SEQUENCES OF REAL NUMBERS

The notion of a sequence is central in Analysis. Among sequences, convergent se-

quences form a special important class. They are introduced and studied in this chapter.

1 Generality
Definition 1. • A sequence is an application u : N → R.
• For n ∈ N, u(n) is denoted by un and is called the n-th term or general term of the
sequence.

Example 1. ? 1, 21 , 13 , 14 , . . . are the terms of the sequence ( n1 )n∈N∗ .

? The sequence (un )n defined by un = u0 rn (u0 , r ∈ R) is called a geometric sequence where
u0 is the first term and r is the common ratio .
The sum of the n + 1 terms of the (un )n géométric sequence is given by

 u 1−rn+1 , r 6= 1
0 1−r
u0 + · · · + un =
 (n + 1)u , r = 1
0

? The sequence (vn )n defined by vn = v0 + nr (v0 , r ∈ R) is called an arithmetic sequence

of reason r and first term v0 . The sum of the n + 1 terms of the arithmetic sequence (vn )n
is given by
n+1
v0 + · · · + vn = (v0 + vn )
2

1
1.1 Bounded sequences

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

• (un )n∈N is said to be upper bounded if ∃M ∈ R, ∀n ∈ N, un 6 M .
• (un )n∈N is said to be lower bounded if ∃m ∈ R, ∀n ∈ N, un > m.
• (un )n∈N is said to be bounded if

∃M ∈ R, ∀n ∈ N, |un | 6 M.
1
Example 2. 1. The sequence un = is bounded, ∀n ∈ N∗ (M = 1 suffices).
n
2. The sequence ((−1)n )n∈N is bounded because, |(−1)n | = 1.

1.2 Monotone Sequences

Definition 3. A sequence (un )n∈N is called

• strictly increasing if ∀n ∈ N un+1 > un .

• increasing if ∀n ∈ N un+1 ≥ un .

• strictly decreasing if ∀n ∈ N un+1 < un .

• decreasing if ∀n ∈ N un+1 ≤ un .

• A sequence that is either increasing or decreasing is called monotone.

n
Example 3. ∀n ∈ N, un = is a strictly increasing sequence because:
n+1
n+1 n (n + 1)2 − n(n + 2) 1
un+1 − un = − = = > 0.
n+2 n+1 (n + 1)(n + 2) (n + 1)(n + 2)
Remark 1. If (un )n∈N is a sequence with strictly positive terms, it is increasing if and
only if ∀n ∈ N. un+1 /un > 1.
nn
Example 4. Let (un )n be the sequence defined by un = , ∀n ∈ N∗ .
n!
n  n
(n + 1)n+1 (n + 1)n+1 (n + 1)n

un+1 n! 1 n+1 1
= × = × = = = 1+
un (n + 1)! (n)n (n)n n+1 (n)n n n

n
As 1 + n1 > 1, so 1 + n1 > 1, ∀n ∈ N∗ . Thus (un )n is increasing.

2
2 Convergence of a sequence
Definition 4 (Finite limit). A sequence (un )n is said to converge to a number ` if and
only if
∀ε > 0, ∃n0 ∈ N, ∀n ∈ N, (n > n0 ⇒ |un − `| < ε)

This means that for any neighborhood of l from a certain rank, the Un terms are in the
neighborhood.
If the limit ` is unique, we write lim un = `.
n−→+∞

Example 5. Let (un )n be defined by

2n + 1
∀n ∈ N, un =
n+1

Let’s show that lim un = 2. Let ε > 0, show that there exists n0 ∈ N, such that for any
n→+∞
0
integer n, n > n ⇒ |un − 2| < ε. We have

2n + 1 −1 1
|un − 2| = −2 = =
n+1 n+1 n+1

Thus
1 1 1
|un − 2| < ε ⇔ <ε⇔n+1> ⇔n> −1 (1.1)
n+1 ε ε
We set n0 = − 1 + 1 = 1ε . Therefore, for any integer n, if n > n0 = 1ε , we’ll have
1     
ε
1
n> ε
− 1, so from (1), |un − 2| < ε. Hence (un )n is convergent.

3
Remark 2. A sequence that converges is said to be convergent. Otherwise, we say the
sequence diverges or that it is divergent.

Definition 5 (Infinite limit). • The sequence (un )n∈N tends to +∞ if :

∀A > 0 ∃N ∈ N ∀n ∈ N (n > N =⇒ un > A)

• The sequence (un )n∈N tends to −∞ if :

∀A > 0 ∃N ∈ N ∀n ∈ N (n > N =⇒ un 6 −A)

2n2
 
Example 6. The sequence is defined for all n ≥ 2 is divergent. By definition
n−1 n
2n2 2n2
 
lim = +∞ ⇔ ∀A > 0, ∃n0 ∈ N, ∀n ∈ N, n > n0 ⇒ >A
n→+∞ n − 1 n−1
∀n ≥ 2, We have
2n2 A
≥ 2n > A ⇔ n >
n−1 2
A

Let n0 = max 2 + 1, 2 , so if n > n0 then
A 2n2
n> ⇒ >A
2 n−1
Hence
2n2
lim = +∞
n→+∞ n − 1

Proposition 1. A convergent sequence has a unique limit.

Proof. We proceed by contradiction. Let (un )n∈N be a convergent sequence with two limits
|` − `0 |
` 6= `0 . Let’s choose  > 0 such that  < . Since lim un = `, there exists N1 such
2 n→+∞
that n > N1 implies |un − `| < . Similarly limn→+∞ un = `0 , there exists N2 such that
n > N2 implies |un − `0 | < . Let N = max (N1 , N2 ), then for this N we have:

|uN − `| <  and |uN − `0 | < 

So |` − `0 | = |` − uN + uN − `0 | 6 |` − uN | + |uN − `0 | according to the triangle inequality,

we get |` − `0 | ≤  +  = 2 < |` − `0 |.
We’ve just reached the impossible inequality |` − `0 | < |` − `0 |. Conclusion: our initial
hypothesis is false and therefore ` = `0 .

Example 7. The following sequence ((−1)n )n∈N = (1, −1, 1, −1, ...) does not converge to
any real number.

4
3 Further Useful Results
In this section we go over some basic results about the limits of sequences.

Proposition 2. A convergent sequence (un )n∈N is bounded

Proof. Suppose a sequence (un )n converges to `, so ∀ε > 0, ∃n0 ∈ N, ∀n ∈ N, n > n0 ⇒

|un − `| < ε . Let’s take ε = 1 (for example),

∃n1 ∈ N, ∀n ∈ N, n > n1 ⇒ |un − `| < 1

⇒ ∃n1 ∈ N, ∀n ∈ N, n > n1 ⇒ |un | ≤ |un − ` + `| < |un − `| + |`| ≤ 1 + |`|

 
i.e., − 1 − |l| ≤ un ≤ 1 + |l| , for all n ≥ n0

Furthermore, ∀n ∈ {0, 1, . . . , n − 1} such that

min(u0 , u1 , . . . , un−1 ) ≤ un ≤ max(u0 , u1 , . . . , un−1 )

Define M = max (|u0 | , |u1 | , . . . , |un0 | , 1 + |`|), we then have ∀n ∈ N, |un | ≤ M.

Remark 3. The condition given in the previous result is necessary but not sufficient. For
example, the sequence ((−1)n ) is a bounded sequence but it does not converge.

One naturally asks the following question:

Question : Boundedness + (??) ⇒ Convergence.
We now find a condition on a bounded sequence which ensures the convergence of the
sequence.

Theorem 3. Suppose (un ) is a bounded and increasing sequence. Then the least upper
bound of the set {un : n ∈ N} is the limit of (un ).

Proof. Suppose sup un = M . Then for given  > 0, there exists n0 such that M −  ≤ un0 .
n
Since (un ) is increasing, we have un0 ≤ un for all n ≥ n0 . This implies that

M −  ≤ un ≤ M ≤ M +  for all n ≥ n0 .

That is un → M .

5
 For decreasing sequences we have the following result and its proof is similar.

Theorem 4. Suppose (un ) is a bounded and decreasing sequence. Then the greatest lower
bound of the set {un : n ∈ N} is the limit of (un ).

Proposition 5. If lim un = `, lim vn = `0 and un ≤ vn (resp. un < vn ), for all

n→+∞ n→+∞
n ∈ N, then ` ≤ `0 .
1 1
Example 8. 1. let un = and vn = 1+ we have ∀n ∈ N, un ≤ vn with lim un = 0
n n n→+∞
and lim vn = 1.
n→+∞

1 1
2. Let (un ) and (vn ) be the following sequences un = and vn = , we have
n+1 n+2
vn < un for all n ∈ N but the two limits are equal to 0.

Theorem 6 (Squeeze Theorem or Sandwich rule). Suppose that {un }n∈N , {vn }n∈N , and
{ωn }n∈N , are sequences of real numbers such that

1. ∀n ∈ N un ≤ ωn ≤ vn

2. lim un = vn = `.
n→+∞

Then lim ωn = `
n→+∞

Proof. Let ε > 0. Since (un )n and (vn )n converge to `, there exist n1 , n2 ∈ N such that

∀n ∈ N, (n > n2 ⇒ |un − `| < ε)

∀n ∈ N, (n > n3 ⇒ |vn − `| < ε)

Let n3 = max (n1 , n2 ), so




 un ≤ ωn ≤ vn

∀n ∈ N, n > n3 ⇒ −ε < un − ` < ε


 −ε < v − ` < ε

n

This implies
−ε < un − ` ≤ ωn − ` ≤ vn − ` < ε

⇒ |ωn − `| < ε
So (ωn )n converges to `.

6
Example 9. (1) Let us consider the following sequence

cos n
vn = , for all n ∈ N∗ .
n
−1 cos n 1 −1 1
We have −1 ≤ cos n ≤ 1 then ≤ ≤ . Since lim = lim = 0 then
n n n n→+∞ n n→+∞ n

cos n
lim = 0.
n→+∞ n

(2) Let a sequence (un )n be defined by

1! + 2! + · · · + n!
Un =
(n + 2)!

For all k = 1, n,
n
X n
X n
X
k! ≤ n! ⇒ k! ≤ n! = n! 1 = nn!
k=1 k=1 k=1
so
n.n! n
0 < un ≤ =
(n + 2)! (n + 2)(n + 1)
Applying the previous theorem, we obtain

lim un = 0.
n→+∞

Proposition 7. Let (un )n be a sequence. We have

lim un = ` ⇒ lim |un | = |`|.

n→+∞ n→+∞

Proof. Let ε > 0. Since lim un = `, there exists an integer n0 such that
n→+∞

∀n ∈ N, n > n0 ⇒ |un − `| < ε

Now ∀n ∈ N, we have ||un | − |`|| ≤ |un − `|, so

∀n ∈ N, n > n0 ⇒ ||un | − |`|| < |un − `| < ε

Remark 4. The converse is in general wrong. Taking the previous example, let’s consider
the sequence (un )n defined by un = (−1)n . We have lim |un | = 1, but the limit of (un )n
n→+∞
does not exist.

7
Proposition 8. Let (un )n , (vn )n be two sequences such that (un )n is bounded and lim vn =
n→+∞
0 . Then lim un vn = 0.
n→+∞

Proof. The sequence (un )n is bounded, so we can find an réel M > 0 such that ∀n ∈
N, |un | ≤ M . Let ε0 = ε
M
> 0; since vn → 0, there exists n0 ∈ N such that

ε
∀n ∈ N, n > n0 ⇒ |vn | < ε0 =
M

So
ε
∀n ∈ N, n > n0 ⇒ |un vn | ≤ |un | |vn | ≤ M = ε.
M
Hence lim un vn = 0.
n→+∞

n sin(n!)
Example 10. Let (un )n be a sequence defined by un = 2 .
n +1
n
We set vn = sin(n!) and ωn = 2 , since vn = sin(n!) is bouneded and
n +1
n 1
lim ωn = lim = lim 1 = 0.
n→+∞ n→+∞ n2 + 1 n→+∞ n +
n

Therefore lim un = 0.
n→+∞

4 Continuity of algebraic operations

Proposition 9 (Sum Rule, Product Rule and Quotient Rule). Let λ, ` and k ∈ R.
Suppose (un ), (vn ) be convergent sequences such that ` = lim un and k = lim vn .
n→+∞ n→+∞
Then:



 λ` if ` ∈ R


 0 if λ = 0 and ` 6= ±∞

1) lim (λun ) =
n→+∞ 

 +∞ if ` = +∞ and λ > 0 or ` = −∞ and λ < 0


 −∞ if ` = −∞ and λ > 0 or ` = +∞ and λ < 0




 ` + k if `, k ∈ R,

2) lim (un + vn ) = +∞ if ` 6= −∞ and k = +∞ or ` = +∞ and k 6= −∞,
n→+∞ 

 −∞ if ` 6= +∞ and k = −∞ or ` = −∞ and k =

6 +∞.

8
 



 ` · k if `, k ∈ R,





 +∞ if ` > 0 and k = +∞ or ` = +∞ and k > 0


3) lim (un · vn ) = or ` < 0 and k = −∞ or ` = −∞ and k < 0
n→+∞ 


−∞ if ` > 0 and k = −∞ or ` = −∞ and k > 0







or ` < 0 and k = +∞ or ` = +∞ and k < 0


`


 if ` ∈ R and k ∈ R∗ ,



 k
+∞ if ` > 0 and k = 0+ or ` < 0 if k = 0− or





un 
4) lim = ` = +∞ and k > 0 or ` = −∞ if k < 0,
n→+∞ vn 


−∞ if ` < 0 and k = 0+ or ` > 0 if k = 0−








 ` = −∞ and k > 0 or ` = +∞ if k < 0.

There is another useful way we can express all these rules: If (an ) and (bn ) are con-
vergent then

lim (can + dbn ) = c lim an + d lim bn Sum Rule

n→∞ n→∞ n→∞

lim (an bn ) = lim an · lim bn Product Rule

n→∞ n→∞ n→∞

an
 lim (an )
lim = n→∞ , if lim (bn ) 6= 0 Quotient Rule .
n→∞ bn lim (bn ) n→∞
n→∞

9
Example 11. In full detail

1 + n12 6 − n1



(n2 + 1) (6n − 1)
lim = lim
n→∞ 2n3 + 5 n→∞ 2 + n53
using the Quotient Rule
  
1 1
lim 1+ 2 6−
n→∞ n n
=  
5
lim 2 + 3
n→∞ n
using the Product and Sum Rules
     
1 1
1 + lim 2
6 − lim
n→∞ n n→∞ n
=  
1
2 + 5 lim
n→∞ n3
(1 + 0)(6 − 0)
=
2+0
=3

Unless you are asked to show where you use each of the rules you can keep your solutions
simpler. Either of the following is fine:

1 + n12 6 − n1



(n2 + 1) (6n − 1) (1 + 0)(6 − 0)
lim = lim 5 = =3
n→∞ 2n3 + 5 n→∞ 2 + n3 2+0

Remark 5. By induction, the Sum and Product Rules can be extended to cope with any
finite number of convergent sequences. For example, for three sequences:

lim (un · vn · ωn ) = lim un · lim vn · lim ωn .

n→∞

5 Subsequences
Definition 6. Let (un )n∈N be a sequence. A subsequence (vn )n∈N of (un )n∈N is defined by
a function ϕ : N → N such that φ is strictly increasing and vn = uφ(n) for n ∈ N.
In other words, a subsequence is a sequence (vn )n∈N , where each vn is in (un ), and the
terms are in the right order.

Example 12. Let un = n for n ≥ 1. The following are subsequences of (un ).

• 2, 4, 6, 8, . . . the subsequence (u2n )

10
• 2, 4, 8, 16, . . . the subsequence (u2n )
The following are not subsequences of (un ).
• 6, 4, 8, . . . the terms are not in the right order
• 2, 4, 0, . . . not all the terms are in (un )
• 1, 2, 3, . . . , 2020 finite so not a sequence.

Exercise 1. Let un = n2 . Write down the first four terms of the three subsequences
(un+4 ), (u3n−1 ) and (u2n ).

Proposition 10 (Subsequences of a convergent sequence). Let (un ) be a sequence. If (un )

converges, then every subsequence vn = uφ(n) of (un ) converges. Moreover, if un → L
as n → ∞ then every subsequence also converges to L.

Remark 6. (1) So if (un ) is a sequence, and it has two subsequences that tend to different
limits, then (un ) does not converge. This follows from Proposition 16, and can be a useful
strategy for showing that a sequence does not converge.

(2) A sequence (un )n is convergent if and only if the subsequences (u2n ) and (u2n+1 ) are
convergent and have the same limit.

Example 13. The divergent sequence ((−1)n )n∈N admits the following convergent subse-
quences u2n = 1, u2n+1 = −1 for all n ∈ N.

6 Adjacent sequences
Definition 7. Let (un ) (vn ) be two real sequences. We say that the two sequences are
adjacent if the first is increasing, the second is decreasing, and their difference converges
to 0. In other words
1) The sequence (un ) is increasing and the sequence (vn ) is decreasing,
2) The difference (vn − un ) converges to 0, when n approaches +∞.

Theorem 11. Two adjacent sequences converge, and converge to the same limit.

Proof. We have

u0 6 u1 6 u2 6 · · · 6 un 6 · · · · · · 6 vn 6 · · · 6 v2 6 v1 6 v0

11
The sequence (un )n∈N is increasing and bounded above by v0 , so it converges to a limit `.
The sequence (vn )n∈N is decreasing and bounded below by u0 , so it converges to a limit
`0 . Therefore `0 − ` = lim (vn − un ) = 0, hence `0 = `.
n→+∞

Example 14. Let (un )n and (vn )n be two sequences defined by, ∀n ∈ N∗ ,
n
1 1 1 1 X 1
un = + + = . . . =
1! 2! 3! n! k=1 k!

1
textandvn = un +
n.n!
We have a, ∀n ∈ N∗ ,
n+1 n
X 1 X 1
un+1 − un = −
k=1
k! k=1 k!
   
1 1 1 1 1
= 1 + + ··· + + − 1 + + ··· +
2! n! (n + 1)! 2! n!
1
= >0
(n + 1)!

So (un )n is strictly increasing and

1 1
vn+1 − vn = un+1 + − un −
(n + 1) · (n + 1)! n · n!
1 1 1
= + −
(n + 1)! (n + 1) · (n + 1)! n · n!
1 1 1
= + −
(n + 1) · n! (n + 1) · (n + 1) · n! n · n!

n(n + 1) n (n + 1)2
vn+1 − vn = = + −
(n + 1)2 · n · n! (n + 1)2 · n · n! (n + 1)2 · n · n!
n(n + 1) + n − (n + 1)2
=
(n + 1)2 · n · n!
−1
= <0
(n + 1)2 · n · n!
So (vn )n is strictly decreasing. Furthermore, we have

−1
lim (un − vn ) = lim =0
n→+∞ n→+∞ n · n!

1
Moreover, it is clear that un ≤ vn (because we have vn = un + n·n!
). Hence (un )n and
(vn )n are adjacent.

12
7 Recursive Formulas for Sequences
Definition 8. Let f : I ⊂ R → R. A recursive sequence is a sequence in which terms
are defined using one or more previous terms which are given by u0 ∈ I and the relation
∀n ∈ N : un+1 = f (un )
We suppose that f (I) ⊂ I, and so the sequence is well defined.
√ √
Example 15. Let un+1 = un + 1, u0 = 1, we have f (x) = x + 1, D = [−1, +∞[ et
f (I) = [0, +∞ [⊂ I . Thus un is well defined.

Lemma 1. If f is continuous on I and the sequence {un } converges to ` ∈ I, then

` = f (`) “` is called the fixed point of f ”.

Theorem 12. Let {un } be a real sequence recursively defined by un+1 = f (un ).

1. If f is increasing then the sequence (un ) is monotone. More precisely (un ) is mono-
tone:

• If u0 6 u1 , then the sequence is increasing.

• If u0 > u1 , then the sequence is decreasing.

2. If f is decreasing then the sequence un+1 − un , is positive and negative alternatively.

We set g = f of , so g is increasing, we can easily verify that the sequences (u2n ) and
(u2n+1 ) defined by u2n+2 = f (f (u2n )) , u2 = f (u1 ) and u2n+1 = f (f (u2n−1 )) , u1
given, are oppositely monotone such as that g (u1 ) − u1 = f (f (u1 )) − u1 and
g (f (u1 )) − f (u1 ) = f (f (f (u1 ))) − f (u1 ) have different signs.

Example 16 (Tutorials). Consider the sequence (un )n defined by u0 = 1, un+1 =

√
1 + un , n ∈ N.

1. Show that 0 ≤ un ≤ 3, ∀n ∈ N.

2. Study the monotonicity and the convergence of the sequence (un )n . Find its limit.

Exercise 2 (Final Exam Review 2022, Lic Math). Consider the sequence (un )n defined
2
by u0 = 1; un+1 = 1 + , n ∈ N.
1 + un

13
 1. Show that 1 ≤ un ≤ 3, ∀n ∈ N.

2. Is the sequence (un )n monotonic?

3. Study the nature of the subsequences (u2n )n and (u2n+1 )n .

4. Is the sequence (un )n convergent? If so, determine its limit.

Answer. 1) We show that 1 ≤ un ≤ 3 , ∀n ∈ N.

Consider the proposition P(n) : 1 ≤ un ≤ 3, ∀n ∈ N.
Let us show by induction that P(n) is true.
i) we have u0 = 1 and 1 ≤ u0 ≤ 3, so P(0) is true.

ii) we suppose that P(n) is true and show P(n + 1). We have

1 ≤ un ≤ 3 ⇒2 ≤ 1 + un ≤ 4
1 1 1
⇒ ≤ ≤
4 1 + un 2
2 2
⇒ ≤ ≤1
4 1 + un
3 2
⇒ ≤ +1≤2
2 1 + un
2
⇒1≤ +1≤3
1 + un

Hence
1 ≤ un+1 ≤ 3.

Thus P(n) is true ∀n ∈ N.

2) The monotonocity of (un )
2
We have un+1 = f (un ), u0 = 1, where f (x) = 1 + , x ∈ [1; 3]. The function f is
x+1
differentiable on [1; 3] (f ([1; 3]) ⊂ [1; 3]) and we have: ∀x ∈ [1; 3],

−2
f 0 (x) = < 0,
(x + 1)2

so f is decreasing on [1; 3]. Therefore (un ) is not monotonic sequence.

3) Since f is decreasing then the subsequences (u2n ) and (u2n+1 ) are oppositely monotone.
Indeed, since (un )n is bounded, the subsequences (u2n )n and (u2n+1 )n are also bounded.
Then (u2n )n and (u2n+1 )n are convergent sequences.

14
4) The subsequences (u2n )n and (u2n+1 )n converge to a fixed point f ◦ f .
Calculate the fixed points of g = f ◦ f (` = lim u2n and ` = lim u2n+1 )
n→+∞ n→+∞
For all ` ∈ [1; 3] ` = g(`) = (f ◦ f )(`). We have
2 3 + 2`
g(`) = f (f (`)) = 1 + 2 =
1 + (1 + 1+`
) 2+`
So
3 + 2` √ √
g(`) = ` ⇔ = ` ⇔ 3 + 2` = `2 + 2` ⇔ ` = 3 or ` = − 3
2+`
Since the subsequences (u2n ) and (u2n+1 ) converge to the same limit ` then the sequence
(un )n also converges to ` (the two subsequences (u2n )n and (u2n+1 )n are positive, then their
√ √
limits are also positive, so the only possible limit is equal to 3). Hence lim un = 3.
n→+∞

8 Cauchy sequences
Is it possible to know whether a sequence converges (to a finite limit) without knowing
its limit? The notion of a Cauchy sequence answers this question. It reflects the intuitive
idea that the terms of a convergent sequence must be close to each other from a certain
rank

Definition 9. A sequence (un )n is said to be Cauchy if

 
 p>n ,
0
∀ε > 0, ∃n0 ∈ N, ∀p, q ∈ N,  ⇒ |up − uq | < ε
 q>n ,
0

Since p, q ∈ N, we can assume that (for example) p ≥ q, so the previous condition is

equivalent to

∀ε > 0, ∃n0 ∈ N, ∀p, q ∈ N, (p ≥ q > n0 , ⇒ |up − uq | < ε).

1 n



Example 17. 1) The geometric sequence 2 n∈N
is a Cauchy sequence.
We have, for p > q > 0,
 p  q  q  p−q  q  p−q !
1 1 1 1 1 1
− = −1 < because −1 <1
2 2 2 2 2 2

So  q  
1 ln 
<⇒q> .
2 − ln 2

15
Taking  
ln 
n0 = + 1.
− ln 2
1 n



Then 2 n∈N
is a Cauchy sequence.
sin 1 sin 2 sin n
2) The sequence un = + 2 · + n is a Cauchy sequence.
2 2 2
If p > q, we have

sin(q + 1) sin(q + 2) sin n

|up − uq | = q+1
+ q+2
+ . . . + n , (p = q + k)
2 2 2
1 1 1
6 q+1 + q+2 + . . . + n
2 2 2
1 1 1
< q+1 + q+2 + . . . + n
2 2 2
1
(is a sum of a geometric sequence of the common ratio 1/2 with first term q+1 )
2
1 p−q 
−

1 1 ( 2
) 1 1 1
= q+1 1 = q− p ≤ q
2 2
2 2 2

1
Hence ∀ε > 0 choosing n0 ∈ N such that < ε. So (un )n is a Cauchy sequence.
2n0
Remark 7. To show that a sequence (Un )n is not Cauchy, we need to find some ε > 0,
so that for each n0 ∈ N there are some p, q ≥ n0 with |Up − Uq | ≥ 0.

i.e., ∃ > 0, ∀n0 ∈ N, ∃(p, q) ∈ N2 (p > q > n0 et |Up − Uq | ≥ ε).

Example 18. (a) Let (un = (−1)n ) and let ε = 1. For each n0 ∈ N, we put p = 2n0 and
q = 2n0 + 1. Note that |(−1)p − (−1)q | = |1 − (−1)| = 2 > ε. Therefore, ((−1)n ) is not
Cauchy.
(−1)n
(b) Let un = n + n
, and let ε = 1. For each n0 ∈ N, we put p = 2n0 + 2 and
q = 2n0 + 1. Then,
 
1 1
|up − uq | = 2n0 + 2 + − 2n0 + 1 −
2n0 + 2 2n0 + 1
1 1
=1+ +
2n0 + 2 2n0 + 1
>ε

Therefore, (un ) is not a Cauchy sequence.

16
(c) Let xn = ln n, and let ε = ln 2. For each n0 ∈ N, we put p = 2n0 and q = n0 . Then,
we have
|xp − xq | = | ln(2n0 ) − ln n0 | = ln 2 = ε

Therefore, (xn ) is not Cauchy.

Theorem 13 (Bolzano-Cauchy Theorem). A sequence of real numbers is convergent if

and only if it is Cauchy.

Proof. We first prove that every convergent sequence is Cauchy.

Let (un )n be a convergent sequence and let lim un = `.

n→+∞
Then for any ε > 0, there exists a positive integer n0 such that

|un − `| < ε/2 for all n ≥ n0

For any p, q ≥ n0 , one has

|up − uq | = |up − ` + ` − uq | ≤ |up − `| + |uq − `| ≤ ε/2 + ε/2 = ε.

Thus, (un )n is a Cauchy sequence.

• The proof of the converse of the theorem is very long and technical, we refer the reader
to “Eléments d’analyse de Kada Allab” for a proof.

Exercise 3. Let (un )n be a sequence defined by:


 u0 = α, ,α > 0
√ (1.2)
 un+1 = 3un + 4, ∀n ∈ N.

1) We assume that 0 < α < 4.

1. Show that 0 < un < 4, ∀n ∈ N.

2. Study the monotonicity of the sequence (un )n .

3. Is the sequence (un )n convergent?

4. Let A = {un , n ∈ N}. Determine, whenever these exist, sup A, inf A, min A and
max A.

2) We assume that α = 4. Study the nature pf the sequence (un )n .

17