, 1), (2, 3), (3, 5)
2.3. represent different points in a plane
CARTESIAN PRODUCT OF SETSs
Let A and B be
any two
is the set of non-empty sets.
all
ordered pairs («, y) such that xeThen the cartesian
A and product of the sets A and 5
AxB= {(x, y): x yeB and it is denoted by
For example, e Aand y e B}. A xB. Thus
consider A =(1, 2), B =(3, 4, 5). We
Here AxB {(1, 3), (1, 4), (1,
= find AxB, Bx A, Ax A, B XB.
5), (2, 3), (2, 4), (2,
BxA ={3, 1), (8, 2), (4, 1), 5)}
(4, 2), (5, 1), (5, 2)}
AxA={(1, 1), (1, 2), (2, 1), (2,
BxB={3, 3), 2)}
(8, 4), (8, 5), (4,
2.4. RELATION (or Binary 3), (4, 4), (4, 5), (5, 3), (5, 4), (0:
Relation)
Let A and B be
any two
from A to B is a subset of
A x B. non-empty sets. Then
R is relation from Ato
a
Thus, biunary relationa
or ply aa
simply relalion
j

Notation. If R is a relationB from

iffRCAxB.
we write xRy (read as t is a
related to y non-empty set A to a
by the relation
) E R

R). non-emnty set B and iI

50
RELATIONS

51
2.5. TOTAL NUMBER OF RELATIONS
A
Let and B be two
Then the number non-empty finite sets containing and n elements respeely
m
of ordered pairs in Ax B=
is 2mn mu, Therefore, total number or ubo
from AtoB
A 2mn (inel
is 2"
toB1s includingAB
the
defines a relation from A to B, so total number of relations
empty relation o and the universal relation A D:
Similarly, if a set A *

contains nt elements, then the

total number of relations on A 1sS
2.6. DOMAIN AND 2
RANGE OF A RELATION
Let R be a relation from the set A
co-ordinates of the ordered to the set B,
Then. the domain of Ris the set of all
second co-ordinates of the pairs belonging to R. Similarly, the first
orderedpairs belonging to R. For range of R is the set ot all
from Consider
A to B.
A={1,3, 5, 7}, B= {2, 4,6, 8, 10} e.g.
Then and R =(1, 8), (3, 6). (5, 2), (1, 4) be a relationn
Domain of R = {1,
3, 5}
Range of R ={2, 4, 6, 8.
2.7. INVERSE RELATION
Let R be a relation from
a set A to

denoted by R-1. Thus, ifanother

from B to A. It is set B. Then, the
inverse relation is a
R {(a, b): a e A, be B},
=

Then relation
Also,
R-1 {(b, a): be B, ae A}
Domain of R-1 Range of R =

Range ofR1 Domain of R.=

2.8. TYPES OF RELATIONS

The following are the types of relations.
(aEmpty void relation. Let A be a set. Then
or
¢ c Ax A is true
relation on A. This relation is called void relation or and hence it is a
empty relation.
(b) Universal relation. Let Abe any set. Then Ax
relation on A. This relation is called universal relation AcAxA is true and hence it is a
on the set A.
Remark. The void relation is the smallest relation on A and the
relation on A. universal relation is the
largest
(c) Identity relation. Let A be any set. Then
called the identity relation or diagonal relation on A and the relation R {t, x): x e A} on A is
=

is denoted by A,or A
For example, consider A {a, b, c} and define relations R, and R, as follows,
=

R {(a, a) (6, 6), (¢, c)}

=

R {(a, a), 6, b), (c, c), (a, c)}

Then R. is an identity relation on A, but R, 1s not an identity relation on A as the
element a is related to a and c.

2.9. PROPERTIES OF RELATIONS

(a) Reflexive relation. A relation R on a set A1s saidtobea reflexive relation if every
element of A is related to itself. Thus, R is retlexive iff (, 1) E R for all re A ery
A relation R on a set A is not reflexive if there is an element x E A such that (t, ) e R.
 52 DISCRETE STRUCTURE
For example, consider A = (1, 2, 3). Then the relation R, defined by R, = {(1, 1). (2,

8, is a reflexive relation on A.
3), (1. 3). (2, 1)}
The relation R, defined by R, = {(1, 1), (3, 3), (2, 1). (3, 2)} is not a reflexive relation on A

sinee (2, 2) e R,.

The identity relation and uninersal relation on a non-empty set is a reflexite relation
b) Irreflexive relation. A relation R on a set A is said to be irreflexive if(a, )e Ry
aE A. In other words, a relation R on a set A is not irreflexive if there exists at least one a eA

such that (a, a) e R.

the relation of
Example. Consider R {(a, b) : a # b} on the set A {a, b} i.e., R is
= =

R.
inequality on the set A, thus R is irreflexive since for every a E A, (a, a)£
Example. non-empty set and R be a relation
Let A be a on A ={a:ais a computer
science student) defined by R={(a, b): a scored less than b}
Then for all a e A, (a, a) e R i.e., R is an irreflexive relation.
reflexive relation, but not conversely.
Every identity relation on a non-empty set Ais
a
Remark.
b), (c, c), (a, b)}. Then R is a reflexive
{a, b, c} and define a relation R by R {a, a), (b,
=
Consider A =

the element (a, b) in R.

relation on A but not an identity relation on A due to
relation R on the set A is called
A be any set. Then a
(c) Symmetric relation: Let
symmetric iff
A.
(x, ) ¬ R » 0, 1) e R for allx, ye set A are
relation and universal relation on a non-empty
For example, the identity
A
symmetric relations on

Consider A =(1, 2, 3, 4) and define relations R, and R, as follows

1)}
R, {(1, 3), (1, 4), (3, 1), (2, 2), (4,
=

R={(1, 1), (2, 2), (3, 3), (1, 3)}

(1, 3) e R ( 3 , 1) e R
In R
(1. 4) E R, 4, 1) e R
(2, 2) e R, (2, 2) e R
symmetric relation on A.
R, is a
Thus, the relation relation.
E R2 (3, 1) e R. Thus R, is not a symmetric
In R,, (1, 3) iff
to be transitive
relation: Let A be any set. Then a relation R is said
(d) Transitive
(d) z E A.
and , 2)e R= (r, 2) e R for all x, y,
(a, y) E R relation on any
the identity relation and universal relation are transitive
For example,
non-empty ser. on A.
lines in a plane. Define a relation "is parallel to'"
ConsiderA=The set of all straight that , 15

Take . lg. lz E A. Then l, is parallel l, and l, is parallel to l. It implies

to
relation.
the relation "is parallel to" is a transitive
parallel to l also. Hence Let A be any
set. A

Asymmetrie o r Non-symmetric relation:

(e) Antisymmetrie o r
iff
relation R on A is said to be a antisymmetric
A
(a, ) E R and y, x) e R>x=y for all x, ye
 53
RELATIONS

whenever xRy and yk,Then x=y,

or
Arelanon k on a
non-emptv set A is Antisvmmetric if
ie
ifwhenever (x, y), (y, ) E R then x =
On the other hand, a relation R on a non-empty set A is not antisyme
exists , E A Such that (r, y) and (y, x) E R but x+ y.
since w
Example. The relation"
asb andb Sa, then a = b.
(read as less than or equal to) is antisymmetric

2.10. EQUIVALENCE RELATION (P.TU.B.Tech. May 2013, May 2012, May 2010, Dec. 2007)
set. Then A iff it
LetA be any
satisfies the following
a relation R on A is said to be an equivalence relation on

) It is reflexive
(ii) It is symmetric
(ii) It is transitive.
Theorem I. IfR a d S are tuwo equivalence relations on a setA, then
() RoS is also an equirvalence
relation o A. P.T.U. M.C.A. Dec. 2006)
(6) RUS may or may not be an equivalence relation on A.
(P.T.U. B.Tech. Dec. 2013; P.T.U. M.C.A. Déc. 2006)
Proof. Given R is a relation on A. It means R is a subset of Ax A. It implies
RCAxA ... (1)
Also, SCAXA ...(2
(1) and (2) gives R o ScAxAie., RnS is a subset of A x A. Hence R nS is a relation
on A.
We show RnS is an equivalencerelation on A.
() Reflexivity. Let xe A be any arbitrary element of A.
As R and S are equivalence relations on A. It implies R and S are reflexive relations on A.

IfxEA= (r, x) e R and (x, x) E S.

x, x) e RnS for allx e A.
Hence RnS is reflexive on A.
that (r, y) RoS. It means (, ) e Rand (x, y) E S.
(ii) Symmetry. Let x, ye Asuch
e

symmetric relations on A.
As R and S are

o, x) e R and ), 1) ES
y, a) e RnS.
Hence (x, y) E RnS 0, 1) ¬ RnS for all x, y E A.
Therefore R n S is symmetric on A. Rn S.
z e A such
that (x, y) e RnS, y, 2) ¬

(ii) Transitivity., Let x, y,

We show z) ¬ Rn S.
(x, R and (v, 2) e S.
R and (, ) e S. Also, (v, z) E

Now, (x, y) e
on A
As R, S are transitive

, y) E R, V, z) e R (r, 2) ¬R
(r, 2) E S
(x, y) E S, (y, z) S
¬
Also,
(x, z) E Rn S. relation on A.
Sis an equivalence relation.
Hence Rn
or may not be an equivalence
show RUS may relations R andS by
Further, we Define the
A= {a, b, c}. (6, a)}
Consider b), (¢, c), (a, b),
R={(a, a), (6, b)}
S {(a, a), (6, b), (c, c), (6, c), (c,
 DISCRETE STR
54 UCTURES
relations on A.
Clearly. R and S a r e equivalence (c, c), (a, b), (6, a), (6, c), (c, b)}
a), (b, b,),
But RUS={(a, E RUS.
Here (a, b) e R US, (6, c)
RUSis not transitive.
But (a,c)e RUS A.
equivalence relation on
Hence RUS cannot be
2.11. COMPATIBLE RELATION

which is reflexive and symmetric is called a compatibla r o l

A binary relation on a set
tion. Every equivalence relation is a compatible relation. But every compatible relation may

not be an equivalence relation.

Let A is the set ofpersons and define R on A by R such that for a, b e A, (a, b) E Rifaisis
a friend of b. then R is reflexive and symmetric. Hence R is compatible.
Let a, b. ce A such that (a, b) e R, (6, c) E R. Here, a is a friend ofb, b is a friend ofc but
a may or may not be a friend of c. Hence R may or may not be transitive.
Example. Let A = {t: xis a English word}. Let R be a relation on A such that
R=
{(a, b): a, be A and a and b has one or more letters common}. We
show R is a compatible relation.
Reflexive. Every English word has letters same as itself. Therefore R is reflexive.
(ii) Symmetric. If a has one or more letters same as that of b, then b also has one or
more letters same as a. Hence R is symmetric.
Therefore, R is a compatible relation. But R is not transitive.
Take b= kite, c = ten (any English
a=book, word).
Here (a, b) = (book, kite) e R (as book and kite has'k common)
(b, c) = (kite, ten) e R. But (a, c) = (book, ten) e R.
. R is not an equivalence relation.

2.12. PARTIAL ORDER RELATION P.T.U. B. Tech. Dec. 2013, May 2006)
A relation R on a set A is called a partial relation if it is
) reflexive
(ii) anti-symmetrie
(iii) transitive.
For example, define a relation 'C on A. We show the relation ' is a partial order
relaticn on A.
) Let B e P(A). Then BcB is true. Therefore, ' i s reflexive.
(ii) Let B,, B, e P(A) and if B, SB,, B,cB,. Then B, = B, is true.
' S is anti-symmetric.
(iii) Let B, B, B E P(A) and if B, c B, B, c B,. then B, cB, is true.
Hence the relation 'C is transitive.
The relation 'C is a partial order relation.
2.13. PRODUCT OF SETSs
Let A, Az.., A,, are n sets. Then the set of all ordercd n-tuples («1, ag
E A, u E A,.., a, E
A is called the product of the
..,
a w A
sels A, Ag, . . .

A,, and is denoted 5y Ai

xAg X.. xA, or | A.
i=1
 55
RELATIONs

Hence, we write
AxA=A2
Ax Ax A = A3

**************:***

Ax Ax... A = A"

2.14. TERNARY RELATION

Let S be any set. Then, a subset of S x Sx S is called a ternary relation an S. e g
Consider A = [1, 2, 3, 4,.. 15]. Let R be the ternary relation on A defined by the equau
+5y= z. We write R as a set of ordered pairs.
3
Forx> 3, x> 15. Hence, equation x2 + 5y=z
we find only solutions of the for
For x= 1, we have 1+ 5y = z and the values of y, z are 1, 6; 2, 11.
For x = 2, we have 4 + 5y = z and the values of y, z are 1, 9; 2, 14

For = 3, we have 9+ 5y = z and the values of y, z are 1, 14. Hence

R={(1, 1, 6), (1, 2, 11), (2, 1, 9), (2, 2, 14), (3, 1, 14)}.
Example 9. Shou, how a binary operation, say, addition +), may be vieued as a teriuary
relation.
Sol. Let the binary operation + may be defined as a set of ordered triples as follows:
+ {(x, y, 2): xty=2}
+ defined above is relation. For e.g., (2, 5, 7) E t, but
Then, the relation a ternary
(2,4, 8) E+
Example. Let A = {1, 2, 3,... 15. Let R be the 4-ary relation on A defined by
R = {(r, y, z, t): 4x +3y +z = t}. Write R as a set of 4-tuples.
Sol. Let (x, y, z, ) e R such that
4x+3y +z=,.
Then x can assume the values 1, 2 and 3 only for all y, z, te A.
Thus, R={(1, 1, 1, 8), (1, 1, 2, 11), (1, 2, 1, 11), (1, 2, 2, 14), (1, 3, 1, 14),
(2, 1, 1, 12), (2, 1, 2, 15), (2, 2, 1, 15)}.

2.15. CLOSURE PROPERTIES OF RELATIONS

the relation R does not satisfy the
Consider a relation R on a given set A. Suppose
relation R, if R satisfies
the least number of ordered pairs to
desired property. After addingdesired of relation R.
he desired property, then the property is called closure
is called reflexive closure of the relation R ifR, is
(a) Reflexive closure. A relation R reflexive property.
he smallest relation containing R having the
{7, 8, 10. Define a relation by
R
For example, consider A
=

is not reflexive. We find the reflexive closure of R.

R={(7; 8), (7, 10), (8, 8), (10, 7)}. Here R We add the ordered pairs (7, 7),
if (x, 1) E R for all x e A.
The relation R is reflexive closure of R is given by
, 1 0 ) inR. The required reflexive
8), (7, 10), (8, 8), (10, 7)}
R {(7, 7), (10, 10), (7, is called symmetric closure of R if Ra is the
() Symmetric closure. A relation Rs The smallest symmetric relation
Smallest relation containing R having the symmetric property.
Containing R is Rg = RUR,
5, 6;. Define a relation R by R =
{(4, 5), (5, 5), (5, 6)}.
For example, consider A ={4,
e find the symmetric closure of R.
 56
CTURE
ere
R'
{(6, 1), (5, 5), (G, 5)}
The required symmetric closure of R
is given by
R RuR' ={(1, 5), (5, 5), (6, 6), (5, 4), (6, 5)}
() relation R, is called transitive
Transitive closure. A the
closure RififR,
of R R, is the.
the small.
est relation having
containing R transitive property.
For eample, consider A =
{4, 6, 8, 10} and define a relation R on A by
R= {(4, 4), (4, 10), (6, 6), (6, 8), (8, 10)}. We find the transitive closureoe
Here (6, 8) e R, (8, 10) e R, but (6, 10) e R. So we add the ordered pair (6, 10) to R
Hence the required transitive closure of R is given by
R,(4, 4), (4, 10), (6, 6). (6, 8), (8, 10) (6, 10)%.
2.16. COMPOSITION OF RELATIONS (P.T.U. B. Tech. May 2007
Let R and S be two relations from sets A to B and B to C respectively. Then a relatiom
ation
RoS is called composite relation from A to C where (a, c)e RoS iffwe can find be Bsuch that
(a, b) e R and (b, c) e S.
The relation RoS is read as composition of R and S.

Let A, B. C be any three sets. Let R be a relation from A to B and S be a relation from B
to C. ie., R is a subset of A x B and S is a subset of B x C. Then R and S give rise to a relation
from A to C denoted by RoS and defined by
a(RoS)e if for some be B, we have aRb and bSc. Thus,
RoS {(a, c): There exists b e B for which (a, b) e R and (b, e) e S}
A= (1, 2, 3), B = (a, b, e), C = (x, y, 2).
e.g, Let
Consider the following relation: R from A to B and S from B to C, given by
R {(1, b), (2, ), (2, e)};S={«, ), (6, ), (C, ), (c, 2)}
We find the composition relation RoS
Draw the arrow diagram of R and S as shown Fig. 2.1. There is an arrow from 1 to b
which is followed by an arrow from b to x, Thus 1(RoS)x or (1, )e RoS.
Similarly, 2, ), 2, 2) E RoS. No
other pairs belong to RoS. Thus, RoS =
{(1, ), C,
(2,

Fig. 2.1
Remark. In genoral RoS SoR. Also, (RoS)-l S-loR-1 =

Theorem II. Let A, B and C be three sets and R be a relation from A to B, S are , S a r e l a t i o n

from B to C. Then (RoS)-l = S-'oR-1

Proof. Let a e A, be B, ce C. Then
(a, b) e R, (b, c) e S
(a, c) e RoS
But
c, a) e (RoS)-1
(b, c) e S
Also
(c, (c, b)eS-1,
(a, b) e
R » (b, a) E R-l » (c, a) e S-1oR-1
RELATIONS
57
From (1) and (2), (Ros)1 cs-loR-1
and SoR-1 c (RoS)-1
Hence (RoS)-1 =SoR-1.
2.17. DIRECTED GRAPH OR DIGRAPH OF A
RELATIONN
Consider a relation R on A {1, 2, 3, defined
=
4} by
R {(1, 2), (2, 2), (2, 4), (3, 2), (3, 4), (4, 1), (4, 3]
To find the directed
graph of R, write down the elements of A.
Draw an arrow from an element x to
an element y whenever
(,y)E R. The Fig. 2.2 so obtained is known as directed
relation R. graph for the
We emphasize that a directed graph is not defined for
from a set to another set. a relation Fig. 2.2 Directed
graph of R
 DISCKEICOIRUCTUR
IRUCTUREs
78
PARTITIONS
2.18. EQUIVALENCE RELATIONS AND
fA,
Let S be non-empty set. By partition
P of S, we mean
a finite collection
Ajo
non-empty subsets of S such that, disjoint
S are mutually
AA=¢Vi*jie, the subsets A/s of
U A =s
P of S is a subdivision
of S into disjoint non-empty sets
n words,
other partition
a
S, let la denotes the set
relation on S. For eachae
Further, suppose that R be related under R. Thus,
an equivalence
of elements of S to which a is
lal={r e S: (a, ) e R} or
{re S: aRr}
=

The set [a] is called equualence class if a in S.

an equivalence relation
The collection of all equivalence classes of elements of S under
R is denoted by S/R. Thus,
S/R = {a]: ae S} where
S/R (to be read as the quotient set of S by R) is called the quotient set.
Theorem. Let R be an equivalence relation on a set S and let S/R denotes the quotient
set of S by R. Then, the quotient set S/R is a partition of S. Also,
) ae [alVae S
) la] {b] iff (a, b) E R
=

(i) Iflol b}, then (a] amd b] are

disjoint.
Proof () Since R is an equivalence relation on S, it must be reflexive. Thus, for every
aE S, (a, a) e R ae la | by definition of lal
(i) Let (a, b) ¬ R and we show [a) = [b]

Let xe [b] ( b , )e R. Also (a, b) E R and since R 1s an equivalence relation on S, it

must be transitive.
Hence (a, b) e R, (6, ) ¬R » (a, x) e R

To show la)c [b], let (a, b) ¬ R ...(1)

(b, a)e R
ae [b] = lalc [b] Ris symmetric
Hence, combining (1) and (2), we have ..(2)
al = [b]

Conversely, if lal [b], then by using part (),

be [b]=[a) be la
we have
(a, b) e R
(ii) We prove that if [a]n [b] *0, then
al = [b]

3 element * e S such that

Since [a] o b]#¢ = an

*E l a 6 ] *¬ la} and x e [b].

Now x e [a] (a, x) e R
Similarly, (b, ) e R
RELATIONNS

79
But R is symmetric, therefore.
(b, 1) e R (x, b) e R
Aaain, (a, x) E R, (*, b) ER and R is
transitive, we must have
(a, b) e R
la]= [b] I Using Part (i)