Chapter 3

Sequences of real numbers

3.1 Definitions and examples

Definition 1
A sequence of real numbers is a real-valued function whose domain is the set of natural numbers N
or an infinite subset N1 ⊂ N to the real numbers i.e:
u : N −→ R u : N1 −→ R
or
n 7−→ u(n) n 7−→ u(n)

Notations:

• For n ∈ N, u(n) is denoted by un and is called the general term or n-th term of the sequence.

• The sequence u is denoted by (un )n∈N or (un )n∈N1 .

Example 1
1 1 1
1 The sequence (un )n∈N∗ defined by: un = , starts with u1 = 1, and u2 = , u3 = ,......
n 2 3

u1 = 1
2 The recurrent sequence defined by: 1 starts with u1 = 1, and u2 = 2, u3 =
un = 1 +
un−1
3
,....
2

Remark
The ways in which a sequence can be defined.

• By an explicit definition of the general term of the sequence (un ) i.e.: Express un in terms of n.
2n + 1
For example, un = .
n+7
• By a recurrence formula, i.e. a relationship that links any term in the sequence to the one that
precedes it. In this case, to calculate un , you need to calculate all the terms that precede it. For
example :

33
 (
u0 = 2
un+1 = 3un − 1

3.2 Bounded sequences

Definition 2
Let (un )n∈N be a real sequence.

• A sequence (un )n∈N is bounded from above iff: ∃M ∈ R, ∀n ∈ N; un ≤ M

• A sequence (un )n∈N is bounded from below iff: ∃m ∈ R, ∀n ∈ N; m ≤ un

• A sequence (un )n∈N is bounded iff: it is bounded from above and bounded from below which
means :
∃M ∈ R+ , ∀n ∈ N; |un | ≤ M

3.3 Increasing and decreasing sequences

Definition 3
Let (un )n∈N be a sequence

• (un )n∈N is an increasing sequence iff: ∀n ∈ N; un ≤ un+1

• (un )n∈N is a strictly increasing sequence iff: ∀n ∈ N; un < un+1

• (un )n∈N is a decreasing sequence iff: ∀n ∈ N; un ≥ un+1

• (un )n∈N is a strictly decreasing sequence iff: ∀n ∈ N; un > un+1

• (un )n∈N is monotonic if it is increasing or decreasing.

• (un )n is strictly monotonic if it is strictly increasing or strictly decreasing.

• (un )n∈N is a constant sequence iff ∀n ∈ N; un+1 = un

3.4 Finite and infinite limit of a numerical sequence

Definition 4: Convergent sequences
Let (un )n∈N be a real sequence. We say that the sequence (un )n∈N converges to l iff:

∀ε > 0, ∃n0 ∈ N, ∀n ∈ N; n ≥ n0 =⇒ |un − l| ≤ ε

In this case, we say that the sequence (un )n∈N is convergent to the limit l and we note lim un = l
n→+∞

34
 Remark

|un − l| ≤ ε ⇔ l − ε ≤ un ≤ l + ε ⇔ un ∈ [l − ε, l + ε]
The above definition means that for any strictly positive real ε, there exists an integer n0 (rank) such
that: all terms un0 , un0 +1 , un0 +2 .... are in the interval [l − ε, l + ε].

Example 2
n
• The sequence un = converges to 1
n+1
Using the definition of convergence, we show that lim un = 1
n→+∞
Let ε > 0 we have:

|un − 1| ≤ ε
n
⇔| − 1| ≤ ε
n+1
n
⇔| − 1| ≤ ε
n+1
1
⇔ |1 − − 1| ≤ ε
n+1
1
⇔ ≤ε
n+1
1
⇔ −1 ≤ n
ε
1 1
By setting n0 = ⌊ ⌋ > − 1, we obtain :
ε ε
1
∀ε > 0, ∃n0 ∈ N (n0 = ⌊ ⌋), ∀n ∈ N; n ≥ n0 =⇒ |un − 1| ≤ ε
ε
=⇒ (un )n∈N converges to l = 1
Using Maple, we get the following graph:

Figure 3.1: ε = 0.1

35
 Definition 5
1 We say that the sequence (un )n∈N tends to +∞ as n tends to infinity and we note lim un = +∞
n→+∞
iff:
∀A > 0, ∃n0 ∈ N, ∀n ∈ N; n ≥ n0 =⇒ un ≥ A

2 We say that the sequence (un )n∈N tends to −∞ as n tends to infinity and we note lim un = −∞
n→+∞
iff:
∀A > 0, ∃n0 ∈ N, ∀n ∈ N; n ≥ n0 =⇒ un ≤ −A

Example 3
• Let be the following sequences : (
un = 2n + 1
vn = −3n + 4

We show that lim un = +∞ and lim vn = −∞

n→+∞ n→+∞

1 Let A > 0 we have:

un ≥ A
⇔ 2n + 1 ≥ A
⇔ 2n ≥ A − 1
A−1
⇔ 2n ≥
2

A−1 A−1
Let’s put n0 = [ ]+1 >
2 2
A−1
=⇒ (∀A > 0, ∃n0 ∈ N (n0 = [ ] + 1), ∀n ∈ N; n ≥ n0 =⇒ un ≥ A)
2

2 The same method used for the sequence (vn )n∈N

Definition 6: divergent sequences

Let (un )n∈N be a sequence of real numbers. We say that the sequence (un )n∈N is divergent if it is not
convergent, i.e

∀l ∈ R, ∃ε > 0, ∀n0 ∈ N, ∃n ∈ N; (n ≥ n0 ) ∧ (|un − l| > ε)

Remark
here are two types of divergence

1 Divergence of infinite type: in this case the sequence converges to +∞ or −∞. For example the
sequence with general term un = 2n + 4.

2 Divergence of type limit does not exist: in this case the sequence has no finite or infinite limit.

36
 For example, the sequence with general term un = (−1)n

Proof:
We will show that the sequence (−1)n does not have a finite or infinite limit.

1 By contradiction, suppose that: lim (−1)n = l /l ∈ R. According to the convergence definition with
n→+∞
1
ε = we get:
4
1 1
∃n0 ∈ N, ∀n ∈ N; n ≥ n0 =⇒ un ∈ [l − , l + ]
4 4
1 1
=⇒ −1, 1 ∈ [l − , l + ]
4 4
1 1


 l− ≤1≤l+
 4
 4
=⇒
l − 1 ≤ −1 ≤ l + 1



4 4
1 1


 l− ≤1≤l+
 4
 4
=⇒
−l − 1 ≤ 1 ≤ −l + 1



4 4

1 1
=⇒ − ≤ 2 ≤
2 2
It’s a contradiction.

2 By contradiction, suppose that: lim (−1)n = +∞. According to the convergence definition with
n→+∞
A = 4 we get:
∃n0 ∈ N, ∀n ∈ N; n ≥ n0 =⇒ un ≥ 4
=⇒ un ∈ [4, +∞[ =⇒ −1, 1 ∈ [4, +∞[
It’s a contradiction.

3 We use the same method for the case: lim (−1)n = −∞

n→+∞

Proposition 1:
If a sequence of real numbers (un )n∈N has a limit, then this limit is unique.

Proof:

By contradiction

 lim un = l1
Suppose that:; n→+∞
 lim un = l2
n→+∞
|l1 − l2 |
Taking ε = with l1 ̸= l2 which implies
4
(
∃n1 ∈ N, ∀n ∈ N; n ≥ n1 =⇒ |un − l1 | ≤ ε
∃n2 ∈ N, ∀n ∈ N; n ≥ n2 =⇒ |un − l2 | ≤ ε

37
Putting n0 = max(n1 , n2 )

=⇒ (∀n ∈ N; n ≥ n0 =⇒ |un − l1 | + |un − l2 | ≤ 2ε)

With n ≥ n0 we get

|l1 − l2 | ≤ |un − l1 | + |un − l2 | ≤ 2ε

=⇒ |11 − l2 | ≤ 2ε
|11 − l2 | ε
=⇒ ≤
4 2
ε
=⇒ ε ≤ it’s a contradiction
2

Proposition 2
If (un )n∈N is a convergent sequence, then (un )n∈N is a bounded sequence.

Proof:

We’ll show the following implication:

(un )n∈N is a convergent sequence =⇒ (un )n∈N is bounded

Suppose that (un )n∈N is convergent, then for ε = 1 we have:

∃n0 ∈ N, ∀n ∈ N; n ≥ n0 =⇒ |un − l| ≤ 1
=⇒ m = l − 1 ≤ un ≤ l + 1 = M

So the set {un0 , un0 +1 , .......} is bounded.

On the other hand A = {u0 ....., un0 −2 , un0 −1 } is bounded (because Card(A) < +∞). Then the set of values
of (un ) is: {u0 ....., un0 −2 , un0 −1 , un0 , un0 +1 , .......} is bounded, this means (un ) is bounded.

3.5 Finding Limits: Properties of Limits

Theorem 1
Let (un )n∈N and (vn )n∈N two convergent sequences with: lim un = l and lim vn = l ′ . The
n→+∞ n→+∞
properties of limits are summarized as follows:

1 lim λ un = λ l with λ ∈ R
n→+∞

2 lim (un + vn ) = l + l ′
n→+∞

3 lim un vn = ll ′
n→+∞

1 1
4 If un ̸= 0 for n ≥ n0 and l ̸= 0 then lim =
n→+∞ un l
un l
5 If vn ̸= 0 for n ≥ n0 and l ′ ̸= 0 then lim = ′
n→+∞ vn l

38
 Remark
lim un = l =⇒ lim |un | = |l|. Be careful the reverse is not true. For example, if we take the
n→+∞ n→+∞
sequence un = (−1)n we have lim |un | = 1 but lim un doesn’t exist.
n→+∞ n→+∞

Proposition 3: Infinite limit’s operations

Let (un )n∈N , (vn )n∈N two sequences with: lim un = +∞ and lim vn = +∞ then:
n→+∞ n→+∞

1 lim (un + vn ) = +∞
n→+∞

1
2 If ∀n ≥ n0 , un ̸= 0 then lim =0
n→+∞ un

3.6 Limits and inequalities

Theorem 2
1 Let (un )n∈N , (vn )n∈N be two convergent sequences, then:
If ∃n0 ∈ N, ∀n ≥ n0 ; un ≤ vn this implies lim un ≤ lim vn
n→+∞ n→+∞


∃n0 ∈ N, ∀n ≥ n0 ; un ≤ vn


2 If, we have (un )n∈N and (vn )n∈N two sequences which verify:- and

 lim un = +∞

n→+∞
this implies lim vn = +∞
n→+∞

3 Squeeze Theorem : If (un )n∈N , (vn )n∈N and (wn )n∈N three sequences with:

∃n0 ∈ N, ∀n ≥ n0 ; un ≤ vn ≤ wn


and

 lim un = lim wn = l

n→+∞ n→+∞

then the sequence (vn )n∈N is convergent and lim vn = l

n→+∞

3.7 Convergence theorems

Theorem 3: Convergence of monotonic sequences
• If a sequence of real numbers is increasing and bounded from above, then it converges.

• If a sequence of real numbers is decreasing and bounded from below, then it converges.

39
Example 4

u0 = 1
Let (un )n∈N be a numerical sequence defined by: 2 .
un+1 = 1 + un
2

1 Prove that ∀n ∈ N; un ≤ 1
2 Deduce that the sequence (un )n∈N is convergent.
1
• by using proof by induction, we have for n = 0, u0 = ≤ 1 so the proposition is true. Let’s
2
assume that the proposition is true for k ∈ {1, ..., n} and we’ll show that un+1 ≤ 1. According
to the assumption we have:

1 + u2n
un ≤ 1 =⇒ u2n ≤ 1 =⇒ 1 + u2n ≤ 2 =⇒ ≤ 1 =⇒ un+1 ≤ 1
2

So, assertion ∀n ∈ N; un ≤ 1 is true.

1 + u2n (un − 1)2
• On a ∀n ∈ N; un+1 − un = − un = ≥0
2 2
• Since (un )n∈N is increasing and bounded from above so (un )n∈N is convergent.

Definition 7: Adjacent sequences

Let (un )n∈N and (vn )n∈N be two real sequences. We say that (un )n∈N and (vn )n∈N are adjacent iff:


(un )n∈N is increasing




 and
(vn )n∈N is decreasing



 and

 lim (un − vn ) = 0

n→+∞

Theorem 4:
If the sequences (un )n∈N and (vn )n∈N are adjacent then they converge to the same limit.

Example 5
n 1 2
The sequences un = ∑ 2
and vn = un + are adjacent :
k=1 k n

n+1 1 n 1 1
• un+1 − un = ∑ 2
− ∑ 2= ≥ 0 =⇒ (un )n∈N is increasing
k=1 k k=1 k (n + 1)2
1 2 2 (n + 2)
• vn+1 − vn = 2
+ − =− ≤ 0 =⇒ (vn )n∈N is decreasing
(n + 1) n+1 n n(n + 1)2

40
 2
• lim (un − vn ) = lim (− ) = 0
n→+∞ n→+∞ n
Therefore the sequences (un )n∈N and (vn )n∈N are convergent to the same limits.

Figure 3.2: (un ) and (vn ) are adjacent

Definition 8: Cauchy sequence

Let (un )n∈N be a sequence of real numbers.
(un )n∈N is called a Cauchy sequence in R iff:

∀ε > 0, ∃n0 ∈ N, ∀p, q ∈ N; p, q ≥ n0 =⇒ |u p − uq | ≤ ε

Remark
|u p − uq | ≤ ε ⇔ the distance between u p , and uq is less than ε.
So the definition above means that:- for any strictly positive real ε, there exists n0 (rank), such that
the distance between each two terms u p ,uq (with p, q ≥ n0 ) is less than ε.

Using Maple, we obtain the following graph of a Cauchy sequence:

41
 ‘
cos(n) + sin(n) + n
Figure 3.3: un = , ε = 0.08
n

Example 6
1
• un = is a Cauchy sequence
n

Let p, q ∈ N∗ with p ≤ q then we have:

1 1 1 1
|u p − uq | = − ≤ + according to the triangular inequality
p q p q
2 1 1
=⇒ |u p − uq | ≤ ( because: ≤ )
p q p
 
2 2
Let ε > 0, we put n0 = +1 >
ε ε
So, ∀ε > 0, ∃n0 ∈ N∗ , ∀p, q ∈ N∗ ; p, q ≥ n0 =⇒ |u p − uq | ≤ ε

Theorem 5:
Let (un )n∈N be a real sequence then:
(un )n∈N is a Cauchy sequence ⇐⇒ (un )n∈N is convergent

3.8 Subsequence
Definition 9
The sequence (uφ (n) )n∈N is a subsequence of the sequence (un )n∈N if φ : N −→ N is a strictly
increasing sequence of of natural numbers.

42
Example 7

2n
u2n = (−1) = 1

1 un = (−1)n −→

u2n+1 = (−1)2n+1 = −1

(u2n )n∈N and (u2n+1 )n∈N are subsequences taken from (un )n∈N

2 vn = cos( nπ3 ) −→ v3n = cos(nπ) = (−1)n

(v3n )n∈N is a sub-sequence of (vn )n∈N

Proposition 4:
Let (un )n∈N be a sequence of real numbers:

1 If lim un = l, then for any subsequence (uφ (n) )n∈N ; lim uφ (n) = l
n→+∞ n→+∞

2 If (un )n∈N admits a divergent subsequence then (un )n∈N is divergent

3 If (un )n∈N has two subsequences converging to distinct limits then (un )n∈N is divergent.

Example 8
the sequence with general term un = (−1)n is divergent:
We have:

lim u = 1

u2n = 1 n→+∞ 2n
 

and =⇒ and
 
u2n+1 = −1  lim u2n+1 = −1
 
n→+∞

So,(u2n )n∈N and (u2n+1 )n∈N are two subsequences of (un )n∈N which converge to distinct limits,
therefore (un )n∈N is divergent.

Theorem 6: Bolzano-Weierstrass Property

Every bounded sequence has a convergent sub-sequence.

Definition 9: Cluster Points of the sequence

A cluster Point of a numerical sequence (un )n∈N is any scalar which is the limit of a subsequence of
(un )n∈N .

Example 9
• Let’s consider the sequence (un )n∈N defined by: un = cos(n π2 )

43
 
u4n = cos(2nπ) = 1 =⇒ lim u4n = 1
n→+∞




u4n+1 = cos( π2 ) = 0 =⇒ lim u4n+1 = 0

n→+∞
u4n+2 = cos(π) = −1 =⇒ lim u4n+2 = −1
n→+∞



 π
u4n+3 = cos(3 2 ) = 0 =⇒ lim u4n+3 = 0

n→+∞

So the sequence (un )n∈N is divergent. The numbers 1, −1, 0 are the cluster points of the sequence
(un )n∈N .

3.9 Limit inferior and limit superior

Definition 10
Let (un )n∈N be a sequence of real numbers.
Denoting by S = The set of cluster points of the sequence (un )n∈N .
We define the limit superior (resp. inferior) of (un )n∈N as

lim sup un = sup S



lim inf un = inf S


Example 10
Let (un )n∈N defined by: un = (−1)n
The set of all cluster points of the sequence (un )n∈N is S = {1, −1}
so, lim sup un = 1, and lim inf un = −1

44