Real Analysis

– in a nutshell –

This book was created and used for the lecture Math-
ematical Analysis at Hamburg University of Tech-
nology in the summer term 2019 for General En-
gineering Science and Computer Science students.

Hamburg, 24th June 2024, Version 1.3

Julian P. Großmann
Hamburg University of
Technology
[email protected]
ii

The author would like give special thanks

• to Timo Reis, Florian Bünger, Anton Schiela and Francisco Hoecker-Escuti for an
excellent documentation of the Mathematical Analysis course held at Hamburg Uni-
versity of Technology before 2019,
• to Fabian Gabel and Jan Meichsner for many corrections and remarks,
• to all students who pointed out typos and other problems in this script.

Hamburg, 24th June 2024 J.P.G.

 Contents

1 Sequences and Limits 5

1.1 Just Numbers . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 5
1.2 Convergence of Sequences . . . . . . . . . . . . . . . . . . . . . . . . . . . 11
1.3 Subsequences and accumulation values . . . . . . . . . . . . . . . . . . . . 24
1.4 Cauchy Sequences . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 29
1.5 Open, Closed and Compact Sets . . . . . . . . . . . . . . . . . . . . . . . . 30

2 Infinite Series 35
2.1 Basic Definitions, Convergence Criteria and Examples . . . . . . . . . . . . 36
2.2 Absolute Convergence and Criteria . . . . . . . . . . . . . . . . . . . . . . 41
2.3 The Cauchy Product of Series . . . . . . . . . . . . . . . . . . . . . . . . . 49

3 Continuous Functions 53
3.1 Bounded Functions, Pointwise and Uniform Convergence . . . . . . . . . . 53
3.2 Continuity . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 57

4 Elementary Functions 65
4.1 Exponential Function . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 65
4.2 Logarithm . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 68
4.3 Hyperbolic and Trigonometric Functions . . . . . . . . . . . . . . . . . . . 71
4.3.1 Hyperbolic Functions . . . . . . . . . . . . . . . . . . . . . . . . . . 71
4.3.2 Area Functions . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 75
4.3.3 Trigonometric Functions . . . . . . . . . . . . . . . . . . . . . . . . 77
4.4 Arcus functions . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 81
4.5 Polynomials and Rational Functions . . . . . . . . . . . . . . . . . . . . . . 82
4.5.1 Polynomials . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 82
4.5.2 Rational Functions . . . . . . . . . . . . . . . . . . . . . . . . . . . 88
4.6 Power Series . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 94

5 Differentiation of Functions 99
5.1 Differentiability and Derivatives . . . . . . . . . . . . . . . . . . . . . . . . 99
5.2 Mean Value Theorems and Consequences . . . . . . . . . . . . . . . . . . . 107
5.3 Higher Derivatives, Curve Sketching . . . . . . . . . . . . . . . . . . . . . . 113
5.4 Taylor’s Formula . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 118

iii
iv Contents

5.5 Simple methods for the numerical solution of equations . . . . . . . . . . . 122

6 The Riemann Integral 129

6.1 Differentiation and Integration . . . . . . . . . . . . . . . . . . . . . . . . . 137
6.2 Integration Rules . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 139
6.2.1 Integration by Substitution . . . . . . . . . . . . . . . . . . . . . . 139
6.2.2 Integration by Parts . . . . . . . . . . . . . . . . . . . . . . . . . . 143
6.3 Integration of Rational Functions . . . . . . . . . . . . . . . . . . . . . . . 145
6.4 Integration on Unbounded Domains and Integration of Unbounded Functions146
6.4.1 Unbounded Interval . . . . . . . . . . . . . . . . . . . . . . . . . . . 146
6.4.2 Unbounded Integrand . . . . . . . . . . . . . . . . . . . . . . . . . 151
6.5 Parameter-dependent integrals . . . . . . . . . . . . . . . . . . . . . . . . . 154
6.6 Solids of revolution, path integrals . . . . . . . . . . . . . . . . . . . . . . . 156
6.7 Fourier series . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 165

Index 175
 Some words

This text should help you to understand the course Real Analysis. It goes hand in hand
with a video course you find on YouTube. Therefore, you will always find links and QR
codes that bring you to the corresponding videos.

The whole playlist can be found here::

https://jp-g.de/bsom/rab/ra-list/

To expand your knowledge even more, you can look into the following books:
• Jonathan Lewin: An Interactive Introduction to Mathematical Analysis,
• A. N. Kolmogorov: Introductory Real Analysis,
• Claudio Canute, Anita Tabacco: Mathematical Analysis I.
• Vladimir A. Zorich: Mathematical Analysis I.
Real Analysis is also known as Calculus with real numbers. It is needed for a lot of other
topics in mathematics and the foundation of every new career in mathematics or in fields
that need mathematics as a tool. We1 discuss simple examples later. Some important
bullet points are limits, continuity, derivatives and integrals. In order to describe these
things, we need a good understanding of the real numbers. They form the foundation of
a real analysis course.

complete number line

For this reason, the first step in a Real Analysis is to define the real number line. After
this, we will be able to work with these numbers and understand the field as a whole. Of
course, this is not an easy task and it will be a hiking tour that we will do together. The
1
In mathematical texts, usually, the first-person plural is used even if there is only one author. Most of
the time it simply means “we” = “I (the author) and the reader”.

1
2 Contents

summit and goal is to understand why working with real numbers is indeed a meaningful
mathematical theory.

We start in the valley of mathematics and will shortly scale the first hills. Always stay
in shape, practise and don’t hesitate to ask about the ways up. It is not an easy trip but
you can do it. Maybe the following tips can guide you:
• You will need a lot of time for this course if you really want to understand everything
you learn. Hence, make sure that you have enough time each week to do mathematics
and keep these time slots clear of everything else.
• Work in groups, solve problems together and discuss your solutions. Learning math-
ematics is not a competition.
• Explain the content of the lectures to your fellow students. Only the things you can
illustrate and explain to others are really understood by you.
• Learn the Greek letters that we use in mathematics:
α alpha β beta γ gamma Γ Gamma
δ delta � epsilon ε epsilon ζ zeta
η eta θ theta Θ Theta ϑ theta
ι iota κ kappa λ lambda Λ Lambda
µ mu ν nu ξ xi Ξ Xi
π pi Π Pi ρ rho σ sigma
Σ Sigma τ tau υ upsilon Υ Upsilon
φ phi Φ Phi ϕ phi χ chi
ψ psi Ψ Psi ω omega Ω Omega

This video may help you there:

https://jp-g.de/bsom/la/greek/

• Choosing a book is a matter of taste. Look into different ones and choose the book
that really convinces you.
Contents 3

• Keep interested, fascinated and eager to learn. However, do not expect to under-
stand everything at once.
DON’T PANIC J.P.G.
 Sequences and Limits
1
I’m a gym member. I try to go four times a week, but I’ve missed the
last twelve hundred times.
Chandler Bing

1.1 Just Numbers

Before we start with the Real Analysis course, we need to lie down some foundations. You
only need some knowledge about working with sets and maps to get started. From this,
we will introduce all the number sets we will need in this course in a quick way. Hence,
we quickly have the real numbers R we work with throughout this course.
However, if you interested in a more detailed discussion, I can recommend you my video
series about the foundations of mathematics:

Video: Start Learning Mathematics

https: // jp-g. de/ bsom/ la/ slm/

In order to construct the real number line, we need to generalise the equality sign. We
get a more abstract notion that we can use for sets to put similar elements into the same
box. In the end, we want to calculate with these boxes.
It turns out that we just need three properties from the equality sign to get the general
concept of an equivalence relation.

5
6 1 Sequences and Limits

Definition 1.1. Equivalence relation

Let X be a set. A subset R∼ ⊂ X × X is called a relation on X. We write x ∼ y
if (x, y) ∈ R∼ . A relation R∼ is called an equivalence relation if it satisfies the
following:

(a) x ∼ x for all x ∈ X. (reflexive)

(b) If x ∼ y, then also y ∼ x. (symmetric)

(c) If x ∼ y and y ∼ z, then also x ∼ z. (transitive)

By having this, we now can put equivalent elements in the corresponding boxes. These
boxes ared called equivalent classes.

Proposition & Definition 1.2. Equivalent classes

An equivalence relation ∼ on X gives a partition of X into disjoint subsets. For all
a ∈ X, we define
[a]∼ := {x ∈ X : x ∼ a}
and call it an equivalent class. We have the disjoint union:
�
X= [a]∼ .
a∈X

In the same way as we generalised the equality sign, we can also generalise the greater or
equal sign you might have seen often for numbers. It turns out that we just need some
defining properties there to get an abstract notion of such an ordering.

Definition 1.3. Ordering

Let X be a set. Let R≤ ⊂ X × X be a relation on X where we write x ≤ y if
(x, y) ∈ R≤ . A relation R≤ is called a partial order if it satisfies the following:

(a) x ≤ x for all x ∈ X. (reflexive)

(b) If x ≤ y and y ≤ x, then x = y. (antisymmetric)

(c) If x ≤ y and y ≤ z, then also x ≤ z. (transitive)

If, in addition,

(d) for all x, y ∈ X, we have either x ≤ y or y ≤ x. (total)

then ≤ is called a total order or chain.

Remark: Notation
If one has an ordering relation ≤, one usually also defines the following symbols:

x ≥ y : ⇐⇒ y ≤ x
� �
x < y : ⇐⇒ x ≤ y and x �= y .
1.1 Just Numbers 7

Reminder: Natural numbers

The natural numbers with zero N0 = {0, 1, 2, 3, . . .} originate from counting. There
is an operation + that is associative and commutative, called addition. Then we
can define a total order ≤. Instead of 3 + 3 + 3 + 3 + 3, one writes 5 · 3. This defines
also an associative and commutative operation on N 0 , called multiplication.
The element 0 is the neutral element with respect to the addition + and the element
1 is the neutral element with respect to the multiplication ·.

Example 1.4. Sequence of numbers

An ordered infinite list of numbers is called a sequence. We will define it later in
more detail. For example:

1, 2, 4, 8, 16, 32, 64, 128, 256, 512, 1024, . . .

Here, one would shorten that to (an )n∈N0 with an = 2n .

What can we do with this? Fibonacci numbers! Collatz conjecture!

Box 1.5. Construction of Z

We do not have an element x ∈ N0 that fulfils x + 5 = 0. This is what we would
call the inverse of 5 with respect to the addition. This element is what we recognise
inside the following elements of N0 × N0 :

(0, 5) (1, 6) (101, 106) (56, 61) (77, 82) (91, 96)

We define an equivalence relation on the set N 0 × N0 by setting:

(a, b) ∼ (c, d) : ⇐⇒ a + d = b + c

The equivalent classes are excactly what we need:

[(a, b)] := {(x, y) : (x, y) ∼ (a, b)} .

We can also define an addition:

[(a, b)] + [(c, d)] := [(a + c, b + d)] .

It is well-defined, associative and commutative and [(0, 0)] defines the neutral ele-
ment. The set of all equivalence classes with this new addition is called the integers
and denoted by Z.

Also extend the multiplication to Z. Moreover, we can also define a multiplication:

[(a, b)] · [(c, d)] := [(ac + bd, ad + bc)] .

8 1 Sequences and Limits

Definition 1.6. Commutative ring

The integers Z form a so-called commutative ring. This means:

(A) Addition
(A1) associative: x + (y + z) = (x + y) + z
(A2) neutral element: There is a (unique) element 0 with x + 0 = x for all x.
(A3) inverse element: For all x there is a (unique) y with x + y = 0. We write
for this element simply −x.
(A4) commutative: x + y = y + x

(M) Multiplication
(M1) associative: x · (y · z) = (x · y) · z
(M2) neutral element: There is a (unique) element 1 with 1 · x = x for all x.
(M3) commutative: x · y = y · x

(D) Distributivity: x · (y + z) = x · y + x · z.

Box 1.7. Construction of Q

We do not have an element x ∈ Z \ {0} that fulfils x · 5 = 1. This is what we
would call the inverse of 5 with respect to the multiplication. This element is what
we recognise inside the following elements of Z \ {0} × Z \ {0}:

(1, 5) (2, 10) (101, 505) (50, 250) (11, 55) (500, 2500)

We define an equivalence relation on the set Z \ {0} × Z \ {0} by setting:

(a, b) ∼ (c, d) : ⇐⇒ a · d = b · c

The equivalent classes are excactly what we need:

[(a, b)] := {(x, y) : (x, y) ∼ (a, b)} .

We can also define a multiplication:

[(a, b)] · [(c, d)] := [(a · c, b · d)] .

It is well-defined, associative and commutative. The set of all equivalence classes

with this new multiplication is called the rational numbers (without zero).

[(a, b)] + [(c, d)] := [(ad + bc, bd)]

We can visualise these numbers in the number line:

1.1 Just Numbers 9

Definition 1.8. Real numbers

The real numbers are a non-empty set R together with the operations + : R×R → R
and · : R × R → R and an ordering relation ≤: R × R → {True, False} that fulfil
the following rules

(A) Addition
(A1) associative: x + (y + z) = (x + y) + z
(A2) neutral element: There is a (unique) element 0 with x + 0 = x for all x.
(A3) inverse element: For all x there is a (unique) y with x + y = 0. We write
for this element simply −x.
(A4) commutative: x + y = y + x

(M) Multiplication
(M1) associative: x · (y · z) = (x · y) · z
(M2) neutral element: There is a (unique) element 1 �= 0 with x·1 = x for all x.
(M3) inverse element: For all x �= 0 there is a (unique) y with x · y = 1. We
write for this element simply x−1 .
(M4) commutative: x · y = y · x

(D) Distributivity: x · (y + z) = x · y + x · z.

(O) Ordering
(O1) x ≤ x is true for all x.
(O2) If x ≤ y and y ≤ x, then x = y.
(O3) transitive: x ≤ y and y ≤ z imply x ≤ z.
(O4) For all x, y ∈ X, we have either x ≤ y or y ≤ x.
(O5) x ≤ y implies x + z ≤ y + z for all z.
(O6) x ≤ y implies x · z ≤ y · z for all z ≥ 0.
(O7) x > 0 and ε > 0 implies x < ε + · · · + ε for sufficiently many summands.

(C) Let X, Y ⊂ R be two non-empty subsets with the property x ≤ y for all x ∈ X
and y ∈ Y . Then there is a c ∈ R with x ≤ c ≤ y for all x ∈ X and y ∈ Y .

Remark:
We will later reformulate the completeness axiom with the help of sequences. Then
it sounds like:
Completeness: Every sequence (an )n∈N with the property [For all ε > 0 there is an
N ∈ N with |an − am | < ε for all n, m > N ] has a limit.

Exercise 1.9.
Use the axioms to show:

(1) : 0 · x = 0
10 1 Sequences and Limits

(2) : −x = (−1) · x

(3) : (−1) · (−1) = 1

(4) : 1 > 0

(A2) (D) (A2)

(1): 0 · x = (0 + 0)x = 0x + 0x =⇒ 0x = 0
(∗) (A3) (D) (A2−4)
(2): −x = 0x + (−x) = (1 + (−1))x + (−x) = x + (−1)x + (−x) = (−1)x
(∗∗) (A3) (M 2)
(3): x(−1)(−1) = −(−x) = x =⇒ (−1) · (−1) = 1
(4): Exercise!

Definition 1.10. Absolute value for real numbers

The absolute value of a number x ∈ R is defined by
�
x if x ≥ 0,
|x| :=
−x if x < 0.

Proposition 1.11. Two important properties

For any two real numbers x, y ∈ R, one has

(a) |x · y| = |x| · |y|, (|·| is multiplicative),

(b) |x + y| ≤ |x| + |y|, (|·| fulfils the triangle inequality).

1.2 Convergence of Sequences 11

1.2 Convergence of Sequences

Now we start with sequences and the important definition of convergence.

Definition 1.12.
Let M be a set. A sequence in M is a map a : N → M or a : N0 → M .

We use the following symbols for sequences:

(an )n∈N , (an ), (an )∞

n=1 , (a1 , a2 , a3 , . . .).

Remark:
M is usually a real subset (M ⊂ R), but M can also be a complex subset (M ⊂ C)
or a subset of some normed space (or M = R, M = C, M a normed space itself ).

Example 1.13. (a) an = (−1)n , then (an )n∈N = ((−1)n )n∈N = (−1, 1, −1, 1, −1, 1, . . .);
(b) an = n1 , then (an )n∈N = ( n1 )n∈N = (1, 12 , 13 , 14 , 15 , 16 , . . .);
(c) an = in (i is the imaginary unit), then (an )n∈N = (in )n∈N = (i, −1, −i, 1, i, −1, . . .);
(d) an = 1
2n
, then (an )n∈N = ( 21n )n∈N = (1, 12 , 14 , 18 , 16
1 1
, 32 , . . .);
(e) Approximation of π:
Consider a circle with radius r = 1.
The area is given by A = πr 2 = π.

Now: Approximation of the circle by a regular polygon (all edges have equal length):
Area of a “piece of cake”: Acn = sin( πn ) cos( πn ) = 12 sin( 2π
n
) (the latter equation holds

Figure 1.1: Circle approximated by a polygon

true due to the general equality sin(2x) = 2 sin(x) cos(x) (will be treated later)).
The area of the polygon is therefore given by
� �
n 2π
An = n · Acn = · sin .
2 n
12 1 Sequences and Limits

p1
p2

Figure 1.2: “Piece of cake”

Now consider the sequence (An )n∈N . Some values for An are listed in the following
table:

n An π − An
3 1.299038 1.84255
6 2.598076 0.54351
12 3.000000 0.14159
3072 3.14159046 0.00000219
50331648 3.141592653589 8.16 · 10−16

(f) Some linear algebra:

Consider the linear system

Ax = b, A ∈ Rn×n , b ∈ Rn .

Then o = −Ax + b and thus

x = x − Ax + b = (1 − A)x + b.

Consider the iteration

x1 = o
“Richardson iteration”
xn+1 = (1 − A)xn + b for n ≥ 1

The sequence (xn )n∈N is a sequence of approximate solutions. The method is “good”
if (xn )n∈N “converges”.

Since most (but not all) results stay the same in the real and complex case, we make the
following definition:

Definition 1.14.
The symbol F stands for either the real or complex numbers, i.e., F ∈ {R, C}.

Next we define the notions of convergence and limits:

1.2 Convergence of Sequences 13

Definition 1.15. Convergence/divergence of sequences

Let (an )n∈N be a sequence in F. We say that

– (an )n∈N is convergent to a ∈ F if for all ε > 0 there exists some N = N (ε) ∈ N
such that for all n ≥ N holds |an − a| < ε. In this case, we write

lim an = a.
n→∞

– (an )n∈N is divergent if it is not convergent, i.e., for all a ∈ F holds: There
exists some ε > 0 such that for all N there exists some n > N with |a n −a| ≥ ε.

Convergence for real sequences means that if you give any small distance ε, one finds that
all sequence members an lie in the interval (a − ε, a + ε) with the exception of only finitely
many.

a1 a2 aN a
a−ε a+ε

Example 1.16. • Show: (an )n∈N with an = (1/n) is convergent with limit 0.
√
• Show: (bn )n∈N with bn = (1/ n) is convergent with limit 0.

Proof. Let ε > 0. Choose N > 1

ε2
. Then for all n ≥ N , we have

1 1
|bn − 0| = √ ≤ √ < ε
n N
This means bn is arbitrarily close to 0, eventually.

Remark:
(a) It can be shown that for a complex sequence (a n )n∈N , convergence to a ∈ C holds
true if, and only if, (Re(an ))n∈N converges to Re(a) and (Im(an ))n∈N converges
to Im(a)

(b) In fact, convergence can also be defined for sequences in some arbitrary normed
vector space V (for the definition of a normed space, e.g. consult the linear
algebra script). Then one has to replace the absolute value by the norm (e.g.,
“�an − a� < ε”.)

(c) Due to the fact that for any N ∈ N, we can find some x ∈ R with x > N , we
can equivalently reformulate the convergence definition as follows: “(a n )n∈N is
convergent to a ∈ R if for all ε > 0 there exists some N = N (ε) ∈ R such that
for all n ≥ N holds |an − a| < ε”. In the following, we just write “there exists
some N ”.

What does convergence mean?

14 1 Sequences and Limits

a1 a2 aN a aN +1
a−ε a+ε

Outside any ε-neighbourhood of a only finitely many elements of the sequence exist.

Example 1.17. (a) The real sequence ( n1 )n∈N converges to 0.

Proof: Let ε > 0 (be arbitrary): Choose N = 1ε + 1 = 1+ε
ε
.
Then for all n ≥ N holds
1 1 ε
≤ = < ε.
n N 1+ε
Therefore � �
�1 �
� − 0� = 1 < ε.
�n � n

(b) The real sequence ((−1)n )n∈N is divergent.

Proof by contradiction:
Assume that ((−1)n )n∈N is convergent to a ∈ R. Then take e.g. ε = 1
10
. Due to
convergence, we should have some N such that for all n ≥ N holds

|(−1)n − a| < 1
10
.

Thus we have that | − 1 − a| < 1

10
and |1 − a| < 1
10
. As a consequence,
1 1
2 = |1 + a − a + 1| ≤ |1 + a| + | − a + 1| = |1 + a| + |a − 1| < 10
+ 10
= 15 .

This is a contradiction. �
(c) For q ∈ C\{0} with |q| < 1 the complex sequence (q n )n∈N converges to 0.
Proof: |q| < 1 gives rise to |q|
1
> 1, whence |q|
1
− 1 > 0. Therefore, we are able to apply
Bernoulli’s inequality (see tutorial) in the following way:
� � ��n � � ��n � �
1 1 1 − |q| 1 − |q|
= 1+ −1 = 1+ ≥1+n· ,
|q|n |q| |q| |q|
and thus
1 |q|
|q|n ≤ � �= .
1 + n · 1−|q| |q| + n · (1 − |q|)
|q|

Now let ε > 0 (be arbitrary):

Choose
|q| |q|
N= − +1
ε · (1 − |q|) 1 − |q|
Then for all n ≥ N holds
|q| |q|
n> −
ε · (1 − |q|) 1 − |q|
and thus
|q|
n · (1 − |q|) > − |q|.
ε
1.2 Convergence of Sequences 15

This leads to
|q|
|q| + n · (1 − |q|) > ,
ε
whence
|q|
< ε.
|q| + n · (1 − |q|)
The above calculations now imply

|q|
|q n − 0| = |q|n ≤ < ε.
|q| + n · (1 − |q|)

Remark:
The choice of the N often seems “to appear from nowhere”. However, there is
a systematic way to formulate the proof. For instance in a), we need to end up with
the equation | n1 − 0| < ε or, equivalently, n1 < ε. Inverting this expression leads to
n > 1ε . Therefore, if N is chosen as N = 1ε + 1 = 1+εε
, the desired statement follows.
If one has to formulate such a proof (for instance, in some exercise), then first
these above calculations have to be done “on some extra sheet” and then formulate
the convergence proof in the style as in a) or c).

Definition 1.18. Boundedness of sequences

Let (an )n∈N be a sequence in F. Then (an )n∈N is called

– bounded if there exists some c ∈ R such that for all n ∈ N holds |a n | ≤ c;

– unbounded if it is not bounded, i.e., for all c ∈ R, there exists some n ∈ N

with |an | > c.

Theorem 1.19.
Let (an )n∈N be a convergent sequence in F. Then (an )n∈N is bounded.

Proof. Suppose that limn→∞ an = a. Take ε = 1. Then there exists some N such that for
all n ≥ N holds |an − a| < 1. Thus, for all n ≥ N holds

|an | = |an − a + a| ≤ |an − a| + |a| < 1 + |a|.

Now choose
c = max{|a1 |, |a2 |, . . . , |aN −1 |, |a| + 1}

and consider some arbitrary sequence element ak .

If k < N , then |ak | ≤ max{|a1 |, |a2 |, . . . , |aN −1 |} ≤ c.
In the case k ≥ N , the above calculations lead to |ak | < |a| + 1 ≤ c.
Altogether, this implies that |ak | ≤ c for all k ∈ N, so (an )n∈N is bounded by c.