Equivalence Relations
In working with mappings that are not one-to-one, it is very useful to introduce the notion of an equivalence relation. The basic idea of an equivalence relation is to collect together elements that are in some way related, and then treat them as a single entity. The main goals of this topic are to show that every equivalence relation on a set can be realized as a partition of the set, and conversely; and that every equivalence relation can be defined in a natural way by a mapping.
Definition (Equivalence Relation) A relation
on a nonempty set
is an equivalence relation on
if it satisfies the following three properties:
(i) (Reflexive) If
then
![]()
(ii) (Symmetric) If
and
then
![]()
(iii) (Transitive) If
and
and
then
![]()
Definition (Equivalence Class) Let
be an equivalence relation on a set
and
Then the equivalence class of
is
The set of all equivalence classes of
under
is denoted by
and the set consisting of exactly one element from each equivalence class from
is called a complete set of equivalence class representatives.
Example (Equivalence Relation)
(i) Let
be a permutation group on
and define a relation
on
by
if and only if
for some
Then
is an equivalence relation on
(ii) Let
be a positive integer. For integers
we define
if and only if
Then
is an equivalence relation on
and is denoted instead by
The set of equivalence classes is denoted
![]()
(iii) Let
be the set of all integers and let
be the set of all nonzero integers. On the set
of ordered pairs, define
if and only if
Then
is an equivalence relation and the set of equivalence classes is
![]()
(iv) Let two complex numbers be equivalent if their real parts are equal. Then this is an equivalence relation and the set of equivalence classes is
![]()
(v) Let
be a group and define
if and only if
or
Then
is an equivalence relation on
![]()
(vi) Let
be a group and define
if and only if
for some
Then
is an equivalence relation on
and
and
are called conjugates.
(vii) Let
be a set and let
For subsets
and
of
define
if and only if
Then
is an equivalence relation on
▪
Definition (Partition) A collection
of nonempty subsets of a nonempty set
forms a partition of
provided
(i)
and
(ii)
![]()
Proposition (Equivalence Relations and Partitions)
(i) Any partition of a set determines an equivalence relation.
(ii) Any equivalence relation on a set determines a partition.
(iii) There is a one-to-one correspondence between the equivalence relations on a set and the partitions of the set.
Proof. (i): Assume that we are given a partition
of the set
Then the given partition yields an equivalence relation on
by defining
if there is some element in the partition
to which both
and
belong. To prove that this is an equivalence relation first let
then
because
belongs to one of the elements of the partition
To show symmetric law: let
and
Then there is a element of the partition in which both
and b both belong; and hence
To show transitivity, let
and
Then
and
belong to the same subset say
and
and
belong to the same subset say
where
Since
is a partition
Whence,
and so
is an equivalence relation on
![]()
(ii): Let
be the set of all equivalence classes of
on a nonempty set
We will show that
is a set of mutually disjoint subsets of
whose union is
Suppose
Then
and
for some
Since
no member of
is empty, and the union of the members of
is
Suppose that
Then
and
Thus,
and in fact
To see this, let
then
and hence
Thus,
so that
Similiarily,
and so
Therefore, the sets in
are mutually disjoint, and
is a partition of
as desired.
(iii): Beginning with an equivalence relation on a set
(ii) shows that the equivalence class defines a partition on the set and (i) shows that this partition determines the original equivalence relation
On the other hand, if a partition on a set is given, then the equivalence classes of the corresponding equivalence relation are the subsets in the original partition. Thus, there is a natural one-to-one correspondence between the equivalence relations on a set and the partitions of the set.
Definition (Natural Projection) Let
be any set and let
be an equivalence relation on
Then the mapping
defined by
for all
is called the natural projection from
to
For any
we have
if and only if
Thus these are the same equivalence relations, that is;
and
are the same. Therefore, every equivalence relation can be realized as the equivalence relation defined in a natural way by a mapping. Notice that the natural projection is an onto mapping.
Proposition (Mapping Factorization) If
is any mapping, and
is the equivalence relation on
by letting
if and only if
for
then the mapping
defined by
for all
is bijective; and hence we have a factorization
where
is the inclusion mapping.
![equivalence relations _gr_161.gif]](pages/equivalence-relations/Images/equivalence-relations_gr_161.gif)
Proof. The mapping
is well-defined since if
and
belong to the same equivalence class in
then by definition of the equivalence relation we must have
In addition,
is onto since if
then
for some
and thus we have
Finally,
is one-to-one since if
then
and so by definition
which implies that
Thus we have a one-to-one correspondence as desired; and we have a factorization of
into a composition of better-behaved mappings since
where
is the inclusion mapping, as shown.
Equivalence Relations
Published by Library of Math -- Online math organized by subject into topics.
Written by Smith, David A.
http://www.libraryofmath.com/equivalence-relations.html


