Lebanese University

Lebanese University Faculty of Sciences

Faculty of Sciences

M1102
Algebra II
Department of Mathematics

Spring Semester
2019 - 2020

©opyright Reserved
Preface

This book is designed to be compatible with the content of the course M1102
of the new curriculum at the faculty of sciences of the Lebanese University.

Ideas are presented in a clear style and the exercises were chosen delicately
to deepen the understanding of the required concepts of each semester.

It is a ring in the chain of mathematical subjects required in our curriculum

performed according to the new vision.

We hope the students find this text convenient with their understanding skills
and enable them to go around elegantly between the ideas and the proofs
presented.

i
Contents

Preface i

1 Binary relations 1
1.1 Definitions and Notations . . . . . . . . . . . . . . . . . . . . 1
1.2 Properties . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 2
1.3 Order Relations . . . . . . . . . . . . . . . . . . . . . . . . . . 3
1.4 Equivalence Relations . . . . . . . . . . . . . . . . . . . . . . . 5
Exercises and problems . . . . . . . . . . . . . . . . . . . . . . . . . 9

2 Relations and Graphs 15

2.1 Symmetric Relations . . . . . . . . . . . . . . . . . . . . . . . 15
2.2 Bipartite graph . . . . . . . . . . . . . . . . . . . . . . . . . . 23
2.3 Digraphs . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 25
Exercises and problems . . . . . . . . . . . . . . . . . . . . . . . . . 28

3 Complement on the Natural Integers and the Cardinals of

Countable Sets 31
3.1 The Construction of N . . . . . . . . . . . . . . . . . . . . . . 31
3.2 Cardinal of a Finite Set . . . . . . . . . . . . . . . . . . . . . 33
3.3 Countable sets . . . . . . . . . . . . . . . . . . . . . . . . . . . 35
3.4 Combinatorial Analysis . . . . . . . . . . . . . . . . . . . . . . 39
Exercises and problems . . . . . . . . . . . . . . . . . . . . . . . . . 41

iii
Chapter 1

Binary relations

1.1 Definitions and Notations

After defining sets, collections of objects, it is important to study the kinds of
binary relations which can exist between the elements of a set. So we are led
to introduce an interesting notion in mathematics, the relations, which are
the basic elements of quotient structures, in particular, those defined from Z,
and are strongly used in arithmetic and in other applications1 . The binary
relations between the elements of a set E and the elements of a set F may
be viewed as properties defined on the set E × F . We are limited in this
chapter on binary relations defined on one set.

Definition 1.1.1. Let E be a non empty set and let (p) be a property defined
on the set E × E. (p) defines on E a binary relation R as follows: For every
x and y in E, we say that x has a relation R with y, we write xRy, if and
only if (x, y) verifies (p) in E × E.
xRy ⇔ (x, y) verifies (p).
The subset of E × E, {(x, y); (x, y) verifies (p)}, denoted by GR , is said to be
the graph of R. A more precise definition and details of graph are presented
in chapter 2. We have:
xRy ⇔ (x, y) ∈ GR
Remark 1.1.1. Two binary relations R and S defined on a set E are said
to be equal, we write R = S, if they have the same graph. We have in this
case the following equivalence:
xRy ⇔ xSy ∀x, y ∈ E.
1
We use quotient structure of Z in the creation of advanced cryptographic systems.
2 CHAPTER 1. BINARY RELATIONS

A relation R defined on a set E is clearly defined on any part A of E. This

is the restriction of R on A. The negation of R, denoted by R, is the relation
defined on E by:

xRy ⇔ (x, y) 6∈ GR for every x, y ∈ E.

Example 1.1.1. 1. The equality (or the identity) defined trivially on a

set. (This relation is not always obviously verified!)
s r
√ √ q
√
q √ 
2+ 3 = 5 + 2 6, 2 − 3 − x + 2x + 1 + x4 = 0, x ≤ 0.

√
2. The usual order defined on R. π ≤ 4, 2 > 1.

3. Let p ∈ Z. We define on Z the binary relation Rp by:

xRp y ⇔ p divides x − y.

This is the congruence relation modulo p.

4. Let f be a mapping defined from a set E into a set F . We define on E

the binary relation Rf by:

∀x, y ∈ E; xRf y ⇔ f (x) = f (y).

1.2 Properties
Definition 1.2.1. A relation R defined on a set E is said to be reflexive if
xRx for every element x in E.

Example 1.2.1. The equality, the relations Rp and Rf defined in the above
examples are reflexive. The relation < (less than strictly) defined on R is not
reflexive. It is a relation R verifying xRx ∀x ∈ R. Such a relation is said to
be antireflexive2 .

Exercise 1.2.1. R is a relation defined on a set E. We define the subset F

of E formed by the elements y ∈ E such that the elements x ∈ E having
relation with y are the elements verifying xRx. Show that F = ∅.

Definition 1.2.2. A relation R defined on a set E is said to be symmetric

if for every elements x and y of E, we have xRy ⇒ yRx.
2
Do not confuse between a non reflexive relation and an antireflexive one.
1.3. ORDER RELATIONS 3

Example 1.2.2. The equality, the relation Rp and Rf are all symmetric. ≤
is not a symmetric relation.
Definition 1.2.3. A relation R defined on a set E is said to be antisymmetric
if for every elements x and y of E, we have: (xRy and yRx) ⇒ x = y
Example 1.2.3. ≤ is an antisymmetric relation.
Remark 1.2.1. Note that a relation can be symmetric and antisymmetric
at the same time. Also, it can be neither symmetric nor antisymmetric at
the same time.
Definition 1.2.4. A relation R defined on a set E is said to be transitive if
for every elements x, y and z of E, we have: xRy and yRz ⇒ xRz.
Remark 1.2.2. It is important to remark that transitivity does not give
the sense of circulation, that is, we don’t have: zRx instead of xRz in the
definition. If we call a circular relation any relation R verifying xRy and
yRz ⇒ zRx, we get the following fact:
Exercise 1.2.2. Let R be a reflexive relation. Then R is circular if and only
if R is symmetric and transitive.
We treat in the sequel interesting families of relations, the order relations
and the equivalence relations.

1.3 Order Relations

Definition 1.3.1. A relationR defined on a set E is said to be an order
relation if R is reflexive, antisymmetric and transitive.
Example 1.3.1. The relation ≤ defined on R, the relation “division” defined
on N, the inclusion ⊆ defined on the set of parts of a set E, are all order
relations.
To simplify, an arbitrary order relation defined on a set is often denoted by
≤ if there is no any confusion. (Sometimes, we write x < y to say that x ≤ y
and x 6= y).

The couple (E, ≤) denotes a set on which we define an order relation≤, E is

a set ordered by ≤. Two elements x and y of E are said to be comparable by
≤ if we have x ≤ y or y ≤ x. If all the elements of E are pairwise comparable,
≤ is called a total order relation and E is totally ordered by ≤, otherwise,
≤ is called a partial order relation. We give essentially in this paragraph a
collection of notions strongly used in any branch of the order relations theory.
4 CHAPTER 1. BINARY RELATIONS

Definition 1.3.2. Let E be a set endowed with an order relation ≤ and let
a be an element of E.

(i) a is said to be maximal element (resp. minimal) in E if we have:

a ≤ x ⇒ a = x (resp. x ≤ a ⇒ a = x) for every x ∈ E.

(ii) a is said to be greatest element (resp. smallest element) if we have:

x ≤ a (resp. a ≤ x) for every x ∈ E.

Remark 1.3.1. A greatest element (resp. smallest element) of an ordered

set E when it exists, is unique.

Obviously, a greatest element is maximal and a smallest element is minimal.

The reciprocal is not true. These notions coincide in the case of a total
order. To avoid confusion between these notions in the general case, we give
the following example:

Example 1.3.2. Consider the set P (E) ordered by inclusion where E =

{1, 2, 3}.
The parts {1, 2}, {2, 3}, {1, 3} are the maximal elements in P (E)−
{E}, ⊆ .

Remark 1.3.2. (R, ≤) has neither maximal element nor minimal element.

Definition 1.3.3. Let E be a set endowed with an order relation ≤ and let
A be a part of E and a an element of E.

(i) a is said to be upper bound (resp. lower bound) of A in E if we have:

x ≤ a (resp. a ≤ x) for every x ∈ A.

(ii) a is said to be least upper bound (resp. greatest lower bound) if a is

the smallest upper bound (resp. greatest lower bound).

Remark 1.3.3. The least upper bound (resp. greatest lower bound) of a
part A is not necessarily an element of A. (]1, 2[ in (R, ≤).)

Definition 1.3.4. Let f be a mapping from an ordered set (E, ≤) into

another ordered set (F, ≤).

(i) f is said to be an increasing mapping if we have:

x ≤ y ⇒ f (x) ≤ f (y) ∀x, y ∈ E.

(ii) f is said to be a decreasing mapping if:

x ≤ y ⇒ f (y) ≤ f (x) ∀x, y ∈ E.

1.4. EQUIVALENCE RELATIONS 5

(iii) f is said to be isomorphism if f is bijective and if:

x ≤ y ⇔ f (x) ≤ f (y) ∀x, y ∈ E.

Two ordered sets are said to be isomorphic if we may define an isomorphism

between these two sets. It is possible in general to define many isomorphisms
between two isomorphic ordered sets. An isomorphism defines so a sort of
“superposition” between two given structures. This superposition is made
in a unique way in the case of a well order relation. The notion of the well
order relations is the first step towards the ordinals theory.

Definition 1.3.5. An order relation ≤ defined on a set E is said to be a

well order relation if any non empty part of E has a smallest element. We
say that E is well ordered.
Remark 1.3.4. Every well ordered relation is totally ordered. The converse
is not true.
Example 1.3.3. The usual order of N is a well order (See 3).
Exercise 1.3.1. Show that the unique isomorphism from (N, ≤) into (N, ≤)
is the identity.
Proposition 1.3.1. Let (E, ≤) and (F, ≤) be two isomorphic ordered sets.
We suppose that E is well ordered. Then there exists a unique isomorphism
from E into F .
Proof. Let f and g be two isomorphisms from E into F . We define A = {x ∈
E; f (x) 6= g(x)}. Suppose to the contrary that A is not empty. It admits
then a smallest element, set a. We have: f (a) 6= g(a). We suppose,without
loss of generality, that f (a) < g(a). g is surjective and f (a) ∈ F , then there
exists b ∈ E such that g(b) = f (a). We have g(b) = f (a) < g(a), then b < a
and then b 6∈ A, hence f (b) = g(b). We get f (b) = f (a), thus b = a since f
is injective. Consequently, A = ∅ and f = g.

1.4 Equivalence Relations

Definition 1.4.1. A relation R defined on a set E is said to be an equivalence
relation if R is reflexive, symmetric and transitive.
Exercise 1.4.1. Verify that Rp and Rf defined above, are equivalence rela-
tions.
6 CHAPTER 1. BINARY RELATIONS

A well illustration of the notion of an equivalence relation defined on a

set is that of a school. The relation “be in the same class” is an equivalence
relation defined on the set of students of this school. In fact, this relation
is clearly reflexive, symmetric. If in addition, we know that two students
a and b are in the same class and that b and c are in the same class also,
we may deduce (by a subtle reasoning) that a, b and c are all in the same
class, and then the relation is transitive. The class of a student a may be
defined as the set of all students having relation with a. If we choose two
arbitrary students, their classes are either equal or disjoint.By this relation,
the set of students is partitioned into classes, the set of classes will be called
the quotient set by the equivalence relation.To each student is associated
a unique class, this correspondence is called the canonical surjection. Re-
ciprocally, an arbitrary partition of the students into classes gives always
the definition of of the equivalence relation: “be in the same class”. These
obvious remarks are all what we can say on an equivalence relation in general.

Definition 1.4.2. Let R be an equivalence relation defined on a non empty

set E. For every x ∈ E, we define the equivalence class of x by:

x̄ = {y ∈ E; xRy}.

The set of the equivalence classes of the elements of E is said to be the

quotient set of E by R. It is denoted by E/R.

Remark 1.4.1. x ∈ x̄ for every x ∈ E then x̄ 6= ∅. For every y, z ∈ x̄, we

have: yRz. In fact, we have: yRx and xRz then yRz. Similarly, if xRy then
x̄ = ȳ. In fact, it is sufficient to show that x̄ ⊆ ȳ. Let x ∈ x̄, we have zRx,
but xRy, then zRy and z ∈ ȳ.

Exercise 1.4.2. Consider the relation Rf defined from a mapping f from E

into F . Let x ∈ E. Show that x̄ = f −1 (f (x)).

Proposition 1.4.1. Let R be an equivalence relation defined on a non empty

set E. For each x, y ∈ E, we have: x̄ = ȳ or x̄ ∩ ȳ = ∅.

Proof. Suppose that x̄ 6= ȳ. In this case, x̄ ∩ ȳ = ∅, since otherwise, there

exists z ∈ x̄ ∩ ȳ and we have: xRz and zRy, then xRy and x̄ = ȳ, a
contradiction.

Let I be a nonempty set. For all i ∈ I, we suppose that Fi is a well-

defined set. The collection of sets Fi , i ∈ I, is said to be sa family of sets
indexed by I and denoted by Fii∈I .
1.4. EQUIVALENCE RELATIONS 7

Definition 1.4.3. Let E be a set, and let (Fi )i∈I be a family of subsets of
E. (Fi )i∈I is said to be a partition of E if:
(a) Fi 6= ∅ for all i ∈ I.

(b) Fi ∩ Fj = ∅ for all i, j ∈ I.

(c) ∪i∈I Fi = E.
Corollary 1.4.1. Let R be an equivalence relation defined on a non empty
set E. The family {x̄}x∈E/R is a partition of E.
Proof. We may verify easily that this family satisfies the conditions of a
partition, using the above proposition and the fact that x ∈ x̄ for every
x ∈ E.
By the following proposition, we complete the proof of our remark that an
equivalence relation defined on a set E is exactly a partition defined on the
same set.
Proposition 1.4.2. Let E be a non empty set and let {Fi }i∈I be a partition
of E. There exists an equivalence relation R defined on E such that the
classes of R are exactly the members of the partition.
Proof. Let R be the relation defined on E by:

∀x, y ∈ E; xRy ⇔ ∃i ∈ I such that x, y ∈ Fi .

This relation is reflexive, symmetric. To verify that it is transitive, let x, y

and z be three elements of E such that xRy and yRz. ∃i, j ∈ I such that
x, y ∈ Fi and y, z ∈ Fj . Thus y ∈ Fi ∩ Fj and Fi ∩ Fj 6= ∅, so Fi = Fj .
Consequently, x, y, z ∈ Fi , and xRz. R is then an equivalence relation. Now,
let’s show that E/R = {Fi }i∈I . Let x̄ ∈ E/R, x ∈ E, then ∃i ∈ I such that
x ∈ Fi . We have x̄ = Fi . In fact it is sufficient to to prove that x̄ ⊆ Fi , for
this, let y ∈ x̄. We have xRy, ∃j ∈ I such that x, y ∈ Fj . Then, x ∈ Fi ∩ Fj
and Fi ∩ Fj 6= ∅, so Fi = Fj . Hence, y ∈ Fi and so x̄ ⊆ Fi . In the other hand,
for every i ∈ I, Fi 6= ∅, then ∃x ∈ Fi , we get x̄ = Fi .
We end this paragraph by using equivalence relation in the establishment of
a remarkable property, the canonical decomposition of a mapping.

Definition 1.4.4. Let R be an equivalence relation defined on a non empty

set E. The canonical surjection defined by R is the mapping p which asso-
ciates to each x ∈ E its class x̄ ∈ E/R.
8 CHAPTER 1. BINARY RELATIONS

Proposition 1.4.3. Let f be a mapping defined from a set E into a set F .

Then

ϕ : E/Rf →
7 f (E)
x̄ → 7 ϕ(x̄) = f (x)

is a bijective mapping. Where Rf is the equivalence relation associated to f .

Proof. ϕ is well defined injective mapping. In fact, let x̄, ȳ ∈ E/Rf . We
have: x̄ = ȳ ⇔ xRf y ⇔ f (x) = f (y) ⇔ ϕ(x̄) = ϕ(ȳ). It is surjective by
construction. Then ϕ is bijective.

Proposition 1.4.4. (Canonical decomposition) Let f be a mapping from

a set E into a set F . Then f = i ◦ ϕ ◦ p, where p is the canonical surjection,
ϕ is the bijection defined in the above proposition, and i is the canonical
injection from f (E) into E. Moreover, if f = i ◦ ϕ0 ◦ p, then ϕ = ϕ0 .
Proof. Left to the reader.
 9

Exercises and Problems

1. Let E be a non-empty set and let F be an ordered set. Let Ω be the
set of all mappings of E to F . Define on Ω the relation R by:

uRv ⇔ u(x) ≤ v(x), ∀x ∈ E.

(a) Show that R is an ordering relation on Ω.

(b) Give an example of E and F , such that F is totally ordered but
Ω is not totally ordered by R.

2. Let
2
A = {xn = + (−1)n ; n ∈ N∗ }.
n
Show that SupQ (A) = 2 and InfQ (A) = −1.

3. Does the set

2
B = {xn = ; n ∈ N∗ }
n
have a least upper bound or a greatest lower bound in Q∗ ?

4. Let E = N − {0, 1} and let R be the ordering defined on E by

xRy ⇔ x divides y.

Show that E has no maximal elements and that every prime number
is a minimal element of E.

5. Let f : E → F be a mapping of a set E to an ordered set F . Let R

be the relation defined in E by

xRy ⇔ f (x) ≤ f (y).

Show that:

(a) R is reflexive and transitive.

(b) R is an ordering relation on E if and only if f is injective.

6. Let f : E → F be a mapping from a set E to an ordered set F . Define

on F the relation R by

xRy ⇔ ∃a ∈ E, such that x ≤ f (a) and f (a) ≤ y.

(a) Show that R is antisymmetric and transitive.

10

(b) Prove that R is an ordering relation on F if and only if f is

surjective.

7. Consider a mapping f of a set E to a set F . Let R be the relation

defined on P (E) by

ARB ⇔ A = f −1 (f (B)).

(a) Show that R is reflexive if and only if f is injective.

(b) Show that if A, B ∈ P (E) and ARB, then B ⊆ A. Deduce that
R is antisymmetric.
(c) Show that if A, U ∈ P (E) and A = f −1 (U ), then ARA.
(d) Show that R is transitive.
(e) Deduce that if f is injective, then R is an ordering relation.

8. Let A be a set and E be a non-empty subset of P(A). Show that the

relation R defined on E by

XRY ⇔ there exists a bijection from Xonto Y

is an equivalence relation on E.

9. Let f : R → R be the mapping defined by

f (x) = 2x2 − 3.

Let Rf be the equivalence relation associated with f . Find the classes

0 and 1.

10. Let f : N → N be defined by f (x) = x2 − 2x + 3, ∀x ∈ N. Let R be the

equivalence relation associated with f .

(a) Find the classes 0 and 1 modulo R.

(b) Find a if a is a singleton.

11. Let E and F be two sets and let R (resp. S) be an equivalence relation
on E (resp. F ). Let p : E → E/R (resp. q : F → F/S) be the
canonical surjection of E (resp. F ) onto E/R (resp. F/R). Show that
if f : E → F is a mapping of E to F , then the following properties are
equivalent:

(a) There exists a mapping g : E/R → F/S, such that g ◦ p = q ◦ f .

 11

(b) The implication

xRy → f (x)Sf (y)
is true, ∀x, y ∈ E.
12. A relation R on a set E is said to be circular if the following implication
is true, ∀a, b, c ∈ E
[aRb and bRc] ⇒ cRa.
Show that if R is reflexive and circular, then R is an equivalence relation
on E. Is the converse true?
13. Let E be a non-empty set and let ∆ be the diagonal of E 2 , i.e

∆ = {(x, x); x ∈ E}.

For each subset A of E × E, set

A−1 = {(x, y); (y, x) ∈ A}

and

A ◦ A = {(x, y); ∃z ∈ E, such that (x, z) ∈ Aand (z, y) ∈ A}.

Let R be a binary relation on E and let G be the graph of R. Show

that
(a) If R is reflexive and transitive, then G ◦ G = G.
(b) R is an ordering relation on E if and only if

G ◦ G = G and G ∩ G−1 = ∆.

(c) R is an equivalence relation on E if and only if

G ◦ G = G, G = G−1 and P r1 (G) = E.

14. Let R be the relation defined in Q by

xRy ⇔ (x − y) ∈ Z.

Show that R is an equivalence relation on Q and find the class 0 modulo

R.
15. Let P be the set of all the points M (x, y) in the plane Oxy. Let R be
the relation defined in P by

M (x, y) R M 0 (x0 , y 0 ) ⇔ x2 + y 2 = x02 + y 02 .

12

(a) Show that R is an equivalence relation on P .

(b) Find the class modulo R of the point A(a, b) of the plane.

16. In R∗+ , we define the relation R by

xRy ⇔ xln(y) = yln(x).

(a) Show that R is an equivalence relation.

(b) Describe the quotient set.
(c) find the canonical decomposition of the following mapping:

f : R∗+ 7→ R
ln(x)
x 7→
x

17. A relation R defined on a set E is said to be connected if:

(a) ∀x ∈ E; ∃u ∈ E such that xRy.

(b) R is symmetric.
(c) ∀x, y, z ∈ E we have: xRy and yRz ⇒ zRx.

Show that R is an equivalence relation if and only if R is connected.

18. Let E be a set endowed with an equivalence relation R. Let f be a

mapping defined from E into E such that f (x̄) ⊆ x̄ for every x ∈ E.

(a) Show that xRf (x) for every x ∈ E.

¯
(b) Show that f (x̄) ⊆ f (x).
(c) Suppose that f is surjective.
i. Prove that f (x̄) = x̄.
ii. Deduce that the mapping

ϕ : E/R 7→ E/R
x̄ ¯
7→ ϕ(x̄) = f (x)

is bijective.

19. Let f and g be two mappings defined from a set E into a set F such
that f (x) 6= g(x) for every x ∈ E. We define on E the relation R by
xRy ⇔ f (x) = g(y).
 13

(a) A part L of E is said to be free if xRy for every x and y in L.

i. Let A ⊆ E. Show that A is free if and only if f (A)∩g(A) = ∅.
ii. Let z ∈ F and let A = f −1 (z). Show that A is a free part.
iii. Let x, y, z ∈ E be such that xRz and yRz. Show that {x, y}
is a free part.
(b) A part C of E is said to be complete if {x, y} is not free whenever
x and y are two distinct elements of C. Let A be a complete part
of E.
i. Let x, y ∈ A be such that xRy. Show that ∀z ∈ A, we have
yRz and zRx.
ii. Deduce that a complete part of E contains at most 3 elements.

20. Let (E, ≤) be an ordered non empty set and let f be an increasing
mapping from E to E verifying x ≤ f (x) = f (f (x)) for every x ∈ E.
Set F = {x ∈ E, x = f (x)}.

(a) Show that for every x ∈ E the set Fx = {y ∈ F, x ≤ y} is not

empty.
(b) Show that f (x) is a smallest element of Fx .
(c) Let g be a mapping from E to E and let G be a part of E such
that the set Gx = {y ∈ G, x ≤ y} is not empty for every x ∈ E
and admits g(x) as smallest element.
i. Prove that x ≤ g(x) for every x ∈ E.
ii. Prove that G = {x ∈ E, x = g(x)}.
iii. Let x, y ∈ G be such that x ≤ y. Show that Gy ⊆ Gy . Deduce
that g is increasing.
iv. Show that ∀x ∈ E we have g(g(x)) ⊆ Gx . Deduce that
g(g(x)) = g(x).
Chapter 2

Relations and Graphs

2.1 Symmetric Relations

Many structures in many domains rise from symmetric relations. Some of
these structures are deep and rich in theoretical mathematical properties,
even though a symmetric relation in its own state is a simple concept.

A very simple illustration of these structures is the friendship example: given

a set V of persons, we define the symmetric relation R: “being a friend”. At
the first glance, V endowed with this relation is a very simple structure.
But, through applying on it some tools from algebraic theories, the following
elegant result, called the friendship theorem, emerges:

Theorem 2.1.1. If any two persons in V have exactly one common friend,
then there is a person in V who is the friend of everyone in V .

Definition 2.1.1. Let R be a symmetric relation defined on a finite set V . V

endowed with the relation R is said to be a graph. A graph is often denoted
by the capital letter G. It is defined by the two sets: V = V (G), called the
set of vertices of G, and E = E(G) = {{x, y}; xRy}, called the set of edges
of G. So a vertex v of G is simply an element of V , while an edge is a pair
{x, y} ⊆ V such that xRy. To simplify, we write e = xy instead of {x, y}
where e is an edge of G.

Graphs, in general, are supposed to be defined by antireflexive relations, that

is xRx ∀x ∈ V . The order of a graph G, denoted by v(G), is the number of
its vertices and the size of G, denoted by e(G), is the number of its edges.
Graphs may be illustrated by points and lines.
16 CHAPTER 2. RELATIONS AND GRAPHS

Example 2.1.1. Given the graph G:

a c f

b d

V (G) = {a, b, c, d, f }
E(G) = {ab, ad, bc, bd, cd, df }
v(G) = 5, e(G) = 6, e = ab = ba
Example 2.1.2. G is a complete graph if xy ∈ E(G) ∀x, y ∈ V (G).

G=K4

Definition 2.1.2. Let G be a graph. A graph H is said to be subgraph of

G if
V (H) ⊆ V (G)
E(H) ⊆ E(G)
H is said to be an “induced subgraph of G” if
xy ∈ E(H) ⇔ xy ∈ E(G) ∀x, y ∈ V (H)
H is also called the subgraph of G induced by A = V (H); we write:
H =< A >
. H is said to be a “spanning subgraph of G” if V (H) = V (G).
Example 2.1.3.
x z v x z

y t y t
G H: subgraph of G
2.1. SYMMETRIC RELATIONS 17

z v x z v

y t y t
H1 : induced subgraph of G H2 : spanning subgraph of G

Notation 2.1.1. Let G be a graph, v ∈ G, e = ab ∈ E(G). G − v is the

subgraph of G induced by V (G) \ {v}. G − e is the subgraph of G obtained
by deleting the edge e. Let x, y ∈ G such that xy ∈/ E(G), then G + xy is
the graph obtained from G by adding the edge xy.
Example 2.1.4.
x y z x z

t v t v
G G−y

x y z x y z

t v t v
G − vz G + xt

Definition 2.1.3. A path P is a graph verifying:

V (P) = {v1 , v2 , . . . , vn }

E(P) = {v1 v2 , v2 v3 , . . . , vn−1 vn }

We write P = v1 v2 . . . vn−1 vn .
v1 v2 v3 vn−1 vn

v1 is the origin of P and vn is its end. P is a v1 vn -path. The length of a path

is the number of its edges.
18 CHAPTER 2. RELATIONS AND GRAPHS

l(P)=4 l(Q)=3 l(R)=0

Definition 2.1.4. Let G be a graph. A path P in G is a subgraph of G. G

is said to be connected if it contains an xy-path ∀x, y ∈ V (G). A connected
component of a graph G is an induced connected subgraph maximal with
respect to the inclusion.

Definition 2.1.5. Let x, y ∈ V (G). The distance between x and y in a

connected graph G, denoted by d(x, y), is the length of the shortest xy-path
in G. The xy-path P such that l(P) = d(x, y) is called an xy-geodesic. The
diameter of G is d(G) = max{d(x, y); x, y ∈ G}.

Example 2.1.5. Consider the following graph G:

Then,

C1 : connected component of G C2 : connected component of G

Example 2.1.6. G is a connected graph:

x v

d(x, y) = 2 and (P ) = xvy ia a geodesic in G. d(G) = 2.

Notation 2.1.2. Let P be an xy-path and v ∈ P. P[x,v] is the subpath of

P of ends x and v.
2.1. SYMMETRIC RELATIONS 19

P[x,v]

y
x v
P

Exercise 2.1.1. Let P be an xy-geodesic of a connected graph G (x, y ∈ G)

and let v ∈ P. Show that P[x,v] is an xv-geodesic.
Exercise 2.1.2. Let G be a connected graph and let x, y, z ∈ G. Show that
d(x, y) ≤ d(x, z) + d(z, y).
Definition 2.1.6. A cycle is a graph C verifying:
V (C) = {v1 , v2 , . . . , vn } n≥3
E(C) = {v1 v2 , v2 v3 , . . . , vn−1 vn , vn v1 }
We write C = v1 v2 . . . vn . The length of C, l(C), number of edges of C.
Example 2.1.7. C: triangle, l(C) = 3:
v3

v1 v2

C: square, l(C) = 4:

l(C) = n:
vn v1 v2
vn−1 v3
20 CHAPTER 2. RELATIONS AND GRAPHS

A cycle is said to be even if its length is even and odd if its length is odd.

Definition 2.1.7. A cycle of a graph G is a subgraph of G which is a cycle.

x t
v

y z
G

ytvz is a square in G, xytvz is a cycle in G.

A connected graph without cycles is said to be a tree.

Tree

Definition 2.1.8. Let G be a graph and let v ∈ G. The neighborhood of v

is the set N (v) = {x ∈ G; vx ∈ E(G)}. For more precision, we may write
NG (v). The degree of v is d(v) = dG (v) = |NG (v)|. G is said to be regular
or k-regular if d(x) = k ∀x ∈ V (G).

Example 2.1.8.

x y z
t

u v
G

N (v) = {u, x, y, z}, d(v) = 4

N (t) = ∅, d(t) = 0 (isolated vertex)
2.1. SYMMETRIC RELATIONS 21

3-regular graph

Notation 2.1.3. G is a graph.

δ(G) = min{d(v); v ∈ G}.
∆(G) = max{d(v); v ∈ G}.
Theorem 2.1.2. (Euler,1736) Let G be a graph. We have:
X
dG (v) = 2e(G) (i.2)
v∈G

Proof. We argue by induction on e(G). If e(G) = 0, then G has no edges

and so d(v) = 0 ∀x ∈ G, the equality (2.1) holds. Suppose that the equality
holds for graphs of size s and suppose that e(G) = s + 1. Let e = xy ∈ E(G).
Set G0 = G − e. e(G0 ) = s, then the equality (2.1) holds for G0 . We have:
X
dG0 (v) = 2e(G0 )
v∈G0
X
⇒ dG0 (v) + dG0 (x) + dG0 (y) = 2(e(G) − 1)
v6=xy
X
⇒ dG (v) + (dG (x) − 1) + (dG (y) − 1) = 2e(G) − 2
v∈G
v6=xy
X
⇒ dG (v) − 2 = 2e(G) − 2
v∈G
P
Thus v∈G dG (v) = 2e(G).
Corollary 2.1.1. Let G be a graph. Then |{v; d(v) is odd}| is even.
Proof. Remark that
X
d(v) = 2e(G) is even.
v∈G

So, X X
d(v) + d(v) is even.
v∈G v∈G
d(v) odd d(v) even
22 CHAPTER 2. RELATIONS AND GRAPHS

Thus X
d(v) is even.
v∈G
d(v) odd

Then |{v; d(v) is odd}| is even.

Definition 2.1.9. Let G be a graph. A subset ζ ⊆ V (G) is said to be stable

if xy ∈
/ E(G) ∀x, y ∈ ζ. The stability of G is α(G) = max{|ζ|; ζ stable in
G}. The chromatic number of G, denoted by χ(G), is the minimum number
of stables in G whose union is V (G). χ(G) = t means ∃ ζ1 , ζ2 , . . . , ζt stables
in G such that V (G) = ζ1 ∪ ζ2 ∪ . . . ∪ ζt and t is minimal for this property.

Example 2.1.9.

y
x• • •z

• • •
t v u

∗ {x, y} is a stable in G.
∗ {t, z, u} is a maximal stable in G.
∗ α(G) = 3.
∗ χ(G) = 3. V (G) = S1 ∪ S2 ∪ S3 , S1 = {x, t}, S2 = {y, z}, S3 = {v, u}.
∗ χ(G) 6= 2.

Definition 2.1.10. Let G be a graph. The complement of G, denoted by

Ḡ, is the graph defined by:

V (Ḡ) = V (G)
E(Ḡ) = {xy; xy 6∈ E(G)}
2.2. BIPARTITE GRAPH 23

Example 2.1.10.

y y
x• • •z x• • •z

• • • •
t v t v
G Ḡ

2.2 Bipartite graph

Definition 2.2.1. A graph G is said to be bipartite if:
(
V (G) = X ∪ Y, where X and Y are disjoint.
e ∈ E(G) ⇔ e = xy, with x ∈ X, y ∈ Y

We write G=G(X,Y).
Remark: X and Y are stables; χ(G) ≤ 2.

Example 2.2.1.

• • • X

• • • • Y

Remark 2.2.1.

x• •y x• •z X

t• •z y• •t Y

G : bipartite
24 CHAPTER 2. RELATIONS AND GRAPHS

v1
•
x v2 •
• • v5

y• •z v3 • • v4

G is not bipartite G is not bipartite

Proposition 2.2.1. An odd cycle is not bipartite.

Proof. Let C = v1 v2 . . . v2p+1 be an odd cycle. Suppose to the contrary that
C is bipartite and let X, Y be two stables such that V (C) = X ∪Y . Suppose,
without loss of generality, that v1 ∈ X. We have

v ∈ X ⇒ v2 ∈ Y ⇒ v3 ∈ X.

By induction, if v2k−1 ∈ X, then v2k ∈ Y ; and so v2k+1 ∈ X. Thus v1 , v2p+1 ∈

X and v1 v2p+1 ∈ E(C). A contradiction.

Remark 2.2.2. Let G = G(X, Y ) be a bipartite graph.

• Any subgraph of G is bipartite.

• x ∈ X, y ∈ Y , xy 6∈ E(G); then G0 = G + xy is bipartite.

• x ∈ X, y ∈ Y , P is an xy-path, then l(P ) is odd.

P P
• e(G) = x∈X d(x) = y∈Y d(y).

Proposition 2.2.2. A graph G is bipartite if and only if any connected

component of G is bipartite.
Proof. N.C. Suppose that G is bipartite; G = G(X, Y ) and let C be c.c. of
G. We have C = C(X ∩ C, Y ∩ C). So, C is bipartite.

S.C. Suppose that C1 , C2 , . . . , Cs are the c.c. of G. Each one is bipartite.

Set Ci = Ci (Xi , Yi ). Then X = ∪ni=1 Xi and Y = ∪ni=1 Yi are both stable
and

V (G) = ∪i V (Ci ) = ∪i (Xi ∪ Yi ) = (∪i Xi )(∪i Yi ) = X ∪ Y.

Then G is bipartite; G = G(X, Y ).

2.3. DIGRAPHS 25

Theorem 2.2.1. [(König - 1928)] A connected graph G is bipartite if and

only if it contains no odd cycle.

Admitted without proof.

Remark 2.2.3. Based on proposition 2.2.2, we may easily remark that an

arbitrary graph is bipartite if and only if it contains no odd cycles.

Proposition 2.2.3. Let G be a k-regular bipartite graph G = G(X, Y ).

Then |X| = |Y |.

Proof. We have
X X
e(G) = d(x) = d(y).
x∈X y∈Y

Then,
X X
k= k.
x∈X y∈Y

And so k|X| = k|Y |, |X| = |Y |.

2.3 Digraphs
When a relation R is not symmetric, it defines what we call oriented graph
or digraph. Pairs are replaced by couples, and edges become arcs.

Definition 2.3.1. Let V be a finite set, and let R be an arbitrary relation

defined on V . The set V endowed with the relation R is said to be a digraph.
Thus a digraph D is defined by the sets:
(
V = V (D) : the set of vertices of D
E(D) = {(x, y) ∈ V 2 ; xRy} : the set of arcs of D

The underlying graph of D, G(D), is the graph obtained from D by ignoring

the orientations of the arcs.
(
V (G(D)) = V (D)
E(G(D)) = {xy; (x, y) ∈ E(D) or (y, x) ∈ E(D)}

Example 2.3.1.
26 CHAPTER 2. RELATIONS AND GRAPHS

x y z
• • •

D:

• •
t v

V (D) = {x, y, z, t, v}.

E(D) = {(x, t), (z, v), (t, y), (t, v), (v, x), (v, y)}.

Definition 2.3.2. Let D be a digraph, v ∈ D. The out-neighborhood of v is

the set:
N + (v) = {x; (v, x) ∈ E(D)}
The out-degree of v is d+ (v) = |N + (v)|.

By a similar approach, we define the in-neighbor and the in-degree of v.

Remark 2.3.1. • As we defined for graphs, similarly we define sub-

graphs, induced subgraphs and spanning subgraphs of a digraph D.
• The stability of a digraph D is the stability of its underlying graph
G(D). Also, χ(D) = χ(G(D)).
• A digraph D is connected if G(D) is connected.
• An oriented path (resp. cycle) is a digraph whose underlying graph is
a path (resp. cycle).

Definition 2.3.3. (i) A directed path P is a path verifying:

(
V (P ) = {v1 , v2 , . . . , vn }
E(P ) = {(v1 , v2 ), (v2 , v3 ), . . . , (vn−1 , vn )}

• • • • •

We write P = v1 v2 . . . vn . P is a v1 vn -directed path, l(P ) = n − 1.

(ii) A circuit or a directed cycle C is a digraph verifying:
(
V (C) = {v1 , v2 , . . . , vn }
E(C) = {(v1 , v2 ), (v2 , v3 ), . . . , (vn−1 , vn ), (vn , v1 )}
2.3. DIGRAPHS 27

•
• •

• •
•

We write C = v1 v2 . . . vn , l(C) = n.
Definition 2.3.4. A digraph is said to be strongly connected or strong if
∀x, y ∈ D, D contains an xy-directed path.
Example 2.3.2. The digraph below is a strong digraph.
•
• • • •

• • •

• • • •
•

Definition 2.3.5. A tournament T is a digraph such that ∀x, y ∈ V (T ) we

have (x, y) ∈ E(T ) or (y, x) ∈ E(T ). (It is not allowed that both (x, y) and
(y, x) be in E(T ))
Example 2.3.3. T : Tournament
•

• •

Proposition 2.3.1. Any tournament T contains a directed path passing

through all the vertices of T .
Proof. Let P = v1 v2 . . . vs be a longest directed path in T . If s < v(T ), then
let v ∈ T with v ∈ / P . If (v, v1 ) ∈ E(T ), then vv1 v2 . . . vs is a directed path
longer than P ; a contradiction. Then, (v1 , v) ∈ E(T ). Similarly, we show
that (v, vs ) ∈ E(T ). So, ∃ i ∈ [1, s] such that (vi , v), (v, vi+1 ) ∈ E(T ). Thus,
v1 . . . vi vvi+1 . . . vs is a directed path in T longer than P ; a contradiction.
28

Exercises and Problems

1. Consider the graph defined by:

V (G) = {v1 , v2 , v3 , v4 , v5 , v6 , v7 }

E(G) = {v1 v3 , v1 v4 , v2 v3 , v2 v4 , v3 v4 , v4 v5 , v6 v7 }

(a) Draw G.
(b) Find δ(G), ∆(G), α(G), d(G), χ(G).
(c) Is G connected? Find the connected components of G.
(d) Find a longest path of G.
/ E(G) and G0 = G + xy verifies
(e) Find x, y ∈ V (G) such that xy ∈
that G is connected and ∆(G0 ) = ∆(G). Show that δ(G0 ) = δ(G).

2. Given the Petersen graph:

G : Petersen graph

(a) Find the stability of G.

(b) Find the chromatic number of G.
(c) Find a longest cycle in G.

3. Let G be a connected graph. Let x, y ∈ G such that xy ∈ E(G) and

G0 = G − xy.

(a) Show that if G0 is connected, then G has a cycle C containing xy.

(b) Show that if G0 is disconnected, then G0 contains exactly 2 distinct
connected components, C1 and C2 , such that x ∈ C1 and y ∈ C2 .

4. Let G be a connected graph.

(a) Let v ∈ G such that d(v) = 1. Show that G0 = G − v is connected.

 29

(b) Suppose that G has no cycles. Let P be a longest path in G and

let x be an end of P . Show that d(x) = 1.
(c) Deduce that G is a tree if and only if v(G) = e(G) + 1.

5. Let G be a connected graph with δ(G) ≥ 2.

(a) Show that e(G) ≥ v(G). Deduce that G contains a cycle.

(b) Show that G contains a cycle C with a vertex v ∈ C such that
N (v) ⊆ C.

6. Let G be a connected graph and let P be a longest path in G. Set

P = v1 v2 . . . vn .

(a) Show that N (v1 ) ⊆ P , N (vn ) ⊆ P .

(b) Show that if v1 vn ∈ E(G), then v(G) = n.
(c) Suppose that v1 vn ∈
/ E(G). Set

N (vn ) = {vi1 , vi2 , . . . , vis }; 2 < i1 < i2 < . . . < is ≤ n − 1

Set X = {vi1 +1 , vi2 +1 , . . . , vis +1 }. Show that if N (v1 ) ∩ X 6= ∅,

then v(G) = n.
(d) Deduce that if d(v1 ) + d(vn ) ≥ n, then v(G) = n.

7. Let G be a connected graph.

(a) Let S, S 0 ⊆ V (G), S ∩ S 0 = ∅. Show that there is an xy-path P

in G such that P ∩ S = {x} and P ∩ S 0 = {y}.
(b) Two paths, P1 and P2 , of G are said to be parallel if P1 ∩ P2 = ∅.
Let R be an xy-path such that x ∈ P1 and y ∈ P2 (P1 and P2 are
two parallel paths of G). Show that P1 ∪ R ∪ P2 contains a path
P such that l(P ) > l(P1 ) and l(P ) > l(P2 ).
(c) Deduce that two longest paths of G are not parallel.

8. Let G be a graph.

(a) Show that v(G) ≤ α(G)χ(G).

(b) Show that e(G) ≥ δ(G)α(G).
v(G) e(G)
(c) Deduce that χ(G)
≤ δ(G)
.

9. Let G be a graph with v ∈ G.

(a) Show that α(G − v) ≥ α(G) − 1.

30

(b) Show that χ(G − v) ≥ χ(G) − 1.

(c) Show that:

χ(G − v) = χ(G) − 1 ⇒ d(v) ≥ χ(G) − 1.

10. Let G be a graph. Set ω(G) = max{v(H); H is a complete subgraph

of G}.

(a) Show that ω(G) = α(Ḡ) and α(G) = ω(Ḡ).

Deduce that v(G) ≤ χ(G)χ(Ḡ).
(b) Let v ∈ G. Show that dG (v) + dḠ (v) = v(G) − 1.

11. Let D be a diagraph. Suppose that D contains no circuits.

(a) Show that D contains a vertex v such that d+ (v) = 0. Show that
S = {v; d+ (v) = 0} is a stable.
(b) Let v ∈ D and set c(v) = max{l(P); P is a divided path in D of
origin v}. Show that if c(v) = c(v 0 ) then vv 0 ∈
/ E(G(D)).
(c) Deduce that D contains a directed path P such that l(P) ≥
χ(D) − 1.

12. Let T be a tournament of order n ≥ 4.

(a) Show that T contains a non directed triangle.

(b) Show that if T is strong then it contains a circuit. Deduce that T
contains a directed triangle.

13. Let H be a group of neutral element e. Let a 6= e ∈ H. We define a

diagraph D by:
V (D) = H
(x, y) ∈ E(H) ⇔ y = ax

(a) Show that d+ (v) = d− (v) = 1 ∀v ∈ G.

(b) Show that any cycle in D is a circuit.
(c) Show that any cycle in D is of length p such that p is the order of
a (p = min{m; am = e}).
(d) Deduce that an = e where n = |G|.
Chapter 3

Complement on the Natural

Integers and the Cardinals of
Countable Sets

3.1 The Construction of N

The mathematical teaching programs allow us to admit the existence of N,
the properties of the natural numbers and the classical operations defined on
these numbers1 .

However, it is interesting to reintroduce N by a minimal number of axioms

and establish then the fundamental properties of this set.

N is a non empty ordered set verifying the following axioms.

Axiom 3.1.1. (A1 ) Any non empty part of N admits a smallest element.

The smallest element of N is denoted by the symbol 0 (called zero).

Axiom 3.1.2. (A2 ) N has no greatest element.

Axiom 3.1.3. (A3 ) Any element a 6= 0 of N has a predecessor a0 (i.e. a0 < a

and a0 ≤ b ≤ a ⇒ b = a0 or b = a2 .

The set N may be introduced using another family of axioms, like Peano’s
axioms, the above axioms will be then simple consequences, they are chosen
1
From Michel Queysanne: Algèbre (1er cycle et classes préparatoires).
2
We define the successor in the same way: a0 is the successor of a if a < a0 and if
a ≤ b ≤ a0 then b = a0 or b = a.
 CHAPTER 3. COMPLEMENT ON THE NATURAL INTEGERS AND
32 THE CARDINALS OF COUNTABLE SETS
here for a simplification reason: these axioms permit us to present the in-
duction as a proved fact. This notion may be presented as an axiom after
defining the successor function S(n) as follows:

Axiom 3.1.4. Let K be a set verifying:

1. 0 ∈ K.

2. n ∈ K ⇒ S(n) ∈ K.

Then K = N.

Induction is a fundamental principle justifying the apparent passage from

particular to general in mathematics. It can be exposed in two forms:

Theorem 3.1.1. (First form) Let p be a property defined on N such that

0 verifies p and if k verifies p, then k + 1 verifies p. Then p is true for every
n ∈ N.

Proof. We have to show that the set A = {k ∈ N; k doesn’t not verify p}

is empty. Suppose to the contrary that A 6= ∅, then A admits a smallest
element s by A1 . We have s 6= 0 since 0 ∈ / A, then s admits a predecessor
0 0 0
s . We have s < s, then s ∈ / A and then s0 verifies p, thus by hypothesis
0
s = s + 1 verifies p, a contradiction. Then A = ∅.

Theorem 3.1.2. (Second form) Let p be a property defined on N such

that 0 verifies p and if s verifies p for every s < k, then k verifies p. Then p
is true for every n ∈ N.

Proof. A and s are defined as in the above theorem, for every s0 < s, we have
s0 ∈
/ A and then s0 verifies p, then by hypothesis s verifies p, a contradiction.
So A = ∅.
The induction principle is based then on the axiom of the smallest element
satisfied in N. It is also founded on what we call “mathematical definition”.
In fact, the ambiguity in the definition of many properties -in a mathematical
point of view- raises up the belief that some of our fundamental logical laws
are built over a weak base. The “sorites” are among the more ancient para-
doxicons that go back to the Greek philosophers of the Megare school. These
paradoxicons consist of proving that it is impossible to construct a handful of
grains by adding one grain at a time. Other examples show in the same way
that there are no rich men. In fact, a man possessing a franc is not rich and
he will not be so if one franc is added. If we apply induction, we are forced to
conclude that this man will remain poor no matter how many pieces of money
3.2. CARDINAL OF A FINITE SET 33

he receives. For certain, the same argument is valid for the example of grains.

Clearly, one of the pragmatic mean to destroy this argument is to unload

containers full of grains. But this counter-example does not serve to discover
the weak logical point in this reasoning.

The problem resides in the fact that we have no precise borders permitting to
decide if something is a “handful” or not, if a man is rich or not, etc. Then,
it is the mathematical ambiguity in the definitions of “handful”, “rich”, etc
which leads to contradictory results3 .

Finally, it remains to remark that the first step 0 p(0)0 must be carefully
verified. To realize this see exercise 3 (A group of persons containing a
woman contains only women).

3.2 Cardinal of a Finite Set

To simplify, the set {1, . . . , n} is denoted by [1, n] for every n ∈ N∗ .

Proposition 3.2.1. Let m and n be two non zero positive integers. If there
exists an injection from [1, m] into [1, n], then m ≤ n.

Proof. We argue by induction on m; if m = 1, the inequality is trivial since

n ≥ 1. Suppose that the property holds for m and let’s prove it for m + 1 the
successor of m. Let f be an injection from [1, m + 1] into [1, n] (Note that
n > 1 in this case, since otherwise f (1) = f (m + 1) and 1 6= m + 1). Let s
be the element of [1, n] such that f (m + 1) = s and let ϕ the permutation
(s n) (if s = n, ϕ is the identical mapping), then ϕ ◦ f is an injection from
[1, m + 1] into [1, n] such that ϕ ◦ f (m + 1) = n, so the restriction of f ◦ ϕ on
[1, m] is an injection into [1, n∗ ] where n∗ is the predecessor of n, we deduce,
by induction, that m ≤ n∗ and consequently m + 1 ≤ n.

Corollary 3.2.1. If there exists a bijection from [1, m] into [1, n], then m =
n.

Proposition 3.2.2. Let n be a non zero natural integer. Any injection from
[1, n] into [1, n] is a bijection.

Proof. We argue by induction on n; if n = 1, the property is trivial. Suppose

that the property holds for n and let’s prove it for n + 1. Let f be the
3
By Nicholas Falletta: The paradoxicons.
 CHAPTER 3. COMPLEMENT ON THE NATURAL INTEGERS AND
34 THE CARDINALS OF COUNTABLE SETS

injection from [1, n + 1] into [1, n + 1]. Let s be the element of [1, n + 1] such
that f (n + 1) = s and let ϕ be the permutation (s n + 1) (if s = n + 1, ϕ
is then the identity mapping), then ϕ ◦ f is an injection from [1, n + 1] into
[1, n + 1] such that ϕ ◦ f (n + 1) = n + 1, then it is a bijection by induction
hypothesis. We deduce that ϕ ◦ f is a bijection, then f is a bijection since ϕ
is so (f = ϕ−1 ◦ (ϕ ◦ f )).
Definition 3.2.1. A non empty set E is said to be finite of cardinal n (we
write card(E) = n) if there exists a bijection from E into [1, n] where n ∈ N∗ .
A non finite set is said to be infinite. By definition ∅ is a finite set of cardinal
zero, card(∅) = 0.
Remark 3.2.1. The above definition is well justified, in fact, if there exist
bijections from E to [1, n] and [1, m], then these two last sets are equipollent
and m = n.
The verification of the following remarks is left to the reader:
Proposition 3.2.3. Let E and F be two given sets. We have:
1. If F is finite and E ⊆ F , then E is finite and card(E) ≤ card(F ).

2. If E and F are finite, then E × F is finite and we have card(E × F ) =

card(E) × card(F ).

3. If E and F are finite and if E ∩ F = ∅, then E ∪ F is finite and we

have card(E ∪ F ) = card(E) + card(F ).
(To establish these three facts, argue by induction on the cardinal of F )
Exercise 3.2.1. Let E and F be two finite sets. Show that card(E ∪ F ) =
card(E) + card(F ) − card(E ∩ F ).
Proposition 3.2.4. Let f be a bijection from a set E into a set F . If F is
finite, then E is finite and we have card(E) = card(F ).
Proof. Set card(F ) = n and let ϕ be a bijection from F into [1, n]. ϕ ◦ f is
a bijection from E into [1, n] and then card(E) = n = card(F ).
Remark 3.2.2. If f is injective, then f : E 7→ f (E) is bijective and we have
card(E) = card(f (E)) ≤ card(F ) since f (E) ⊆ F . If f is surjective, we may
find a part A in E such that f : A 7→ F is a bijection. We have in this case
card(F ) = card(A) ≤ card(E).
Proposition 3.2.5. Let E of a finite set F such that card(E) = card(F ),
then E = F .
3.3. COUNTABLE SETS 35

Proof. It is sufficient to remark that the canonical injection i : E 7→ F is

bijective. We have E = i(E) = F . Set card(E) = card(F ) = n and let ϕ
(resp. ϕ0 ) be a bijection from E (resp. F ) in [1, n]. We have ϕ0 ◦ i ◦ ϕ−1 is an
injection from [1, n] into [1, n], then it is a bijection and so i is a bijection.
Corollary 3.2.2. Let E and F be two finite set such that card(E) =
card(F ). For every mapping f : E 7→ F the following properties are equiva-
lent:
1. f is injective.

2. f is surjective.

3. f is bijective.
Proof. If f is injective, then f : E 7→ f (E) is bijective and we have card(f (E)) =
card(E) = card(F ), then E = f (E) and f is bijective. If f is surjective, we
may find a part A of E such that f : A 7→ F is a bijection. We have in this
case card(A) = card(F ) = card(E), then A = E and f is bijective.

3.3 Countable sets

Definition 3.3.1. A set E is said to be countable if E is finite or E is
equipollent to N.
Remark 3.3.1. If a countable set E is equipollent to a set F , we write
E ∼ F , then F is countable.
As in the finite case, we are able to show that a subset of a countable set
is countable, the intersection and union of 2 countable sets are countable.
Moreover, we shall prove that the cartesian product of two countable sets is
countable.
Proposition 3.3.1. Let B be an infinite subset of N. Then B can be ar-
ranged into a strictly increasing sequence (xn )n≥0 .
Proof. Let x0 be the smallest element of B; it exists due to the axioms of N.
Suppose by induction that x0 , . . . , xs are defined, s ≥ 0. Since B is infinite,
then B − {x0 , . . . , xs } =
6 ∅. Let xs+1 be its smallest element. Then, the
sequence is well defined and is obviously increasing. Set H = {xn , n ≥ 0}.
Trivially, H ⊆ B. Conversely, let m ∈ B. If m = 0, then m = x1 and so
m ∈ H. Otherwise, A = B ∩[1, m−1] ⊆ [1, m], then A is finite. Let xk be the
greatest element of A. We have A = {x0 , x2 , . . . , xk }. Since m is the smallest
element of B − A, then m = xk+1 ∈ B. Thus B ⊆ H, and B = H.
 CHAPTER 3. COMPLEMENT ON THE NATURAL INTEGERS AND
36 THE CARDINALS OF COUNTABLE SETS

Corollary 3.3.1. Any subset of N is countable.

Proof. Let B ⊆ N. If B is finite, then B is countable. If B is infinite, then
B = {xn , n ≥ 0} where (xn )n≥1 is strictly increasing. And so,

f :N → B
s → f (s) = xs . is a bijection.

B is countable.
Corollary 3.3.2. A set E is countable if there exists an injection from E to
N.
Proof. Let f : E → N be an injective mapping. Then E ∼ f (E) which is
countable. So, E is countable.
Corollary 3.3.3. A subset E of a countable set F is countable.
Proof. Let

iE : E → F ; f : F → N injection.
x → iE (x) = x.

Then f ◦ iE : E → N is an injective mapping. So, E is countable.

Corollary 3.3.4. Let E be a countable set. Then for every set F we have:
E ∩ F and E − F are countable.
Proposition 3.3.2. Let E and F be two countable sets with E ∩ F = ∅.
Then E ∪ F is countable.
Proof. E and F are countable, then ∃f : E → N, g : F → N which are
injective. Define

ϕ:E∪F →N
(
2f (x) if x ∈ E.
x → ϕ(x) =
2g(x) + 1 if x ∈ F.

We may easily verify that ϕ is injective. So, E ∪ F is countable.

Remark 3.3.2. We conclude that the union of two arbitrary countable sets
E and F is countable since E ∪ F = (E − F ) ∪ F . Also, if E1 , E2 , . . . , En are
countable sets, then E1 ∪ . . . ∪ En is countable.
Proposition 3.3.3. N × N is countable.
3.3. COUNTABLE SETS 37

Proof. Let

f :N×N→N
(a, b) → 2a × 3b

f is injective. In fact, let (a, b), (a0 , b0 ) ∈ N × N such that f (a, b) = f (a0 , b0 ).
We have:
0 0
f (a, b) = f (a0 , b0 ) ⇒ 2a 3b = 2a 3b .
0 0 0 0
2a divides 2a 3b = 2a 3b and 2a ∧ 3b = 1 ⇒ 2a divides 2a . So, a ≤ a0 .
Similarly, we show that a0 ≤ a. So a = a0 . We prove by a similar way that
b = b0 .
Corollary 3.3.5. The cartesian product of 2 countable sets is countable.
We prove now that the set R is not countable. For this, we prove first
that P(N) is not countable.

Let χ = {(xn )n≥0 ; xi ∈ {0, 1}}. ∀A ⊆ N we define the characteristic

function of A by

χA : N → {0, 1}
(
0 if t ∈
/ A.
t → χA (t) =
1 if t ∈ A.

Proposition 3.3.4. P(N) ∼ χ.

Proof. Define

ϕ : P(N) → χ
A → ϕ(A) = (xn )n≥0

where xn = χA (n), ∀n ≥ 0.
• ϕ is injective: Let A, A0 ⊆ N/ ϕ(A) = ϕ(A0 ). Set ϕ(A) = (xn )n≥0 ,
ϕ(A0 ) = (x0n )n≥0 . We have:

ϕ(A) = ϕ(A0 ) ⇒ xn = x0n ∀n ≥ 0.

Let t ∈ A. We have xt = χA (t) = 1.

x0t = xt = 1 ⇒ χA0 (t) = 1 ⇒ t ∈ A0

So, A ⊆ A0 . Similarly, A0 ⊆ A. So, A = A0 and ϕ is injective.

 CHAPTER 3. COMPLEMENT ON THE NATURAL INTEGERS AND
38 THE CARDINALS OF COUNTABLE SETS

• ϕ is surjective: Let (xn )n≥0 ∈ χ. Set A = {xn ; xn = 1}. We have

A ⊆ N, and ϕ(A) = (xn )n≥0 . ϕ is surjective.

Therefore, ϕ is bijective and P(N) ∼ χ.

Theorem 3.3.1. P(N) is not countable.

Proof. Suppose to the contrary that P(N) is countable. Then χ is countable.

Let f : N → χ be a bijection. Set f (i) = (xin )n≥0 . Define (un )n≥0 by:

ui = 1 − xii , i ≥ 0.

Clearly (un )n≥0 ∈ χ. Then ∃ s such that f (s) = (un )n≥0 . We have f (s) =
(xsn )n≥0 = (un )n≥0 . So, us = xss , a contradiction.

Let (xn )n≥0 ∈ χ. We define the sequence:

si = 0, x1 x2 . . . xi i ≥ 0.

This sequence of positive number is strictly increasing and bounded by 1.

So, it converges in R to a limit s. We write: s = 0, x1 x2 . . . xn . . .

Theorem 3.3.2. R is not countable.

Proof. Let

ϕ:χ→R
(xn )n≥ 0 → ϕ((xn )n≥0 ) = 0, x1 x2 . . . xn . . .

ϕ is injective: Let ϕ((xn )n≥0 ) = ϕ((x0n )n≥0 ). Suppose to the contrary that
(xn )n≥0 6= (x0n )n≥0 and let n0 be the first integer such that xn0 6= x0n0 . Suppose
that xn0 = 0, x0n0 = 1.

10n0 . 0, x1 x2 . . . xn . . . = xn0 , xn0 +1 . . . xn . . .

10n0 . 0, x01 x02 . . . x0n . . . = x0n0 , x0n0 +1 . . . . . .

Then, ϕ((xn )n≥0 ) 6= ϕ((x0n )n≥0 ), a contradiction. So, ϕ(χ) ⊆ R is equipollent

to χ. So, it is not countable, then R is not countable.
3.4. COMBINATORIAL ANALYSIS 39

3.4 Combinatorial Analysis

Definition 3.4.1. Let E be a set of cardinal n ≥ 1. For every p ≤ n, a p-
arrangement (x1 , x2 , . . . , xp ) of E is an ordered family formed by p elements
of E pairwise distinct. The number of all p-arrangement of E depends only
upon p and n. This number will be denoted by Apn . The n-arrangements are
called the permutation of E, its number Ann is denoted by Pn .

Example 3.4.1. Let E = {1, 2, 3}. The 3-arrangements (permutations) of

E are (1, 2, 3), (2, 3, 1), (3, 1, 2), (1, 3, 2), (2, 1, 3), (3, 2, 1).

Notation 3.4.1. The number of parts of E formed by p elements is denoted

by Cnp .

In this section, we calculate Apn , Pn and Cnp in terms of p and n.

n!
Theorem 3.4.1. Apn = (n−p)!
.

Proof. Apn can be considered as the number of the p-arrangements of the

set E = {1, 2, . . . , n}. We establish the equality by induction on p ≤ n.
n!
For p = 1, A1n is the number of the elements of E, then A1n = n = (n−1)! .
Suppose that the equality holds for p < n. Let Hk (k = p, p + 1) be the
set of k-arrangements of E and let f be the surjective mapping associat-
ing to each element (x1 , x2 , . . . , xp , xp+1 ) of Hp+1 the element (x1 , x2 , . . . , xp )
of Hp . Consider the equivalence relation Rf . Let z ∈ Hp+1 /Rf , z =
(x1 , x2 , . . . , xp , xp+1 ), the only difference between two arbitrary elements of
z is the (p + 1)th element (x1 , x2 , . . . , xp are fixed), then card(z) = n − p.
So card(Hp+1 ) = card(Hp+1 /Rf ) × (n − p) (The classes of Rf form a parti-
tion of Hp+1 ). But card(Hp+1 /Rf ) = card(Hp ) since Hp+1 /Rf and Hp are
n!
equipollent (proposition 3.4.3). But card(Hp ) = Apn = (n−p)! by the induction
hypothesis. Consequently:

n! n!
card(Hp+1 ) = card(Hp ) × (n − p) = × (n − p) =
(n − p)! (n − (p + 1))!

Corollary 3.4.1. Pn = n!.

n!
Proof. We have Pn = Ann = (n−n)!
= n! (0! = 1).

n!
Corollary 3.4.2. Cnp = p!(n−p)!
.
 CHAPTER 3. COMPLEMENT ON THE NATURAL INTEGERS AND
40 THE CARDINALS OF COUNTABLE SETS

Proof. Let Sp be the set of parts of E = {1, 2, . . . , n} formed by p ele-

ments and let f be the surjective mapping which associate to each element
(x1 , x2 , . . . , xp ) of Hp the element {x1 , x2 , . . . , xp } of Sp . We consider the
equivalence relation Rf . Let z ∈ Hp /Rf , z = (x1 , x2 , . . . , xp ), the elements
of z are the permutations of the set {x1 , x2 , . . . , xp }, so card(z) = p!. Thus
card(Hp ) = card(Hp /Rf ) × p!. But card(Hp /Rf ) = card(Sp ) since Hp /Rf
n!
and Sp is equipollent (proposition 3.4.3). But card(Hp ) = Apn = (n−p)! . Con-
sequently:
card(Hp ) = card(Sp ) × p!
So
card(Hp ) n!
Cnp = =
p! p!(n − p)!
 41

Exercises and Problems

1. Use induction to establish the following:

(a) 1 + 12 + 1
22
+ ... + 1
2n
=2− 1
2n
; n ≥ 1.
(b) (1 + 2 + . . . + n) = 1 + 2 + . . . + n3 .
2 3 3
 n
(c) n+1n
≤ 3.

2. Let E be a set of cardinal n ≥ 1 and let a ∈ E.

(a) Find the number of parts of E containing a.

(b) Show that E is the union of its parts containing a if and only if
n ≥ 3.

3. Someone wants to show that if a group of persons contains a woman

then it contains only women. He argues by induction as follows: If the
group is formed by one person, the property is trivial. Suppose that
we have n persons in the group, n ≥ 2 and consider the parts of n − 1
persons containing the woman of the group, by induction these parts
contain only women. So the group itself contains only women. Is this
argument true? Use the above exercise to justify your answer.

4. Let E be a finite set and let F = E ∪ {a} where a ∈

/ E. We define the
subset G of P (F ) by

G = {X ∈ P (F ) such that a ∈ X}

(a) Show that G and P (E) form a partition of P (F ). Deduce that

card(P (F )) = card(P (E)) + card(G).
(b) We define the mapping f from P (E) into G by:

f : P (E) 7→ G
X 7→ X ∪ {a}

Show that f is bijective. Deduce that card(P (F )) = 2card(P (E)).

(c) Conclude that if a set is of cardinal n, then its power set is of
cardinal 2n .

5. Show that if p is prime, then

(a) If p divides a1 × a2 × . . . × an , then ∃1 ≤ i ≤ n such that p divides

ai .
42

(b) If p divides an , where n ∈ N∗ , then p divides a.

(c) p divides pn if and only if n ≥ 1.
(d) pt divides pn if and only if t ≤ n.
(e) If q is prime and q t divides pn , for some t ∈ N∗ , then q = p and
t ≤ n.

6. Let p1 , p2 , . . . , pr be pairwisely distinct prime numbers.

(a) Show that if q is a prime dividing (p1 × p2 × . . . × pr ) + 1, then

q∈/ {p1 , p2 , . . . pr }.
(b) Deduce that the set of prime numbers is infinite.

7. Let p be prime.

(a) Show that p is relatively prime with s!, for all s ≤ p − 1.

(b) Deduce that p divides Cpt , for all 1 ≤ t ≤ p − 1.

8. Show that if p is prime, then p divides np − n, for every natural number

n. (Hint: argue by induction on n and use exercise 7.)

9. Show that if p is prime and p does not divide n, then p divides np−1 − 1.
√
10. Let n ≥ 3. Show that if p is a prime, such that n < p and every
prime ≤ p does not divide n, then n is prime.

11. Let n ≥ 3. Show that if p is a prime, such that every prime ≤ p does
not divide n, and the quotient q of the division of n by p is < p, then
n is prime.

12. Show that if n ∈ N, then a and b are coprime in the following cases:

(a) a = 3n + 2 and b = 2n + 1.
(b) a = 14n + 3 and b = 5n + 1.

13. Let a, b ∈ N∗ . Show that if a

b
= αβ , where α ∧ β = 1, then

a b
a∧b= = and a ∨ b = aβ = bα.
α β
Application: Compute a ∧ b and a ∨ b in the following cases:

(a) a = 18 and b = 42.

(b) a = 1224 and b = 216.
 43

14. Find the natural numbers x and y, satisfying

x2 + y 2 = 25.

Deduce that we can find non-zero natural numbers x, y, and z, such

that
x2 + y 2 = z 2 .

15. Show that if x and y are two natural numbers, such that 3 divides
x2 + y 2 , then 3 divides x and y. Deduce that the equation x2 + y 2 =
7500000 has no solution in N.

16. Show that there exists no natural numbers x and y, such that x and y
are non-zero and
x3 + y 3 = 93 .
Remark: This is a particular case of Fermat’s Last Theorem which
states that if n is any natural number, such that n ≥ 3, then the
equation
xn + y n = z n
has no solutions for which x, y and z are non-zero natural numbers.
Pierre de Fermat, who lived from 1601 to 1665 was a French math-
ematician. He wrote in the margin of a notebook “I have found an
admirable proof of this theorem, but the margin is too narrow to con-
tain it.” Mathematicians have been searching for this proof ever since.
Using computers, they had proved the theorem for all n ≤ 30000, but
the year 1993 witnessed the end of this problem which had been fac-
ing mathematicians since the 17th century, when Professor Andrew
Wiles, who is a British mathematician at the University of Cambridge,
announced a proof for Fermat’s Last Theorem. He used ideas of many
mathematicians, but one of the crucial ingredients was the work of the
German mathematician Dr.Matthias Flach of Heidelberg University.

17. Show that if A and B are non-empty sets, then the following are equiv-
alent:

(a) there exists an injection from A to B,

(b) there exists a surjection from B onto A.

Deduce that if A is countable and f : A → B is surjective, then B is

countable.
44

18. Show that if A is a non-empty set, then A is countable if and only if

there exists a surjection from N onto A.

19. Show that if E1 , . . . , En are countable sets, then so are the sets E1 ×
. . . × En and ∪ni=1 Ei .

20. Show that Z is infinitely countable and give a bijection from Z onto N.

21. Let f : N × N → N be defined by

1
f (m, n) = (m + n + 1)(m + n) + n.
2
(a) Show that if m ≥ 1, then

f (m − 1, n + 1) = f (m, n) + 1, f or all n ∈ N.

(b) Show that

2f (a, b) = (a + b + 1)2 + (b − a − 1), f or all a, b ∈ N.

(c) Let a, b, n ∈ N, Show that if a 6= 0, then

i. if a + b < n, then

(a + b + 1)2 + (b − a) < (n + 1)2 + n,

ii. if a + b > n, then

(a + b + 1)2 − (n + 1)2 > n + a − b.

(d) Deduce that if f (a, b) = f (0, n), where a, b, n ∈ N, then a = 0 and

b = n.
(e) Prove by induction on m that

f (a, b) = f (m, n) ⇒ a = m and b = n.

where m, n, a, b ∈ N. Deduce that f is injective.

(f) Show that if b ∈ N and k = f (0, b), then k + 1 = f (b + 1, 0).
(g) Show by induction on k, where k ∈ N, that ∃m, n ∈ N, such that
f (m, n) = k.
(h) Deduce that f is bijective.

22. Show that Q is countable.

 45

23. Let A be a set.

(a) Show that if f : A → P (A) is a mapping and

B = {x ∈ A; x ∈
/ f (x)}

then B 6= f (a), for all a ∈ A.

(b) Deduce that there is no surjection from A onto P (A).

24. A is said to be uncountable if it is not countable. Show that P (N) is

uncountable.

25. We admit that R is uncountable. Show that the set I of irrational

numbers is uncountable.

26. Let n and k be non zero positive integers such that k ≤ 2n.
k+1 k (2n−2k−1) k+1 k k+1
(a) Show that C2n − C2n = (2n+1)
C2n+1 . Deduce that C2n ≤ C2n
if and only if k < n.
k n
(b) Conclude that C2n ≤ C2n ∀ k ≤ 2n.
(c) Using the binomial formula, calculate the expression
0 1 2n
C2n + C2n + . . . + C2n
4n n
Deduce that 2n+1
≤ C2n .

27. Let f be a mapping defined from a set E into E such that f (x) 6= x
for every x ∈ E. Set H = {A ⊆ E, f (A) ∩ A = ∅}.

(a) Let (An )n beSan increasing sequence (for inclusion) of elements in

H. Set A = n≥1 An . Show that A ∈ H.
(b) Suppose that E is finite.
i. Show that H admits an element of maximal cardinal N .
ii. Let z ∈ E such that z ∈ / f (N ) and set L = N ∪ {z}. Show
that f (L) ∩ L = {f (z)} ∩ N .
iii. Let x ∈ E such that x ∈ / f (N ) ∪ N . Show that f (x) ∈
N . Deduce that f (N ) ∪ N ∈ H, where f (N ) ∪ N is the
complement of f (N ) ∪ N in E.
iv. Conclude that there exists in E a subset F such that card(F ) ≥
card(E)
3
and f (F ) ∩ F = ∅.