Topology Polytechnique

MAA302
Topology and differential calculus
Kleber Carrapatoso
kleber.carrapatoso@polytechnique.edu
Bachelor of Sciences
École Polytechnique
2023 – 2024
Contents
1 Topological and metric spaces 1

1.1 Topological spaces . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 1
1.2 Metric spaces . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 2
2 Neighborhoods, basis, interior and closure 5

2.1 Neighborhoods . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 5
2.2 Basis . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 7
2.3 Interior and closure . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 8
2.4 Separation and countability . . . . . . . . . . . . . . . . . . . . . . . . . . 9
3 Continuity and limits 13

3.1 Continuous maps . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 13
3.2 Construction of topologies . . . . . . . . . . . . . . . . . . . . . . . . . . 15
3.3 Limits and cluster points of a sequence . . . . . . . . . . . . . . . . . . . 19
4 Completeness 21
4.1 Cauchy sequences . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 21
4.2 Complete metric spaces . . . . . . . . . . . . . . . . . . . . . . . . . . . . 22
4.3 Uniformly continuous and Lipschitz maps . . . . . . . . . . . . . . . . . 23
4.4 Fixed-point theorem . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 23
4.5 Baire’s category theorem . . . . . . . . . . . . . . . . . . . . . . . . . . . 24
5 Compactness 25
5.1 Open covers . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 25
5.2 Compact spaces . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 25
5.3 Compact metric spaces . . . . . . . . . . . . . . . . . . . . . . . . . . . . 27
5.4 Product of compact spaces . . . . . . . . . . . . . . . . . . . . . . . . . . 27
5.5 Local compactness . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 29
6 Connectedness 31
6.1 Connected spaces . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 31
6.2 Path-connected spaces . . . . . . . . . . . . . . . . . . . . . . . . . . . . 32
7 Function spaces 35
7.1 Pointwise and uniform convergences . . . . . . . . . . . . . . . . . . . . 35
7.2 Space of continuous functions on a compact set . . . . . . . . . . . . . . 36
7.3 Equicontinuity . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 37
7.4 Arzelà-Ascoli theorem . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 38
7.5 Stone-Weierstrass theorem . . . . . . . . . . . . . . . . . . . . . . . . . . 39
iii
iv Contents
8 Banach spaces and continuous linear maps 43

8.1 Normed vector spaces and Banach spaces . . . . . . . . . . . . . . . . . . 43
8.2 Finite-dimensional normed vector spaces . . . . . . . . . . . . . . . . . . 45
8.3 Continuous linear maps . . . . . . . . . . . . . . . . . . . . . . . . . . . . 46
8.4 Applications of Baire’s theorem . . . . . . . . . . . . . . . . . . . . . . . 50
8.5 Integration of vector-valued functions . . . . . . . . . . . . . . . . . . . . 52
9 Differentiable maps 55
9.1 Fréchet-differentiable maps . . . . . . . . . . . . . . . . . . . . . . . . . . 55
9.2 Operations on differentiable maps . . . . . . . . . . . . . . . . . . . . . . 57
9.3 Maps with values in a product space . . . . . . . . . . . . . . . . . . . . . 59
9.4 Partial derivatives . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 60
10 Mean-value theorem and applications 63

10.1 Mean-value theorems . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 63
10.2 Relation between partial differentiability and differentiability . . . . . . . 64
10.3 Limits of differentiable maps . . . . . . . . . . . . . . . . . . . . . . . . . 65
10.4 High-order derivatives . . . . . . . . . . . . . . . . . . . . . . . . . . . . 66
10.5 Taylor formulas . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 69
11 Inverse function and implicit function theorems 71

11.1 Diffeomorphisms . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 71
11.2 Inverse function theorem . . . . . . . . . . . . . . . . . . . . . . . . . . . 71
11.3 Implicit function theorem . . . . . . . . . . . . . . . . . . . . . . . . . . . 73
12 Optimization 75
12.1 Extremum points . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 75
12.2 First-order conditions . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 75
12.3 Optimization with constraints . . . . . . . . . . . . . . . . . . . . . . . . 76
12.4 Second-order conditions . . . . . . . . . . . . . . . . . . . . . . . . . . . 77
A Reminders: Sets and maps 79

A.1 Sets . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 79
A.2 Maps . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 83
Chapter 1
Topological and metric spaces
1.1 Topological spaces

Definition 1.1. A topology on a set 𝑋 is a family 𝒪 ⊆ 𝔓(𝑋 ) of subsets of 𝑋 such that:
(i) the empty set ∅ and 𝑋 belong to 𝒪 ;
(ii) every finite intersection of elements of 𝒪 is an element of 𝒪 ;
(iii) every arbitrary union of elements of 𝒪 is an element of 𝒪.
We say that (𝑋 , 𝒪) is a topological space.
Remark 1.2. Axiom (i) can actually be deduced from the others, and it was added in
the definition for the sake of clarity. Indeed, with our convention, taking an empty
intersection in (ii) implies that 𝑋 ∈ 𝒪; and taking an empty union in (iii) implies that
∅ ∈ 𝒪.
Remark 1.3. One can verify that (ii) is equivalent to: “𝑋 ∈ 𝒪 and every intersection of
two elements of 𝒪 is an element of 𝒪”.
Elements of 𝑋 are called points. Elements of 𝒪 are called open sets in 𝑋 (with respect
to the topology 𝒪). We often denote by 𝑋 the couple (𝑋 , 𝒪).
Definition 1.4. A subset 𝑈 of a topological space 𝑋 is closed if its complement 𝑋 ⧵ 𝑈 is

open.
It easily follows from Definition 1.1 that:

• ∅ and 𝑋 are closed;
• any arbitrary intersection of closed sets is closed;
• any finite union of closed sets is closed.
Definition 1.5. Let 𝒪1 and 𝒪2 be two topologies on a set 𝑋 . If 𝒪1 ⊆ 𝒪2 one says that
𝒪1 is coarser than 𝒪2 , or that 𝒪2 is finer than 𝒪1 . If moreover 𝒪1 ≠ 𝒪2 , then 𝒪1 is said to
be strictly coarser than 𝒪2 , or 𝒪2 is said to be strictly finer than 𝒪1 .
Example 1.1. The collection {∅, 𝑋 } is a topology on any set 𝑋 , called the trivial topology.
Example 1.2. The powerset 𝔓(𝑋 ) of 𝑋 is a topology on any set 𝑋 , called the discrete
topology. A topological space endowed with the discrete topology is called a discrete
space.
1
2 Chapter 1. Topological and metric spaces
Example 1.3. The standard topology on R is given by the collection of arbitrary unions of
open intervals. We shall always consider this topology on R (unless otherwise specified).
Example 1.4. (Extended real number line) Consider the set R ∶= R ∪ {−∞, +∞}, where
+∞ and −∞ are two sets that do not belong to R. This set is equipped with the topology
given by the collection of unions of sets of the form {−∞} ∪ (−∞, 𝑎), (𝑏, +∞) ∪ {+∞} and
(𝑎, 𝑏), with 𝑎, 𝑏 ∈ R.
Example 1.5. Let (𝑋 , 𝒪) be a topological space and 𝑌 ⊆ 𝑋 a subset. The collection of
subsets {𝑈 ∩ 𝑌 ∶ 𝑈 ∈ 𝒪} is a topology on 𝑌 , called the subspace topology.
Example 1.6. Let 𝑋 be a set, 𝑌 a topological space and 𝑓 ∶ 𝑋 → 𝑌 a map. The collection
of subsets {𝑓 −1 (𝑈 ) ∶ 𝑈 open in 𝑌 } is a topology on 𝑋 , called the preimage topology.
Example 1.7. The intersection of an arbitrary family of topologies on a set is a topology
on this set.
1.2 Metric spaces

Definition 1.6. A metric (or distance) on a set 𝑋 is a map 𝑑 ∶ 𝑋 × 𝑋 → R+ such that
for all 𝑥, 𝑦, 𝑧 ∈ 𝑋 one has:
(i) separation: 𝑑(𝑥, 𝑦) = 0 if and only if 𝑥 = 𝑦;
(ii) symmetry: 𝑑(𝑥, 𝑦) = 𝑑(𝑦, 𝑥);
(iii) triangle inequality: 𝑑(𝑥, 𝑧) ⩽ 𝑑(𝑥, 𝑦) + 𝑑(𝑦, 𝑧).
The pair (𝑋 , 𝑑) is called a metric space.
Remark 1.7. If (i) is replaced by the weaker assumption “𝑑(𝑥, 𝑥) = 0”, then 𝑑 is called a
pseudo-metric.
Remark 1.8. If 𝑑 satisfies 𝑑(𝑥, 𝑧) ⩽ sup(𝑑(𝑥, 𝑦), 𝑑(𝑦, 𝑧)) for any 𝑥, 𝑦, 𝑧 ∈ 𝑋 , which is
stronger than the triangle inequality, then we say that 𝑑 is an ultrametric and (𝑋 , 𝑑) an
ultrametric space.
Remark 1.9. For any 𝑥, 𝑦, 𝑧 ∈ 𝑋 we also have |𝑑(𝑥, 𝑧) − 𝑑(𝑦, 𝑧)| ⩽ 𝑑(𝑥, 𝑦).
Let (𝑋 , 𝑑) be a metric space. For any 𝑥 ∈ 𝑋 and 𝑟 > 0, we define the open ball centred
at 𝑥 of radius 𝑟 by
𝐵(𝑥, 𝑟) ∶= {𝑦 ∈ 𝑋 ∣ 𝑑(𝑥, 𝑦) < 𝑟},
the closed ball centred at 𝑥 of radius 𝑟 by
𝐵(𝑥, 𝑟) ∶= {𝑦 ∈ 𝑋 ∣ 𝑑(𝑥, 𝑦) ⩽ 𝑟},
and the sphere centred at 𝑥 of radius 𝑟 by
𝑆(𝑥, 𝑟) ∶= {𝑦 ∈ 𝑋 ∣ 𝑑(𝑥, 𝑦) = 𝑟}.
Definition 1.10. A subset 𝑌 of a metric space (𝑋 , 𝑑) is said to be bounded if there are

𝑥 ∈ 𝑋 and 𝑟 > 0 such that 𝑌 ⊆ 𝐵(𝑥, 𝑟).
Equivalently, 𝑌 is bounded if and only if sup𝑥,𝑦∈𝑌 𝑑(𝑥, 𝑦) < ∞.

1.2 Metric spaces 3
Example 1.8. On any non-empty set 𝑋 , we can define the distance 𝑑(𝑥, 𝑦) = 1 if 𝑥 ≠ 𝑦
and 𝑑(𝑥, 𝑦) = 0 if 𝑥 = 𝑦. This is called the discrete distance.
Example 1.9. The set of real numbers R (resp. complex numbers C) is endowed with the
metric 𝑑(𝑥, 𝑦) = |𝑥 − 𝑦| for any 𝑥, 𝑦 in R (resp. in C) where | ⋅ | denotes the absolute value
(resp. the modulus).
Example 1.10. If (𝑋 , 𝑑) is as metric space and 𝑌 is a subset of 𝑋 , we define the map
𝑑𝑌 ∶ 𝑌 × 𝑌 → R+ as the restriction of 𝑑 to 𝑌 × 𝑌 . Then 𝑑𝑌 is a metric on 𝑌 and hence
(𝑌 , 𝑑𝑌 ) is a metric space, called a (metric) subspace of (𝑋 , 𝑑).
Example 1.11. Any normed vector space (𝐸, ‖ ⋅ ‖) is a metric space with the distance
𝑑(𝑥, 𝑦) ∶= ‖𝑥 − 𝑦‖, for any 𝑥, 𝑦 ∈ 𝐸.
Example 1.12. Let (𝑋 , 𝑑𝑋 ) and (𝑌 , 𝑑𝑌 ) be metric spaces. On the product set 𝑋 × 𝑌 we can
define the following distances: for any (𝑥, 𝑦), (𝑥 ′ , 𝑦 ′ ) ∈ 𝑋 × 𝑌
𝑑1 ((𝑥, 𝑦), (𝑥 ′ , 𝑦 ′ )) ∶= 𝑑𝑋 (𝑥, 𝑥 ′ ) + 𝑑𝑌 (𝑦, 𝑦 ′ ),
and
𝑑∞ ((𝑥, 𝑦), (𝑥 ′ , 𝑦 ′ )) ∶= max (𝑑𝑋 (𝑥, 𝑥 ′ ), 𝑑𝑌 (𝑦, 𝑦 ′ )).
1.2.1 Topology associated to a metric

Definition 1.11. Let (𝑋 , 𝑑) be a metric space. We define a topology 𝒪𝑑 on 𝑋 by
𝒪𝑑 ∶= {𝑈 ⊆ 𝑋 ∣ ∀𝑥 ∈ 𝑈 , ∃𝑟 > 0, 𝐵(𝑥, 𝑟) ⊆ 𝑈 },
called the topology associated to or induced by the metric 𝑑.
Let us verify that 𝒪𝑑 is indeed a topology on 𝑋 . It is clear that ∅ and 𝑋 belong to

𝒪𝑑 . Let 𝑈1 , … , 𝑈𝑛 ∈ 𝒪𝑑 and 𝑥 ∈ ⋂1⩽𝑖⩽𝑛 𝑈𝑖 , then for each 1 ⩽ 𝑖 ⩽ 𝑛 we have 𝑥 ∈ 𝑈𝑖 ∈ 𝒪𝑑 ,
hence there is 𝑟𝑖 > 0 such that 𝐵(𝑥, 𝑟𝑖 ) ⊆ 𝑈𝑖 . Consider 𝑟 = inf1⩽𝑖⩽𝑛 𝑟𝑖 , we have 𝑟 > 0
since there are finitely many 𝑟𝑖 , then 𝐵(𝑥, 𝑟) ⊆ 𝑈𝑖 for each 1 ⩽ 𝑖 ⩽ 𝑛, which implies that
𝐵(𝑥, 𝑟) ⊆ ⋂1⩽𝑖⩽𝑛 𝑈𝑖 , hence ⋂1⩽𝑖⩽𝑛 𝑈𝑖 ∈ 𝒪𝑑 . Now consider a family {𝑈𝑖 }𝑖∈𝐼 of elements of
𝒪𝑑 . Let 𝑥 ∈ ⋃𝑖∈𝐼 𝑈𝑖 , then there is 𝑖0 ∈ 𝐼 such that 𝑥 ∈ 𝑈𝑖0 ∈ 𝒪𝑑 , hence there is 𝑟 > 0 such
that 𝐵(𝑥, 𝑟) ⊆ 𝑈𝑖0 ⊆ ⋃𝑖∈𝐼 𝑈𝑖 , which implies ⋃𝑖∈𝐼 𝑈𝑖 ∈ 𝒪𝑑 .
In this way, we shall always consider a metric space as a topological space endowed
with the topology associated to the metric.
Is is easy to check that open balls (resp. closed balls) are open sets (resp. closed sets)
for the topology associated to the metric.
Example 1.13. The topology on R associated to its standard metric (see Example 1.9)
coincides with the standard topology (see Example 1.3).
Definition 1.12. A topological space (𝑋 , 𝒪) is metrizable if there is a metric 𝑑 on 𝑋

such that the topology associated to 𝑑 coincides with 𝒪.
1.2.2 Equivalent metrics

Definition 1.13. We say that two metrics 𝑑1 and 𝑑2 on a set 𝑋 are:
(i) topologically equivalent if they induce the same topology ;
4 Chapter 1. Topological and metric spaces
(ii) equivalent if there exist constants 𝑐1 , 𝑐2 > 0 such that for any 𝑥, 𝑦 ∈ 𝑋 one has
𝑐1 𝑑1 (𝑥, 𝑦) ⩽ 𝑑2 (𝑥, 𝑦) ⩽ 𝑐2 𝑑1 (𝑥, 𝑦).
It is easy to verify that “equivalent” implies “topologically equivalent”.

Chapter 2
Neighborhoods, basis, interior and

closure
2.1 Neighborhoods
Let 𝑋 be a topological space, 𝐴 ⊆ 𝑋 a subset and 𝑥 ∈ 𝑋 a point.
Definition 2.1. A neighborhood of 𝑥 (resp. 𝐴) is a subset 𝑉 ⊆ 𝑋 that contains an open

set containing 𝑥 (resp. 𝐴).
In other words, denoting by 𝒪 the topology of 𝑋 , a subset 𝑉 of 𝑋 is a neighborhood

of 𝑥 if there is 𝑈 ∈ 𝒪 such that 𝑥 ∈ 𝑈 ⊆ 𝑉 .
We denote by 𝒱 (𝑥) the family of all neighborhoods of 𝑥, which is called the neigh-
borhood system for 𝑥. Similarly, we denote by 𝒱 (𝐴) the family of all neighborhoods of
𝐴, which is called the neighborhood system for 𝐴.
Definition 2.2. A neighborhood basis or local basis for 𝑥 (resp. 𝐴) is a family 𝒮 of neigh-
borhoods of 𝑥 (resp. 𝐴) such that every neighborhood of 𝑥 (resp. 𝐴) contains an element
of 𝒮 .
Example 2.1. Consider a point 𝑥 in R endowed with its standard topology. Each of the
following families are neighborhood basis for 𝑥:
{(𝑥 − 𝑟, 𝑥 + 𝑟) ∶ 𝑟 > 0}, {[𝑥 − 𝑟, 𝑥 + 𝑟] ∶ 𝑟 > 0},
{(𝑥 − 𝑛1 , 𝑥 + 𝑛1 ) ∶ 𝑛 ∈ N∗ }, {[𝑥 − 𝑛1 , 𝑥 + 𝑛1 ] ∶ 𝑛 ∈ N∗ }.
One easily verifies the following properties:
Proposition 2.3. A subset of a topological space 𝑋 is open if and only if it is a neighbor-

hood of each of its points.
Proof. Let 𝑈 be a subset of 𝑋 . If 𝑈 is open then it is clear by definition that 𝑈 is a

neighborhood of each of its points. Conversely, assume that 𝑈 is a neighborhood of
each of its points. For each 𝑥 ∈ 𝑈 there is an open set 𝑈𝑥 in 𝑋 such that 𝑥 ∈ 𝑈𝑥 ⊆ 𝑈 .
This implies that 𝑈 = ⋃𝑥∈𝑈 𝑈𝑥 , and ⋃𝑥∈𝑈 𝑈𝑥 is open as the union of open sets.
Proposition 2.4. Let 𝑋 be a topological space. For any 𝑥 ∈ 𝑋 , the neighborhood system
𝒱 (𝑥) satisfies:
5
6 Chapter 2. Neighborhoods, basis, interior and closure
(i) any subset of 𝑋 containing an element of 𝒱 (𝑥) belongs to 𝒱 (𝑥);

(ii) any finite intersection of elements of 𝒱 (𝑥) belongs to 𝒱 (𝑥);
(iii) 𝑥 belongs to any element of 𝒱 (𝑥);
(iv) for all 𝑉 ∈ 𝒱 (𝑥), there exists 𝑊 ∈ 𝒱 (𝑥) such that, for all 𝑦 ∈ 𝑊 , 𝑉 ∈ 𝒱 (𝑦).
Proof. Consider 𝑈 ⊆ 𝑋 such that there is 𝑉 ∈ 𝒱 (𝑥) that satisfies 𝑉 ⊆ 𝑈 . Then there
exists an open set 𝑊 ⊆ 𝑋 such that 𝑥 ∈ 𝑊 ⊆ 𝑉 , hence 𝑥 ∈ 𝑊 ⊆ 𝑈 and 𝑈 is a neigh-
borhood of 𝑥, which proves (i). Now let 𝑉1 , … , 𝑉𝑛 ∈ 𝒱 (𝑥), then there exist open sets
𝑈1 , … , 𝑈𝑛 such that 𝑥 ∈ 𝑈𝑗 ⊆ 𝑉𝑗 for any 𝑗 ∈ {1, … , 𝑛}. The set ⋂1⩽𝑗⩽𝑛 𝑈𝑗 is open, included
in ⋂1⩽𝑗⩽𝑛 𝑉𝑗 , and contains 𝑥, thus ⋂1⩽𝑗⩽𝑛 𝑉𝑗 is a neighborhood of 𝑥, and this gives (ii).
Property (iii) is trivial. Finally, let 𝑉 ∈ 𝒱 (𝑥). Then there exists an open set 𝑈 such that
𝑥 ∈ 𝑈 ⊆ 𝑉 , and 𝑈 is also a neighborhood of 𝑥. We hence obtain that, for any 𝑦 ∈ 𝑈 , we
have 𝑦 ∈ 𝑈 ⊆ 𝑉 , thus 𝑉 ∈ 𝒱 (𝑦). This proves (iv).
We also observe that, thanks to the above properties on the neighborhood system, it
is possible to define a topology on a set by 𝑋 by prescribing the neighborhood system
𝒱 (𝑥) for all points 𝑥 ∈ 𝑋 , provided it satisfies properties (i) to (iv) of Proposition 2.4.
Proposition 2.5. Let 𝑋 be a set and 𝒱 ∶ 𝑋 → 𝔓(𝑋 ) a map that sends a point 𝑥 ∈ 𝑋 to a
collection 𝒱 (𝑥) of subsets of 𝑋 satisfying properties (i) to (iv) of Proposition 2.4. Then there
exists a unique topology on 𝑋 such that, for any 𝑥 ∈ 𝑋 , 𝒱 (𝑥) is the neighborhood system
of 𝑥 associated to this topology.
Proof. The uniqueness follows from the fact that a set is open if and only if it is a neigh-
borhood of each of its points (Proposition 2.3). Indeed, assume that there exists a topol-
ogy verifying the properties of the statement, then it must be equal to
𝒪 = {𝑈 ⊆ 𝑋 ∣ ∀𝑥 ∈ 𝑈 , 𝑈 ∈ 𝒱 (𝑥)}.
We easily verify that 𝒪 is a topology on 𝑋 : ∅, 𝑋 ∈ 𝒪 is trivial; the fact that finite

intersections of elements of 𝒪 belong to 𝒪 follows from property (i) of Proposition 2.4;
and the fact that arbitrary unions of elements of 𝒪 belong to 𝒪 follows from property
(ii) of Proposition 2.4.
It remains to verify that, for any 𝑥 ∈ 𝑋 , 𝒱 (𝑥) corresponds to the collection of neigh-
borhoods of 𝑥 for the topology 𝒪. Thanks to property (i) of Proposition 2.4, every neigh-
borhood of 𝑥 belongs to 𝒱 (𝑥). Conversely, let 𝑉 ∈ 𝒱 (𝑥) and 𝑈 be the set of points
𝑦 ∈ 𝑋 such that 𝑉 ∈ 𝒱 (𝑦). We shall prove that 𝑥 ∈ 𝑈 ⊆ 𝑉 and 𝑈 ∈ 𝒪 (i.e. 𝑉 is a
neighborhood of 𝑥) which completes the proof. We have 𝑥 ∈ 𝑈 since 𝑉 ∈ 𝒱 (𝑥). Let
𝑦 ∈ 𝑈 , then 𝑉 ∈ 𝒱 (𝑦). On the one hand, from property (iii) of Proposition 2.4 we deduce
𝑦 ∈ 𝑉 and hence 𝑈 ⊆ 𝑉 . On the other hand, from property (iv) of Proposition 2.4 there
is 𝑊 ∈ 𝒱 (𝑦) such that for any 𝑧 ∈ 𝑊 we have 𝑉 ∈ 𝒱 (𝑧), which implies 𝑧 ∈ 𝑈 and hence
𝑊 ⊆ 𝑈 , which finally yields 𝑈 ∈ 𝒱 (𝑦) thanks to (i), that is 𝑈 ∈ 𝒪.
We also remark that a neighborhood basis 𝒮 (𝑥) for 𝑥 characterizes the collection
𝒱 (𝑥) of neighborhoods for 𝑥 by
𝒱 (𝑥) = {𝑉 ⊆ 𝑋 ∣ ∃𝐵 ∈ 𝒮 (𝑥), 𝐵 ⊆ 𝑉 }.
In view of this, one can also define a topology on a set 𝑋 by prescribing the sets 𝒮 (𝑥) of
neighborhood basis for all points of 𝑋 , provided that, for any 𝑥 ∈ 𝑋 , 𝒮 (𝑥) is non-empty
and verifies:
2.2 Basis 7
(i) 𝑥 belongs to all elements of 𝒮 (𝑥) ;

(ii) for any 𝑈 , 𝑉 ∈ 𝒮 (𝑥), there is 𝑊 ∈ 𝒮 (𝑥) such that 𝑊 ⊆ 𝑈 ∩ 𝑉 ;
(iii) for any 𝑉 ∈ 𝒮 (𝑥), there is 𝑊 ∈ 𝒮 (𝑥) such that, for any 𝑦 ∈ 𝑊 , there is 𝑈 ∈ 𝒮 (𝑦)
such that 𝑈 ⊆ 𝑉 .
2.2 Basis
Let (𝑋 , 𝒪) be a topological space.
Definition 2.6. A family ℬ ⊆ 𝒪 of open sets in 𝑋 is a basis for the topology 𝒪 if every
open set in 𝑋 can be written as a union of elements of ℬ.
Equivalently, we can caractherise a basis for a topology in the following way.
Proposition 2.7. A subset ℬ ⊆ 𝒪 is a basis for the topology 𝒪 if and only if for any open
set 𝑈 ∈ 𝒪 and any 𝑥 ∈ 𝑈 , there exists 𝐵 ∈ ℬ such that 𝑥 ∈ 𝐵 ⊆ 𝑈 .
Proof. Suppose that ℬ is a basis for 𝒪, namely that every open set 𝑈 ∈ 𝒪 is a union of
elements of ℬ. Let 𝑈 ∈ 𝒪 and 𝑥 ∈ 𝑈 . Then there is ℬ1 ⊆ ℬ such that 𝑈 = ⋃𝐵∈ℬ1 𝐵,
therefore there is 𝐵 ∈ ℬ1 such that 𝑥 ∈ 𝐵 ⊆ 𝑈 .
Conversely, let ℬ ⊆ 𝒪 satisfy
∀ 𝑈 ∈ 𝒪, ∀ 𝑥 ∈ 𝑈 , ∃ 𝑉 ∈ ℬ, 𝑥 ∈ 𝑉 ⊆ 𝑈 .
Let 𝑈 ∈ 𝒪, then for each 𝑥 ∈ 𝑈 , there is 𝐵𝑥 ∈ ℬ such that 𝑥 ∈ 𝐵𝑥 ⊆ 𝑈 . This implies that
𝑈 = ⋃𝑥∈𝑈 𝐵𝑥 , that is, 𝑈 is a union of elements of ℬ and hence ℬ is a basis for 𝒪.
Example 2.2. In a metric space (𝑋 , 𝑑), the family ℬ = {𝐵(𝑥, 𝑟) ∶ 𝑥 ∈ 𝑋 , 𝑟 > 0} is a basis
for the topology associated to the metric 𝑑.
2.2.1 Sub-basis and topology generated by a family of subsets

Recall that the intersection of a family {𝒪𝑖 }𝑖∈𝐼 of topologies on a set 𝑋 still is a topology
on 𝑋 . Moreover, given a family 𝒮 of subsets of 𝑋 there exists at least one topology on
𝑋 containing 𝒮 , namely the discrete topology 𝔓(𝑋 ).
Definition 2.8. Let 𝒮 be a family of subsets of a set 𝑋 . The topology generated by 𝒮

is the intersection of all topologies on 𝑋 containing 𝒮 , which also corresponds to the
coarsest topology on 𝑋 containing 𝒮 . One says that 𝒮 is a sub-basis for this topology.
It is easy to verify that the topology 𝒪𝒮 generated by 𝒮 corresponds to the collection

of arbitrary unions of finite intersections of elements of 𝒮 . In other words, the family
of finite intersections of elements of 𝒮 is a basis for 𝒪𝒮 .
The following result gives a condition under which a collection of subsets of 𝑋 (that
is a sub-basis) is actually a basis for the topology it generates.
Proposition 2.9. Let ℬ be a family of subsets of a set 𝑋 satisfying

(i) ⋃ ℬ ∶= ⋃𝐵∈ℬ 𝐵 = 𝑋 ;
(ii) for any 𝑈 , 𝑉 ∈ ℬ and any 𝑥 ∈ 𝑈 ∩ 𝑉 , there is 𝐵 ∈ ℬ such that 𝑥 ∈ 𝐵 ⊆ 𝑈 ∩ 𝑉 .
Then the family 𝒪 of arbitrary unions of elements of ℬ coincides with the topology gener-
ated by ℬ.
Proof. One easily observes that if 𝒪 is a topology then it coincides with the topology
generated by ℬ. Therefore, we only need to show that 𝒪 is a topology.
It is clear that ∅ ∈ 𝒪 and 𝑋 ∈ 𝒪. Let 𝒰 be a family of elements of 𝒪, then each
member of 𝒪 is a union of elements of ℬ, therefore the union ⋃ 𝒰 is also a union of
elements of ℬ. Let 𝑈 , 𝑉 ∈ 𝒪, then there is ℬ1 , ℬ2 ⊆ ℬ such that 𝑈 = ⋃ ℬ1 and
𝑉 = ⋃ ℬ2 . Therefore
𝑈 ∩ 𝑉 = ( ⋃ 𝐴) ∩ ( ⋃ 𝐵) = ⋃ 𝐴 ∩ 𝐵,
𝐴∈ℬ1 𝐵∈ℬ2 𝐴∈ℬ1 , 𝐵∈ℬ2
and the proof will be complete if we show that any intersection of two elements of ℬ is
a union of elements of ℬ. Consider then 𝐴, 𝐵 ∈ ℬ. For each 𝑥 ∈ 𝐴 ∩ 𝐵, there is 𝑊𝑥 ∈ ℬ
such that 𝑥 ∈ 𝑊𝑥 ⊆ 𝐴 ∩ 𝐵. Thus 𝐴 ∩ 𝐵 = ⋃𝑥∈𝐴∩𝐵 𝑊𝑥 , which completes the proof.
2.3 Interior and closure

Let 𝑋 be a topological space, 𝐴 a subset of 𝑋 , and 𝑥 ∈ 𝑋 a point.
Definition 2.10. One says that 𝑥 is an interior point of 𝐴 if 𝐴 is a neighborhood of 𝑥.

The set of all interior points of 𝐴 is called the interior of 𝐴, and is denoted by 𝐴̊ or int(𝐴).
It is easy to verify that 𝐴̊ is the union of all open sets contained in 𝐴, therefore also
the largest open set contained in 𝐴. Indeed let 𝐵 be the union of all open sets contained
in 𝐴. Since a union of open sets is open, 𝐵 is also the largest open set contained in 𝐴.
We now check that 𝐴̊ = 𝐵. Let 𝑥 ∈ 𝐴,̊ then there is an open set 𝑈 such that 𝑥 ∈ 𝑈 ⊆ 𝐴,
but 𝑈 ⊆ 𝐵 thus 𝑥 ∈ 𝐵. Let 𝑥 ∈ 𝐵, then there is an open set 𝑈 such that 𝑥 ∈ 𝑈 ⊆ 𝐴, hence
𝑥 ∈ 𝐴.̊
Definition 2.11. One says that 𝑥 is an adherent point (or closure point) of 𝐴 if every
neighborhood of 𝑥 intersects 𝐴. The set of all adherent points of 𝐴 is called the closure
of 𝐴, and is denoted by 𝐴 or cl(𝐴).
It is easy to verify that 𝐴 is the intersection of all closed sets containing 𝐴, there-
fore also the smallest closed set containing 𝐴. The proof of this fact follows by taking
complement sets in the above argument for proving that 𝐴̊ is the union of all open sets
contained in 𝐴.
Definition 2.12. One says that 𝑥 is a limit point (or cluster point) of 𝐴 if every neigh-
borhood of 𝑥 intersects 𝐴 ⧵ {𝑥}.
Definition 2.13. One says that 𝑥 is a boundary point of 𝐴 if every neighborhood of 𝑥

intersects both 𝐴 and its complement 𝑋 ⧵ 𝐴. The set of all boundary points of 𝐴 is called
the boundary of 𝐴, and is denoted by 𝜕𝐴.
It is easy to verify that 𝜕𝐴 = 𝐴 ⧵ 𝐴̊ = 𝐴 ∩ 𝑋 ⧵ 𝐴.
Definition 2.14. One says that 𝑥 is an isolated point of 𝐴 if 𝑥 ∈ 𝐴 and there exists a
neighborhood of 𝑥 which does not contain any other points of 𝐴.
2.4 Separation and countability 9
Equivalently, 𝑥 is an isolated point of 𝐴 if and only it is not a limit point of 𝐴. Another

equivalent formulation is the following: the singleton {𝑥} is open in the topological space
𝐴 endowed with the subspace topology (see Example 1.5).
Definition 2.15. One says that 𝐴 is dense in 𝑋 if 𝐴 = 𝑋 .
Equivalently, a subset 𝐴 ⊆ 𝑋 is dense if and only if every non-empty open set in 𝑋

intersects 𝐴. Indeed, assume that 𝐴 is dense and let 𝑈 be a non-empty open set in 𝑋 .
Let 𝑥 ∈ 𝑈 , then, since 𝑈 is a neighborhood of 𝑥 in 𝑋 and 𝑥 ∈ 𝑋 = 𝐴, one obtains that
𝑈 ∩ 𝐴 is non-empty. Conversely, assume that every non-empty open set in 𝑋 intersects
𝐴. Let 𝑥 ∈ 𝑋 and 𝑉 be a neighborhood of 𝑥 in 𝑋 . Since 𝑉 ̊ ∩ 𝐴 is non-empty, one deduces
that 𝑉 ∩ 𝐴 is non-empty, thus 𝑥 ∈ 𝐴.
One can easily verify the following properties, which are left as an exercise.
Proposition 2.16. Let 𝐴 and 𝐵 be subsets of a topological space 𝑋 , then:

(i) 𝐴 is open if and only if 𝐴 = 𝐴.̊
(ii) 𝐴 is closed if and only if 𝐴 = 𝐴.
(iii) If 𝐴 ⊆ 𝐵 then 𝐴̊ ⊆ 𝐵̊ and 𝐴 ⊆ 𝐵.
∘ ∘
(iv) 𝐴̊ ∩ 𝐵̊ = 𝐴
⏞ ∩ 𝐵 and 𝐴̊ ∪ 𝐵̊ ⊆ 𝐴
⏞ ∪ 𝐵.
(v) 𝐴 ∪ 𝐵 = 𝐴 ∪ 𝐵 and 𝐴 ∩ 𝐵 ⊆ 𝐴 ∩ 𝐵.
(vi) 𝑋 ⧵ 𝐴̊ = 𝑋 ⧵ 𝐴.
∘
⏞
(vii) 𝑋 ⧵ 𝐴 = 𝑋 ⧵ 𝐴.
2.4 Separation and countability

2.4.1 Axioms of separation
We list below some of the axioms of separation.
Definition 2.17. We say that a topological space 𝑋 is

(i) T0 (or Kolmogorov) if for any two distinct points there exist a neighborhood of one
of the points that does not contain the other.
(ii) T1 (or Fréchet) if for any two distinct points 𝑥, 𝑦 ∈ 𝑋 there exist a neighborhood
𝑈 of 𝑥 not containing 𝑦, and a neighborhood 𝑉 of 𝑦 not containing 𝑥.
(iii) Hausdorff (or T2 ) if for any two distinct points 𝑥, 𝑦 ∈ 𝑋 there exist a neighborhood
𝑈 of 𝑥 and a neighborhood 𝑉 of 𝑦 such that 𝑈 ∩ 𝑉 = ∅.
(iv) Regular (or T3 ) if it is Hausdorff and if for any point 𝑥 ∈ 𝑋 and any closed set
𝐹 ⊆ 𝑋 such that 𝑥 ∉ 𝐹 , there exist a neighborhood 𝑈 of 𝑥 and a neighborhood 𝑉
of 𝐹 such that 𝑈 ∩ 𝑉 = ∅.
(v) Normal (or T4 ) if it is Hausdorff and if for any disjoint closed sets 𝐸, 𝐹 ⊆ 𝑋 , there
exist a neighborhood 𝑈 of 𝐸 and a neighborhood 𝑉 of 𝐹 such that 𝑈 ∩ 𝑉 = ∅.
One can verify that 𝑋 is T0 if and only if for any two distinct points there is one of
them that is not an adherent point of the singleton formed by the other. Moreover, 𝑋 is
T1 if and only if every singleton is closed. Furthermore a Hausdorff space is regular if
and only if for any 𝑥 ∈ 𝑋 and any neighborhood 𝑈 of 𝑥, there exists a neighborhood 𝑉
of 𝑥 such that 𝑥 ∈ 𝑉 ⊆ 𝑉 ⊆ 𝑈 ; and it is normal if and only if for any closed set 𝐹 in 𝑋 and
any neighborhood 𝑈 of 𝐹 , there exists a neighborhood 𝑉 of 𝐹 such that 𝐹 ⊆ 𝑉 ⊆ 𝑉 ⊆ 𝑈 .
2.4.2 Axioms of countability

We list below some of the axioms of countability.
Definition 2.18. We say that a topological space 𝑋 is

(i) first-countable if every point has a countable neighborhood basis;
(ii) separable if there exists a countable dense subset of 𝑋 ;
(iii) second-countable if there is a countable basis for the topology of 𝑋 .
One can also remark that in a first-countable space every point admits a neighbor-
hood basis composed by a decreasing sequence of open sets. Indeed let 𝑥 be a point in a
first-countable space, then there exists a neighborhood basis {𝑊𝑖 }𝑖∈N of 𝑥. For any 𝑛 ∈ N
let 𝑉𝑛 = ⋂0⩽𝑖⩽𝑛 𝑊̊ 𝑖 , therefore the family {𝑉𝑛 }𝑛∈N is a decreasing sequence of open sets
and it is a neighborhood basis for 𝑥.
Proposition 2.19. Every second-countable space is first-countable and separable.
Proof. Let 𝑋 be a second-countable space, then there is a countable basis ℬ = {𝐵𝑛 }𝑛∈N
for the topology of 𝑋 .
We first show that 𝑋 is separable. For any 𝑛 ∈ N we pick some 𝑥𝑛 ∈ 𝐵𝑛 , and then
define the countable subset 𝐴 = {𝑥𝑛 ∶ 𝑛 ∈ N} of 𝑋 . We now verify that 𝐴 = 𝑋 . Let
𝑈 be a non-empty open set in 𝑋 , then there is some non-empty set 𝐽 ⊆ N such that
𝑈 = ⋃𝑛∈𝐽 𝐵𝑛 . Therefore 𝑈 ∩ 𝐴 = {𝑥𝑛 ∶ 𝑛 ∈ 𝐽 } is non-empty, thus 𝐴 = 𝑋 .
We now show that 𝑋 is first-countable. For any 𝑥 ∈ 𝑋 , define 𝒮 (𝑥) = {𝐵𝑛 ∈ ℬ ∣ 𝑥 ∈
𝐵𝑛 } which is a countable family of neighborhoods of 𝑥. Let 𝑉 be a neighborhood of 𝑥,
then there is a open set 𝑈 in 𝑋 such that 𝑥 ∈ 𝑈 ⊆ 𝑉 . Since ℬ is a basis, 𝑈 is a union
of elements of ℬ, therefore there is 𝑛 ∈ N such that 𝑥 ∈ 𝐵𝑛 ⊆ 𝑈 ⊆ 𝑉 , hence 𝒮 (𝑥) is a
neighborhood basis for 𝑥.
Proposition 2.20. Every metric space is first-countable.
Proof. Let (𝑋 , 𝑑) be a metric space. For any 𝑥 ∈ 𝑋 define 𝒮 (𝑥) = {𝐵(𝑥, 𝑛1 )}𝑛∈N∗ , which
is a countable family of neighborhoods of 𝑥. Let 𝑉 be a neighborhood of 𝑥, then there
exists a open set 𝑈 in 𝑋 such that 𝑥 ∈ 𝑈 ⊆ 𝑉 . Since 𝑈 is open, there is 𝑛 ∈ N∗ large
enough such that 𝐵(𝑥, 𝑛1 ) ⊆ 𝑈 ⊆ 𝑉 , thus 𝒮 (𝑥) is a neighborhood basis for 𝑥.
Proposition 2.21. For metric spaces second-countability and separability are equivalent.
Proof. Let (𝑋 , 𝑑) be a metric space. Thanks to Proposition 2.19 we only need to show that
separable implies second-countable. Assume 𝑋 is separable, then there is a countable
subset 𝐴 ⊆ 𝑋 such that 𝐴 = 𝑋 . Define ℬ = {𝐵(𝑦, 𝑛1 )}𝑦∈𝐴,𝑛∈N∗ , which is a countable
family of open subsets of 𝑋 .
2.4 Separation and countability 11
Let us show that ℬ is a basis for the topology of 𝑋 . Let 𝑈 be a open set in 𝑋 and
𝑥 ∈ 𝑈 . Since 𝑈 is open, there is 𝜀 > 0 such that 𝐵(𝑥, 𝜀) ⊆ 𝑈 , hence there exists 𝑛 ∈ N∗
large enough such that 𝐵(𝑥, 𝑛2 ) ⊆ 𝑈 . Since 𝐴 = 𝑋 and 𝐵(𝑥, 𝑛1 ) is a neighborhood of 𝑥,
there is a point 𝑦 ∈ 𝑋 such that 𝑦 ∈ 𝐴 ∩ 𝐵(𝑥, 𝑛1 ). Therefore 𝐵(𝑦, 𝑛1 ) ⊆ 𝐵(𝑥, 𝑛2 ) ⊆ 𝑈 .
Chapter 3
Continuity and limits
3.1 Continuous maps

Definition 3.1. A map 𝑓 ∶ 𝑋 → 𝑌 between two topological spaces is continuous at a
point 𝑎 ∈ 𝑋 if for any neighborhood 𝑉 of 𝑓 (𝑎) in 𝑌 , there is a neighborhood 𝑈 of 𝑎 in 𝑋
such that 𝑓 (𝑈 ) ⊆ 𝑉 .
We say that 𝑓 is continuous on 𝑋 if it is continuous at every point of 𝑋 .
Remark that 𝑓 (𝑈 ) ⊆ 𝑉 is equivalent to 𝑈 ⊆ 𝑓 −1 (𝑉 ), and thus we can reformulate

the definition of continuity of 𝑓 at the point 𝑎 by: for any neighborhood 𝑉 of 𝑓 (𝑎) in 𝑌 ,
𝑓 −1 (𝑉 ) is a neighborhood of 𝑎 in 𝑋 .
Thanks to the definition of a neighborhood basis, one can observe that if 𝒰 is a
neighborhood basis for 𝑎 in 𝑋 and 𝒱 a neighborhood basis for 𝑓 (𝑎) in 𝑌 , then 𝑓 is
continuous at 𝑎 if and only if for any 𝑉 ∈ 𝒱 there exists 𝑈 ∈ 𝒰 such that 𝑓 (𝑈 ) ⊆ 𝑉 .
If (𝑌 , 𝑑𝑌 ) is a metric space, then 𝑓 ∶ 𝑋 → 𝑌 is continuous at 𝑎 ∈ 𝑋 if and only
if for any 𝜀 > 0, there is a neighborhood 𝑈 of 𝑎 such that, for any 𝑥 ∈ 𝑈 , one has
𝑑𝑌 (𝑓 (𝑥), 𝑓 (𝑎)) < 𝜀.
Moreover, if both (𝑋 , 𝑑𝑋 ) and (𝑌 , 𝑑𝑌 ) are metric spaces, then 𝑓 ∶ 𝑋 → 𝑌 is continuous
at 𝑎 ∈ 𝑋 if and only if
∀𝜀 > 0, ∃𝛿 > 0, ∀𝑥 ∈ 𝑋 , 𝑑𝑋 (𝑥, 𝑎) < 𝛿 ⟹ 𝑑𝑌 (𝑓 (𝑥), 𝑓 (𝑎)) < 𝜀.
Example 3.1. A constant map between topological spaces is continuous.

Example 3.2. The identity map from a topological space into itself is continuous.
Example 3.3. Every map from a discrete space into an arbitrary topological space is
continuous.
Example 3.4. Every map from an arbitrary topological space into a space with the trivial
topology is continuous.
Example 3.5. A linear map between finite-dimensional normed spaces is continuous.
It is easy to obtain that the composition of continuous maps is continuous.
Proposition 3.2. Let 𝑋 , 𝑌 and 𝑍 be topological spaces. Suppose that 𝑓 ∶ 𝑋 → 𝑌 is

continuous at 𝑎 ∈ 𝑋 and that 𝑔 ∶ 𝑌 → 𝑍 is continuous at 𝑓 (𝑎) ∈ 𝑌 . Then 𝑔 ∘ 𝑓 ∶ 𝑋 → 𝑍
is continuous at 𝑎 ∈ 𝑋 .
13
14 Chapter 3. Continuity and limits
Proof. Let 𝑉 be a neighborhood of (𝑔 ∘ 𝑓 )(𝑎) = 𝑔(𝑓 (𝑎)) in 𝑍 . Since 𝑔 is continuous at

𝑓 (𝑎), one has that 𝑔 −1 (𝑉 ) is a neighborhood of 𝑓 (𝑎) in 𝑌 . Since 𝑓 is continuous at 𝑎,
𝑓 −1 (𝑔 −1 (𝑉 )) = (𝑔 ∘ 𝑓 )−1 (𝑉 ) is a neighborhood of 𝑎 in 𝑋 .
We prove now equivalent formulations of continuity.
Theorem 3.3. Let 𝑓 ∶ 𝑋 → 𝑌 be a map between two topological spaces. The following
are equivalent:
(i) 𝑓 is continuous on 𝑋 ;
(ii) for any open set 𝑈 in 𝑌 , 𝑓 −1 (𝑈 ) is an open set in 𝑋 ;
(iii) for any closed set 𝐹 in 𝑌 , 𝑓 −1 (𝐹 ) is a closed set in 𝑋 ;
(iv) for any 𝐴 ⊆ 𝑋 , one has 𝑓 (𝐴) ⊆ 𝑓 (𝐴).
If ℬ is a basis for the topology of 𝑌 and 𝒮 is a sub-basis for the topology of 𝑌 ,

then one can easily verify that (ii) is equivalent to each one of the following equivalent
properties:
(ii’) for any 𝑈 ∈ ℬ, 𝑓 −1 (𝑈 ) is an open set in 𝑋 ;
(ii”) for any 𝑈 ∈ 𝒮 , 𝑓 −1 (𝑈 ) is an open set in 𝑋 .
Indeed the implications (ii) ⇒ (ii’) ⇒ (ii”) are trivial. The converse follows from the fact
that an open set in 𝑋 is a union of elements of ℬ, and also a union of finite intersections
of elements of 𝒮 , together with the property that the preimage under 𝑓 commutes with
unions and intersections.
Proof of Theorem 3.3. (i) ⇔ (ii) Assume that 𝑓 is continuous on 𝑋 . Let 𝑉 be an open set
in 𝑌 , we shall prove that 𝑓 −1 (𝑉 ) is a neighborhood of each of its points, which implies
that 𝑓 −1 (𝑉 ) is open in 𝑋 . Let 𝑥 ∈ 𝑓 −1 (𝑉 ), then 𝑓 (𝑥) ∈ 𝑉 and thus 𝑉 is a neighborhood of
𝑓 (𝑥). Since 𝑓 is continuous at 𝑥, there is a neighborhood 𝑈 of 𝑥 in 𝑋 such that 𝑓 (𝑈 ) ⊆ 𝑉 ,
that is 𝑈 ⊆ 𝑓 −1 (𝑉 ), and hence 𝑓 −1 (𝑉 ) is a neighborhood of 𝑥. Conversely, assume that
for any open set 𝑈 in 𝑌 , 𝑓 −1 (𝑈 ) is open in 𝑋 . Let 𝑥 ∈ 𝑋 and 𝑉 be a neighborhood of
𝑓 (𝑥) in 𝑌 . There is an open set 𝑈 ⊆ 𝑌 such that 𝑓 (𝑥) ∈ 𝑈 ⊆ 𝑉 , which implies that
𝑥 ∈ 𝑓 −1 (𝑈 ) ⊆ 𝑓 −1 (𝑉 ) and 𝑓 −1 (𝑈 ) is open in 𝑋 , hence 𝑓 −1 (𝑉 ) is a neighborhood of 𝑥 in
𝑋.
(ii) ⇔ (iii) It follows from the relation 𝑋 ⧵ 𝑓 −1 (𝐵) = 𝑓 −1 (𝑌 ⧵ 𝐵) for any subset 𝐵 of 𝑌 .
(ii) ⇒ (iv) Assume that for any open set 𝑈 in 𝑌 , 𝑓 −1 (𝑈 ) is open in 𝑋 . Let 𝑦 ∈ 𝑓 (𝐴)
and 𝑉 be an open neighborhood of 𝑦 in 𝑌 . Consider 𝑥 ∈ 𝐴 such that 𝑓 (𝑥) = 𝑦, hence
𝑓 −1 (𝑉 ) is an open neighborhood of 𝑥 in 𝑋 . By definition, 𝑓 −1 (𝑉 ) intersects 𝐴 and thus
𝑉 intersects 𝑓 (𝐴), which implies that 𝑦 ∈ 𝑓 (𝐴).
(iv) ⇒ (iii) Assume that for any 𝐴 ⊆ 𝑋 , one has 𝑓 (𝐴) ⊆ 𝑓 (𝐴). Let 𝐹 be a closed set
in 𝑌 and denote 𝐴 = 𝑓 −1 (𝐹 ). One has 𝑓 (𝐴) ⊆ 𝑓 (𝐴) = 𝑓 (𝑓 −1 (𝐹 )) ⊆ 𝐹 = 𝐹 , therefore
𝐴 ⊆ 𝑓 −1 (𝐹 ) = 𝐴 and 𝐴 is closed.
3.1.1 Open and closed maps, homeomorphisms

Definition 3.4. Let 𝑋 and 𝑌 be topological spaces. A map 𝑓 ∶ 𝑋 → 𝑌 is said to be open
(resp. closed) if the image under 𝑓 of every open (resp. closed) set in 𝑋 is open (resp.
closed) in 𝑌 .
3.2 Construction of topologies 15
Definition 3.5. A bijective map 𝑓 ∶ 𝑋 → 𝑌 between topological spaces is a homeomor-

phism if both 𝑓 and its inverse map 𝑓 −1 are continuous. Two topological spaces 𝑋 and
𝑌 are said to be homeomorphic if there exists a homeomorphism between them.
In other words, a bijection 𝑓 ∶ 𝑋 → 𝑌 is an homeomorphism if and only if the image
under 𝑓 of every open set in 𝑋 is a open set in 𝑌 and the preimage of every open set in
𝑌 is a open set in 𝑋 , that is, if and only if 𝑓 is continuous and open.
In the framework of metric spaces one defines the following:
Definition 3.6. A map 𝑓 ∶ 𝑋 → 𝑌 between metric spaces (𝑋 , 𝑑𝑋 ) and (𝑌 , 𝑑𝑌 ) is called an
isometry if for any 𝑥, 𝑦 ∈ 𝑋 one has 𝑑𝑌 (𝑓 (𝑥), 𝑓 (𝑦)) = 𝑑𝑋 (𝑥, 𝑦). An isometric isomorphism
is a bijective isometry. Two metric spaces 𝑋 and 𝑌 are said to be isometric if there exists
an isometric isomorphism between them.
3.2 Construction of topologies

3.2.1 Initial topology
Definition 3.7. Let 𝑋 be a set, {𝑌𝑖 }𝑖∈𝐼 a collection of topological spaces and 𝑓𝑖 ∶ 𝑋 → 𝑌𝑖
a map, for any 𝑖 ∈ 𝐼 . The initial topology on 𝑋 defined by {𝑓𝑖 }𝑖∈𝐼 is the coarsest topology
on 𝑋 for which every map 𝑓𝑖 is continuous. This topology is generated by {𝑓𝑖−1 (𝑈𝑖 ) ∶ 𝑖 ∈
𝐼 , 𝑈𝑖 open in 𝑌𝑖 }.
The properties below easily follow from the definition:
Proposition 3.8. Let 𝑋 be endowed with the initial topology defined by the family of maps
{𝑓𝑖 ∶ 𝑋 → 𝑌𝑖 }𝑖∈𝐼 as above.
(i) If ℬ𝑖 is a basis for the topology on 𝑌𝑖 for all 𝑖 ∈ 𝐼 , then the family ℬ of finite
intersection of elements of {𝑓𝑖−1 (𝐵𝑖 ) ∶ 𝑖 ∈ 𝐼 , 𝐵𝑖 ∈ ℬ𝑖 } is a basis for the initial topology
on 𝑋 . In particular, if 𝐼 is countable and 𝑌𝑖 is second-countable for all 𝑖 ∈ 𝐼 ,then 𝑋 is
second-countable.
(ii) If 𝑥 ∈ 𝑋 and 𝒮𝑖 is a neighborhood basis for 𝑓𝑖 (𝑥) on 𝑌𝑖 , then the family 𝒮 of finite
intersection of elements of {𝑓𝑖−1 (𝑉𝑖 ) ∶ 𝑖 ∈ 𝐼 , 𝑉𝑖 ∈ 𝒮𝑖 } is a neighborhood basis for 𝑥
on 𝑋 . In particular, if 𝐼 is countable and 𝑌𝑖 is first-countable for all 𝑖 ∈ 𝐼 , then 𝑋 is
first-countable.
(iii) Let 𝑍 be a topological space and 𝑔 ∶ 𝑍 → 𝑋 be a map. Then 𝑔 is continuous if and
only if every 𝑓𝑖 ∘ 𝑔 is continuous.
Proof. Let us denote by 𝒪 the initial topology on 𝑋 and by 𝒪𝑖 the topology of 𝑌𝑖 , for all
𝑖 ∈ 𝐼.
(i) Let 𝑈 ∈ 𝒪 and 𝑥 ∈ 𝑈 . By definition of 𝒪, 𝑈 is a union of finite intersections of elements
of {𝑓𝑖−1 (𝑈𝑖 ) ∶ 𝑖 ∈ 𝐼 , 𝑈𝑖 ∈ 𝒪𝑖 }, hence there is a finite set 𝐽 ⊆ 𝐼 and, for each 𝑖 ∈ 𝐽 , an open
set 𝑈𝑖 ∈ 𝒪𝑖 such that
𝑥 ∈ ⋂ 𝑓𝑖−1 (𝑈𝑖 ) ⊆ 𝑈 .
𝑖∈𝐽
For any 𝑖 ∈ 𝐽 one has 𝑥 ∈ 𝑓𝑖−1 (𝑈𝑖 ), thus 𝑓𝑖 (𝑥) ∈ 𝑈𝑖 . Since ℬ𝑖 is a basis for the topology 𝒪𝑖 ,
there is 𝐵𝑖 ∈ ℬ𝑖 such that 𝑓𝑖 (𝑥) ∈ 𝐵𝑖 ⊆ 𝑈𝑖 , hence 𝑥 ∈ 𝑓𝑖−1 (𝐵𝑖 ) ⊆ 𝑓 −1 (𝑈𝑖 ). Therefore,
𝑥 ∈ ⋂ 𝑓𝑖−1 (𝐵𝑖 ) ⊆ ⋂ 𝑓𝑖−1 (𝑈𝑖 ) ⊆ 𝑈 ,
𝑖∈𝐽 𝑖∈𝐽
and the set in the left-hand side belongs to ℬ.

(ii) Let 𝑥 ∈ 𝑋 and 𝑉 be a neighborhood of 𝑥. Then there is 𝑈 ∈ 𝒪 such that 𝑥 ∈ 𝑈 ⊆ 𝑉 .
By definition, there is a finite set 𝐽 ⊆ 𝐼 and, for each 𝑖 ∈ 𝐽 , an open set 𝑈𝑖 ∈ 𝒪𝑖 such that
𝑥 ∈ ⋂ 𝑓𝑖−1 (𝑈𝑖 ) ⊆ 𝑉 .
𝑖∈𝐽
For any 𝑖 ∈ 𝐽 one has 𝑥 ∈ 𝑓𝑖−1 (𝑈𝑖 ), thus 𝑓𝑖 (𝑥) ∈ 𝑈𝑖 . Since 𝒮𝑖 is a neighborhood basis for
𝑓𝑖 (𝑥) in 𝑌𝑖 and 𝑈𝑖 is a neighborhood of 𝑓𝑖 (𝑥), there is some 𝑉𝑖 ∈ 𝒮𝑖 such that 𝑓𝑖 (𝑥) ∈ 𝑉𝑖 ⊆ 𝑈𝑖 ,
hence 𝑥 ∈ 𝑓𝑖−1 (𝑉𝑖 ) ⊆ 𝑓 −1 (𝑈𝑖 ). Therefore,
𝑥 ∈ ⋂ 𝑓𝑖−1 (𝑉𝑖 ) ⊆ ⋂ 𝑓𝑖−1 (𝑈𝑖 ) ⊆ 𝑉 ,

and the set in the left-hand side belongs to 𝒮 .

(iii) Suppose 𝑔 is continuous. Since 𝑓𝑖 is continuous for each 𝑖 ∈ 𝐼 , by composition each
𝑓𝑖 ∘ 𝑔 is continuous.
Conversely, suppose that 𝑓𝑖 ∘ 𝑔 is continuous for each 𝑖 ∈ 𝐼 . Let 𝑧 ∈ 𝑍 and 𝑉 be a
neighborhood of 𝑔(𝑧) in 𝑋 . Hence there is a finite set 𝐽 ⊆ 𝐼 and for each 𝑖 ∈ 𝐽 an open
set 𝑈𝑖 ∈ 𝒪𝑖 such that
𝑔(𝑧) ∈ ⋂ 𝑓𝑖−1 (𝑈𝑖 ) ⊆ 𝑉 .

𝑖∈𝐽
Therefore
𝑧 ∈ ⋂ 𝑔 −1 (𝑓𝑖−1 (𝑈𝑖 )) = ⋂(𝑓𝑖 ∘ 𝑔)−1 (𝑈𝑖 ) ⊆ 𝑔 −1 (𝑉 ),
and the set in the left-hand side is open in 𝑋 since it is a finite intersection of open sets,
hence 𝑔 −1 (𝑉 ) is a neighborhood of 𝑧 in 𝑍 , which completes the proof.
Two particular examples of initial topologies are the subspace topology and the prod-
uct topology that we describe below.
3.2.1.1 Subspace topology

Definition 3.9. Let (𝑋 , 𝒪) be a topological space and 𝑌 ⊆ 𝑋 . We define the subspace
topology 𝒪𝑌 on 𝑌 as the trace on 𝑌 of the topology of 𝑋 , namely
𝒪𝑌 ∶= {𝑈 ∩ 𝑌 ∣ 𝑈 ∈ 𝒪}.
The topological space (𝑌 , 𝒪𝑌 ) is called a (topological) subspace of (𝑋 , 𝒪).

Let us verify that 𝒪𝑌 is indeed a topology on 𝑌 . Since ∅, 𝑋 ∈ 𝒪, we get ∅ = ∅ ∩ 𝑌 ,
𝑌 = 𝑋 ∩ 𝑌 ∈ 𝒪𝑌 . Let {𝑉𝑖 }1⩽𝑖⩽𝑛 be a finite family of elements of 𝒪𝑌 , then there exist
𝑈𝑖 ∈ 𝒪 such that 𝑉𝑖 = 𝑈𝑖 ∩ 𝑌 for any 𝑖 ∈ {1, … , 𝑛}. We have 𝑈 ∶= ⋂1⩽𝑖⩽𝑛 𝑈𝑖 ∈ 𝒪 and
𝑉 ∶= ⋂1⩽𝑖⩽𝑛 𝑉𝑖 = (⋂1⩽𝑖⩽𝑛 𝑈𝑖 ) ∩ 𝑌 , whence 𝑉 ∈ 𝒪𝑌 . Let {𝑉𝑖 }𝑖∈𝐼 be an arbitrary family of
elements of 𝒪𝑌 , then there exist 𝑈𝑖 ∈ 𝒪 such that 𝑉𝑖 = 𝑈𝑖 ∩ 𝑌 for any 𝑖 ∈ 𝐼 . We have
𝑈 ∶= ⋃𝑖∈𝐼 𝑈𝑖 ∈ 𝒪 and 𝑉 ∶= ⋃𝑖∈𝐼 𝑉𝑖 = (⋃𝑖∈𝐼 𝑈𝑖 ) ∩ 𝑌 , whence 𝑉 ∈ 𝒪𝑌 .
In this way, given a topological space 𝑋 and a subset 𝑌 ⊆ 𝑋 , we shall always consider
𝑌 endowed with the subspace topology, unless otherwise specified.
The following properties are easy to verify.
3.2 Construction of topologies 17
Proposition 3.10. Let (𝑋 , 𝒪) be a topological space and (𝑌 , 𝒪𝑌 ) a subspace of 𝑋 .

(i) The topology 𝒪𝑌 is the coarsest topology on 𝑌 for which the inclusion map 𝜄 ∶ 𝑌 → 𝑋 ,
𝑥 ↦ 𝑥 is continuous. In other words, the subspace topology is the initial topology on
𝑌 defined by the inclusion map.
(ii) For any 𝑥 ∈ 𝑌 , 𝑉 ⊆ 𝑌 is a neighborhood of 𝑥 in 𝑌 if and only if there is a neighborhood
𝑉 ′ of 𝑥 in 𝑋 such that 𝑉 = 𝑉 ′ ∩ 𝑌 . In particular, if 𝑋 is Hausdorff then so is 𝑌 .
(iii) If ℬ is a basis for the topology on 𝑋 , then ℬ𝑌 ∶= {𝑈 ∩ 𝑌 ∣ 𝑈 ∈ ℬ} is a basis for the
subspace topology on 𝑌 . In particular, if 𝑋 is second-countable then so is 𝑌 .
(iv) For any 𝑥 ∈ 𝑌 , if 𝒮 is a neighborhood basis for 𝑥 in 𝑋 , then 𝒮𝑌 ∶= {𝑈 ∩ 𝑌 ∣ 𝑈 ∈ 𝒮 }
is a neighborhood basis for 𝑥 in 𝑌 . In particular, if 𝑋 is first-countable then so is 𝑌 .
Proof. (i) It is clear that 𝜄 ∶ (𝑌 , 𝒪𝑌 ) → (𝑋 , 𝒪) is continuous, indeed if 𝑈 ∈ 𝒪, then
𝜄 −1 (𝑈 ) = 𝑈 ∩ 𝑌 ∈ 𝒪𝑌 . Now let 𝒯 be a topology on 𝑌 such that 𝜄 ∶ (𝑌 , 𝒯 ) → (𝑋 , 𝒪) is
continuous, then for any 𝑈 ∈ 𝒪 we have 𝑈 ∩ 𝑌 ∈ 𝒯 , hence 𝒯 ⊇ 𝒪𝑌 .
(ii) Let 𝑉 be a neighborhood of 𝑥 in 𝑌 , then there exists 𝑈 ∈ 𝒪𝑌 such that 𝑥 ∈ 𝑈 ⊆ 𝑉 .
Hence there is 𝑈 ′ ∈ 𝒪 such that 𝑈 = 𝑈 ′ ∩ 𝑌 and 𝑈 ′ is a neighborhood of 𝑥 in 𝑋 . Thus
the set 𝑉 ′ = 𝑉 ∪ 𝑈 ′ is a neighborhood of 𝑥 in 𝑋 and satisfies 𝑉 = 𝑉 ′ ∩ 𝑌 .
Conversely, let 𝑉 ′ be a neighborhood of 𝑥 in 𝑋 , then there is 𝑈 ′ ∈ 𝒪 such that
𝑥 ∈ 𝑈 ′ ⊆ 𝑉 ′ . Since 𝑥 ∈ 𝑌 , one has 𝑥 ∈ 𝑈 ′ ∩ 𝑌 ⊆ 𝑉 ′ ∩ 𝑌 and 𝑈 ′ ∩ 𝑌 ∈ 𝒪𝑌 , thus 𝑉 ′ ∩ 𝑌 is a
neighborhood of 𝑥 in 𝑌 .
(iii) Let 𝑈 ∈ 𝒪𝑌 and 𝑥 ∈ 𝑈 . We want to show that there exists 𝑊 ∈ ℬ𝑌 such that
𝑥 ∈ 𝑊 ⊆ 𝑈 . There is 𝑈 ′ ∈ 𝒪 such that 𝑈 = 𝑈 ′ ∩ 𝑌 . Since ℬ is a basis for 𝒪, there is
𝑊 ′ ∈ ℬ such that 𝑥 ∈ 𝑊 ′ ⊆ 𝑈 ′ . Hence 𝑊 ′ ∩ 𝑌 ∈ ℬ𝑌 and 𝑥 ∈ 𝑊 ′ ∩ 𝑌 ⊆ 𝑈 ′ ∩ 𝑌 .
(iv) Let 𝑉 be a neighborhood of 𝑥 in 𝑌 . We want to show that there exists 𝑊 ∈ 𝒮𝑌 such
that 𝑥 ∈ 𝑊 ⊆ 𝑉 . There is a neighborhood 𝑉 ′ of 𝑥 in 𝑋 such that 𝑉 = 𝑉 ′ ∩ 𝑌 . Since 𝒮
is a neighborhood basis for 𝑥 in 𝑋 , there is 𝑊 ′ ∈ 𝒮 such that 𝑥 ∈ 𝑊 ′ ⊆ 𝑉 ′ . Therefore
𝑊 ′ ∩ 𝑌 ∈ 𝒮𝑌 and 𝑥 ∈ 𝑊 ′ ∩ 𝑌 ⊆ 𝑉 ′ ∩ 𝑌 = 𝑉 .
3.2.1.2 Product topology

Definition 3.11. Consider a collection {𝑋𝑖 }𝑖∈𝐼 of topological spaces. On the product set
𝑋 = ∏𝑖∈𝐼 𝑋𝑖 , we define the product topology as the coarsest topology on 𝑋 for which the
canonical projection pr𝑗 ∶ 𝑋 → 𝑋𝑗 , (𝑥𝑖 )𝑖∈𝐼 ↦ 𝑥𝑗 is continuous for all 𝑗 ∈ 𝐼 .
In other words, the product topology is the initial topology on 𝑋 defined by the
canonical projections {pr𝑖 }𝑖∈𝐼 .
The properties below follow from the properties of the initial topology (see Proposi-
tion 3.8).
(i) A basis for this topology is given by the collection of sets of the form
{(𝑥𝑖 )𝑖∈𝐼 ∈ 𝑋 ∣ ∀𝑗 ∈ 𝐽 , 𝑥𝑗 ∈ 𝑈𝑗 } = ⋂ pr−1

𝑗 (𝑈𝑗 )
𝑗∈𝐽
where 𝐽 ⊆ 𝐼 is finite and 𝑈𝑗 is open in 𝑋𝑗 for any 𝑗 ∈ 𝐽 .

(ii) If ℬ𝑖 is a basis for the topology of 𝑋𝑖 for all 𝑖 ∈ 𝐼 , then the family of sets of the
form
{(𝑥𝑖 )𝑖∈𝐼 ∈ 𝑋 ∣ ∀𝑗 ∈ 𝐽 , 𝑥𝑗 ∈ 𝑈𝑗 } = ⋂ pr−1
𝑗 (𝑈𝑗 ),
𝑗∈𝐽
where 𝐽 ⊆ 𝐼 is finite and 𝑈𝑗 ∈ ℬ𝑗 for any 𝑗 ∈ 𝐽 , is a basis for the product topology.
(iii) Let 𝑥 = (𝑥𝑖 )𝑖∈𝐼 ∈ 𝑋 and 𝒱𝑖 be a neighborhood basis for 𝑥𝑖 in 𝑋𝑖 for any 𝑖 ∈ 𝐼 , then
the family of sets of the form
{(𝑦𝑖 )𝑖∈𝐼 ∈ 𝑋 ∣ ∀𝑗 ∈ 𝐽 , 𝑦𝑗 ∈ 𝑉𝑗 } = ⋂ pr−1

𝑗 (𝑉𝑗 ),
𝑗∈𝐽
where 𝐽 ⊆ 𝐼 is finite and 𝑉𝑗 ∈ 𝒱𝑗 for any 𝑗 ∈ 𝐽 , is a neighborhood basis for 𝑥 in 𝑋

for the product topology.
The next result shows that any product of Hausdorff spaces is also Hausdorff.
Proposition 3.12. Let {𝑋𝑖 }𝑖∈𝐼 be a family of Hausdorff topological spaces. Then the product
space 𝑋 = ∏𝑖∈𝐼 𝑋𝑖 is Hausdorff.
Proof. Let 𝑥 = (𝑥𝑖 )𝑖∈𝐼 ≠ (𝑦𝑖 )𝑖∈𝐼 = 𝑦 belong to 𝑋 . There is 𝑘 ∈ 𝐼 such that 𝑥𝑘 ≠ 𝑦𝑘 thus,
since 𝑋𝑘 is Hausdorff, there are disjoint open sets 𝑈𝑘 and 𝑉𝑘 in 𝑋𝑘 such that 𝑥𝑘 ∈ 𝑈𝑘 and
𝑦𝑘 ∈ 𝑉𝑘 . Therefore the sets pr−1𝑘
(𝑈𝑘 ) and pr−1
𝑘
(𝑉𝑘 ) are disjoint open sets in 𝑋 containing
𝑥 and 𝑦, respectively.
3.2.2 Final topology

Definition 3.13. Let 𝑋 be a set, {𝑌𝑖 }𝑖∈𝐼 a collection of topological spaces and 𝑓𝑖 ∶ 𝑌𝑖 → 𝑋
a map, for any 𝑖 ∈ 𝐼 . The final topology on 𝑋 defined by {𝑓𝑖 }𝑖∈𝐼 is the finest topology on
𝑋 for which every map 𝑓𝑖 is continuous. This topology is given by the collection of all
subsets 𝑈 ⊆ 𝑋 such that 𝑓𝑖−1 (𝑈 ) is open in 𝑌𝑖 for any 𝑖 ∈ 𝐼 .
We easily verify the following property:
Proposition 3.14. Let 𝑋 be endowed with the final topology associated to the family of
maps {𝑓𝑖 ∶ 𝑌𝑖 → 𝑋 }𝑖∈𝐼 as above. Let 𝑍 be a topological space and 𝑔 ∶ 𝑋 → 𝑍 a map. Then
𝑔 is continuous if and only if 𝑔 ∘ 𝑓𝑖 is continuous for every 𝑖 ∈ 𝐼 .
Proof. Assume that 𝑔 is continuous. Since 𝑓𝑖 is continuous for every 𝑖 ∈ 𝐼 , we deduce
that 𝑔 ∘ 𝑓𝑖 is continuous by composition.
Conversely, assume that 𝑔 ∘ 𝑓𝑖 is continuous for every 𝑖 ∈ 𝐼 . Let 𝑉 be open in 𝑍 , then
(𝑔 ∘ 𝑓𝑖 )−1 (𝑉 ) = 𝑓𝑖−1 (𝑔 −1 (𝑉 )) is open in 𝑌𝑖 for every 𝑖 ∈ 𝐼 , hence by definition 𝑔 −1 (𝑉 ) is
open for the final topology. This implies that 𝑔 is continuous.
A particular example of final topology is the quotient topology described below.
3.2.2.1 Quotient topology

Let 𝑋 be a topological space and ℛ be an equivalence relation on 𝑋 . Denote by 𝑋 /ℛ
the quotient set and consider the canonical projection 𝜋 ∶ 𝑋 → 𝑌 , 𝑥 ↦ [𝑥], that sends
every point to its equivalence class.
Definition 3.15. We define the quotient topology on 𝑋 /ℛ as the finest topology on
𝑋 /ℛ for which 𝜋 is continuous.
In other words, the quotient topology is the final topology on 𝑋 /ℛ defined by the
map 𝜋. Therefore, this topology is given by the collection of subsets 𝑈 ⊆ 𝑋 /ℛ such
that 𝜋 −1 (𝑈 ) is open in 𝑋 .
3.3 Limits and cluster points of a sequence 19
3.3 Limits and cluster points of a sequence

Let 𝑋 be a topological space. A sequence (𝑥𝑛 )𝑛⩾0 in 𝑋 corresponds to a map 𝑛 ↦ 𝑥𝑛
from N into 𝑋 .
One says that a sequence (𝑦𝑛 )𝑛⩾0 in 𝑋 is a subsequence of (𝑥𝑛 )𝑛⩾0 if there is a strictly
increasing function 𝜑 ∶ N → N such that
𝑦𝑛 = 𝑥𝜑(𝑛) for all 𝑛 ∈ N.
The function 𝜑 is called an extraction map and it verifies 𝜑(𝑛) ⩾ 𝑛 for all 𝑛 ∈ N.
Let (𝑥𝑛 )𝑛⩾0 be a sequence in 𝑋 and 𝑥 ∈ 𝑋 a point.
Definition 3.16. One says that (𝑥𝑛 )𝑛⩾0 converges to 𝑥 when 𝑛 goes to +∞ if for any
neighborhood 𝑉 of 𝑥 in 𝑋 , there is 𝑁0 ∈ N such that, for any 𝑛 ⩾ 𝑁0 , one has 𝑥𝑛 ∈ 𝑉 .
When (𝑥𝑛 )𝑛⩾0 converges to 𝑥, we denote it by
lim 𝑥𝑛 = 𝑥 or 𝑥𝑛 −−−−−→ 𝑥.
𝑛→+∞ 𝑛→+∞
We observe that if (𝑋 , 𝑑) is a metric space, then (𝑥𝑛 )𝑛⩾0 converges to 𝑥 if and only if the
real sequence (𝑑(𝑥𝑛 , 𝑥))𝑛⩾0 converges to 0.
Definition 3.17. One says that 𝑥 is a cluster point of the sequence (𝑥𝑛 )𝑛⩾0 if for any
neighborhood 𝑉 of 𝑥 in 𝑋 and any 𝑁 ∈ N, there exists an integer 𝑛 ⩾ 𝑁 such that
𝑥𝑛 ∈ 𝑉 .
It easily follows from the definition that the set of all cluster points of a sequence
(𝑥𝑛 )𝑛⩾0 is given by
⋂ {𝑥𝑝 ∶ 𝑝 ⩾ 𝑛}.
𝑛∈N
We now prove some basic properties concerning the sequential caracterization of

closed sets and the closure of a set.
Proposition 3.18. Let 𝐴 be a subset of 𝑋 . If there is a sequence of elements of 𝐴 converging
to 𝑥 ∈ 𝑋 , then 𝑥 ∈ 𝐴. If 𝑋 is first-countable then the converse also holds.
Proof. Let (𝑥𝑛 )𝑛⩾0 be a sequence of elements of 𝐴 that converges to 𝑥 ∈ 𝑋 . Then for any
neighborhood 𝑉 of 𝑥, there is 𝑁 ∈ N such that 𝑥𝑁 ∈ 𝑉 and 𝑥𝑁 ∈ 𝐴, hence 𝑉 ∩ 𝐴 ≠ ∅.
This proves that 𝑥 ∈ 𝐴.
Assume now that 𝑋 is first-countable and let 𝑥 ∈ 𝐴. Let {𝑉𝑛 }𝑛∈N be a countable
neighborhood basis for 𝑥 composed by a decreasing sequence of opens sets. For any
𝑛 ∈ N there exists 𝑥𝑛 ∈ 𝑉𝑛 ∩ 𝐴 ≠ ∅. We now verify that the sequence (𝑥𝑛 )𝑛⩾0 converges
to 𝑥. Let 𝑉 be a neighborhood of 𝑥 in 𝑋 , then there is 𝑁0 ∈ N such that 𝑥 ∈ 𝑉𝑁0 ⊆ 𝑉 .
Therefore for any 𝑛 ⩾ 𝑁0 one has 𝑥𝑛 ∈ 𝑉𝑛 ⊆ 𝑉𝑁0 ⊆ 𝑉 .
As a consequence, closed sets in a first-countable space 𝑋 can be characterized by
sequences: A subset 𝐴 ⊆ 𝑋 is closed if and only if for every sequence of elements of 𝐴
that converges in 𝑋 , the limit belongs to 𝐴.
Proposition 3.19. Let (𝑥𝑛 )𝑛⩾0 be a sequence in 𝑋 . If there exists some subsequence of
(𝑥𝑛 )𝑛⩾0 converging to 𝑥 ∈ 𝑋 , then 𝑥 is a cluster point of (𝑥𝑛 )𝑛⩾0 . If 𝑋 is first-countable,
then the converse also holds.
Proof. Let (𝑥𝜑(𝑛) )𝑛⩾0 be a subsequence of (𝑥𝑛 )𝑛⩾0 that converges to 𝑥. Let 𝑉 be a neigh-
borhood of 𝑥 and 𝑁0 ∈ N. Since (𝑥𝜑(𝑛) )𝑛⩾0 converges to 𝑥, there is 𝑁1 ∈ N such that for
all integer 𝑛 ⩾ 𝑁1 one has 𝑥𝜑(𝑛) ∈ 𝑉 . Choosing 𝑛 = max(𝑁0 , 𝑁1 ) ∈ N, one has 𝜑(𝑛) ⩾ 𝑁0
and 𝑥𝜑(𝑛) ∈ 𝑉 . This proves that 𝑥 is a cluster point of (𝑥𝑛 )𝑛⩾0 .
Assume now that 𝑋 is first-countable and let 𝑥 be a cluster point of the sequence
(𝑥𝑛 )𝑛⩾0 . Let {𝑉𝑛 }𝑛∈N be a countable neighborhood basis for 𝑥 composed by a decreasing
sequence of opens sets. Since 𝑉0 is an open neighborhood of 𝑥, there is 𝑛0 ∈ N such
that 𝑥𝑛0 ∈ 𝑉0 . By induction one can therefore construct a sequence (𝑥𝑛𝑘 )𝑘⩾0 such that
𝑛𝑘 > 𝑛𝑘−1 and 𝑥𝑛𝑘 ∈ 𝑉𝑘 for all 𝑘 ∈ N∗ , thus (𝑥𝑛𝑘 )𝑘⩾0 is a subsequence of (𝑥𝑛 )𝑛⩾0 .
Let us now verify that (𝑥𝑛𝑘 )𝑘⩾0 converges to 𝑥. Let 𝑉 be a neighborhood of 𝑥, then
there is 𝑁0 ∈ N such that 𝑥 ∈ 𝑉𝑁0 ⊆ 𝑉 . Hence for any 𝑘 ⩾ 𝑁0 one has 𝑛𝑘 ⩾ 𝑘 and thus
𝑥𝑛𝑘 ∈ 𝑉𝑛𝑘 ⊆ 𝑉𝑁0 ⊆ 𝑉 .
As a consequence, in a first-countable space 𝑋 the closure of a subset can be charac-
terized using sequences: If 𝐴 is a subset of 𝑋 , then the closure 𝐴 is the set of all limits
of all convergent sequences of points in 𝐴.
Chapter 4
Completeness
In this chapter we shall always consider metric spaces.
4.1 Cauchy sequences

Definition 4.1. A sequence (𝑥𝑛 )𝑛⩾0 in a metric space (𝑋 , 𝑑) is a Cauchy sequence if
∀𝜀 > 0, ∃𝑁0 ∈ N, ∀𝑛, 𝑚 ⩾ 𝑁0 , 𝑑(𝑥𝑛 , 𝑥𝑚 ) < 𝜀.
One easily verifies that every convergent sequence is a Cauchy sequence, and that
every Cauchy sequence is bounded.
Proposition 4.2. Let (𝑋 , 𝑑) be a metric space, then the following properties hold:
(i) Every convergent sequence is a Cauchy sequence.
(ii) Every Cauchy sequence is bounded.
Proof. (i) Let (𝑥𝑛 )𝑛⩾0 be a sequence in 𝑋 that converges to 𝑥 ∈ 𝑋 . Let 𝜀 > 0, then there
is 𝑁0 ∈ N such that 𝑑(𝑥𝑛 , 𝑥) < 𝜀/2 for all 𝑛 ⩾ 𝑁0 . Thanks to the triangle inequality, for
all 𝑛, 𝑚 ⩾ 𝑁0 one obtains
𝑑(𝑥𝑛 , 𝑥𝑚 ) ⩽ 𝑑(𝑥𝑛 , 𝑥) + 𝑑(𝑥𝑚 , 𝑥) < 𝜀.
(ii) Let (𝑥𝑛 )𝑛⩾0 be a Cauchy sequence in 𝑋 . We want to show that there is some 𝑎 ∈ 𝑋
and 𝑀 > 0 such that 𝑑(𝑥𝑛 , 𝑎) ⩽ 𝑀 for all 𝑛 ∈ N. There is 𝑁0 ∈ N such that 𝑑(𝑥𝑛 , 𝑥𝑁0 ) ⩽ 1
for all 𝑛 ⩾ 𝑁0 . Denote
𝑀 ′ ∶= max {𝑑(𝑥0 , 𝑥𝑁0 ), … , 𝑑(𝑥𝑁0 −1 , 𝑥𝑁0 )}
then for all 𝑛 ∈ N we get 𝑑(𝑥𝑛 , 𝑥𝑁0 ) ⩽ max{𝑀 ′ , 1}.
We now prove another property concerning the convergence of Cauchy sequences.
Proposition 4.3. If a Cauchy sequence in a metric space (𝑋 , 𝑑) has a cluster point, then
it converges to it.
21
22 Chapter 4. Completeness
Proof. Let (𝑥𝑛 )𝑛⩾0 be a Cauchy sequence in 𝑋 that has a cluster point 𝑥 ∈ 𝑋 . From
Proposition 3.19, there is some subsequence (𝑥𝜑(𝑛) )𝑛⩾0 that converges to 𝑥. Let 𝜀 > 0. On
the one hand, since (𝑥𝜑(𝑛) )𝑛⩾0 converges to 𝑥, there is 𝑁0 ∈ N such that 𝑑(𝑥𝜑(𝑛) , 𝑥) < 𝜀/2
for any integer 𝑛 ⩾ 𝑁0 . On the other hand, since (𝑥𝑛 )𝑛⩾0 is a Cauchy sequence, there
is 𝑁1 ∈ N such that 𝑑(𝑥𝑛 , 𝑥𝑚 ) < 𝜀/2 for any integers 𝑛, 𝑚 ⩾ 𝑁1 . Therefore for any
𝑛 ⩾ max(𝑁0 , 𝑁1 ) it follows
𝑑(𝑥𝑛 , 𝑥) ⩽ 𝑑(𝑥𝑛 , 𝑥𝜑(𝑛) ) + 𝑑(𝑥𝜑(𝑛) , 𝑥) < 𝜀.
4.2 Complete metric spaces

Definition 4.4. A metric space (𝑋 , 𝑑) is said to be complete if every Cauchy sequence
converges.
Example 4.1. (R, | ⋅ |) is complete.

Example 4.2. (Q, | ⋅ |) is not complete.
Example 4.3. If (𝑋1 , 𝑑1 ), … , (𝑋𝑛 , 𝑑𝑛 ) are complete metric spaces, then the product space
𝑛
𝑋 = ∏𝑖=1 𝑋𝑖 is complete for the metrics
𝑛 1/𝑝
𝑑𝑝 ((𝑥1 , … , 𝑥𝑛 ), (𝑦1 , … , 𝑦𝑛 )) = {∑ 𝑑𝑖 (𝑥𝑖 , 𝑦𝑖 )𝑝 } , 1 ⩽ 𝑝 < +∞,
𝑖=1
and also for the metric
𝑑∞ ((𝑥1 , … , 𝑥𝑛 ), (𝑦1 , … , 𝑦𝑛 )) = max 𝑑𝑖 (𝑥𝑖 , 𝑦𝑖 ).

1⩽𝑖⩽𝑛
Proposition 4.5. A closed subset of a complete metric space is complete.
Proof. Let (𝑋 , 𝑑) be a complete metric space and 𝑌 ⊆ 𝑋 be closed. Let (𝑥𝑛 )𝑛⩾0 be a
Cauchy sequence in 𝑌 . Then (𝑥𝑛 )𝑛⩾0 is a Cauchy sequence in 𝑋 and hence there is
𝑦 ∈ 𝑋 such that (𝑥𝑛 )𝑛⩾0 converges to 𝑦 in 𝑋 . Since 𝑌 is closed, one concludes that 𝑦 ∈ 𝑌
by Proposition 3.18.
Proposition 4.6. A complete subspace of a metric space is closed.
Proof. Let (𝑋 , 𝑑) be a metric space and 𝑌 ⊆ 𝑋 be a complete subspace. Let 𝑦 ∈ 𝑌 , then

there is a sequence (𝑥𝑛 )𝑛⩾0 in 𝑌 that converges to 𝑦 in 𝑋 by Proposition 3.18. Therefore
(𝑥𝑛 )𝑛⩾0 is a Cauchy sequence in 𝑌 that is complete, hence it converges to some 𝑦 ̄ ∈ 𝑌 .
By uniqueness of the limit, one obtains 𝑦 = 𝑦 ̄ ∈ 𝑌 .
Definition 4.7. A topological space (𝑋 , 𝒪) is completely metrizable if there is a metric

𝑑 on 𝑋 that induces the topology 𝒪 and for which (𝑋 , 𝑑) is complete.
4.3 Uniformly continuous and Lipschitz maps 23
4.3 Uniformly continuous and Lipschitz maps

Let (𝑋 , 𝑑𝑋 ) and (𝑌 , 𝑑𝑌 ) be metric spaces.
Definition 4.8. A map 𝑓 ∶ 𝑋 → 𝑌 is uniformly continuous if
∀𝜀 > 0, ∃𝛿 > 0, ∀𝑥, 𝑥 ′ ∈ 𝑋 , 𝑑𝑋 (𝑥, 𝑥 ′ ) < 𝛿 ⇒ 𝑑𝑌 (𝑓 (𝑥), 𝑓 (𝑥 ′ )) < 𝜀.
Definition 4.9. A map 𝑓 ∶ 𝑋 → 𝑌 is called 𝑘-Lipschitz, with 𝑘 ⩾ 0, if
∀𝑥, 𝑥 ′ ∈ 𝑋 , 𝑑𝑌 (𝑓 (𝑥), 𝑓 (𝑥 ′ )) ⩽ 𝑘𝑑𝑋 (𝑥, 𝑥 ′ ).
A map 𝑓 ∶ 𝑋 → 𝑌 is Lipschitz if it is 𝑘-Lipschitz for some 𝑘 ⩾ 0, and it is a contraction if

it is 𝑘-Lipschitz with 0 ⩽ 𝑘 < 1.
It is clear that “Lipschitz” implies “uniformly continuous”, which implies “continu-

ous”.
4.4 Fixed-point theorem

Theorem 4.10 (Banach-Picard fixed-point theorem). Let (𝑋 , 𝑑) be a complete metric
space and consider a contraction map 𝑓 ∶ 𝑋 → 𝑋 . Then 𝑓 has a unique fixed point,
that is, there is a unique 𝑥 ̄ ∈ 𝑋 such that 𝑓 (𝑥)̄ = 𝑥.̄
Proof. Since 𝑓 is a contraction map, there is 0 ⩽ 𝑘 < 1 such that 𝑑(𝑓 (𝑥), 𝑓 (𝑦)) ⩽ 𝑘𝑑(𝑥, 𝑦)
for any 𝑥, 𝑦 ∈ 𝑋 .
We first prove uniqueness. Assume 𝑥,̄ 𝑦 ̄ ∈ 𝑋 are fixed points of 𝑓 , then 𝑑(𝑥,̄ 𝑦)̄ =
𝑑(𝑓 (𝑥),
̄ 𝑓 (𝑦))
̄ ⩽ 𝑘𝑑(𝑥,̄ 𝑦),̄ which implies 𝑑(𝑥,̄ 𝑦)̄ = 0 and hence 𝑥 ̄ = 𝑦.̄
We shall now construct a convergent sequence in 𝑋 whose limit is a fixed point. Let
𝑥0 ∈ 𝑋 and define 𝑥𝑛 = 𝑓 (𝑥𝑛−1 ) for any 𝑛 ∈ N∗ . Let us verify that (𝑥𝑛 )𝑛⩾0 is a Cauchy
sequence. For any 𝑛 ∈ N∗ one has 𝑑(𝑥𝑛+1 , 𝑥𝑛 ) = 𝑑(𝑓 (𝑥𝑛 ), 𝑓 (𝑥𝑛−1 )) ⩽ 𝑘𝑑(𝑥𝑛 , 𝑥𝑛−1 ), hence
one gets by induction 𝑑(𝑥𝑛+1 , 𝑥𝑛 ) ⩽ 𝑘 𝑛 𝑑(𝑥1 , 𝑥0 ), therefore for any integers 𝑛, 𝑝 ∈ N one
has
𝑑(𝑥𝑛+𝑝 , 𝑥𝑛 ) ⩽ 𝑑(𝑥𝑛+𝑝 , 𝑥𝑛+𝑝−1 ) + 𝑑(𝑥𝑛+𝑝−1 , 𝑥𝑛+𝑝−2 ) + ⋯ + 𝑑(𝑥𝑛+1 , 𝑥𝑛 )
⩽ (𝑘 𝑛+𝑝−1 + 𝑘 𝑛+𝑝−2 + ⋯ + 𝑘 𝑛 ) 𝑑(𝑥1 , 𝑥0 )
𝑘𝑛
⩽ 𝑑(𝑥1 , 𝑥0 ).
1−𝑘
Fron this inequality we deduce that (𝑥𝑛 )𝑛⩾0 is a Cauchy sequence in 𝑋 . Indeed, let 𝜀 > 0,
𝑘 𝑁0
then we choose 𝑁0 ∈ N large enough such that 1−𝑘
𝑑(𝑥1 , 𝑥0 ) < 𝜀, then for all integers
𝑛, 𝑚 ⩾ 𝑁0 one has
𝑘 min(𝑛,𝑚) 𝑘 𝑁0
𝑑(𝑥𝑚 , 𝑥𝑛 ) ⩽ 𝑑(𝑥1 , 𝑥0 ) ⩽ 𝑑(𝑥1 , 𝑥0 ) < 𝜀.
1−𝑘 1−𝑘
Since 𝑋 is complete one deduces that there is 𝑥 ̄ ∈ 𝑋 such that (𝑥𝑛 )𝑛⩾0 converges to 𝑥.̄
Since 𝑓 is continuous at 𝑥 ̄ and 𝑋 is Hausdorff (so that limits are unique), one can pass
to the limit 𝑛 → +∞ in the relation 𝑥𝑛+1 = 𝑓 (𝑥𝑛 ), which yields 𝑥 ̄ = 𝑓 (𝑥).
̄
24 Chapter 4. Completeness
4.5 Baire’s category theorem

Definition 4.11. A topological space is said to be a Baire space if the intersection of any
countable family of dense open sets is dense.
Equivalently, by passing to complement sets, a topological space is a Baire space if

the union of any countable family of closed sets with empty interior has empty interior.
In other words, in a Baire space, if the union of a countable collection of closed sets has
non-empty interior, then at least one of these closed sets has non-empty interior.
Theorem 4.12 (Baire’s category theorem, first version). Every complete metric space is
a Baire space.
Proof. Let 𝑋 be a complete metric space and {𝑈𝑛 }𝑛∈N be a countable family of dense open
sets in 𝑋 . Denoting 𝑈 ∶= ⋂𝑛∈N 𝑈𝑛 the intersection, we want to show that 𝑈 is dense in
𝑋 , that is, every non-empty open set 𝑊 in 𝑋 intersects 𝑈 .
Let 𝑊 be a non-empty open set in 𝑋 . Since 𝑈0 is dense, 𝑊 intersects 𝑈0 , hence there
is 𝑥0 ∈ 𝑈0 ∩ 𝑊 . Therefore there is 0 < 𝜀0 < 1 such that 𝐵(𝑥0 , 𝜀0 ) ⊆ 𝑈0 ∩ 𝑊 . By induction,
1
for any 𝑛 ∈ N there is 𝑥𝑛 and 0 < 𝜀𝑛 < 𝑛+1 such that 𝐵(𝑥𝑛 , 𝜀𝑛 ) ⊆ 𝐵(𝑥𝑛−1 , 𝜀𝑛−1 ) ∩ 𝑈𝑛 . The
sequence (𝑥𝑛 )𝑛⩾0 is a Cauchy sequence in 𝑋 which is complete, hence it converges to
some 𝑦 ∈ 𝑋 . For any 𝑛 ∈ N, (𝑥𝑘 )𝑘⩾𝑛 is a sequence in 𝐵(𝑥𝑛 , 𝜀𝑛 ) which is closed, and it also
is a subsequence of (𝑥𝑛 )𝑛⩾0 , therefore 𝑦 ∈ 𝐵(𝑥𝑛 , 𝜀𝑛 ) ⊆ 𝑊 ∩ 𝑈𝑛 .
Chapter 5
Compactness
5.1 Open covers

Let 𝑋 be a topological space and 𝐴 a subset of 𝑋 .
Definition 5.1. A family {𝑈𝑖 }𝑖∈𝐼 of subsets of 𝑋 is said to be a cover of 𝑋 if its union
is the whole space 𝑋 , that is 𝑋 = ⋃𝑖∈𝐼 𝑈𝑖 . If moreover every 𝑈𝑖 is open in 𝑋 , then we
say that {𝑈𝑖 }𝑖∈𝐼 is an open cover of 𝑋 . Finally, a sub-cover of {𝑈𝑖 }𝑖∈𝐼 is any sub-collection
{𝑈𝑗 }𝑗∈𝐽 , with 𝐽 ⊆ 𝐼 , that still is a cover of 𝑋 .
Similarly, if 𝐴 is a subset of 𝑋 and {𝑈𝑖 }𝑖∈𝐼 a family of subsets of 𝑋 , we say that {𝑈𝑖 }𝑖∈𝐼
is a cover of 𝐴 if its union contains 𝐴, that is 𝐴 ⊆ ⋃𝑖∈𝐼 𝑈𝑖 .
Definition 5.2. A collection {𝐹𝑖 }𝑖∈𝐼 of subsets of 𝑋 is said to have the finite intersection
property if the intersection over any finite sub-collection of {𝐹𝑖 }𝑖∈𝐼 is non-empty.
More precisely, {𝐹𝑖 }𝑖∈𝐼 has the finite intersection property if for every finite subset
𝐽 ⊆ 𝐼 the intersection ⋂𝑗∈𝐽 𝐹𝑗 is non-empty.
5.2 Compact spaces

Definition 5.3. A topological space 𝑋 is quasi-compact if every open cover of 𝑋 has a
finite sub-cover. A topological space 𝑋 is compact if it is Hausdorff and quasi-compact.
By passing to complement sets, a topological space 𝑋 is quasi-compact if and only
if every family of closed sets in 𝑋 with the finite intersection property has non-empty
intersection.
A subset 𝐴 ⊆ 𝑋 is said to be compact if the topological subspace 𝐴 is compact (en-
dowed with the subspace topology). This is equivalent to the fact that the topological
subspace 𝐴 is Hausdorff and every cover of 𝐴 by open sets in 𝑋 has a finite sub-cover.
By passing to the complement sets, it is still equivalent to the fact that the topologi-
cal subspace 𝐴 is Hausdorff and that for every family of closed sets in 𝑋 such that the
intersection of any finite sub-family intersects 𝐴, the intersection of the whole family
intersects 𝐴.
Example 5.1. Any bounded closed interval [𝑎, 𝑏] ⊆ R is compact.
Example 5.2. R𝑁 is not compact.
Let us now prove some basic properties of compact spaces.
25
26 Chapter 5. Compactness
Proposition 5.4. Let 𝑋 be a Hausdorff topological space.

(i) If 𝐴 is a compact subset of 𝑋 , then 𝐴 is closed in 𝑋 .
(ii) A finite union of compact subsets of 𝑋 is compact.
Proof. (i) We shall prove that 𝑋 ⧵ 𝐴 is open, so let 𝑥 ∈ 𝑋 ⧵ 𝐴. Since 𝑋 is Hausdorff, for
any 𝑦 ∈ 𝐴 there exist an open neighborhood 𝑈𝑦 of 𝑥 in 𝑋 and an open neighborhood 𝑉𝑦
of 𝑦 in 𝑋 such that 𝑈𝑦 ∩ 𝑉𝑦 = ∅. Hence 𝐴 ⊆ ⋃𝑦∈𝐴 𝑉𝑦 and there exist {𝑦1 , … , 𝑦𝑁 } such
that 𝐴 ⊆ ⋃1⩽𝑖⩽𝑁 𝑉𝑦𝑖 . Then the set 𝑈 = ⋂1⩽𝑖⩽𝑁 𝑈𝑦𝑖 is open in 𝑋 and 𝑈 ∩ 𝐴 = ∅, hence
𝑥 ∈ 𝑈 ⊆ 𝑋 ⧵ 𝐴.
(ii) Let 𝐴1 , … , 𝐴𝑛 be compact in 𝑋 . The union 𝐴 ∶= ⋃1⩽𝑘⩽𝑛 𝐴𝑘 is Hausdorff, since it
is a subset of a Hausdorff space. Let {𝑈𝑖 }𝑖∈𝐼 be a cover of 𝐴 by open sets in 𝑋 . For
any 𝑘 ∈ {1, … , 𝑛} we have 𝐴𝑘 ⊆ ⋃𝑖∈𝐼 𝑈𝑖 , hence there exists a finite 𝐽𝑘 ⊆ 𝐼 such that
𝐴𝑘 ⊆ ⋃𝑖∈𝐽𝑘 𝑈𝑖 . Then 𝐽 = ⋃1⩽𝑘⩽𝑛 𝐽𝑘 is a finite set contained in 𝐼 and 𝐴 ⊆ ⋃𝑖∈𝐽 𝑈𝑖 , hence
𝐴 is compact.
Proposition 5.5. A closed subset of a compact topological space is compact.
Proof. Let 𝐴 be a closed set of a compact topological space 𝑋 . Recall that we already
know that 𝐴 is Hausdorff. Let {𝑈𝑖 }𝑖∈𝐼 be a cover of 𝐴 by open sets in 𝑋 . Let 𝑘 ∉ 𝐼 and
define 𝑈𝑘 = 𝑋 ⧵ 𝐴 which is open since 𝐴 is closed. Therefore {𝑈𝑖 }𝑖∈𝐼 ∪{𝑘} is a open cover
of 𝑋 , and since 𝑋 is compact, there exists a finite set 𝐽 ⊆ 𝐼 ∪ {𝑘} such that ⋃𝑖∈𝐽 𝑈𝑖 is a
finite cover of 𝑋 . This implies that ⋃𝑖∈𝐽 ⧵{𝑘} 𝑈𝑖 is a finite cover of 𝐴.
Proposition 5.6 (Cantor’s intersection theorem). Let 𝑋 be a Hausdorff topological space.

The intersection of a decreasing sequence of non-empty compact sets in 𝑋 is non-empty.
Proof. Let {𝐾𝑛 }𝑛∈N be a decreasing sequence of non-empty compact sets in 𝑋 . By way
of contradiction, suppose that ⋂𝑛∈N 𝐾𝑛 = ∅. Consider, for any 𝑛 ∈ N, the open set 𝑈𝑛 =
𝑋 ⧵ 𝐾𝑛 in 𝑋 . The collection {𝑈𝑛 }𝑛∈N is a open cover of 𝐾0 , which is compact, therefore
there is a finite set 𝐽 ⊆ N such that {𝑈𝑛 }𝑛∈𝐽 is a open cover of 𝐾0 . Denoting 𝑝 = max 𝐽 ,
it follows that 𝐾0 ⊆ 𝑈𝑝 which in turn implies 𝐾𝑝 ⊆ 𝑈𝑝 , hence a contradiction.
Theorem 5.7. Let 𝑓 ∶ 𝑋 → 𝑌 be a continuous maps from a compact topological space 𝑋

into a Hausdorff topological space 𝑌 . Then 𝑓 (𝑋 ) is compact in 𝑌 .
Proof. One has that 𝑓 (𝑋 ) is Hausdorff since it is a subspace of 𝑌 which is Hausdorff. Let
{𝑈𝑖 }𝑖∈𝐼 be an open cover of 𝑓 (𝑋 ) by open sets in 𝑌 . Therefore, since 𝑋 = 𝑓 −1 (𝑓 (𝑋 )) and
𝑓 is continuous, one obtains that {𝑓 −1 (𝑈𝑖 )}𝑖∈𝐼 is an open cover of 𝑋 . Since 𝑋 is compact,
there is a finite 𝐽 ⊆ 𝐼 such that {𝑓 −1 (𝑈𝑖 )}𝑖∈𝐽 is a finite cover of 𝑋 , which implies that
{𝑈𝑖 }𝑖∈𝐽 is a finite sub-cover of 𝑓 (𝑋 ).
Proposition 5.8. In a compact topological space, every sequence has at least one cluster
point.
Proof. Let (𝑥𝑛 )𝑛⩾0 be a sequence in a compact topological space 𝑋 . Recall that the set of
cluster points of (𝑥𝑛 )𝑛⩾0 is given by 𝐹 = ⋂𝑛∈N {𝑥𝑝 ∶ 𝑝 ⩾ 𝑛}. Denote 𝐹𝑛 ∶= {𝑥𝑝 ∶ 𝑝 ⩾ 𝑛}
and consider the family of closed sets {𝐹𝑛 }𝑛∈N . The intersection of any finite family of
{𝐹𝑛 }𝑛∈N is non-empty, indeed if 𝐽 ⊆ N is finite then ⋂𝑛∈𝐽 𝐹𝑛 contains 𝑥max 𝐽 . Therefore,
since 𝑋 is compact, we deduce that 𝐹 = ⋂𝑛∈N 𝐹𝑛 is non-empty.
5.3 Compact metric spaces 27
5.3 Compact metric spaces

Theorem 5.9. In a metric space (𝑋 , 𝑑) the following are equivalent:
(i) 𝑋 is compact ;
(ii) 𝑋 sequentially compact, i.e. every sequence has a convergent subsequence ;
(iii) 𝑋 is complete and totally bounded, i.e. for any 𝜀 > 0 there exists a finite family of
open balls of radius 𝜀 that covers 𝑋 .
Proof. (i) ⇒ (ii) It follows from Proposition 5.8 and Proposition 3.19.
(ii) ⇒ (iii) Assume that 𝑋 is sequentially compact. We want to show that 𝑋 is com-
plete and totally bounded. Let (𝑥𝑛 )𝑛⩾0 be a Cauchy sequence in 𝑋 , then it admits a
subsequence that converges to some 𝑥 ∈ 𝑋 , and one deduces that 𝑥 is a cluster point of
(𝑥𝑛 )𝑛⩾0 from Proposition 3.19. Therefore, by Proposition 4.3, one concludes that (𝑥𝑛 )𝑛⩾0
converges to 𝑥.
Let 𝜀 > 0. We argue by contradiction and assume that 𝑋 does not admit any finite
cover by a finite family of open balls of radius 𝜀. We fix 𝑥0 ∈ 𝑋 , then by hypothesis
we can chose some 𝑥1 ∈ 𝑋 ⧵ 𝐵(𝑥0 , 𝜀). By induction, we can then construct a sequence
𝑛−1
(𝑥𝑛 )𝑛⩾0 of elements of 𝑋 such that 𝑥𝑛 ∉ ⋃𝑖=1 𝐵(𝑥𝑖 , 𝜀) for all 𝑛 ⩾ 1. This sequence satisfies
𝑑(𝑥𝑖 , 𝑥𝑗 ) ⩾ 𝜀 for any integers 𝑖 ≠ 𝑗 and hence does not admit any convergent subsequence,
which is a contradiction.
(ii) ⇒ (i) Assume that 𝑋 is sequentially compact. We want to show that 𝑋 is compact.
We already know that 𝑋 is Hausdorff, so we need to prove that 𝑋 is quasi-compact.
Let {𝑈𝑖 }𝑖∈𝐼 be an open cover of 𝑋 . We now show that there is 𝛿 > 0 such that, for any
𝑥 ∈ 𝑋 , there is 𝑖 ∈ 𝐼 such that 𝐵(𝑥, 𝛿) ⊆ 𝑈𝑖 . We argue by contradiction and assume that it
is not the case, hence for any 𝑛 ∈ N∗ there is 𝑥𝑛 ∈ 𝑋 such that 𝐵(𝑥𝑛 , 𝑛1 ) ⊈ 𝑈𝑖 for any 𝑖 ∈ 𝐼 .
The sequence (𝑥𝑛 )𝑛⩾1 has a subsequence that converges to some 𝑥 ∈ 𝑋 , which implies
that 𝑥 is a cluster point of (𝑥𝑛 )𝑛⩾1 . Let 𝑖0 ∈ 𝐼 such that 𝑥 ∈ 𝑈𝑖0 , thus for 𝑛 large enough
one obtains 𝐵(𝑥𝑛 , 𝑛1 ) ⊆ 𝑈𝑖0 , which is a contradiction.
By repeating the above argument used to show that (ii) implies (iii), we show that 𝑋
can be recovered by a finite number of ball of radius 𝛿.
(iii) ⇒ (ii) Assume now that 𝑋 is complete and totally bounded. We wan to prove that 𝑋
is sequentially compact. Let x = (𝑥𝑛 )𝑛⩾0 be a sequence in 𝑋 . For 𝑗 = 1 there exists a finite
family of open ball of radius 1 covering 𝑋 , hence there is a subsequence x1 = (𝑥𝜑1 (𝑛) )𝑛⩾0
of x that is included in some open ball of radius 1. By induction, for any integer 𝑗 ⩾ 2,
one can then find a subsequence x𝑗 = (𝑥𝜑1 ∘⋯∘𝜑𝑗 (𝑛) )𝑛⩾0 of x𝑗−1 (hence also a subsequence
of x) that is included in some open ball of radius 1𝑗 . By a diagonal extraction argument,
one defines the sequence (𝑦𝑛 )𝑛⩾1 with 𝑦𝑛 = 𝑥𝜑1 ∘⋯∘𝜑𝑛 (𝑛) , which is a subsequence of (𝑥𝑛 )𝑛⩾0
and satisfies 𝑑(𝑦𝑛 , 𝑦𝑚 ) ⩽ 𝑛1 for any positive integers 𝑛 ⩽ 𝑚. Therefore (𝑦𝑛 )𝑛⩾1 is a Cauchy
sequence, thus it converges since 𝑋 is complete.
5.4 Product of compact spaces

Lemma 5.10 (Alexander’s sub-basis theorem). Let 𝑋 be a topological space and 𝒮 be
a sub-basis for the topology of 𝑋 . If every open cover of 𝑋 by elements of 𝒮 has a finite
sub-cover, then 𝑋 is quasi-compact.
28 Chapter 5. Compactness
Proof. Assume that every open cover of 𝑋 by elements of 𝒮 has a finite sub-cover. By
way of contradiction, suppose that there is a open cover of 𝑋 with no finite sub-cover.
Consider the collection 𝔓 of open covers of 𝑋 with no finite sub-cover, which is
non-empty, partially ordered by inclusion Let 𝔘 be a chain (totally ordered subset) in
𝔓. We now define 𝒰 = ⋃ 𝔘 which is an upper bound of 𝔘, and we shall verify that
𝒰 has no finite sub-cover. Consider any finite sub-collection 𝑈1 , … , 𝑈𝑛 of 𝒰, then for
any 1 ⩽ 𝑖 ⩽ 𝑛 there is 𝒰𝑖 ∈ 𝔘 such that 𝑈𝑖 ∈ 𝒰𝑖 . Since 𝔘 is totally ordered, there is
some 𝑘 ∈ {1, … , 𝑛} such that 𝒰𝑘 contains all the sets of { 𝑈1 , … , 𝑈𝑛 }, therefore this finite
sub-collection cannot cover 𝑋 . By Zorn’s Lemma (Lemma A.4), one deduces that 𝔓 has
a maximal element ℳ. Since ℳ is a maximal element, for any open 𝑈 ∉ ℳ the cover
ℳ ∪ {𝑈 } has a finite sub-cover, necessarily of the form ℳ ′ ∪ {𝑈 } for some finite subset
ℳ ′ ⊆ ℳ.
We claim that if 𝑈 , 𝑉 are open sets in 𝑋 such that 𝑈 , 𝑉 ∉ ℳ, then 𝑈 ∩ 𝑉 ∉ ℳ. Indeed,
there is finite subsets ℳ ′ ⊆ ℳ and ℳ ″ ⊆ ℳ such that ℳ ′ ∪ {𝑈 } and ℳ ″ ∪ {𝑉 } both
cover 𝑋 . Therefore {𝑈 ∩ 𝑉 } ∪ ℳ ′ ∪ ℳ ″ is a finite cover of 𝑋 , hence 𝑈 ∩ 𝑉 ∉ ℳ since ℳ
has no finite sub-cover.
We also claim that if 𝑈 , 𝑉 are open sets in 𝑋 such that 𝑈 ∉ ℳ and 𝑈 ⊆ 𝑉 , then
𝑉 ∉ ℳ. Indeed, if ℳ ′ ∪ {𝑈 } is a finite cover, with ℳ ′ ⊆ ℳ finite, then also is ℳ ′ ∪ {𝑉 },
thus 𝑉 ∉ ℳ.
We now show that ℳ ∩ 𝒮 covers 𝑋 . Let 𝑥 ∈ 𝑋 , since ℳ is a cover, there is some
open set 𝑈 ∈ ℳ such that 𝑥 ∈ 𝑈 . Since 𝒮 is a sub-basis, there are 𝑆1 , … , 𝑆𝑛 ∈ 𝒮 such that
𝑥 ∈ 𝑆1 ∩ ⋯ ∩ 𝑆𝑛 ⊆ 𝑈 . Thanks to the two claims above, there is some 𝑖 such that 𝑆𝑖 ∈ ℳ,
whence 𝑥 ∈ 𝑆𝑖 ⊆ ℳ ∩ 𝒮 .
Since ℳ ∩ 𝒮 is a cover of 𝑋 , it follows that ℳ ∩ 𝒮 has a finite sub-cover, which
contradicts the fact that ℳ has no finite sub-cover. Therefore the original collection 𝔓
is empty and thus every open cover of 𝑋 has a finite sub-cover.
Theorem 5.11 (Tychonoff’s theorem). Any product of compact topological spaces is com-
pact.
Proof. Let {𝑋𝑖 }𝑖∈𝐼 be a family of compact topological spaces and consider the product
space 𝑋 = ∏𝑖∈𝐼 𝑋𝑖 , equipped with the product topology, and denote by pr𝑗 ∶ 𝑋 → 𝑋𝑗 ,
(𝑥𝑖 )𝑖∈𝐼 ↦ 𝑥𝑗 , the canonical projections. Since every 𝑋𝑖 is Hausdorff, we already know
that 𝑋 is also Hausdorff (Proposition 3.12). Recall that the collection 𝒮 of subsets of the
form pr−1𝑖 (𝑈𝑖 ), where 𝑖 ∈ 𝐼 and 𝑈𝑖 is open in 𝑋𝑖 , is a sub-basis for the product topology
on 𝑋 .
We argue by contradiction and assume that 𝑋 is not compact. Thanks to Lemma 5.10,
there is a cover 𝒰 of 𝑋 by elements of 𝒮 with no finite sub-cover. For each 𝑖 ∈ 𝐼 , define
the collection 𝒮𝑖 of open sets 𝑈 ⊆ 𝑋𝑖 such that pr−1 𝑖 (𝑈 ) ∈ 𝒮 .
We now show that 𝒮𝑖 does not cover 𝑋𝑖 . Indeed, since 𝑋𝑖 is compact, if 𝒮𝑖 covers 𝑋𝑖
then there is a finite collection 𝑈1 , … , 𝑈𝑛 ∈ 𝒮𝑖 that covers 𝑋𝑖 . If follows that
pr−1 −1 −1
𝑖 (𝑈1 ) ∪ ⋯ ∪ pr𝑖 (𝑈𝑛 ) = pr𝑖 (𝑋𝑖 ) = 𝑋 ,
which contradicts the fact that 𝒮 has no finite sub-cover.

Then let 𝑥𝑖 ∈ 𝑋𝑖 such that 𝑥𝑖 ∉ ⋃ 𝒮𝑖 . Define 𝑥 = (𝑥𝑖 )𝑖∈𝐼 , then, since 𝒮 is a cover, there
is 𝑖 ∈ 𝐼 and an open set 𝑈 in 𝑋𝑖 such that 𝑥 ∈ pr−1 𝑖 (𝑈 ), but this contradicts the fact that
𝑥𝑖 ∉ ⋃ 𝒮𝑖 .
5.5 Local compactness 29
5.5 Local compactness

Definition 5.12. A topological space 𝑋 is locally compact if it is Hausdorff and if every
point admits a compact neighbourhood.
Proposition 5.13. In a locally compact space 𝑋 , every point has a neighbourhood basis
composed by compact sets.
Proof. Let 𝑥 ∈ 𝑋 and 𝑉 be a neighborhood of 𝑥. We want to find a compact neighborhood

of 𝑥 contained in 𝑉 .
By definition, there exists a compact 𝐾 ⊆ 𝑋 that is a neighborhood of 𝑥 in 𝑋 . Since
𝑉 ∩ 𝐾 is a neighborhood of 𝑥, there is an open set 𝑈 such that 𝑥 ∈ 𝑈 ⊆ 𝑉 ∩ 𝐾 ⊆ 𝐾 .
Remark that the set 𝐾 ⧵ 𝑈 is compact, since it is closed and contained in the compact
𝐾 . Since 𝑋 is Hausdorff, there are disjoint open sets 𝑊1 and 𝑊2 such that 𝑥 ∈ 𝑊1 and
𝐾 ⧵ 𝑈 ⊆ 𝑊2 . The set 𝐹 ∶= 𝑈 ∩ 𝑊1 contains 𝑥, is compact by construction (since it is a
closed set included in the compact 𝐾 ) and is contained in 𝑈 , hence in 𝑉 ∩ 𝐾 ⊆ 𝑉 .
Example 5.3. R is locally compact.
Example 5.4. A discrete space is locally compact.
Example 5.5. A finite product of locally compact spaces is locally compact.
Theorem 5.14 (Baire’s category theorem, second version). Every locally compact topo-
logical space is a Baire space.
Proof. Let 𝑋 be a locally compact topological space and {𝑈𝑛 }𝑛∈N be a countable family
of dense open sets in 𝑋 . Denote 𝑈 ∶= ⋂𝑛∈N 𝑈𝑛 the intersection.
Let 𝑊 be a non-empty open set in 𝑋 . Since 𝑈0 is dense, 𝑊 intersects 𝑈0 , which implies
that there is 𝑥0 ∈ 𝑈0 ∩ 𝑊 . Since 𝑋 is locally compact, there exists a compact set 𝐾0 that
is a neighborhood of 𝑥0 , hence with non-empty interior, such that 𝑥0 ∈ 𝐾0 ⊆ 𝑊 ∩ 𝑈0 .
Since 𝑈1 is dense, one has 𝑈1 ∩ 𝐾̊0 ≠ ∅, which implies that there is 𝑥1 ∈ 𝑈1 ∩ 𝐾̊0 . As
before, there exists a compact set 𝐾1 with non-empty interior such that 𝐾1 ⊆ 𝑈1 ∩ 𝐾0̊ .
By induction, for any 𝑛 ∈ N, there is a compact set 𝐾𝑛 with non-empty interior such
that 𝐾𝑛 ⊆ 𝑈𝑛 ∩ 𝐾̊𝑛−1 . Therefore ⋂𝑛∈N 𝐾𝑛 ⊆ 𝑊 ∩ ⋂𝑛∈N 𝑈𝑛 , and the intersection ⋂𝑛∈N 𝐾𝑛 is
non-empty by Proposition 5.6.
Chapter 6
Connectedness
6.1 Connected spaces

Definition 6.1. A topological space 𝑋 is said to be connected if there is no subset of 𝑋 ,
other than th empty set ∅ and 𝑋 itself, that is both open and closed.
We say that a subset 𝐴 of a topological space 𝑋 is connected if the topological sub-

space 𝐴, equipped with the subspace topology, is connected.
Definition 6.2. A topological space 𝑋 is locally connected if every point has a neighbor-
hood basis formed by connected sets.
Theorem 6.3. Let 𝑋 be a topological space. The following are equivalent:

(i) 𝑋 is connected ;
(ii) 𝑋 cannot be written as the union of two non-empty disjoint open sets ;
(iii) 𝑋 cannot be written as the union of two non-empty disjoint closed sets ;
(iv) Any continuous map from 𝑋 into {0, 1} equipped with the discrete topology is con-
stant ;
(v) Any continuous map from 𝑋 into a discrete topological space is constant.
Proof. (i) ⇔ (ii) ⇔ (iii) It follows easily by taking complement sets.

(v) ⇒ (iv) Trivial.
(i) ⇒ (v) Let 𝑌 be a discrete topological space and assume that there is a continuous
function 𝑓 ∶ 𝑋 → 𝑌 that is not constant. Since 𝑓 is not constant, there is 𝑦 ∈ 𝑌 such
that 𝑓 −1 ({𝑦}) is a non-empty proper subset of 𝑋 . But 𝑓 −1 ({𝑦}) is both open and closed
in 𝑋 .
(iv) ⇒ (ii) Now assume that there are two non-empty disjoint open sets 𝐴, 𝐵 such that
𝑋 = 𝐴 ∪ 𝐵. Define the map 𝑓 ∶ 𝑋 → {0, 1} by 𝑓 (𝑥) = 0 if 𝑥 ∈ 𝐴, and 𝑓 (𝑥) = 1 if 𝑥 ∈ 𝐵.
Then 𝑓 is continuous and non-constant.
Example 6.1. Any interval of R is connected.

Example 6.2. R ⧵ {0} is not connected.
Example 6.3. Q is not connected.
Example 6.4. If 𝑓 ∶ 𝑋 → 𝑌 is continuous and 𝑋 is connected, then 𝑓 (𝑋 ) is connected.
31
32 Chapter 6. Connectedness
Example 6.5. If 𝑋 and 𝑌 are connected, then 𝑋 × 𝑌 is connected.

Let us now prove some basic properties of connected spaces.
Proposition 6.4. Consider subsets 𝐴 ⊆ 𝐵 ⊆ 𝐴 of a topological space 𝑋 and assume that

𝐴 is connected. Then 𝐵 is also connected.
Proof. Consider a continuous map 𝑓 from 𝐵 to the discrete space {0, 1}. The restriction
𝑓|𝐴 of 𝑓 to 𝐴 is continuous, hence constant since 𝐴 is connected. Let 𝑎 ∈ {0, 1} be
the image of 𝐴, then 𝑓 −1 ({𝑎}) is a closed set in 𝐵 that contains 𝐴. This implies that
𝐵 = 𝐴 ∩ 𝐵 ⊆ 𝑓 −1 ({𝑎}) ⊆ 𝐵, therefore 𝐵 = 𝑓 −1 ({𝑎}) and 𝑓 is constant.
Proposition 6.5. Let {𝐴𝑖 }𝑖∈𝐼 be a collection of connected subsets of a topological space 𝑋
such that 𝐴𝑖 ∩ 𝐴𝑗 ≠ ∅ for any 𝑖, 𝑗 ∈ 𝐼 . Then 𝐴 = ⋃𝑖∈𝐼 𝐴𝑖 is connected.
Proof. Consider a continuous map 𝑓 from 𝐴 to the discrete space {0, 1}. For any 𝑖 ∈ 𝐼 , the
restriction 𝑓|𝐴𝑖 of 𝑓 to 𝐴𝑖 is continuous hence constant. For any 𝑖, 𝑗 ∈ 𝐼 , since 𝐴𝑖 ∩ 𝐴𝑗 ≠ ∅,
it follows that 𝑓 is constant on 𝐴𝑖 ∪ 𝐴𝑗 . Therefore 𝑓 is constant on 𝐴 and hence 𝐴 is
connected.
From the above result, given a topological space 𝑋 and a point 𝑥 ∈ 𝑋 , we define
the set C𝑥 , called the connected component of 𝑥, as the union of all connected subsets
containing 𝑥, which is also the largest connected subset containing 𝑥. Remark that,
given 𝑦 ∈ 𝑋 , 𝑦 ∈ C𝑥 if and only if C𝑦 = C𝑥 , hence the connected components form a
partition of 𝑋 .
A topological space is said to be totally disconnected if all connected components are
singletons.
6.2 Path-connected spaces

Definition 6.6. A topological space 𝑋 is path-connected if for any points 𝑥, 𝑦 ∈ 𝑋 there
is a path joining 𝑥 to 𝑦, that is, a continuous map 𝛾 ∶ [0, 1] → 𝑋 such that 𝛾 (0) = 𝑥 and
𝛾 (1) = 𝑦.
A subset 𝐴 ⊆ 𝑋 is said to be path-connected if every points 𝑥, 𝑦 ∈ 𝐴 can de joined
by a path taking values in 𝐴, that is, a continuous map 𝛾 ∶ [0, 1] → 𝐴 with 𝛾 (0) = 𝑥 and
𝛾 (1) = 𝑦.
Definition 6.7. A topological space 𝑋 is locally path-connected if every point has a
neighborhood basis formed by path-connected sets.
As above, given a topological space 𝑋 and a point 𝑥 ∈ 𝑋 , we define the path-connected
component of 𝑥 as the largest path-connected subset containing 𝑥.
Theorem 6.8. Every path-connected space is connected.
Proof. Let 𝑋 be a path-connected topological space. By way of contradiction, assume
that 𝑋 is not connected. Then there are non-empty open disjoint sets 𝐴, 𝐵 ⊆ 𝑋 such that
𝑋 = 𝐴 ∪ 𝐵. Let 𝑎 ∈ 𝐴 and 𝑏 ∈ 𝐵. There exists a path 𝛾 joining 𝑎 and 𝑏. Therefore [0, 1] =
𝛾 −1 (𝐴) ∪ 𝛾 −1 (𝐵) is the union of non-empty disjoint open sets, which is a contradiction
since [0, 1] is connected (see Example 6.1).
Proposition 6.9. If 𝑋 is connected and locally path-connected, then 𝑋 is path-connected.
6.2 Path-connected spaces 33
Proof. Let 𝑎 ∈ 𝑋 and denote by 𝐴 the set of all points that can be joined to 𝑎 by a path.
We will show that 𝐴 is both open and closed, which concludes the proof since 𝑋 is
connected.
For any 𝑥 ∈ 𝑋 there is a open neighborhood 𝐵𝑥 of 𝑥 such that every point 𝑦 ∈ 𝐵𝑥
can be joined to 𝑥 by a path. Therefore, for any 𝑥 ∈ 𝐴 and 𝑦 ∈ 𝐵𝑥 there is a path joining
𝑎 to 𝑦, this implies 𝑦 ∈ 𝐴 and thus 𝐴 is open.
Let 𝑧 ∈ 𝐴. Let 𝑦 ∈ 𝐴 ∩ 𝐵𝑧 , then there is a path joining 𝑦 to 𝑧 and a path joining 𝑎 to 𝑦,
hence the existence of a path joining 𝑎 to 𝑧. Therefore 𝑧 ∈ 𝐴 and hence 𝐴 is closed.
Chapter 7
Function spaces
If 𝑋 and 𝑌 are sets, we denote by ℱ (𝑋 , 𝑌 ) = 𝑌 𝑋 the set of maps from 𝑋 to 𝑌 . If 𝑋 and

𝑌 are topological space, we then denote by 𝒞 (𝑋 , 𝑌 ) the set of continuous maps from 𝑋
to 𝑌 .
7.1 Pointwise and uniform convergences

Let 𝑋 be a set and 𝑌 a topological space. Consider a sequence (𝑓𝑛 )𝑛⩾0 of maps from 𝑋
into 𝑌 and let 𝑓 be a map from 𝑋 to 𝑌 .
One says that the the sequence (𝑓𝑛 )𝑛⩾0 converges pointwisely to 𝑓 on 𝑋 if for any 𝑥 ∈
𝑋 the sequence (𝑓𝑛 (𝑥))𝑛⩾0 converges to 𝑓 (𝑥) in 𝑌 . The map 𝑓 is called the pointwise limit
of (𝑓𝑛 )𝑛⩾0 . In particular, if (𝑌 , 𝑑) is a metric space, then (𝑓𝑛 )𝑛⩾0 converges pointwisely to
𝑓 on 𝑋 if
∀𝑥 ∈ 𝑋 , ∀𝜀 > 0, ∃𝑁 ∈ N, ∀𝑛 ⩾ 𝑁 , 𝑑(𝑓𝑛 (𝑥), 𝑓 (𝑥)) < 𝜀.
When (𝑌 , 𝑑) is a metric space, then one says that (𝑓𝑛 )𝑛⩾0 converges uniformly to 𝑓 on
𝑋 if
∀𝜀 > 0, ∃𝑁 ∈ N, ∀𝑛 ⩾ 𝑁 , ∀𝑥 ∈ 𝑋 , 𝑑(𝑓𝑛 (𝑥), 𝑓 (𝑥)) < 𝜀.
The map 𝑓 is called the uniform limit of (𝑓𝑛 )𝑛⩾0 .
It is clear that the uniform convergence implies pointwise convergence.
7.1.1 Topology of pointwise convergence

Consider a set 𝑋 and a topological space 𝑌 . The topology associated to the pointwise
convergence described is called the topology of pointwise convergence and it is nothing
but the product topology on the set 𝑌 𝑋 . Indeed let us verify that (𝑓𝑛 )𝑛⩾0 converges to 𝑓
in 𝑌 𝑋 endowed with the product topology if and only if (𝑓𝑛 )𝑛⩾0 converges pointwisely
to 𝑓 on 𝑋 .
Assume that (𝑓𝑛 )𝑛⩾0 converges to 𝑓 for the product topology. Let 𝑥 ∈ 𝑋 . Let 𝑉 be
a neighbourhood of 𝑓 (𝑥) in 𝑌 , then there is an open set 𝑈 in 𝑌 such that 𝑓 (𝑥) ∈ 𝑈 ⊆ 𝑉 .
Consider the set pr−1 𝑥 (𝑈 ) = {𝑔 ∈ 𝑌
𝑋 ∣ 𝑔(𝑥) ∈ 𝑈 }, which is open in 𝑌 𝑋 and contains
−1
𝑓 , then pr𝑥 (𝑈 ) is a neighbourhood of 𝑓 . Therefore, there is 𝑛0 ∈ N such that, for any
integer 𝑛 ⩾ 𝑛0 one has 𝑓𝑛 ∈ pr−1𝑥 (𝑈 ). Hence for any 𝑛 ⩾ 𝑛0 one has 𝑓𝑛 (𝑥) ∈ 𝑈 ⊆ 𝑉 .
Conversely, now assume that (𝑓𝑛 )𝑛⩾0 converges pointwisely to 𝑓 . Let 𝑉 be a neigh-
bourhood of 𝑓 in 𝑌 𝑋 . Then there is a finite number of points 𝑥1 , … , 𝑥𝑘 ∈ 𝑋 and a finite
number of open sets 𝑈1 , … , 𝑈𝑘 in 𝑌 such that 𝑓 ∈ pr−1 −1
𝑥1 (𝑈1 ) ∩ ⋯ ∩ pr𝑥𝑘 (𝑈𝑘 ) ⊆ 𝑉 . Therefore,
35
36 Chapter 7. Function spaces
for any 𝑖 = 1, … , 𝑘, one has 𝑓 (𝑥𝑖 ) ∈ 𝑈𝑖 and hence there is 𝑛𝑖 ∈ 𝑁 such that 𝑓𝑛 (𝑥𝑖 ) ∈ 𝑈𝑖
for any integer 𝑛 ⩾ 𝑛𝑖 . Hence, taking 𝑁 = max1⩽𝑖⩽𝑘 𝑛𝑖 , for any integer 𝑛 ⩾ 𝑁 one has
𝑓𝑛 ∈ pr−1 −1
𝑥1 (𝑈1 ) ∩ ⋯ ∩ pr𝑥𝑘 (𝑈𝑘 ) ⊆ 𝑉 .
7.1.2 Topology of uniform convergence

Consider a set 𝑋 and a metric space (𝑌 , 𝑑). For any 𝑓 , 𝑔 ∈ ℱ (𝑋 , 𝑌 ) define the map
𝑑∞ (𝑓 , 𝑔) ∶= min (sup 𝑑(𝑓 (𝑥), 𝑔(𝑥)), 1) .

𝑥∈𝑋
One can easily check that 𝑑∞ is a metric on ℱ (𝑋 , 𝑌 ), and that a sequence of maps (𝑓𝑛 )𝑛⩾0
converges uniformly to 𝑓 on 𝑋 if and only if (𝑓𝑛 )𝑛⩾0 converges to 𝑓 with respect the
metric 𝑑∞ . The metric 𝑑∞ is called the metric of uniform convergence or the uniform
metric, and its associated topology is called the topology of uniform convergence.
Proposition 7.1. Let 𝑋 be a set and (𝑌 , 𝑑) a complete metric space. Then (ℱ (𝑋 , 𝑌 ), 𝑑∞ )
is complete.
Proof. Let (𝑓𝑛 )𝑛⩾0 be a Cauchy sequence in (ℱ (𝑋 , 𝑌 ), 𝑑∞ ). Then for any 𝑥 ∈ 𝑋 , (𝑓𝑛 (𝑥))𝑛⩾0
is a Cauchy sequence in (𝑌 , 𝑑), since 𝑑(𝑓𝑛 (𝑥), 𝑓 (𝑥)) ⩽ sup𝑦∈𝑋 (𝑓𝑛 (𝑦), 𝑓 (𝑦)) if 𝑑∞ (𝑓𝑛 , 𝑓 ) < 1.
Therefore, for any 𝑥 ∈ 𝑋 , (𝑓𝑛 (𝑥))𝑛⩾0 converges to some point 𝑓 (𝑥) in 𝑌 . We have hence
constructed a map 𝑓 ∶ 𝑋 → 𝑌 , 𝑥 ↦ 𝑓 (𝑥), and it remains to show that (𝑓𝑛 )𝑛⩾0 converges
to 𝑓 with respect to the metric 𝑑∞ .
Let 𝜀 > 0, then there is 𝑛0 ∈ N such that one has, for any for any integers 𝑛, 𝑚 ⩾ 𝑛0
and any 𝑥 ∈ 𝑋 , 𝑑(𝑓𝑛 (𝑥), 𝑓𝑚 (𝑥)) < 𝜀. Hence, with 𝑥 ∈ 𝑋 fixed and letting 𝑚 → +∞, we
obtain 𝑑(𝑓𝑛 (𝑥), 𝑓 (𝑥)) ⩽ 𝜀 for all 𝑛 ⩾ 𝑛0 . Thus (𝑓𝑛 )𝑛⩾0 converges uniformly to 𝑓 on 𝑋 .
7.2 Space of continuous functions on a compact set

Let 𝑋 be a compact space and (𝑌 , 𝑑) a metric space, and consider the space 𝒞 (𝑋 , 𝑌 ) of
continuous maps from 𝑋 to 𝑌 . We define, for any 𝑓 , 𝑔 ∈ 𝒞 (𝑋 , 𝑌 ), the map
𝑑∞ (𝑓 , 𝑔) = sup 𝑑(𝑓 (𝑥), 𝑔(𝑥)).

𝑥∈𝑋
As before, one can easily check that 𝑑∞ is a metric on 𝒞 (𝑋 , 𝑌 ). Furthermore, a sequence

of maps (𝑓𝑛 )𝑛⩾0 in 𝒞 (𝑋 , 𝑌 ) converges uniformly on 𝑋 to a function 𝑓 ∶ 𝑋 → 𝑌 if and
only if (𝑓𝑛 )𝑛⩾0 converges to 𝑓 with respect the metric 𝑑∞ . The metric 𝑑∞ is called the
metric of uniform convergence or the uniform metric, and its associated topology is called
the topology of uniform convergence.
Theorem 7.2. Let 𝑋 be compact and (𝑌 , 𝑑) a complete metric space. Then the space
(𝒞 (𝑋 , 𝑌 ), 𝑑∞ ) is complete.
Proof. Let (𝑓𝑛 )𝑛⩾0 be a Cauchy sequence in (𝒞 (𝑋 , 𝑌 ), 𝑑∞ ). Then for any 𝑥 ∈ 𝑋 , (𝑓𝑛 (𝑥))𝑛⩾0
is a Cauchy sequence in (𝑌 , 𝑑), since 𝑑(𝑓𝑛 (𝑥), 𝑓𝑚 (𝑥)) ⩽ 𝑑∞ (𝑓𝑛 , 𝑓𝑚 ). Therefore, for any
𝑥 ∈ 𝑋 , (𝑓𝑛 (𝑥))𝑛⩾0 converges to some 𝑓 (𝑥) in 𝑌 . We have hence constructed a map
𝑓 ∶ 𝑋 → 𝑌 , 𝑥 ↦ 𝑓 (𝑥).
It remains to show that (𝑓𝑛 )𝑛⩾0 converges to 𝑓 with respect to the metric 𝑑∞ , and that
𝑓 ∈ 𝒞 (𝑋 , 𝑌 ). Let 𝜀 > 0, then there is 𝑛0 ∈ N such that one has, for any integers 𝑛, 𝑚 ⩾ 𝑛0
7.3 Equicontinuity 37
and any 𝑥 ∈ 𝑋 , 𝑑(𝑓𝑛 (𝑥), 𝑓𝑚 (𝑥)) < 𝜀. Hence, with 𝑥 ∈ 𝑋 fixed and letting 𝑚 → +∞, we
obtain 𝑑(𝑓𝑛 (𝑥), 𝑓 (𝑥)) ⩽ 𝜀 for all 𝑛 ⩾ 𝑛0 . Thus (𝑓𝑛 )𝑛⩾0 converges uniformly to 𝑓 on 𝑋 .
We now prove that 𝑓 is continuous on 𝑋 . Let 𝑎 ∈ 𝑋 and 𝜀 > 0. There is 𝑛0 ∈ N such
that sup𝑥∈𝑋 𝑑(𝑓𝑛0 (𝑥), 𝑓 (𝑥)) < 𝜀/3. Since 𝑓𝑛0 is continuous at 𝑎, there is a neighbourhood
𝑉 of 𝑎 in 𝑋 such that for any 𝑥 ∈ 𝑉 one has 𝑑(𝑓𝑛0 (𝑥), 𝑓𝑛0 (𝑎)) < 𝜀/3. Therefore for any
𝑥 ∈ 𝑉 one has
𝑑(𝑓 (𝑥), 𝑓 (𝑎)) ⩽ 𝑑(𝑓 (𝑥), 𝑓𝑛0 (𝑥)) + 𝑑(𝑓𝑛0 (𝑥), 𝑓𝑛0 (𝑎)) + 𝑑(𝑓𝑛0 (𝑎), 𝑓 (𝑎)) < 𝜀,
thus 𝑓 is continuous at 𝑎. We conclude that 𝑓 is continuous on 𝑋 since 𝑎 is arbitrary.
7.3 Equicontinuity
Let 𝑋 be a topological space and (𝑌 , 𝑑) a metric space.
Definition 7.3. Let ℋ be a subset of ℱ (𝑋 , 𝑌 ) and 𝑥 ∈ 𝑋 a point. One says that ℋ is
equicontinuous at 𝑥 if for any 𝜀 > 0, there is a neighbourhood 𝑈 of 𝑥 in 𝑋 such that,
∀𝑦 ∈ 𝑈 , ∀𝑓 ∈ ℋ , 𝑑(𝑓 (𝑦), 𝑓 (𝑥)) < 𝜀.
One says that ℋ is equicontinuous on 𝑋 if ℋ is equicontinuous at every point of 𝑋 .
Example 7.1. A finite family ℋ = {𝑓1 , … , 𝑓𝑝 } of functions that are continuous at a point
𝑥 ∈ 𝑋 is equicontinuous at 𝑥.
Example 7.2. If 𝑋 is a metric space and 𝑘 > 0, the set of all 𝑘-Lipschitz functions from 𝑋
to 𝑌 is equicontinuous on 𝑋 .
Proposition 7.4. Let (𝑓𝑛 )𝑛⩾0 be a sequence of maps from 𝑋 to 𝑌 . Suppose that {𝑓𝑛 ∶ 𝑛 ∈ N}
is equicontinuous at 𝑎 ∈ 𝑋 and that (𝑓𝑛 )𝑛⩾0 converges pointwisely on 𝑋 to some map
𝑓 ∶ 𝑋 → 𝑌 . Then 𝑓 is continuous at 𝑎.
Proof. Let 𝜀 > 0. Since {𝑓𝑛 ∶ 𝑛 ∈ N} is equicontinuous at 𝑎, there is a neighbourhood 𝑈
of 𝑎 such that for all 𝑥 ∈ 𝑈 and any 𝑛 ∈ N one has 𝑑(𝑓𝑛 (𝑥), 𝑓𝑛 (𝑎)) < 𝜀/3. Since (𝑓𝑛 )𝑛⩾0
converges pointwisely to 𝑓 , for any 𝑥 ∈ 𝑈 there is 𝑁0 ∈ N such that for any 𝑛 ⩾ 𝑁0 one
has 𝑑(𝑓𝑛 (𝑥), 𝑓 (𝑥)) < 𝜀/3. Hence, for any 𝑥 ∈ 𝑈 and any 𝑛 ⩾ 𝑁0 , one obtains
𝑑(𝑓 (𝑥), 𝑓 (𝑎)) ⩽ 𝑑(𝑓 (𝑥), 𝑓𝑛 (𝑥)) + 𝑑(𝑓𝑛 (𝑥), 𝑓𝑛 (𝑎)) + 𝑑(𝑓𝑛 (𝑎), 𝑓 (𝑎)) < 𝜀,
thus 𝑓 is continuous at 𝑎.
Proposition 7.5. Let (𝑓𝑛 )𝑛⩾0 be a sequence of maps from 𝑋 to 𝑌 and suppose that 𝐴 is
dense in 𝑋 and that 𝑌 is complete. Assume that {𝑓𝑛 ∶ 𝑛 ∈ N} is equicontinuous on 𝑋 and
that (𝑓𝑛 )𝑛⩾0 converges pointwisely on 𝐴. Then (𝑓𝑛 )𝑛⩾0 converges pointwisely on 𝑋 .
Proof. We show that, for any 𝑥 ∈ 𝑋 , the sequence (𝑓𝑛 (𝑥))𝑛⩾0 is a Cauchy sequence
in 𝑌 . Let 𝑥 ∈ 𝑋 and 𝜀 > 0. Since {𝑓𝑛 ∶ 𝑛 ∈ N} is equicontinuous at 𝑥, there is a open
neighbourhood 𝑈 of 𝑥 such that for all 𝑦 ∈ 𝑈 and any 𝑛 ∈ N one has 𝑑(𝑓𝑛 (𝑥), 𝑓𝑛 (𝑦)) < 𝜀/3.
By density of 𝐴, there exists a point 𝑧 ∈ 𝐴 ∩ 𝑈 and in particular one has 𝑑(𝑓𝑛 (𝑥), 𝑓𝑛 (𝑧)) <
𝜀/3. Thanks to the pointwise convergence, the sequence (𝑓𝑛 (𝑧))𝑛⩾0 is a Cauchy sequence
in 𝑌 , therefore there is 𝑁0 ∈ N such that for any 𝑛, 𝑚 ⩾ 𝑁0 one has 𝑑(𝑓𝑛 (𝑧), 𝑓𝑚 (𝑧)) < 𝜀/3.
One finally gets, for any 𝑛, 𝑚 ⩾ 𝑁0 , that
𝑑(𝑓𝑛 (𝑥), 𝑓𝑚 (𝑥)) ⩽ 𝑑(𝑓𝑛 (𝑥), 𝑓𝑛 (𝑧)) + 𝑑(𝑓𝑛 (𝑧), 𝑓𝑚 (𝑧)) + 𝑑(𝑓𝑚 (𝑧), 𝑓𝑚 (𝑥)) < 𝜀,
hence (𝑓𝑛 (𝑥))𝑛⩾0 is a Cauchy sequence in 𝑌 .
Proposition 7.6. Let (𝑓𝑛 )𝑛⩾0 be a sequence of maps from 𝑋 to 𝑌 and suppose that 𝑋 is
compact. Assume that {𝑓𝑛 ∶ 𝑛 ∈ N} is equicontinuous on 𝑋 and that (𝑓𝑛 )𝑛⩾0 converges
pointwisely on 𝑋 to some map 𝑓 ∶ 𝑋 → 𝑌 . Then (𝑓𝑛 )𝑛⩾0 converges uniformly on 𝑋 to 𝑓 .
Proof. We already know that 𝑓 is continuous on 𝑋 by Proposition 7.4.

Let 𝜀 > 0. By equicontinuity of {𝑓𝑛 ∶ 𝑛 ∈ N}, for any 𝑥 ∈ 𝑋 there is a open neigh-
bourhood 𝑈𝑥 of 𝑥 such that for all 𝑦 ∈ 𝑈𝑥 and any 𝑛 ∈ N one has 𝑑(𝑓𝑛 (𝑥), 𝑓𝑛 (𝑦)) < 𝜀/3.
Moreover, since 𝑓 is continuous, for any 𝑥 ∈ 𝑋 there is a open neighbourhood 𝑉𝑥 of 𝑥
such that for all 𝑦 ∈ 𝑈𝑥 one has 𝑑(𝑓 (𝑥), 𝑓 (𝑦)) < 𝜀/3.
Considering the open neighbourhood 𝑊𝑥 ∶= 𝑈𝑥 ∩ 𝑉𝑥 of 𝑥, we write 𝑋 = ⋃𝑥∈𝑋 𝑊𝑥 .
Since 𝑋 is compact, there is a finite set {𝑥1 , … , 𝑥𝑝 } ⊆ 𝑋 such that 𝑋 = 𝑊𝑥1 ∪ ⋯ ∪ 𝑊𝑥𝑝 .
Thanks to the pointwise convergence, there is 𝑁0 ∈ N such that for all 𝑖 = 1, … , 𝑝 and
any 𝑛 ⩾ 𝑁0 one has 𝑑(𝑓𝑛 (𝑥𝑖 ), 𝑓 (𝑥𝑖 )) < 𝜀/3.
Now let 𝑥 ∈ 𝑋 . Then there is 𝑖 ∈ {1, … , 𝑝} such that 𝑥 ∈ 𝑊𝑥𝑖 , therefore for any 𝑛 ⩾ 𝑁0
one has
𝑑(𝑓𝑛 (𝑥), 𝑓 (𝑥)) ⩽ 𝑑(𝑓𝑛 (𝑥), 𝑓𝑛 (𝑥𝑖 )) + 𝑑(𝑓𝑛 (𝑥𝑖 ), 𝑓 (𝑥𝑖 )) + 𝑑(𝑓 (𝑥𝑖 ), 𝑓 (𝑥)) < 𝜀,
which concludes the proof.
7.4 Arzelà-Ascoli theorem

Let 𝑋 be a compact space and (𝑌 , 𝑑) a complete metric space. Consider the space 𝒞 (𝑋 , 𝑌 )
of continuous maps from 𝑋 to 𝑌 endowed with the uniform metric
𝑑∞ (𝑓 , 𝑔) = sup 𝑑(𝑓 (𝑥), 𝑔(𝑥)).

𝑥∈𝑋
The following result characterizes the compact subsets of 𝒞 (𝑋 , 𝑌 ).
Theorem 7.7 (Arzelà-Ascoli). Let ℋ be a subset of 𝒞 (𝑋 , 𝑌 ). Then the closure of ℋ is

compact in 𝒞 (𝑋 , 𝑌 ) if and only if ℋ satisfies:
(i) the closure of the set ℋ (𝑥) ∶= {𝑓 (𝑥) ∶ 𝑓 ∈ ℋ } is compact in 𝑌 for any 𝑥 ∈ 𝑋 ;
(ii) ℋ is equicontinuous on 𝑋 .
Remark 7.8. If 𝑌 = R or C then property (i) just means that ℋ (𝑥) is bounded in R or
C for any 𝑥 ∈ 𝑋 .
Proof of Theorem 7.7. We start with the easier part of the result and assume that ℋ has
compact closure in 𝒞 (𝑋 , 𝑌 ). For any 𝑥 ∈ 𝑋 , the evaluation map 𝑓 ↦ 𝑓 (𝑥) is continuous
from 𝒞 (𝑋 , 𝑌 ) into 𝑌 since 𝑑(𝑓 (𝑥), 𝑔(𝑥)) ⩽ 𝑑∞ (𝑓 , 𝑔), thus the set ℋ (𝑥) is compact as the
image of a compact under a continuous function. This gives (i).
Let 𝑥 ∈ 𝑋 and 𝜀 > 0. Since ℋ is compact it is totally bounded, hus there is a finite
collection {𝑓1 , … , 𝑓𝑝 } ⊆ ℋ such that, for any 𝑓 ∈ ℋ there is a label 𝑖 ∈ {1, … , 𝑝} such that
𝑑∞ (𝑓 , 𝑓𝑖 ) < 𝜀/3. Since the finite family {𝑓1 , … , 𝑓𝑝 } is equicontinuous at 𝑥, there is 𝑈 a
neighbourhood of 𝑥 such that for all 𝑦 ∈ 𝑈 and any 𝑖 ∈ {1, … , 𝑝} one has 𝑑(𝑓𝑖 (𝑦), 𝑓𝑖 (𝑥)) <
𝜀/3. Therefore, for any 𝑦 ∈ 𝑈 and any 𝑓 ∈ ℋ one obtains, with the choice of label
𝑖 ∈ {1, … , 𝑝} described above,
𝑑(𝑓 (𝑦), 𝑓 (𝑥)) ⩽ 𝑑(𝑓 (𝑦), 𝑓𝑖 (𝑦)) + 𝑑(𝑓𝑖 (𝑦), 𝑓𝑖 (𝑥)) + 𝑑(𝑓𝑖 (𝑥), 𝑓 (𝑥)) < 𝜀.
7.5 Stone-Weierstrass theorem 39
This proves (ii).

Conversely, assume that ℋ satisfies (i) and (ii). Since 𝒞 (𝑋 , 𝑌 ) is complete, the subset
ℋ ⊆ 𝒞 (𝑋 , 𝑌 ) is complete thanks to Proposition 4.5. Therefore, thanks to Theorem 5.9,
it remains to prove that ℋ is totally bounded, and for this it suffices to prove that ℋ is
totally bounded.
Let 𝜀 > 0. For any 𝑥 ∈ 𝑋 , there is a open neighbourhood 𝑈𝑥 of 𝑥 in 𝑋 such that, for
any 𝑦 ∈ 𝑈𝑥 and any 𝑓 ∈ ℋ , one has 𝑑(𝑓 (𝑦), 𝑓 (𝑥)) < 𝜀/4. Since 𝑋 = ⋃𝑥∈𝑋 𝑈𝑥 and 𝑋 is
compact, there is a finite number of points 𝑥1 , … , 𝑥𝑝 ∈ 𝑋 such that 𝑋 = 𝑈𝑥1 ∪ ⋯ ∪ 𝑈𝑥𝑝 .
Since 𝑌 is Hausdorff, the set 𝐾 ∶= ℋ (𝑥1 ) ∪ ⋯ ∪ ℋ (𝑥𝑝 ) is compact, thus there is a finite
number of points 𝑦1 , … , 𝑦𝑞 ∈ 𝑌 such that 𝐾 ⊆ 𝐵(𝑦1 , 𝜀/4) ∪ ⋯ ∪ 𝐵(𝑦𝑞 , 𝜀/4).
Define the set Φ of all maps from {1, … , 𝑝} to {1, … , 𝑞}, which is finite, and for any
𝜙 ∈ Φ define
ℋ𝜙 = {𝑓 ∈ ℋ ∣ ∀𝑖 ∈ {1, … , 𝑝}, 𝑓 (𝑥𝑖 ) ∈ 𝐵(𝑦𝜙(𝑖) , 4𝜀 )} .
By construction, the family {ℋ𝜙 }𝜙∈Φ is a finite cover of ℋ . Moreover, for any 𝜙 ∈ Φ, if
𝑓 , 𝑔 ∈ ℋ𝜙 then for any 𝑥 ∈ 𝑋 we consider 𝑖 ∈ {1, … , 𝑝} such that 𝑥 ∈ 𝑈𝑥𝑖 , and thus we get
𝑑(𝑓 (𝑥), 𝑔(𝑥)) ⩽ 𝑑(𝑓 (𝑥), 𝑓 (𝑥𝑖 )) + 𝑑(𝑓 (𝑥𝑖 ), 𝑦𝜙(𝑖) ) + 𝑑(𝑦𝜙(𝑖) , 𝑔(𝑥𝑖 )) + 𝑑(𝑔(𝑥𝑖 ), 𝑔(𝑥))
𝜀 𝜀 𝜀 𝜀
< + + + = 𝜀,
4 4 4 4
therefore ℋ𝜙 is contained in a ball of radius 𝜀 for the uniform metric 𝑑∞ , which concludes
the proof.
7.5 Stone-Weierstrass theorem

Let 𝑋 be a compact space and consider the space 𝒞 (𝑋 , R) of continuous real-valued
maps defined on 𝑋 endowed with the uniform metric.
Theorem 7.9 (Stone-Weierstrass). Let 𝒜 ⊆ 𝒞 (𝑋 , R) satisfy

(i) 𝒜 is a subalgebra: that is, 𝒜 is closed under addition, multiplication, and scalar
multiplication;
(ii) 𝒜 contains the constant maps;
(iii) 𝒜 separates points: that is, for any distinct 𝑥, 𝑦 ∈ 𝑋 , there is 𝑓 ∈ 𝒜 such that
𝑓 (𝑥) ≠ 𝑓 (𝑦).
Then 𝒜 is dense in 𝒞 (𝑋 , R).
The proof of this theorem will use the following:
Lemma 7.10. There exists a sequence of real polynomials (𝑃𝑛 )𝑛⩾0 in 𝑡 that converges uni-
formly to the function 𝑡 ↦ |𝑡| on [−1, 1].
Proof. Define 𝑃0 (𝑡) = 0 for any 𝑡 ∈ [−1, 1] and, for any 𝑛 ∈ N, 𝑃𝑛+1 (𝑡) = 𝑃𝑛 (𝑡) + 12 (𝑡 2 −
𝑃𝑛 (𝑡)2 ) for any 𝑡 ∈ [−1, 1]. By induction we easily obtain that 0 ⩽ 𝑃𝑛 (𝑡) ⩽ 𝑃𝑛+1 (𝑡) ⩽ |𝑡| for
any 𝑡 ∈ [−1, 1]. Therefore, for any 𝑡 ∈ [−1, 1], the real sequence (𝑃𝑛 (𝑡))𝑛⩾0 is increasing
and bounded above, hence it converges to some limit that we denote 𝑃(𝑡). By passing
to the limit 𝑛 → ∞, we obtain that 0 ⩽ 𝑃(𝑡) ⩽ |𝑡| and 𝑃(𝑡) = 𝑃(𝑡) − 12 (𝑡 2 − 𝑃(𝑡)2 ), thus
𝑃(𝑡) = |𝑡|.
This implies that the sequence (𝑃𝑛 )𝑛⩾0 converges pointwisely to 𝑡 ↦ |𝑡| defined on
[−1, 1]. By Dini’s theorem (see Exercise sheet) we deduce that (𝑃𝑛 )𝑛⩾0 converges uni-
formly to 𝑡 ↦ |𝑡| defined on [−1, 1].
Proof of Theorem 7.9. First of all, observe that 𝒜 is also a subalgebra of 𝒞 (𝑋 , R). We
split the proof into three steps.
Step 1. We shall prove that if 𝑓 , 𝑔 ∈ 𝒜 then |𝑓 | ∈ 𝒜 , min(𝑓 , 𝑔) ∈ 𝒜 , and max(𝑓 , 𝑔) ∈ 𝒜 .
Let 𝑓 ∈ 𝒜 . If 𝑓 = 0 then it is clear that 𝑓 ∈ 𝒜 , so we assume that 𝑓 ≠ 0. The sequence
𝑃𝑛 (𝑓 /‖𝑓 ‖∞ ) converges uniformly to |𝑓 |/‖𝑓 ‖∞ thanks to Lemma 7.10. Since 𝑃𝑛 (𝑓 /‖𝑓 ‖∞ ) ∈
𝒜 for any 𝑛 ∈ N, it follows that |𝑓 |/‖𝑓 ‖∞ ∈ 𝒜 and hence |𝑓 | ∈ 𝒜 .
Let 𝑓 , 𝑔 ∈ 𝒜 . Since
1 1
min(𝑓 , 𝑔) = (𝑓 + 𝑔 − |𝑓 − 𝑔|) and min(𝑓 , 𝑔) = (𝑓 + 𝑔 + |𝑓 − 𝑔|),
2 2
we deduce that min(𝑓 , 𝑔) ∈ 𝒜 and max(𝑓 , 𝑔) ∈ 𝒜 .

Step 2. We shall prove now that for any distinct 𝑥, 𝑦 ∈ 𝑋 and any 𝛼, 𝛽 ∈ R, there is ℎ ∈ 𝒜
such that ℎ(𝑥) = 𝛼 and ℎ(𝑦) = 𝛽.
Since 𝒜 separates points, there is 𝑔 ∈ 𝒜 such that 𝑔(𝑥) ≠ 𝑔(𝑦). We define
𝛽 −𝛼 𝛼𝑔(𝑦) − 𝛽𝑔(𝑥)
ℎ(𝑧) = 𝑔(𝑧) + , ∀𝑧 ∈ 𝑋 ,
𝑔(𝑦) − 𝑔(𝑥) 𝑔(𝑦) − 𝑔(𝑥)
so that ℎ ∈ 𝒜 , ℎ(𝑥) = 𝛼 and ℎ(𝑦) = 𝛽.

Step 3. Let 𝑓 ∈ 𝒞 (𝑋 , R) and 𝜀 > 0. Let 𝑥 ∈ 𝑋 . For any 𝑦 ∈ 𝑋 ⧵ {𝑥} there is a function
𝑓𝑦 ∈ 𝒜 such that 𝑓𝑦 (𝑥) = 𝑓 (𝑥) and 𝑓𝑦 (𝑦) = 𝑓 (𝑦). Denote
𝑈𝑦 = {𝑧 ∈ 𝑋 ∣ 𝑓𝑦 (𝑧) < 𝑓 (𝑧) + 𝜀}
which is a non-empty open subset of 𝑋 containing 𝑥. Since 𝑋 = ⋃𝑦∈𝑋 ⧵{𝑥} 𝑈𝑦 and 𝑋

is compact, there are 𝑦1 , … , 𝑦𝑛 ∈ 𝑋 such that 𝑋 = 𝑈𝑦1 ∪ ⋯ ∪ 𝑈𝑦𝑛 . Define the function
𝑔𝑥 = min(𝑓𝑦1 , … , 𝑓𝑦𝑛 ), so that 𝑔𝑥 ∈ 𝒜 and 𝑔𝑥 (𝑥) = 𝑓 (𝑥). Moreover, for any 𝑧 ∈ 𝑋 there
is 𝑖 ∈ {1, … , 𝑛} such that 𝑧 ∈ 𝑈𝑦𝑖 , hence 𝑔𝑥 (𝑧) ⩽ 𝑓𝑦𝑖 (𝑧) < 𝑓 (𝑧) + 𝜀.
We now consider the set
𝑉𝑥 = {𝑧 ∈ 𝑋 ∣ 𝑓 (𝑧) − 𝜀 < 𝑔𝑥 (𝑧)}
which is non-empty and open in 𝑋 . Since 𝑋 = ⋃𝑥∈𝑋 𝑉𝑥 and 𝑋 is compact, there are
𝑥1 , … , 𝑥𝑚 ∈ 𝑋 such that 𝑋 = 𝑉𝑥1 ∪ ⋯ ∪ 𝑉𝑥𝑚 . Consider the function 𝑔 = max(𝑔𝑥1 , … , 𝑔𝑥𝑚 )
so that 𝑔 ∈ 𝒜 . On the one hand we have 𝑔 < 𝑓 + 𝜀 since 𝑔𝑥𝑖 < 𝑓 + 𝜀 for any 𝑖 ∈ {1, … , 𝑚}.
On the other hand, for any 𝑧 ∈ 𝑋 there is 𝑖 ∈ {1, … , 𝑚} such that 𝑧 ∈ 𝑉𝑥𝑖 , hence 𝑔𝑥 (𝑧) ⩾
𝑔𝑥𝑖 (𝑧) > 𝑓 (𝑧) − 𝜀.
Therefore for any 𝑓 ∈ 𝒞 (𝑋 , R) and 𝜀 > 0 there exists 𝑔 ∈ 𝒜 such that 𝑑∞ (𝑓 , 𝑔) < 𝜀,
which means that 𝒜 is dense in 𝒞 (𝑋 , R), hence 𝒜 is also dense in 𝒞 (𝑋 , R).
It is worth noting that the above result regards the space of continuous real-valued
maps and one can prove that it is false in the case of complex-valued maps. One can
however prove a complex version of the Stone-Weierstrass theorem by adding a further
hypothesis.
7.5 Stone-Weierstrass theorem 41
Consider the space 𝒞 (𝑋 , C) of complex-valued continuous maps defined on 𝑋 en-

dowed with the metric of uniform convergence.
If 𝑧 ∈ C, we denote 𝑧 ̄ its complex conjugate. If 𝑓 ∈ 𝒞 (𝑋 , C), we denote by 𝑓 ̄ ∈
𝒞 (𝑋 , C) its complex conjugate function defined by 𝑓 ̄ ∶ 𝑧 ↦ 𝑓 (𝑧). Moreover, any 𝑓 ∈
𝒞 (𝑋 , C) can be decomposed into 𝑓 = ℜ(𝑓 )+iℑ(𝑓 ) where ℜ(𝑓 ) and ℑ(𝑓 ) are real-valued
functions.
Corollary 7.11 (Stone-Weierstrass, complex version). Let ℬ be a subset of 𝒞 (𝑋 , C) that

satisfies
(i) ℬ is a subalgebra: that is ℬ is closed under addition, multiplication, and scalar
multiplication;
(ii) ℬ contains the constant maps;
(iii) ℬ separates points: that is, for any distinct 𝑥, 𝑦 ∈ 𝑋 , there is 𝑓 ∈ ℬ such that
𝑓 (𝑥) ≠ 𝑓 (𝑦);
(iv) ℬ is closed under complex conjugation: that is, if 𝑓 ∈ ℬ then 𝑓 ̄ ∈ ℬ.
Then ℬ is dense in 𝒞 (𝑋 , C).
Proof. Consider the set ℬR ∶= {𝑓 ∈ ℬ ∣ 𝑓 (𝑋 ) ⊆ R}. Therefore ℬR ⊆ 𝒞 (𝑋 , R) is a

subalgebra and contains the constant maps. Furthermore ℬR separates points, indeed
if 𝑥, 𝑦 ∈ 𝑋 are distinct there exists a map 𝑓 ∈ ℬ such that 𝑓 (𝑥) ≠ 𝑓 (𝑦), hence the maps
ℜ(𝑓 ) = 21 (𝑓 + 𝑓 )̄ and ℑ(𝑓 ) = 2𝑖1 (𝑓 − 𝑓 )̄ belong to ℬR and, since 𝑓 = ℜ(𝑓 ) + iℑ(𝑓 ), at
least one of them separates the points 𝑥 and 𝑦. Therefore ℬR is dense in 𝒞 (𝑋 , R) by
the Stone-Weierstrass theorem (Theorem 7.9), and hence we can conclude since ℬ =
ℬR + iℬR .
Chapter 8
Banach spaces and continuous linear

maps
Through this chapter we shall always consider K-vector spaces where K is the field of
real numbers R or complex numbers C. Moreover, by a linear map we shall always mean
a K-linear map.
8.1 Normed vector spaces and Banach spaces

Definition 8.1. A norm on a vector space 𝑋 is a map ‖ ⋅ ‖ ∶ 𝑋 → R+ such that, for all
𝑥, 𝑦 ∈ 𝑋 and 𝜆 ∈ K, one has:
(i) separation: ‖𝑥‖ = 0 if and only if 𝑥 = 0 ;
(ii) homogeneity: ‖𝜆𝑥‖ = |𝜆|‖𝑥‖ ;
(iii) triangle inequality: ‖𝑥 + 𝑦‖ ⩽ ‖𝑥‖ + ‖𝑦‖.
The pair (𝑋 , ‖ ⋅ ‖) is called a normed (vector) space.
Remark 8.2. We deduce that |‖𝑥‖ − ‖𝑦‖| ⩽ ‖𝑥 − 𝑦‖ for any 𝑥, 𝑦 ∈ 𝑋 .
Remark 8.3. A semi-norm on 𝑋 is a map ‖ ⋅ ‖ ∶ 𝑋 → R+ satisfying (ii) and (iii).
If (𝑋 , ‖ ⋅ ‖) is a normed space then we can define a metric on 𝑋 by setting 𝑑(𝑥, 𝑦) =

‖𝑥 − 𝑦‖ for 𝑥, 𝑦 ∈ 𝑋 . And then we can endow 𝑋 with the topology associated to the
metric 𝑑.
If 𝑌 is a vector subspace of a normed space 𝑋 , then 𝑌 is also a normed space endowed
with the restriction to 𝑌 of the norm on 𝑋 .
Proposition 8.4. Let (𝑋 , ‖ ⋅ ‖) be a normed space.

(i) The map 𝑥 ↦ ‖𝑥‖ from 𝑋 to R+ is continuous.
(ii) The map (𝜆, 𝑥) ↦ 𝜆𝑥 from K × 𝑋 to 𝑋 is continuous.
(iii) The map (𝑥, 𝑦) ↦ 𝑥 + 𝑦 from 𝑋 × 𝑋 to 𝑋 is continuous.
Proof. (i) Let 𝑥0 ∈ 𝑋 , we deduce that the map 𝑥 ↦ ‖𝑥‖ is continuous at 𝑥0 thanks to the
inequality
|‖𝑥‖ − ‖𝑥0 ‖| ⩽ ‖𝑥 − 𝑥0 ‖.
43
44 Chapter 8. Banach spaces and continuous linear maps
(ii) Let (𝜆0 , 𝑥0 ) ∈ K × 𝑋 , we deduce that the map (𝜆, 𝑥) ↦ 𝜆𝑥 is continuous at (𝜆0 , 𝑥0 )
since one has
‖𝜆𝑥 − 𝜆0 𝑥0 ‖ ⩽ |𝜆 − 𝜆0 |‖𝑥‖ + |𝜆0 |‖𝑥 − 𝑥0 ‖.
(iii) Let (𝑥0 , 𝑦0 ) ∈ 𝑋 × 𝑋 , we deduce that the map (𝑥, 𝑦) ↦ 𝑥 + 𝑦 is continuous at (𝑥0 , 𝑦0 )
since one has
‖(𝑥 + 𝑦) − (𝑥0 + 𝑦0 )‖ ⩽ ‖𝑥 − 𝑥0 ‖ + ‖𝑦 − 𝑦0 ‖.
Definition 8.5. Let (𝑋 , ‖ ⋅ ‖) be a normed vector space and 𝐴 a subset of 𝑋 . One says
that:
(i) 𝐴 is bounded if sup𝑥∈𝐴 ‖𝑥‖ < ∞;
(ii) 𝐴 is convex if for any 𝑥, 𝑦 ∈ 𝐴 and any 𝑡 ∈ [0, 1] one has 𝑡𝑥 + (1 − 𝑡)𝑦 ∈ 𝐴.
One remarks that 𝐴 is bounded if and only if 𝐴 is contained in some open ball.
Definition 8.6. Two norms ‖ ⋅ ‖1 and ‖ ⋅ ‖2 on a vector space 𝑋 are said to be equivalent
if there are positive constants 𝛼, 𝛽 > 0 such that
∀𝑥 ∈ 𝑋 , 𝛼‖𝑥‖1 ⩽ ‖𝑥‖2 ⩽ 𝛽‖𝑥‖1 .
Definition 8.7. A Banach space is a normed space that is complete with respect to the
metric associated to its norm.
Example 8.1. On K𝑁 we can define the norms, for 𝑥 = (𝑥1 , … , 𝑥𝑁 ) ∈ K𝑁 ,
𝑁 1/𝑝
‖𝑥‖𝑝 = ( ∑ |𝑥𝑘 𝑝
| ) for 1 ⩽ 𝑝 < ∞, ‖𝑥‖∞ = sup |𝑥𝑘 |.
𝑘=1 1⩽𝑘⩽𝑁
Endowed with any of these norms, K𝑁 is a Banach space.

Example 8.2. For any 𝑝 ⩾ 1, the set ℓ𝑝 (N; K) of sequences (𝑥𝑛 )𝑛⩾0 in K such that the
series with general term |𝑥𝑛 |𝑝 is convergent can be endowed with the norm
1
𝑝
‖(𝑥𝑛 )𝑛⩾0 ‖𝑝 = ( ∑ |𝑥𝑛 |𝑝 ) .
𝑛⩾0
The space (ℓ𝑝 (N; K), ‖ ⋅ ‖𝑝 ) is a Banach space.

Example 8.3. The set ℓ∞ (N; K) of bounded sequences in K can be endowed with the
norm
‖(𝑥𝑛 )𝑛⩾0 ‖∞ = sup |𝑥𝑛 |.
𝑛⩾0
The space (ℓ∞ (N; K), ‖ ⋅ ‖∞ ) is a Banach space.
Example 8.4. On the space 𝒞 ([0, 1]; K) of continuous functions from [0, 1] into K, we
can define the following norms:
1
1 𝑝
‖𝑓 ‖𝑝 ∶= (∫ |𝑓 (𝑥)|𝑝 d𝑥) for 1 ⩽ 𝑝 < ∞, ‖𝑓 ‖∞ ∶= sup |𝑓 (𝑥)|.
0 𝑥∈[0,1]
The space (𝒞 ([0, 1]; K), ‖ ⋅ ‖∞ ) is a Banach space, however (𝒞 ([0, 1]; K), ‖ ⋅ ‖1 ) is not.
8.2 Finite-dimensional normed vector spaces 45
Example 8.5. If 𝑋 is a non-empty compact space, we can consider the space 𝒞 (𝑋 ; K) of

continuous functions from 𝑋 to K, endowed with the norm
‖𝑓 ‖∞ ∶= sup |𝑓 (𝑥)|.
𝑥∈𝑋
The space (𝒞 (𝑋 ; K), ‖ ⋅ ‖∞ ) is a Banach space.

As a consequence of results proven in Chapter 4, one has:
Proposition 8.8.
(i) A closed vector subspace of a Banach space is a Banach space.
(ii) A finite product of Banach spaces is a Banach space.
8.2 Finite-dimensional normed vector spaces

In this section we shall focus on finite-dimensional vector spaces and prove some prop-
erties of them.
Theorem 8.9. Let 𝑋 be a finite-dimensional vector space over K. Then all norms on 𝑋 are
equivalent and 𝑋 is complete for any of them.
Proof. Let (𝑒1 , … , 𝑒𝑛 ) be a basis of 𝑋 , and for 𝑥 ∈ 𝑋 we write 𝑥 = 𝑥1 𝑒1 + ⋯ + 𝑥𝑛 𝑒𝑛 . Since
K is complete, it suffices to show that every norm on 𝑋 is equivalent to the norm ‖ ⋅ ‖∞
defined by ‖𝑥‖∞ = sup(|𝑥1 |, … , |𝑥𝑛 |).
Let ‖ ⋅ ‖ be a norm on 𝑋 and 𝑥 ∈ 𝑋 , then we have
‖𝑥1 𝑒1 + ⋯ + 𝑥𝑛 𝑒𝑛 ‖ ⩽ (‖𝑒1 ‖ + ⋯ + ‖𝑒𝑛 ‖) sup(|𝑥1 |, … , |𝑥𝑛 |),
which implies that ‖𝑥‖ ⩽ 𝑐1 ‖𝑥‖∞ , where 𝑐1 = ‖𝑒1 ‖ + ⋯ + ‖𝑒𝑛 ‖ > 0. This gives the first
inequality.
Moreover, last inequality implies that the map ‖ ⋅ ‖ ∶ (𝑋 , ‖ ⋅ ‖∞ ) → R is continuous.
Since 𝑆 ∶= {𝑥 ∈ 𝑋 ∣ ‖𝑥‖∞ = 1} is compact, the image of 𝑆 under ‖ ⋅ ‖ is a compact
subset of R and therefore it possess a minimum, that is, there is some 𝑥0 ∈ 𝑆 such that
‖𝑥‖ ⩾ ‖𝑥0 ‖ > 0 for any 𝑥 ∈ 𝑆. By homogeneity of the norm and setting 𝑐2 = ‖𝑥0 ‖, we
finally get that
∀𝑥 ∈ 𝑋 , 𝑐2 ‖𝑥‖∞ ⩽ ‖𝑥‖,
which completes the proof.
Proposition 8.10. If 𝑋 is a finite-dimensional vector space, then the compact sets in 𝑋
are the closed and bounded sets.
Proof. We will show that the compact sets in (R𝑁 , ‖ ⋅ ‖∞ ) are the closed and bounded
sets, and the case of a finite-dimensional vector space follows from it since all norms are
equivalent.
We already know that if 𝐴 ⊆ R𝑁 is compact then it is closed and bounded. Con-
versely, let 𝐴 ⊆ R𝑁 be bounded and closed, then there is 𝑀 > 0 such that 𝐴 ⊆ [−𝑀, 𝑀]𝑁 ,
and the set [−𝑀, 𝑀]𝑁 is compact as the product of compact sets. Therefore 𝐴 is a closed
subset of a compact set, thus 𝐴 is compact.
Theorem 8.11 (Riesz theorem). The closed unit ball 𝐵(0, 1) of a normed vector space 𝑋
is compact if and only if 𝑋 is finite-dimensional.
Proof. If 𝑋 is finite-dimensional, we already know that the closed unit ball 𝐵(0, 1) is
compact since it is bounded and closed.
Conversely, assume that 𝐵(0, 1) is compact. Since {𝐵(𝑥, 1/2)}𝑥∈𝐵(0,1) is a cover of
𝐵(0, 1) by open sets in 𝑋 , there is a finite subset {𝑒1 , … , 𝑒𝑁 } ⊆ 𝐵(0, 1) such that 𝐵(0, 1) ⊆
𝐵(𝑒1 , 12 ) ∪ ⋯ ∪ 𝐵(𝑒𝑁 , 21 ). Define 𝑌 = span(𝑒1 , … , 𝑒𝑁 ), we will show that 𝑌 = 𝑋 which will
complete the proof. Since 𝑌 is closed, it suffices to show that 𝑌 is dense in 𝑋 . Let 𝑥 ∈ 𝑋 ,
then there are 𝑦 ∈ 𝑌 and 𝑎 ∈ Z such that ‖𝑥 − 𝑦‖ ⩽ 2−𝑎 . Then we have 2𝑎 (𝑥 − 𝑦) ∈ 𝐵(0, 1)
and hence, by definition of 𝑌 , there is 𝑗 ∈ {1, … , 𝑁 } such that ‖2𝑎 (𝑥 − 𝑦) − 𝑒𝑗 ‖ ⩽ 1/2.
Therefore one has that 𝑦1 ∶= 𝑦 + 2−𝑎 𝑒𝑗 ∈ 𝑌 and ‖𝑥 − 𝑦1 ‖ ⩽ 2−𝑎−1 . By induction we can
construct a sequence (𝑦𝑛 )𝑛⩾0 of elements of 𝑌 such that ‖𝑥 − 𝑦𝑛 ‖ ⩽ 2−𝑎−𝑛 , thus 𝑥 belongs
to the closure of 𝑌 .
8.3 Continuous linear maps

Let (𝑋 , ‖ ⋅ ‖𝑋 ) and (𝑌 , ‖ ⋅ ‖𝑌 ) be normed spaces. We recall that a map 𝑇 ∶ 𝑋 → 𝑌 is linear if
for any two vectors 𝑥, 𝑦 ∈ 𝑋 and any scalar 𝜆 ∈ K the following conditions are satisfied
𝑇 (𝑥 + 𝑦) = 𝑇 (𝑥) + 𝑇 (𝑦) and 𝑇 (𝜆𝑥) = 𝜆𝑇 (𝑥).
In that case one usually denotes 𝑇 (𝑥) simply by 𝑇 𝑥.

Theorem 8.12. Let 𝑇 ∶ 𝑋 → 𝑌 be a linear map. The following are equivalent:
(i) 𝑇 is continuous ;
(ii) 𝑇 is continuous at 0 ;
(iii) 𝑇 is bounded: that is, there is 𝐶 > 0 such that for any 𝑥 ∈ 𝑋 one has ‖𝑇 𝑥‖𝑌 ⩽ 𝐶‖𝑥‖𝑋 .
Proof. (i) ⇒ (ii) Trivial.
(ii) ⇒ (iii) Assume that 𝑇 is continuous at 0, then there is 𝛿 > 0 such that
∀𝑥 ∈ 𝑋 , ‖𝑥‖𝑋 ⩽ 𝛿 ⇒ ‖𝑢(𝑥)‖𝑌 ⩽ 1.
‖𝑥‖𝑋 𝛿
Now for any 𝑥 ≠ 0 we write 𝑥 = 𝛿 ‖𝑥‖𝑋
𝑥, hence
‖𝑥‖𝑋 ‖𝑥‖𝑋
‖𝑇 𝑥‖𝑌 = ‖𝑇 ( ‖𝑥‖𝛿 𝑥)‖ ⩽ .
𝛿 𝑋 𝑌 𝛿
(iii) ⇒ (i) Assume that 𝑇 is bounded. Then, for any 𝑥, 𝑦 ∈ 𝑋 , one has
‖𝑇 𝑥 − 𝑇 𝑦‖𝑌 = ‖𝑇 (𝑥 − 𝑦)‖𝑌 ⩽ 𝐶‖𝑥 − 𝑦‖𝑋 ,
which implies that 𝑇 is Lipschitz, hence continuous.
8.3.1 Space of continuous linear maps

We denote by ℒ (𝑋 , 𝑌 ) the vector space of all continuous linear maps from 𝑋 to 𝑌 . For
any 𝑇 ∈ ℒ (𝑋 , 𝑌 ) we define
‖𝑇 𝑥‖𝑌
‖𝑇 ‖ℒ (𝑋 ,𝑌 ) ∶= sup .
𝑥∈𝑋 ⧵{0} ‖𝑥‖𝑋
8.3 Continuous linear maps 47
One can easily check that ‖ ⋅ ‖ℒ (𝑋 ,𝑌 ) is a norm on ℒ (𝑋 , 𝑌 ) and moreover that
‖𝑇 ‖ℒ (𝑋 ,𝑌 ) = sup ‖𝑇 𝑥‖𝑌 = sup ‖𝑇 𝑥‖𝑌 ,

‖𝑥‖𝑋 ⩽1 ‖𝑥‖𝑋 =1
and also
‖𝑇 𝑥‖𝑌 ⩽ ‖𝑇 ‖ℒ (𝑋 ,𝑌 ) ‖𝑥‖𝑋 .
Proposition 8.13. If (𝑌 , ‖ ⋅ ‖𝑌 ) is a Banach space, then (ℒ (𝑋 , 𝑌 ), ‖ ⋅ ‖ℒ (𝑋 ,𝑌 ) ) is a Banach

space.
Proof. Let (𝑇𝑛 )𝑛⩾0 be a Cauchy sequence in ℒ (𝑋 , 𝑌 ). For any 𝜀 > 0 there is 𝑁0 ∈ N such
that for any integers 𝑛, 𝑚 ⩾ 𝑁0 we have ‖𝑇𝑛 − 𝑇𝑚 ‖ℒ (𝑋 ,𝑌 ) < 𝜀. Thus, for any 𝑥 ∈ 𝑋 , we
have
‖𝑇𝑛 𝑥 − 𝑇𝑚 𝑥‖𝑌 ⩽ ‖𝑇𝑛 − 𝑇𝑚 ‖ℒ (𝑋 ,𝑌 ) ‖𝑥‖𝑋 < 𝜀 ‖𝑥‖𝑋 . (⋆)
It follows that, for any 𝑥 ∈ 𝑋 , the sequence (𝑇𝑛 𝑥)𝑛⩾0 is a Cauchy sequence in 𝑌 , which is
complete, hence there is 𝑦𝑥 ∈ 𝑌 such that (𝑇 𝑥𝑛 )𝑛⩾0 converges to 𝑦𝑥 in 𝑌 . Define the map
𝑇 ∶ 𝑥 ↦ 𝑦𝑥 from 𝑋 to 𝑌 , which clearly is linear. It remains to show that 𝑇 is continuous
and that (𝑇𝑛 )𝑛⩾0 converges to 𝑇 in ℒ (𝑋 , 𝑌 ).
From (⋆) we obtain that, for any 𝑥 ∈ 𝑋 ,
‖𝑇 𝑥‖𝑌 ⩽ ‖𝑇 𝑥 − 𝑇𝑁0 𝑥‖𝑌 + ‖𝑇𝑁0 𝑥‖𝑌 ⩽ (𝜀 + ‖𝑇𝑁0 ‖ℒ (𝑋 ,𝑌 ) ) ‖𝑥‖𝑋 ,
which implies that 𝑇 is continuous. Finally, passing to the limit 𝑚 → ∞ in (⋆) we obtain,
for any 𝑥 ∈ 𝑋 and 𝑛 ⩾ 𝑁0 ,
‖𝑇 𝑥 − 𝑇𝑛 𝑥‖𝑌 = lim ‖𝑇𝑛 𝑥 − 𝑇𝑚 𝑥‖𝑌 ⩽ 𝜀‖𝑥‖𝑋 ,

𝑚→∞
which implies that ‖𝑇 − 𝑇𝑛 ‖ℒ (𝑋 ,𝑌 ) ⩽ 𝜀, thus (𝑇𝑛 )𝑛⩾0 converges to 𝑇 in ℒ (𝑋 , 𝑌 ).
We now state a result concernig the extension of continuous linear maps.
Theorem 8.14 (Bounded linear transformation theorem). Let 𝑋 be a normed space, 𝑌

a Banach space, and 𝐷 a dense vector subspace of 𝑋 . Then every continuous linear map
𝑇 ∶ 𝐷 → 𝑌 can be uniquely extended to a continuous linear map 𝑇̂ ∶ 𝑋 → 𝑌 , that is,
𝑇̂ ∈ ℒ (𝑋 , 𝑌 ) and 𝑇̂|𝐷 = 𝑇 .
Proof. Let 𝑥 ∈ 𝑋 , then there is a sequence (𝑥𝑛 )𝑛∈N in 𝐷 converging to 𝑥 in 𝑋 . For all
𝑛, 𝑚 ∈ N we have
‖𝑇 𝑥𝑛 − 𝑇 𝑥𝑚 ‖𝑌 ⩽ ‖𝑇 ‖ℒ (𝐷,𝑌 ) ‖𝑥𝑛 − 𝑥𝑚 ‖𝑋 ,
thus (𝑇 𝑥𝑛 )𝑛∈N is a Cauchy sequence in 𝑌 which is complete, hence it converges to some
limit.
We define the map 𝑇̂ ∶ 𝑋 → 𝑌 , 𝑥 ↦ lim𝑛→∞ 𝑇 𝑥𝑛 , where (𝑥𝑛 )𝑛∈N is a sequence in
𝐷 converging to 𝑥. We first check that 𝑇̂ is well-defined, that is, it does not depend on
the choice of the sequence converging to 𝑥. Indeed if (𝑥𝑛′ )𝑛∈N is another sequence in 𝐷
converging to 𝑥, then
‖𝑇 𝑥𝑛 − 𝑇 𝑥𝑛′ ‖𝑌 ⩽ ‖𝑇 ‖ℒ (𝐷,𝑌 ) ‖𝑥𝑛 − 𝑥𝑛′ ‖𝑋 −−−−→ 0,

𝑛→∞
thus lim𝑛→∞ 𝑇 𝑥𝑛 = lim𝑛→∞ 𝑇 𝑥𝑛′ .

The map 𝑇̂ is clearly linear. For all 𝑥 ∈ 𝐷, we consider the constant sequence (𝑥)𝑛∈N
so that 𝑇̂𝑥 = lim𝑛→∞ 𝑇 𝑥 = 𝑇 𝑥, that is, 𝑇̂ is an extension of 𝑇 .
We now prove that 𝑇̂ is continuous. Let 𝑥 ∈ 𝑋 and (𝑥𝑛 )𝑛∈N a sequence in 𝐷 converg-
ing to 𝑥, then
‖𝑇̂𝑥‖𝑌 = ‖ lim 𝑇 𝑥𝑛 ‖𝑌 = lim ‖𝑇 𝑥𝑛 ‖𝑌 ⩽ lim ‖𝑇 ‖ℒ (𝐷,𝑌 ) ‖𝑥𝑛 ‖𝑋 ⩽ ‖𝑇 ‖ℒ (𝐷,𝑌 ) ‖𝑥‖𝑋 ,

𝑛→∞ 𝑛→∞ 𝑛→∞
which implies that 𝑇̂ is continuous and ‖𝑇̂‖ℒ (𝑋 ,𝑌 ) ⩽ ‖𝑇 ‖ℒ (𝐷,𝑌 ) . Moreover, we have
‖𝑇̂‖ℒ (𝑋 ,𝑌 ) = sup ‖𝑇̂𝑥‖𝑌 ⩾ sup ‖𝑇̂𝑥‖𝑌 = sup ‖𝑇 𝑥‖𝑌 = ‖𝑇 ‖ℒ (𝐷,𝑌 ) ,

𝑥∈𝑋 , ‖𝑥‖𝑋 ⩽1 𝑥∈𝐷, ‖𝑥‖𝑋 ⩽1 𝑥∈𝐷, ‖𝑥‖𝑋 ⩽1
thus ‖𝑇̂‖ℒ (𝑋 ,𝑌 ) = ‖𝑇 ‖ℒ (𝐷,𝑌 ) .

We finally prove that 𝑇̂ is unique. Let 𝑇̃ ∈ ℒ (𝑋 , 𝑌 ) be an extension of 𝑇 . Let 𝑥 ∈ 𝑋
and consider a sequence (𝑥𝑛 )𝑛∈N in 𝐷 converging to 𝑥. Thus
𝑇̂𝑥 = lim 𝑇 𝑥𝑛 = lim 𝑇̃𝑥𝑛 = 𝑇̃ ( lim 𝑥𝑛 ) = 𝑇̃𝑥

𝑛→∞ 𝑛→∞ 𝑛→∞
using the continuity of 𝑇̃, hence 𝑇̂ = 𝑇̃.
8.3.2 Isomorphisms
Definition 8.15. Let 𝑋 and 𝑌 be normed vector spaces.
(i) A map 𝑓 ∶ 𝑋 → 𝑌 is an isometry if it preserves the distance, that is, if ‖𝑓 (𝑥) −
𝑓 (𝑦)‖𝑌 = ‖𝑥 − 𝑦‖𝑋 for all 𝑥, 𝑦 ∈ 𝑋 .
(ii) A map 𝑓 ∶ 𝑋 → 𝑌 is an isomorphism if it is bijective and if both 𝑓 and its inverse
map 𝑓 −1 are linear and continuous. Two normed vector spaces are said to be
isomorphic if there is a isomorphism between them.
(iii) A map 𝑓 ∶ 𝑋 → 𝑌 is an isometric isomorphism if it is an isomorphism that is also
an isometry. Two normed vector spaces are said to be isometrically isomorphic if
there is a isometric isomorphism between them.
Remark 8.16. Remark that if a map 𝑓 ∶ 𝑋 → 𝑌 is linear and bijective, then its inverse
map is also linear.
We shall denote by Isom(𝑋 , 𝑌 ) the set of all isomorphisms from 𝑋 into 𝑌 . We have
the following result concerning the set of isomorphisms and the inversion map, which
is left as an exercise.
Proposition 8.17. Let 𝑋 and 𝑌 be Banach spaces.

(i) If 𝑇 ∈ ℒ (𝑋 , 𝑋 ) with ‖𝑇 ‖ℒ (𝑋 ,𝑋 ) < 1, then id𝑋 − 𝑇 ∈ Isom(𝑋 , 𝑋 ) with inverse
∑𝑛∈N 𝑇 𝑛 .
(ii) Isom(𝑋 , 𝑌 ) is open in ℒ (𝑋 , 𝑌 ).
(iii) The inversion map 𝑇 ↦ 𝑇 −1 from Isom(𝑋 , 𝑌 ) into Isom(𝑌 , 𝑋 ) is continuous.
8.3 Continuous linear maps 49
8.3.3 Dual space

Let (𝑋 , ‖ ⋅ ‖𝑋 ) be a normed vector space.
Definition 8.18. We define the (topological) dual space of 𝑋 as the space 𝑋 ′ = ℒ (𝑋 , K)

of all continuous linear functionals on 𝑋 , endowed with the norm
‖𝑓 ‖𝑋 ′ = ‖𝑓 ‖ℒ (𝑋 ,K) = sup |𝑓 (𝑥)|.

𝑥∈𝑋 , ‖𝑥‖𝑋 ⩽1
Remark that since K is complete, the normed vector space (𝑋 ′ , ‖ ⋅ ‖𝑋 ′ ) is a Banach

space.
8.3.4 Continuous bilinear maps

Let (𝑋1 , ‖ ⋅ ‖𝑋1 ), (𝑋2 , ‖ ⋅ ‖𝑋2 ) and (𝑌 , ‖ ⋅ ‖𝑌 ) be normed vector spaces. The topological space
𝑋1 × 𝑋2 is a normed vector space, and the product topology is associated to the norm
‖(𝑥1 , 𝑥2 )‖𝑋1 ×𝑋2 = sup(‖𝑥1 ‖𝑋1 , ‖𝑥2 ‖𝑋2 )
or
‖(𝑥1 , 𝑥2 )‖𝑋1 ×𝑋2 = ‖𝑥1 ‖𝑋1 + ‖𝑥2 ‖𝑋2 ,
or still any other equivalent norm.
Definition 8.19. A map 𝑇 ∶ 𝑋1 × 𝑋2 → 𝑌 is bilinear if

• for any 𝑥 ∈ 𝑋1 , the map 𝑦 ↦ 𝑇 (𝑥, 𝑦) from 𝑋2 to 𝑌 is linear;
• for any 𝑦 ∈ 𝑋2 , the map 𝑥 ↦ 𝑇 (𝑥, 𝑦) from 𝑋1 to 𝑌 is linear.
Theorem 8.20. Let 𝑇 ∶ 𝑋1 × 𝑋2 → 𝑌 be a bilinear map. The following are equivalent:

(i) 𝑇 is continuous ;
(ii) 𝑇 is continuous at (0, 0) ;
(iii) there is 𝐶 > 0 such that, for any (𝑥1 , 𝑥2 ) ∈ 𝑋1 × 𝑋2 , one has
‖𝑇 (𝑥1 , 𝑥2 )‖𝑌 ⩽ 𝐶‖𝑥1 ‖𝑋1 ‖𝑥2 ‖𝑋2 .
Proof. The proof is similar to the proof of Theorem 8.12, and hence is left as an exercise.
We denote by ℒ2 (𝑋1 , 𝑋2 ; 𝑌 ) the space of bilinear continuous maps from 𝑋1 × 𝑋2 to

𝑌 , endowed with the norm
‖𝑇 (𝑥1 , 𝑥2 )‖𝑌
‖𝑇 ‖ℒ2 (𝑋1 ,𝑋2 ;𝑌 ) ∶= sup
𝑥1 ∈𝑋1 ⧵{0}, 𝑥2 ∈𝑋2 ⧵{0} ‖𝑥1 ‖𝑋1 ‖𝑥2 ‖𝑋2
and thus (ℒ2 (𝑋1 , 𝑋2 ; 𝑌 ), ‖⋅‖ℒ2 (𝑋1 ,𝑋2 ;𝑌 ) ) is a normed vector space. As for continuous linear
maps, one can easily obtain that
‖𝑇 ‖ℒ2 (𝑋1 ,𝑋2 ;𝑌 ) = sup ‖𝑇 (𝑥1 , 𝑥2 )‖𝑌 = sup ‖𝑇 (𝑥1 , 𝑥2 )‖𝑌 .

‖𝑥1 ‖𝑋1 ⩽1,‖𝑥2 ‖𝑋2 ⩽1 ‖𝑥1 ‖𝑋1 =1,‖𝑥2 ‖𝑋2 =1
Proposition 8.21. The spaces ℒ2 (𝑋1 , 𝑋2 ; 𝑌 ) and ℒ (𝑋1 , ℒ (𝑋2 , 𝑌 )) are isomorphic. More
precisely, the map
ℒ2 (𝑋1 , 𝑋2 ; 𝑌 ) → ℒ (𝑋1 , ℒ (𝑋2 , 𝑌 ))
𝑇 ↦ {𝑥 ↦ (𝑦 ↦ 𝑇 (𝑥, 𝑦))}
is an isometric isomorphism between normed spaces, and its inverse map is given by
ℒ (𝑋1 , ℒ (𝑋2 , 𝑌 )) → ℒ2 (𝑋1 , 𝑋2 ; 𝑌 )

𝑆 ↦ {(𝑥, 𝑦) ↦ 𝑆(𝑥)(𝑦)}.
Proof. For any 𝑥 ∈ 𝑋1 and 𝑦 ∈ 𝑋2 one has, denoting 𝑇𝑥 ∶ 𝑦 ↦ 𝑇 (𝑥, 𝑦),
‖𝑇𝑥 (𝑦)‖𝑌 = ‖𝑇 (𝑥, 𝑦)‖𝑌 ⩽ ‖𝑇 ‖ℒ2 (𝑋1 ,𝑋2 ;𝑌 ) ‖𝑥‖𝑋1 ‖𝑦‖𝑋2 ,
thus the linear map 𝑇𝑥 ∶ 𝑋2 → 𝑌 is continuous with
‖𝑇𝑥 ‖ℒ (𝑋2 ,𝑌 ) ⩽ ‖𝑇 ‖ℒ2 (𝑋1 ,𝑋2 ;𝑌 ) ‖𝑥‖𝑋1 .
Therefore the map 𝑥 ↦ 𝑇𝑥 is continuous with norm less or equal than ‖𝑇 ‖ℒ2 (𝑋1 ,𝑋2 ;𝑌 ) .
Since one has
‖𝑇𝑥 ‖ℒ (𝑋2 ,𝑌 ) ‖𝑇 (𝑥, 𝑦)‖𝑌
sup = sup sup = ‖𝑇 ‖ℒ2 (𝑋1 ,𝑋2 ;𝑌 ) ,
𝑥∈𝑋1 ⧵{0} ‖𝑥‖𝑋1 𝑥∈𝑋1 ⧵{0} 𝑦∈𝑋2 ⧵{0} ‖𝑥‖𝑋1 ‖𝑦‖𝑋2
it follows that the map 𝑇 ↦ {𝑥 ↦ 𝑇𝑥 } is an isometry.

Finally, for 𝑆 ∈ ℒ (𝑋1 , ℒ (𝑋2 , 𝑌 )) we define 𝑇 ∶ (𝑥, 𝑦) ↦ 𝑆(𝑥)(𝑦) which is a bilinear
map. Since
‖𝑆(𝑥)(𝑦)‖𝑌 ⩽ ‖𝑆(𝑥)‖ℒ (𝑋2 ,𝐹 ) ‖𝑦‖𝑋2 ⩽ ‖𝑣‖ℒ2 (𝑋1 ,𝑋2 ;𝑌 ) ‖𝑥‖𝑋1 ‖𝑦‖𝑋2 ,
we deduce that 𝑇 is continuous and 𝑆(𝑥) = 𝑇𝑥 , which completes the proof.
8.4 Applications of Baire’s theorem

As a consequence of Baire’s Theorem 4.12, we shall state and prove below three impor-
tant results in functional analysis: the Banach-Steinhaus theorem or uniform bounded-
ness principle, the open mapping thereom, and the closed graph theorem.
8.4.1 Uniform boundedness principle

Theorem 8.22 (Banach-Steinhaus theorem or uniform boundedness principle). Let (𝑋 , ‖⋅
‖𝑋 ) be a Banach space and (𝑌 , ‖ ⋅ ‖𝑌 ) a normed vector space. Let {𝑇𝛼 }𝛼∈𝒜 ⊆ ℒ (𝑋 , 𝑌 ) be an
arbitrary family of continuous linear maps from 𝑋 into 𝑌 . We suppose that for any 𝑥 ∈ 𝑋
one has
sup ‖𝑇𝛼 𝑥‖𝑌 < +∞.
𝛼∈𝒜
Then one has

sup ‖𝑇𝛼 ‖ℒ (𝑋 ,𝑌 ) < +∞.
𝛼∈𝒜
8.4 Applications of Baire’s theorem 51
Proof. Define the set 𝑋𝑛 = {𝑥 ∈ 𝑋 ∣ ∀𝛼 ∈ 𝒜 , ‖𝑇𝛼 𝑥‖𝑌 ⩽ 𝑛} for any 𝑛 ∈ N. Then 𝑋𝑛 is

closed and, by hypothesis, one has 𝑋 = ⋃𝑛∈N 𝑋𝑛 . By Baire’s theorem (Theorem 4.12),
we deduce that there exists 𝑁 ∈ N such that the interior of 𝑋𝑁 is not empty, thus there
are 𝑥0 ∈ 𝑋 and 𝑟 > 0 such that 𝐵𝑋 (𝑥0 , 𝑟) ⊆ 𝑋𝑁 .
This implies that for any 𝛼 ∈ 𝒜 and any 𝑥 ∈ 𝑋 such that ‖𝑥‖𝑋 ⩽ 𝑟, one has ‖𝑇𝛼 𝑥‖𝑌 ⩽
‖𝑢𝛼 (𝑥 + 𝑥0 )‖𝑌 + ‖𝑇𝛼 𝑥0 ‖𝑌 ⩽ 2𝑁 . Finally, one obtains that for any 𝛼 ∈ 𝒜 there holds
2𝑁
‖𝑇𝛼 ‖ℒ (𝑋 ,𝑌 ) = sup ‖𝑇𝛼 𝑥‖𝑌 ⩽ ,
‖𝑥‖𝑋 ⩽1 𝑟
which concludes the proof.
8.4.2 Open mapping theorem

Theorem 8.23 (Open mapping theorem). Let (𝑋 , ‖ ⋅ ‖𝑋 ) and (𝑌 , ‖ ⋅ ‖𝑌 ) be two Banach
spaces, and 𝑇 ∈ ℒ (𝑋 , 𝑌 ) be a surjective continuous linear map from 𝑋 into 𝑌 . Then 𝑇 is
an open map.
Proof. Define 𝐴𝑛 = 𝑛 𝑇 (𝐵𝑋 (0, 1)) for any 𝑛 ∈ N∗ , so that we have 𝑌 = ⋃𝑛∈N∗ 𝐴𝑛 since 𝑇
is surjective. Thanks to Baire’s theorem (Theorem 4.12), we deduce that there is 𝑁 ∈ N∗
such that the interior of 𝐴𝑁 is not empty. It follows that the interior of 𝑇 (𝐵𝑋 (0, 1)) is not
empty, hence there are 𝑟 > 0 and 𝑦0 ∈ 𝑌 such that 𝐵𝑌 (𝑦0 , 𝑟) ⊆ 𝑇 (𝐵𝑋 (0, 1)). In particular
𝑦0 ∈ 𝑇 (𝐵𝑋 (0, 1)) which implies that −𝑦0 ∈ 𝑇 (𝐵𝑋 (0, 1)), hence
𝐵𝑌 (0, 𝑟) = −𝑦0 + 𝐵𝑌 (𝑦0 , 𝑟) ⊆ 𝑇 (𝐵𝑋 (0, 1)) + 𝑇 (𝐵𝑋 (0, 1))
⊆ 𝑇 (𝐵𝑋 (0, 2)) = 2𝑇 (𝐵𝑋 (0, 1)),
and thus 𝐵𝑌 (0, 𝑟/2) ⊆ 𝑇 (𝐵𝑋 (0, 1)). Observe that this implies that for any 𝑦 ∈ 𝑌 such that
‖𝑦‖𝑌 < 𝜆 2𝑟 with 𝜆 > 0, we have
∀𝜀 > 0, ∃𝑧 ∈ 𝑋 , ‖𝑧‖𝑋 < 𝜆 and ‖𝑦 − 𝑇 𝑧‖𝑌 < 𝜀.
Now let 𝑦 ∈ 𝑌 with ‖𝑦‖𝑌 < 𝑟/4, then taking 𝜀 = 𝑟/4 we find 𝑧1 ∈ 𝑋 such that
‖𝑧1 ‖𝑋 < 1/2 and ‖𝑦 − 𝑇 𝑧1 ‖𝑌 < 𝑟/4. By applying the same argument to 𝑦 − 𝑇 𝑧1 that
verifies ‖𝑦 − 𝑇 𝑧1 ‖𝑌 < 𝑟/4 and taking 𝜀 = 𝑟/8, we find 𝑧2 ∈ 𝑋 such that ‖𝑧2 ‖𝑋 < 1/4 and
‖𝑦 − 𝑇 𝑧1 − 𝑇 𝑧2 ‖𝑌 < 𝑟/8. By induction we construct a sequence (𝑧𝑛 )𝑛∈N∗ in 𝑋 such that
for any 𝑛 ∈ N∗ we have
1
‖𝑧𝑛 ‖𝑋 < 𝑛 and ‖𝑦 − 𝑇 (𝑧1 + ⋯ + 𝑧𝑛 )‖𝑌 < 𝑟2−𝑛−1 .
2
Therefore the sequence (𝑥𝑛 )𝑛∈N∗ , defined by 𝑥𝑛 = 𝑧1 + ⋯ + 𝑧𝑛 , is a Cauchy sequence
in 𝑋 , which is complete. Hence there is some 𝑥 ∈ 𝑋 such that 𝑥𝑛 → 𝑥, and we have
‖𝑥‖𝑋 < 1 and 𝑦 = 𝑇 𝑥 since 𝑇 is continuous. We have hence proved that for any 𝑦 ∈ 𝑌
with ‖𝑦‖𝑌 < 𝑟/4 there is 𝑥 ∈ 𝑋 with ‖𝑥‖𝑋 < 1 such that 𝑦 = 𝑇 𝑥, which implies that
𝐵𝑌 (0, 𝑟/4) ⊆ 𝑇 (𝐵𝑋 (0, 1)).
We now prove that the above estimate implies that 𝑇 is an open map. Let 𝑉 be open
in 𝑋 . Let 𝑦 ∈ 𝑇 (𝑉 ), then there is 𝑥 ∈ 𝑉 such that 𝑦 = 𝑇 𝑥. Since 𝑉 is open, there is
𝑟 > 0 so that 𝐵𝑋 (𝑥, 𝑟) ⊆ 𝑉 . Hence 𝑥 + 𝐵𝑋 (0, 𝑟) = 𝐵𝑋 (𝑥, 𝑟) ⊆ 𝑉 from which we obtain
𝑦 + 𝑇 (𝐵𝑋 (0, 𝑟)) ⊆ 𝑢(𝑉 ). By hypothesis we have 𝐵𝑌 (0, 𝑟𝑐) ⊆ 𝑇 (𝐵𝑋 (0, 𝑟)), then it follows
𝐵𝑌 (𝑦, 𝑟𝑐) = 𝑦 + 𝐵𝑌 (0, 𝑟𝑐) ⊆ 𝑦 + 𝑇 (𝐵𝑋 (0, 𝑟)) ⊆ 𝑇 (𝑉 ). This proves that 𝑇 (𝑉 ) is open in
𝑌.
We now state a direct consequence of the open mapping theorem.
Corollary 8.24 (Banach’s isomorphism theorem). Let (𝑋 , ‖ ⋅ ‖𝑋 ) and (𝑌 , ‖ ⋅ ‖𝑌 ) be two

Banach spaces, and 𝑇 ∈ ℒ (𝑋 , 𝑌 ) be a bijective continuous linear map from 𝑋 into 𝑌 . Then
𝑇 is an isomorphism.
Proof. The map 𝑇 is bijective, continuous and open by the Open mapping theorem (The-
orem 8.23), hence 𝑇 is a homeomorphism. Therefore 𝑇 −1 is continuous. Since 𝑇 −1 is
also linear, this implies that 𝑇 is an isomorphism.
8.4.3 Closed graph theorem

Theorem 8.25 (Closed graph theorem). Let (𝑋 , ‖ ⋅ ‖𝑋 ) and (𝑌 , ‖ ⋅ ‖𝑌 ) be two Banach spaces
and 𝑇 ∶ 𝑋 → 𝑌 be a linear map. Assume that the graph Γ = {(𝑥, 𝑦) ∈ 𝑋 × 𝑌 ∣ 𝑦 = 𝑇 𝑥} of
𝑇 is closed in 𝑋 × 𝑌 . Then 𝑇 is continuous.
Recall that the converse is true, indeed we already know that the graph of a contin-
uous map (not necessarily linear) is closed.
We remark that saying that the graph Γ of 𝑇 is closed in 𝑋 × 𝑌 is equivalent to: For
every sequence (𝑥𝑛 )𝑛⩾0 in 𝑋 that converges to 𝑥 ∈ 𝑋 and is such that (𝑇 𝑥𝑛 )𝑛⩾0 converges
to 𝑦 ∈ 𝑌 , then one has 𝑦 = 𝑇 𝑥.
Proof of Theorem 8.25. Recall that the product space 𝑋 × 𝑌 is a Banach space. Since 𝑇 is
linear, it follows that its graph Γ is a closed vector subspace of 𝑋 × 𝑌 , thus also a Banach
space.
The projection 𝜋 ∶ (𝑥, 𝑦) ↦ 𝑥 is a continuous linear bijective map from Γ onto 𝑋 ,
hence by Corollary 8.24, the map 𝜋 −1 ∶ 𝑥 ↦ (𝑥, 𝑇 𝑥) from 𝑋 onto Γ is also continuous.
The projection 𝜃 ∶ (𝑥, 𝑦) ↦ 𝑦 from 𝑋 ×𝑌 into 𝑌 is also a continuous linear map, therefore
one deduces that 𝑇 = 𝜃 ∘ 𝜋 −1 is continuous.
8.5 Integration of vector-valued functions

Let [𝑎, 𝑏] be an interval of R, and (𝑌 , ‖ ⋅ ‖𝑌 ) be a Banach space. We denote by ℬ([𝑎, 𝑏]; 𝑌 )
the Banach space of bounded 𝑌 -valued functions defined on [𝑎, 𝑏] endowed with the
norm ‖𝑓 ‖∞ = sup𝑡∈[𝑎,𝑏] ‖𝑓 (𝑡)‖𝑌 , and by 𝒞 ([𝑎, 𝑏]; 𝑌 ) the subspace of continuous functions.
The aim of this section is to define the integral of (a class of) functions 𝑓 ∶ [𝑎, 𝑏] → 𝑌 .
We recall that a finite partition of the interval [𝑎, 𝑏] is a collection 𝜎 = {𝑡0 , … , 𝑡𝑛+1 }, with
𝑛 ∈ N, such that 𝑎 = 𝑡0 < 𝑡1 < ⋯ < 𝑡𝑛 < 𝑡𝑛+1 = 𝑏.
We start by defining the integral of step functions.
Definition 8.26. One says that a map 𝑓 ∶ [𝑎, 𝑏] → 𝑌 is a step function if there is a
partition 𝜎 = {𝑡0 , … , 𝑡𝑛+1 } of [𝑎, 𝑏] and 𝑦0 , … , 𝑦𝑛 ∈ 𝑌 such that, for all 𝑘 = 0, … , 𝑛, one has
𝑓 (𝑡) = 𝑦𝑘 , ∀𝑡 ∈ (𝑡𝑘 , 𝑡𝑘+1 ).
One then defines the integral of 𝑓 on [𝑎, 𝑏] by

𝑏 𝑛
∫ 𝑓 (𝑡) d𝑡 ∶= ∑ 𝑦𝑘 (𝑡𝑘+1 − 𝑡𝑘 ). (8.1)
𝑎 𝑘=0
8.5 Integration of vector-valued functions 53
One can check that this definition does not depend on the choice of partition, that
is, if 𝜎 ′ is another finite partition of [𝑎, 𝑏] such that 𝑓 is constant on the open intervals
of 𝜎 ′ , then the value of (8.1) associated to 𝜎 coincides with the value of (8.1) associated
to 𝜎 ′ .
Denote by 𝑆([𝑎, 𝑏]; 𝑌 ) ⊆ ℬ([𝑎, 𝑏]; 𝑌 ) the set of all step functions from [𝑎, 𝑏] to 𝑌 . The
map
ℐ ∶ 𝑆([𝑎, 𝑏]; 𝑌 ) → 𝑌
𝑏
𝑓 ↦ ∫ 𝑓 (𝑡) d𝑡
𝑎
is linear and continuous. Moreover we easily obtain from the definition the following
properties:
Lemma 8.27. For any step function 𝑓 ∈ 𝑆([𝑎, 𝑏]; 𝑌 ) there holds:
(i) Let 𝑎 < 𝑐 < 𝑏 then
𝑏 𝑐 𝑏
∫ 𝑓 (𝑡) d𝑡 = ∫ 𝑓 (𝑡) d𝑡 + ∫ 𝑓 (𝑡) d𝑡.
𝑎 𝑎 𝑐
(ii) We have
𝑏 𝑏
‖∫ 𝑓 (𝑡) d𝑡‖ ⩽ ∫ ‖𝑓 (𝑡)‖𝑌 d𝑡 ⩽ (𝑏 − 𝑎)‖𝑓 ‖∞ .
𝑎 𝑌 𝑎
We now define the integral of a larger class of functions by extending the continuous
linear map ℐ .
Definition 8.28. One says that a map 𝑓 ∶ [𝑎, 𝑏] → 𝑌 is a regulated function if it belongs
to 𝑆([𝑎, 𝑏]; 𝑌 ), the closure of 𝑆([𝑎, 𝑏]; 𝑌 ) in ℬ([𝑎, 𝑏]; 𝑌 ).
In other words, 𝑓 is a regulated function if it is the uniform limit of a sequence of step

functions. One can check that 𝑓 is a regulated function if and only if 𝑓 admits left-side
limits on [𝑎, 𝑏) and right-side limits on (𝑎, 𝑏]. In particular 𝒞 ([𝑎, 𝑏]; 𝑌 ) ⊆ 𝑆([𝑎, 𝑏]; 𝑌 ).
Since ℐ is a continuous linear map from 𝑆 into 𝑌 , by the bounded linear transfor-
mation theorem (Theorem 8.14) we can extend ℐ uniquely to a linear continuous map
from 𝑆([𝑎, 𝑏]; 𝑌 ) to 𝑌 . In particular if 𝑓 ∈ 𝑆([𝑎, 𝑏]; 𝑋 ) then its integral is defined by
𝑏 𝑏
∫ 𝑓 (𝑡) d𝑡 ∶= 𝑛→∞
lim ∫ 𝑓𝑛 (𝑡) d𝑡,
𝑎 𝑎
where (𝑓𝑛 )𝑛⩾0 is a sequence of step functions that converges uniformly to 𝑓 , and we
recall that, by the proof of Theorem 8.14, this limit is independent of the chosen sequence.
Thanks to this procedure, properties (i) and (ii) above remain true for 𝑓 ∈ 𝑆([𝑎, 𝑏]; 𝑌 ).
Chapter 9
Differentiable maps
Through this chapter, let (𝑋 , ‖ ⋅ ‖𝑋 ) and (𝑌 , ‖ ⋅ ‖𝑌 ) be Banach spaces over K. When there
is no confusion, we omit the subscripts 𝑋 and 𝑌 of the norms.
9.1 Fréchet-differentiable maps

Let 𝑈 ⊆ 𝑋 be an open set.
Definition 9.1. A map 𝑓 ∶ 𝑈 → 𝑌 is Fréchet-differentiable at a point 𝑎 ∈ 𝑈 if there

exists a continuous linear map 𝐿 ∈ ℒ (𝑋 , 𝑌 ) and a neighborhood 𝑉 of 0 in 𝑋 such that
𝑎 + 𝑉 ⊆ 𝑈 and
𝑓 (𝑎 + ℎ) = 𝑓 (𝑎) + 𝐿ℎ + ‖ℎ‖𝜀(ℎ), ∀ℎ ∈ 𝑉 ,
where 𝜀 ∶ 𝑉 → 𝑌 satisfies limℎ→0 𝜀(ℎ) = 0.
Proposition 9.2. The map 𝐿 in Definition 9.1 is unique.
Proof. Assume that there are two maps 𝐿1 , 𝐿2 ∈ ℒ (𝑋 , 𝑌 ) and two neighborhoods 𝑉1 , 𝑉2
of 0 in 𝑋 such that 𝑎 + 𝑉1 ⊆ 𝑈 , 𝑎 + 𝑉2 ⊆ 𝑈 and
𝑓 (𝑎 + ℎ) = 𝑓 (𝑎) + 𝐿1 ℎ + ‖ℎ‖𝜀1 (ℎ), ∀ℎ ∈ 𝑉1 ,

𝑓 (𝑎 + ℎ) = 𝑓 (𝑎) + 𝐿2 ℎ + ‖ℎ‖𝜀2 (ℎ), ∀ℎ ∈ 𝑉2 ,
where 𝜀1 ∶ 𝑉1 → 𝑌 and 𝜀2 ∶ 𝑉2 → 𝑌 are maps satisfying limℎ→0 𝜀1 (ℎ) = 0 and

limℎ→0 𝜀2 (ℎ) = 0. Then 𝑉1 ∩ 𝑉2 is a neighborhood of 0 in 𝑋 and one has
𝐿1 ℎ − 𝐿2 ℎ = ‖ℎ‖(𝜀2 (ℎ) − 𝜀1 (ℎ)), ∀ℎ ∈ 𝑉1 ∩ 𝑉2 .
This implies that, for any 𝜂 > 0, there is 𝛿 > 0 such that for any ℎ ∈ 𝑉1 ∩ 𝑉2 one has
‖ℎ‖ ⩽ 𝛿 ⇒ ‖(𝐿1 − 𝐿2 )(ℎ)‖ ⩽ 𝜂‖ℎ‖.
By homogeneity of the norm, this implies that ‖𝐿1 − 𝐿2 ‖ ⩽ 𝜂, and we conclude that
𝐿1 = 𝐿2 since 𝜂 > 0 is arbitrary.
Definition 9.3. The map 𝐿 in Definition 9.1 is called the Fréchet derivative of 𝑓 at the
point 𝑎, and it is denoted by
d𝑓 (𝑎) or 𝑓 ′ (𝑎).
55
56 Chapter 9. Differentiable maps
Remark 9.4. The differentiability of 𝑓 and the value of 𝑓 ′ (𝑎) only depend on the topolo-
gies of 𝑋 and 𝑌 . In particular, they do not change if we replace the norms of 𝑋 and 𝑌 by
equivalent ones.
Definition 9.5. Consider a map 𝑓 ∶ 𝑈 → 𝑌 .
(i) One says that 𝑓 is Fréchet-differentiable on 𝑈 if 𝑓 is Fréchet-differentiable at every
point of 𝑈 . In that case we define the Fréchet derivative of 𝑓 , denoted by d𝑓 or 𝑓 ′ ,
as the map
𝑓 ′ ∶ 𝑈 → ℒ (𝑋 , 𝑌 )
.
𝑎 ↦ 𝑓 ′ (𝑎)
(ii) One says that 𝑓 is continuously differentiable or of class 𝒞 1 on 𝑈 if 𝑓 is Fréchet-
differentiable on 𝑈 and the Fréchet derivative 𝑓 ′ is continuous.
We denote by 𝒞 1 (𝑈 , 𝑌 ) the vector space of all continuously differentiable maps on
𝑈.
Hereafter, when there is no possible misunderstanding, Fréchet differentiable shall be
referred to simply as differentiable, and Fréchet derivative to derivative.
Example 9.1. A constant map 𝑓 ∶ 𝑈 → 𝑌 is differentiable on 𝑈 and for any 𝑎 ∈ 𝑈 one
has 𝑓 ′ (𝑎) = 0.
Example 9.2. A continuous linear map 𝑓 ∶ 𝑈 → 𝑌 is differentiable on 𝑈 and for any
𝑎 ∈ 𝑈 one has 𝑓 ′ (𝑎) = 𝑓 .
Example 9.3. When 𝑋 = R, then 𝑓 ′ (𝑎) ∈ ℒ (R, 𝑌 ) is identified with the vector 𝜆 =
𝑓 ′ (𝑎)(1) ∈ 𝑌 , through the canonical isomorphism ℒ (R, 𝑌 ) ≃ 𝑌 , where
𝑓 (𝑥) − 𝑓 (𝑎)
𝜆 ∶= lim
𝑥→𝑎 𝑥 −𝑎
is the usual derivative for functions defined on R.
9.1.1 Gâteaux-differentiable maps

Let 𝑈 ⊆ 𝑋 be an open set.
Definition 9.6. Let 𝑓 ∶ 𝑈 → 𝑌 be a map.
(i) One says that 𝑓 admits a directional derivative along ℎ ∈ 𝑋 at a point 𝑎 ∈ 𝑈 if
𝑓 (𝑎 + 𝑡ℎ) − 𝑓 (𝑎)
𝐷ℎ 𝑓 (𝑎) ∶= lim exists in 𝑌 .
𝑡→0 𝑡
(ii) One says that 𝑓 is Gâteaux-differentiable at 𝑎 ∈ 𝑈 if 𝑓 admits a directional deriva-
tive along any vector ℎ ∈ 𝑋 at 𝑎, and if the map 𝐷G 𝑓 (𝑎) ∶ ℎ ↦ 𝐷ℎ 𝑓 (𝑎) from 𝑋
to 𝑌 is linear and continuous. The map 𝐷G 𝑓 (𝑎) ∈ ℒ (𝑋 , 𝑌 ) is called the Gâteaux-
derivative of 𝑓 at the point 𝑎.
(iii) One says that 𝑓 is Gâteaux-differentiable on 𝑈 if 𝑓 is Gâteaux-differentiable at
every point of 𝑈 .
It is easy to verify that if 𝑓 is Fréchet-differentiable at a point 𝑎, then 𝑓 is Gâteaux-
differentiable at 𝑎 and
𝑓 ′ (𝑎)(ℎ) = 𝐷ℎ 𝑓 (𝑎), ∀ℎ ∈ 𝑋 ,
but the converse is however not true. We also remark that Gâteaux-differentiability does
not imply continuity.
9.2 Operations on differentiable maps 57
9.2 Operations on differentiable maps

We prove some basic results regarding operations on differentiable maps. The first one
shows that the derivative is linear.
Proposition 9.7. Let 𝑈 ⊆ 𝑋 be an open set.
(i) If 𝑓 ∶ 𝑈 → 𝑌 is differentiable at 𝑎 ∈ 𝑈 and 𝜆 ∈ K, then the map 𝜆𝑓 ∶ 𝑥 ↦ 𝜆𝑓 (𝑥)
from 𝑈 to 𝑌 is differentiable at 𝑎, with (𝜆𝑓 )′ (𝑎) = 𝜆𝑓 ′ (𝑎).
(ii) If 𝑓 , 𝑔 ∶ 𝑈 → 𝑌 are differentiable at 𝑎 ∈ 𝑈 , then the map 𝑓 + 𝑔 ∶ 𝑥 ↦ 𝑓 (𝑥) + 𝑔(𝑥)
from 𝑈 to 𝑌 is differentiable at 𝑎, with (𝑓 + 𝑔)′ (𝑎) = 𝑓 ′ (𝑎) + 𝑔 ′ (𝑎).
Proof. (i) There is 𝑉 a neighborhood of 0 in 𝑋 such that 𝑎 + 𝑉 ⊆ 𝑈 and
𝑓 (𝑎 + ℎ) = 𝑓 (𝑎) + 𝑓 ′ (𝑎)(ℎ) + ‖ℎ‖𝜀(ℎ) ∀ℎ ∈ 𝑉 ,
where 𝜀 ∶ 𝑉 → 𝑌 satisfies limℎ→0 𝜀(ℎ) = 0. Then for any ℎ ∈ 𝑉 there holds
(𝜆𝑓 )(𝑎 + ℎ) = 𝜆𝑓 (𝑎 + ℎ) = (𝜆𝑓 )(𝑎) + 𝜆𝑓 ′ (𝑎)(ℎ) + ‖ℎ‖𝜆𝜀(ℎ),
and one has limℎ→0 𝜆𝜀(ℎ) = 0. The map 𝑋 ∋ ℎ ↦ 𝜆𝑓 ′ (𝑎)(ℎ) ∈ 𝑌 is clearly linear and
continuous, thus the proof is complete.
(ii) There are 𝑉1 , 𝑉2 neighborhoods of 0 in 𝑋 such that 𝑎 + 𝑉1 ⊆ 𝑈 , 𝑎 + 𝑉2 ⊆ 𝑈 , and
𝑓 (𝑎 + ℎ) = 𝑓 (𝑎) + 𝑓 ′ (𝑎)(ℎ) + ‖ℎ‖𝜀1 (ℎ) ∀ℎ ∈ 𝑉1 ,

𝑔(𝑎 + ℎ) = 𝑔(𝑎) + 𝑔 ′ (𝑎)(ℎ) + ‖ℎ‖𝜀2 (ℎ) ∀ℎ ∈ 𝑉2 ,
where 𝜀1 ∶ 𝑉1 → 𝑌 and 𝜀2 ∶ 𝑉2 → 𝑌 verify limℎ→0 𝜀1 (ℎ) = 0 and limℎ→0 𝜀2 (ℎ) = 0. For

any ℎ ∈ 𝑉1 ∩ 𝑉2 , which is a neighborhood of 0 on 𝑋 , there holds
(𝑓 + 𝑔)(𝑎 + ℎ) = 𝑓 (𝑎 + ℎ) + 𝑔(𝑎 + ℎ)
= (𝑓 + 𝑔)(𝑎) + 𝑓 ′ (𝑎)(ℎ) + 𝑔 ′ (𝑎)(ℎ) + ‖ℎ‖(𝜀1 (ℎ) + 𝜀2 (ℎ)),
and one has limℎ→0 (𝜀1 (ℎ) + 𝜀2 (ℎ)) = 0. The map 𝑋 ∋ ℎ ↦ 𝑓 ′ (𝑎)(ℎ) + 𝑔 ′ (𝑎)(ℎ) ∈ 𝑌 is
clearly linear and continuous, thus the proof is complete.
In the next result we show that the composition of differentiable maps is also differ-
entiable.
Proposition 9.8. Let 𝑋 , 𝑌 , 𝑍 be Banach spaces, 𝑈 an open subset of 𝑋 , and 𝑉 an open
subset of 𝑌 . Let 𝑓 ∶ 𝑈 → 𝑌 and 𝑔 ∶ 𝑉 → 𝑍 be two maps. Assume that 𝑓 is differentiable
at 𝑎 ∈ 𝑈 and 𝑔 is differentiable at 𝑏 = 𝑓 (𝑎) ∈ 𝑉 . Then the composition map 𝑔 ∘ 𝑓 ∶ 𝑈 → 𝑍
is differentiable at 𝑎 and
(𝑔 ∘ 𝑓 )′ (𝑎) = 𝑔 ′ (𝑓 (𝑎)) ∘ 𝑓 ′ (𝑎).
Proof. By hypothesis there are 𝛿, 𝛿 ′ > 0 such that
𝑓 (𝑎 + ℎ) = 𝑓 (𝑎) + 𝑓 ′ (𝑎)(ℎ) + ‖ℎ‖𝜀(ℎ), ∀ℎ ∈ 𝐵𝑋 (0, 𝛿),

𝑔(𝑏 + 𝑘) = 𝑔(𝑏) + 𝑔 ′ (𝑏)(𝑘) + ‖𝑘‖𝜂(𝑘), ∀𝑘 ∈ 𝐵𝑌 (0, 𝛿 ′ ),
with 𝜀 ∶ 𝐵𝑋 (0, 𝛿) → 𝑌 and 𝜂 ∶ 𝐵𝑌 (0, 𝛿 ′ ) → 𝑍 satisfying limℎ→0 𝜀(ℎ) = 0 and lim𝑘→0 𝜂(𝑘) =
0. We set 𝑘 = 𝑓 (𝑎 + ℎ) − 𝑓 (𝑎) = 𝑓 ′ (𝑎)(ℎ) + ‖ℎ‖𝜀(ℎ) so that
‖𝑘‖ ⩽ (‖𝑓 ′ (𝑎)‖ + ‖𝜀(ℎ)‖)‖ℎ‖.

Let 𝛿 > 0 be small enough such that ‖ℎ‖ < 𝛿 implies ‖𝑘‖ < 𝛿 ′ . Hence one obtains
𝑔 ∘ 𝑓 (𝑎 + ℎ) − 𝑔 ∘ 𝑓 (𝑎) = 𝑔(𝑏 + 𝑘) − 𝑔(𝑏)

= 𝑔 ′ (𝑏)(𝑘) + ‖𝑘‖𝜂(𝑘)
= 𝑔 ′ (𝑏) ∘ 𝑓 ′ (𝑎)(ℎ) + ‖ℎ‖𝑔 ′ (𝑏)(𝜀(ℎ)) + ‖𝑓 ′ (𝑎)(ℎ) + ‖ℎ‖𝜀(ℎ)‖ 𝜂 (𝑓 ′ (𝑎)(ℎ) + ‖ℎ‖𝜀(ℎ))
=∶ 𝑔 ′ (𝑏) ∘ 𝑓 ′ (𝑎)(ℎ) + ‖ℎ‖𝜀(ℎ),
and one observes that limℎ→0 𝜀(ℎ) = 0 since
‖𝜀(ℎ)‖ ⩽ ‖𝑔 ′ (𝑏)‖‖𝜀(ℎ)‖ + (‖𝑓 ′ (𝑎)‖ + ‖𝜀(ℎ)‖)‖𝜂(𝑓 ′ (𝑎)(ℎ) + ‖ℎ‖𝜀(ℎ))‖ −−−−→ 0,

ℎ→0
where we have used the composition of limits in last term.

We also obtain below a result concerning the differentiability of the inverse map.
Proposition 9.9. Let 𝑈 be an open set in 𝑋 and 𝑉 be an open set in 𝑌 . Let 𝑓 ∶ 𝑈 → 𝑉
be a homeomorphism and assume that 𝑓 is differentiable at some point 𝑎 ∈ 𝑈 . Then the
inverse map 𝑓 −1 is differentiable at 𝑏 = 𝑓 (𝑎) ∈ 𝑉 if and only if 𝑓 ′ (𝑎) ∈ Isom(𝑋 , 𝑌 ) is a
isomorphism from 𝑋 into 𝑌 , and thus one has
−1
(𝑓 −1 )′ (𝑏) = (𝑓 ′ (𝑎)) .
Proof. Assume that 𝑔 = 𝑓 −1 is differentiable at 𝑏, then thanks to Proposition 9.8 one

obtains
𝑔 ′ (𝑏) ∘ 𝑓 ′ (𝑎) = id𝑋 and 𝑓 ′ (𝑎) ∘ 𝑔 ′ (𝑏) = id𝑌 ,
which implies that 𝑓 ′ (𝑎) is a isomorphism from 𝑋 into 𝑌 with inverse isomorphism given
by 𝑔 ′ (𝑏).
Conversely, assume that 𝑓 ′ (𝑎) ∈ Isom(𝑋 , 𝑌 ), so in particular one has 𝑐 = ‖(𝑓 ′ (𝑎))−1 ‖ >
0. We have
𝑓 (𝑎 + ℎ) − 𝑓 (𝑎) − 𝑓 ′ (𝑎)(ℎ) = ‖ℎ‖𝜀(ℎ) (∗)
with limℎ→0 𝜀(ℎ) = 0, thus
‖ℎ‖ = ‖(𝑓 ′ (𝑎))−1 (𝑓 ′ (𝑎)(ℎ))‖ ⩽ ‖(𝑓 ′ (𝑎))−1 ‖ (‖𝑓 (𝑎 + ℎ) − 𝑓 (𝑎)‖ + ‖ℎ‖𝜀(ℎ)) .

1
Therefore, if ‖ℎ‖ is small enough so that ‖𝜀(ℎ)‖ < 2𝑐
, one has
‖ℎ‖ ⩽ 2𝑐‖𝑓 (𝑎 + ℎ) − 𝑓 (𝑎)‖.
Denote 𝑏 = 𝑓 (𝑎) and 𝑘 = 𝑓 (𝑎 + ℎ) − 𝑓 (𝑎), then applying (𝑓 ′ (𝑎))−1 to (∗) we obtain
(𝑓 ′ (𝑎))−1 (𝑏 + 𝑘) − (𝑓 ′ (𝑎))−1 (𝑏) − (𝑓 ′ (𝑎))−1 (𝑓 ′ (𝑎)(ℎ)) = ‖ℎ‖(𝑓 ′ (𝑎))−1 (𝜀(ℎ)).
Remarking that ℎ = 𝑎 + ℎ − 𝑎 = 𝑓 −1 (𝑏 + 𝑘) − 𝑓 −1 (𝑏), for ℎ small enough we get
𝑓 −1 (𝑏 + 𝑘) − 𝑓 −1 (𝑏) − (𝑓 ′ (𝑎))−1 (𝑘)

= −‖𝑓 −1 (𝑏 + 𝑘) − 𝑓 −1 (𝑏)‖ (𝑓 ′ (𝑎))−1 ∘ 𝜀 (𝑓 −1 (𝑏 + 𝑘) − 𝑓 −1 (𝑏))
=∶ ‖𝑘‖𝜂(𝑘).
It remains to check that lim𝑘→0 𝜂(𝑘) = 0 and the proof will be complete. Since ‖ℎ‖ ⩽ 2𝑐‖𝑘‖
for ℎ small enough, we obtain ‖𝜂(𝑘)‖ ⩽ 2𝑐 2 ‖𝜀(ℎ)‖ and we conclude by composition of
limits since 𝑘 → 0 implies ℎ → 0.
9.3 Maps with values in a product space 59
9.3 Maps with values in a product space

In this part we consider the case in which 𝑌 = 𝑌1 × ⋯ × 𝑌𝑚 is a finite product of Banach
spaces. We endow 𝑌 with the with the norm
‖𝑦‖ = max(‖𝑦1 ‖𝑌1 , … , ‖𝑦𝑚 ‖𝑌𝑚 )
for any 𝑦 = (𝑦1 , … , 𝑦𝑚 ) ∈ 𝑌 , and we recall that (𝑌 , ‖ ⋅ ‖) is a Banach space.

For any 1 ⩽ 𝑖 ⩽ 𝑚 we define the canonical projection
𝑝𝑖 ∶ 𝑌 → 𝑌𝑖
(𝑦1 , … , 𝑦𝑚 ) ↦ 𝑦𝑖
and the canonical injection
𝜄𝑖 ∶ 𝑌 𝑖 → 𝑌
𝑦𝑖 ↦ (0, … , 𝑦𝑖 , … , 0)
where there are 0’s everywhere except at the 𝑖-th coordinate. One easily checks that 𝑝𝑖
and 𝜄𝑖 are continuous linear maps and verify
𝑚
𝑝𝑖 ∘ 𝜄𝑖 = id𝑌𝑖 and ∑ 𝜄𝑖 ∘ 𝑝𝑖 = id𝑌 .
𝑖=1
We deduce the following result:
Proposition 9.10. Let 𝑈 ⊆ 𝑋 be an open set. A continuous map 𝑓 ∶ 𝑈 → 𝑌 is differen-

tiable at a point 𝑎 ∈ 𝑈 if and only if the map 𝑓𝑖 = 𝑝𝑖 ∘ 𝑓 ∶ 𝑈 → 𝑌𝑖 is differentiable at 𝑎 for
any 1 ⩽ 𝑖 ⩽ 𝑚, and then we have
𝑚
𝑓 ′ (𝑎) = ∑ 𝜄𝑖 ∘ 𝑓𝑖′ (𝑎).
𝑖=1
Proof. If 𝑓 is differentiable at 𝑎, then each 𝑓𝑖 = 𝑝𝑖 ∘ 𝑓 is differentiable at 𝑎 as the compo-

sition of differentiable maps (see Proposition 9.8). Moreover we obtain that
𝑓𝑖′ (𝑎) = 𝑝𝑖 ∘ 𝑓 ′ (𝑎) ∈ ℒ (𝑋 , 𝑌𝑖 ).
Conversely, assume that every 𝑓𝑖 = 𝑝𝑖 ∘ 𝑓 is differentiable at 𝑎. From the above

relation we get
𝑚 𝑚
𝑓 = ∑ 𝜄𝑖 ∘ 𝑝 𝑖 ∘ 𝑓 = ∑ 𝜄𝑖 ∘ 𝑓 𝑖 ,
𝑖=1 𝑖=1
thus 𝑓 is differentiable at 𝑎 as the sum of composition of differentiable maps (see Propo-

sitions 9.7 and 9.8) and moreover
𝑚
𝑓 ′ (𝑎) = ∑ 𝜄𝑖 ∘ 𝑓𝑖′ (𝑎).
𝑖=1
9.4 Partial derivatives

We consider now the case where 𝑋 = 𝑋1 × ⋯ × 𝑋𝑛 is a finite product of Banach spaces.
We endow 𝑋 with the norm
‖𝑥‖ = max(‖𝑥1 ‖𝑋1 , … , ‖𝑥𝑛 ‖𝑋𝑛 )
for any 𝑥 = (𝑥1 , … , 𝑥𝑛 ) ∈ 𝑋 , for which 𝑋 is a Banach space.

For each 𝑎 = (𝑎1 , … , 𝑎𝑛 ) ∈ 𝑋 we define the injection, for any 𝑖 ∈ {1, … , 𝑛},
𝜆𝑖 ∶ 𝑋 𝑖 → 𝑋
.
𝑥𝑖 ↦ (𝑎1 , … , 𝑎𝑖−1 , 𝑥𝑖 , 𝑎𝑖+1 , … , 𝑎𝑛 )
Let 𝑓 ∶ 𝑈 → 𝑌 be a map, where 𝑈 is an open subset of 𝑋 . The composition map 𝑓 ∘ 𝜆𝑖 ,
called the 𝑖-th partial map at the point 𝑎, is then defined on 𝜆𝑖−1 (𝑈 ) ⊆ 𝑋𝑖 , more precisely
one has
𝑓 ∘ 𝜆𝑖 ∶ 𝜆𝑖−1 (𝑈 ) → 𝑌
.
𝑥𝑖 ↦ 𝑓 (𝑎1 , … , 𝑎𝑖−1 , 𝑥𝑖 , 𝑎𝑖+1 , … , 𝑎𝑛 )
Definition 9.11. Let 𝑖 ∈ {1, … , 𝑛}. One says that 𝑓 ∶ 𝑈 → 𝑌 is differentiable with respect
to the 𝑖-th variable at the point 𝑎 ∈ 𝑈 if the map 𝑓 ∘ 𝜆𝑖 is differentiable at 𝑎𝑖 . We then
denote this derivative by
𝜕𝑓
𝜕𝑖 𝑓 (𝑎) or 𝜕𝑥𝑖 𝑓 (𝑎) or (𝑎).
𝜕𝑥𝑖
As usual, one says that 𝑓 ∶ 𝑈 → 𝑌 is differentiable with respect to the 𝑖-th variable
on 𝑈 if it is differentiable with respect to the 𝑖-th variable at every point of 𝑈 , and we
denote
𝜕𝑖 𝑓 ∶ 𝑈 → ℒ (𝑋 , 𝑌 )
𝑎 ↦ 𝜕𝑖 𝑓 (𝑎).
Proposition 9.12. If 𝑓 ∶ 𝑈 → 𝑌 is differentiable at 𝑎 ∈ 𝑈 , then 𝑓 is differentiable with
respect any variable at 𝑎 and one has
𝑛
𝑓 ′ (𝑎)(ℎ1 , … ℎ𝑛 ) = ∑ 𝜕𝑖 𝑓 (𝑎)(ℎ𝑖 ), ∀(ℎ1 , … , ℎ𝑛 ) ∈ 𝑋 .
𝑖=1
Remark 9.13. A map that is differentiable with respect to every variable at some point
is not necessarily differentiable at this point.
Proof of Proposition 9.12. Let 𝜄𝑖 denote the canonical injection
𝜄𝑖 ∶ 𝑋𝑖 → 𝑋
𝑥𝑖 ↦ (0, … , 𝑥𝑖 , … , 0)
where there are 0’s everywhere except at the 𝑖-th coordinate, which clearly is a contin-
uous linear map. Remark that
𝜆𝑖 (𝑥𝑖 ) = 𝑎 + 𝜄𝑖 (𝑥𝑖 − 𝑎𝑖 ) and 𝜆𝑖 (𝑎𝑖 ) = 𝑎,
thus 𝜆𝑖′ (𝑥𝑖 ) = 𝜄𝑖 for all 𝑥𝑖 ∈ 𝑋𝑖 . By composition of differentiable maps, we deduce that
𝑓 ∘ 𝜆𝑖 is differentiable at the point 𝑎𝑖 ∈ 𝑋𝑖 and moreover
(𝑓 ∘ 𝜆𝑖 )′ (𝑎𝑖 ) = 𝑓 ′ (𝑎) ∘ 𝜄𝑖 ,
9.4 Partial derivatives 61
in other words 𝜕𝑖 𝑓 (𝑎) = 𝑓 ′ (𝑎)∘𝜄𝑖 . Denoting by 𝑝𝑖 ∶ 𝑋 → 𝑋𝑖 , (𝑥1 , … , 𝑥𝑛 ) ↦ 𝑥𝑖 the canonical

𝑛
projection and using the identity ∑𝑖=1 𝜄𝑖 ∘ 𝑝𝑖 = id𝑋 , we deduce
𝑛
𝑓 ′ (𝑎) = ∑(𝑓 ′ (𝑎) ∘ 𝜄𝑖 ) ∘ 𝑝𝑖
𝑖=1

Chapter 10
Mean-value theorem and applications
10.1 Mean-value theorems

Theorem 10.1. Let 𝑌 be a Banach space and 𝑎, 𝑏 ∈ R with 𝑎 < 𝑏. Consider two maps
𝑓 ∶ [𝑎, 𝑏] → 𝑌 and 𝑔 ∶ [𝑎, 𝑏] → R that are continuous on [𝑎, 𝑏] and differentiable on (𝑎, 𝑏),
and such that ‖𝑓 ′ (𝑡)‖ ⩽ 𝑔 ′ (𝑡) for any 𝑡 ∈ (𝑎, 𝑏). Then
‖𝑓 (𝑏) − 𝑓 (𝑎)‖ ⩽ 𝑔(𝑏) − 𝑔(𝑎).
Proof. We shall prove that for any 𝜀 > 0 and any 𝑡 ∈ [𝑎, 𝑏] one has ‖𝑓 (𝑡) − 𝑓 (𝑎)‖ ⩽
𝑔(𝑡) − 𝑔(𝑎) + 𝜀(𝑡 − 𝑎 + 1), which concludes the proof by taking 𝑡 = 𝑏 and letting 𝜀 → 0.
By way of contradiction, assume that the set
𝐴 = {𝑡 ∈ [𝑎, 𝑏] ∣ ‖𝑓 (𝑡) − 𝑓 (𝑎)‖ > 𝑔(𝑡) − 𝑔(𝑎) + 𝜀(𝑡 − 𝑎 + 1)}
is non-empty. Then 𝐴 is a non-empty open set (by continuity) bounded by below and
such that 𝑎 ∉ 𝐴, thus 𝐴 has an infimum 𝑚 = inf 𝐴 which satisfies 𝑚 ∈ (𝑎, 𝑏].
By definition of differentiability, there is 𝛿 > 0 such that 𝑚 + 𝛿 ⩽ 𝑏 and for any
𝑡 ∈ (𝑚, 𝑚 + 𝛿] one has
‖𝑓 (𝑡) − 𝑓 (𝑚)‖ 𝜀 𝑔(𝑡) − 𝑔(𝑚) 𝜀
‖𝑓 ′ (𝑚)‖ ⩾ − and 𝑔 ′ (𝑚) ⩽ + .
𝑡 −𝑚 2 𝑡 −𝑚 2
Hence one obtains ‖𝑓 (𝑡) − 𝑓 (𝑚)‖ ⩽ 𝑔(𝑡) − 𝑔(𝑚) + 𝜀(𝑡 − 𝑚). Since 𝑚 ∉ 𝐴, one deduces
‖𝑓 (𝑚) − 𝑓 (𝑎)‖ ⩽ 𝑔(𝑚) − 𝑔(𝑎) + 𝜀(𝑚 − 𝑎 + 1) and thus
‖𝑓 (𝑡) − 𝑓 (𝑎)‖ ⩽ ‖𝑓 (𝑡) − 𝑓 (𝑚)‖ + ‖𝑓 (𝑚) − 𝑓 (𝑎)‖ ⩽ 𝑔(𝑡) − 𝑔(𝑎) + 𝜀(𝑡 − 𝑎 + 1).
This is true for any 𝑡 ∈ (𝑚, 𝑚 + 𝛿], which contradicts the fact that 𝑚 is the infimum of
𝐴.
Theorem 10.2. Let 𝑋 , 𝑌 be Banach spaces, 𝑈 ⊆ 𝑋 an open and convex set, and 𝑓 ∶ 𝑈 → 𝑌
differentiable on 𝑈 . Assume that there is 𝑀 > 0 such that ‖𝑓 ′ (𝑥)‖ ⩽ 𝑀 for any 𝑥 ∈ 𝑈 . Then
∀𝑥, 𝑦 ∈ 𝑈 , ‖𝑓 (𝑦) − 𝑓 (𝑥)‖ ⩽ 𝑀‖𝑦 − 𝑥‖.
Proof. Since 𝑈 is convex, for any 𝑥, 𝑦 ∈ 𝑈 and 𝑡 ∈ [0, 1] one has 𝑥 + 𝑡(𝑦 − 𝑥) ∈ 𝑈 .
Consider the map 𝜙 ∶ 𝑡 ↦ 𝑓 (𝑥 + 𝑡(𝑦 − 𝑥)) from [0, 1] into 𝑌 , which is continuous on
[0, 1] (by composition of continuous maps) and differentiable on (0, 1) (by composition
of differentiable maps). Moreover, for any 𝑡 ∈ (0, 1) one has
𝜙 ′ (𝑡) = 𝑓 ′ (𝑥 + 𝑡(𝑦 − 𝑥))(𝑦 − 𝑥),
hence ‖𝜙 ′ (𝑡)‖ ⩽ ‖𝑓 ′ (𝑥 + 𝑡(𝑦 − 𝑥))‖‖𝑦 − 𝑥‖ ⩽ 𝑀‖𝑦 − 𝑥‖. We then conclude thanks to
Theorem 10.1 applied to the maps 𝜙 and 𝑔 ∶ [0, 1] ∋ 𝑡 ↦ 𝑡𝑀‖𝑦 − 𝑥‖ ∈ R.
63
64 Chapter 10. Mean-value theorem and applications
10.2 Relation between partial differentiability and

differentiability
Let 𝑋 = 𝑋1 × ⋯ × 𝑋𝑛 be a finite product of Banach spaces endowed with the norm
‖𝑥‖ = max(‖𝑥1 ‖𝑋1 , … , ‖𝑥𝑛 ‖𝑋𝑛 ) for any 𝑥 = (𝑥1 , … , 𝑥𝑛 ) ∈ 𝑋 , for which 𝑋 is a Banach space.
Let 𝑌 be a Banach space and 𝑈 be an open set in 𝑋 .
Theorem 10.3. A map 𝑓 ∶ 𝑈 → 𝑌 is of class 𝒞 1 on 𝑈 if and only if 𝑓 is differentiable
with respect to every variable on 𝑈 and, for any 𝑖 ∈ {1, … , 𝑛}, the map
𝜕𝑖 𝑓 ∶ 𝑈 → ℒ (𝑋𝑖 , 𝑌 )
𝑥 ↦ 𝜕𝑖 𝑓 (𝑥)
is continuous.
Proof. Assume that 𝑓 is of class 𝒞 1 on 𝑈 , then by Proposition 9.12 𝑓 is differentiable with
respect to any variable on 𝑈 . Moreover 𝜕𝑖 𝑓 is continuous by composition of continuous
maps.
Conversely, assume that 𝑓 is differentiable with respect to every variable on 𝑈 and
𝜕𝑖 𝑓 is continuous for any 𝑖 ∈ {1, … , 𝑛}. Let 𝑎 = (𝑎1 , … , 𝑎𝑛 ) ∈ 𝑈 . We want to prove that for
any 𝜀 > 0, there is 𝛿 > 0 such that for any ℎ ∈ 𝐵𝑋 (0, 𝛿) one has
𝑛
‖𝑓 (𝑎1 + ℎ1 , … , 𝑎𝑛 + ℎ𝑛 ) − 𝑓 (𝑎1 , … , 𝑎𝑛 ) − ∑ 𝜕𝑖 𝑓 (𝑎)(ℎ𝑖 )‖ < 𝜀‖ℎ‖.
𝑖=1
We hence write
𝑛
𝑓 (𝑎1 + ℎ1 , … , 𝑎𝑛 + ℎ𝑛 ) − 𝑓 (𝑎1 , … , 𝑎𝑛 ) − ∑ 𝜕𝑖 𝑓 (𝑎)(ℎ𝑖 )
𝑖=1
= 𝑓 (𝑎1 + ℎ1 , … , 𝑎𝑛 + ℎ𝑛 ) − 𝑓 (𝑎1 , 𝑎2 + ℎ2 , … , 𝑎𝑛 + ℎ𝑛 ) − 𝜕1 𝑓 (𝑎)(ℎ1 )
+ 𝑓 (𝑎1 , 𝑎2 + ℎ2 , … , 𝑎𝑛 + ℎ𝑛 ) − 𝑓 (𝑎1 , 𝑎2 , 𝑎3 + ℎ3 , … , 𝑎𝑛 ) − 𝜕2 𝑓 (𝑎)(ℎ2 )
+ ⋯ +
+ 𝑓 (𝑎1 , … , 𝑎𝑛−1 , 𝑎𝑛 + ℎ𝑛 ) − 𝑓 (𝑎1 , … , 𝑎𝑛−1 , 𝑎𝑛 ) − 𝜕𝑛 𝑓 (𝑎)(ℎ𝑛 )
so that it suffices to prove that for any 𝜀 > 0 there is 𝛿 > 0 such that ‖ℎ‖ < 𝛿 implies
𝜀
‖𝑓 (𝑎1 + ℎ1 , … , 𝑎𝑛 + ℎ𝑛 ) − 𝑓 (𝑎1 , 𝑎2 + ℎ2 , … , 𝑎𝑛 + ℎ𝑛 ) − 𝜕1 𝑓 (𝑎)(ℎ1 )‖ < ‖ℎ1 ‖,
𝑛
𝜀
‖𝑓 (𝑎1 , 𝑎2 + ℎ2 , … , 𝑎𝑛 + ℎ𝑛 ) − 𝑓 (𝑎1 , 𝑎2 , 𝑎3 + ℎ3 , … , 𝑎𝑛 ) − 𝜕2 𝑓 (𝑎)(ℎ2 )‖ < ‖ℎ2 ‖,
𝑛 (⋆)
⋮
𝜀
‖𝑓 (𝑎1 , … , 𝑎𝑛−1 , 𝑎𝑛 + ℎ𝑛 ) − 𝑓 (𝑎1 , … , 𝑎𝑛−1 , 𝑎𝑛 ) − 𝜕𝑛 𝑓 (𝑎)(ℎ𝑛 )‖ < ‖ℎ𝑛 ‖.
𝑛
Let then 𝜀 > 0. We will show that there is 𝛿 > 0 such that ‖ℎ‖ < 𝛿 implies the first
inequality in (⋆). The other inequalities can be obtained in a similar fashion, hence at
the end we can take 𝛿 > 0 to be the smallest one so that all the above inequalities hold.
Define the map 𝑔 ∶ 𝑡 ↦ 𝑓 (𝑎1 + 𝑡, 𝑎2 + ℎ2 , … , 𝑎𝑛 + ℎ𝑛 ) − 𝜕1 𝑓 (𝑎)(𝑡) for 𝑡 ∈ 𝑋1 small
enough. Then 𝑔 is differentiable and
𝑔 ′ (𝑡) = 𝜕1 𝑓 (𝑎1 + 𝑡, 𝑎2 + ℎ2 , … , 𝑎𝑛 + ℎ𝑛 ) − 𝜕1 𝑓 (𝑎1 , … , 𝑎𝑛 ).

10.3 Limits of differentiable maps 65
Moreover 𝜕1 𝑓 is continuous at 𝑎, hence there is 𝛿 > 0 such that for any ‖ℎ‖ < 𝛿 one has
𝜀
‖𝜕1 𝑓 (𝑎 + ℎ) − 𝜕1 𝑓 (𝑎)‖ <
𝑛
and this implies that, for any ‖𝑡‖ ⩽ ‖ℎ1 ‖, there holds
𝜀
‖𝑔 ′ (𝑡)‖ = ‖𝜕1 𝑓 (𝑎1 + 𝑡, 𝑎2 + ℎ2 , … , 𝑎𝑛 + ℎ𝑛 ) − 𝜕1 𝑓 (𝑎1 , … , 𝑎𝑛 )‖ < .
𝑛
By Theorem 10.2 one deduces that ‖𝑔(ℎ1 ) − 𝑔(0)‖ < 𝑛𝜀 ‖ℎ1 ‖ which is exactly the first
inequality in (⋆).
We have then proved that 𝑓 is differentiable at 𝑎, for any 𝑎 ∈ 𝑈 , and that
𝑛
𝑓 ′ (𝑎)(ℎ 1 , … , ℎ𝑛 ) = ∑ 𝜕𝑖 𝑓 (𝑎)(ℎ𝑖 ),
𝑖=1
and moreover 𝑓 ′ (𝑎) is continuous as the sum of composition of continuous maps.
10.3 Limits of differentiable maps

Let 𝑋 and 𝑌 be Banach spaces.
Theorem 10.4. Let 𝑈 ⊆ 𝑋 be open and connected. Consider a sequence 𝑓𝑛 ∶ 𝑈 → 𝑌 of

differentiable maps on 𝑈 and assume that:
(a) there is 𝑥0 ∈ 𝑈 such that (𝑓𝑛 (𝑥0 ))𝑛⩾0 has a limit in 𝑌 ;
(b) for any 𝑎 ∈ 𝑈 there is 𝑟 > 0 such that (𝑓𝑛′ )𝑛⩾0 converges uniformly on 𝐵𝑋 (𝑎, 𝑟) to
some map 𝑔 ∶ 𝑈 → ℒ (𝑋 , 𝑌 ).
Then there holds:
(i) (𝑓𝑛 )𝑛⩾0 converges pointwisely on 𝑈 , and we denote 𝑓 (𝑥) = lim𝑛→∞ 𝑓𝑛 (𝑥) for any
𝑥 ∈ 𝑈;
(ii) for any 𝑎 ∈ 𝑈 there is 𝑟 > 0 such that (𝑓𝑛 )𝑛⩾0 converges uniformly to 𝑓 on 𝐵𝑋 (𝑎, 𝑟);
(iii) 𝑓 is differentiable on 𝑈 and 𝑓 ′ (𝑎) = 𝑔(𝑎) for any 𝑎 ∈ 𝑈 .
Proof. Let 𝑎 ∈ 𝑈 . For any 𝑥 ∈ 𝐵𝑋 (𝑎, 𝑟) and 𝑛, 𝑚 ∈ N there holds, thanks to Theorem 10.2,
‖𝑓𝑛 (𝑥) − 𝑓𝑚 (𝑥) − (𝑓𝑛 (𝑎) − 𝑓𝑚 (𝑎))‖ ⩽ ‖𝑥 − 𝑎‖ sup ‖𝑓𝑛′ (𝑦) − 𝑓𝑚′ (𝑦)‖
𝑦∈𝐵𝑋 (𝑎,𝑟)
(∗)
⩽𝑟 sup ‖𝑓𝑛′ (𝑦) − 𝑓𝑚′ (𝑦)‖.
𝑦∈𝐵𝑋 (𝑎,𝑟)
Therefore, if (𝑓𝑛 (𝑎))𝑛⩾0 converges in 𝑌 it follows that (𝑓𝑛 )𝑛⩾0 is a Cauchy sequence for
the uniform norm on 𝐵𝑋 (𝑎, 𝑟), hence (𝑓𝑛 )𝑛⩾0 converges uniformly on 𝐵𝑋 (𝑎, 𝑟) to some
map 𝑓 which is continuous.
Moreover, let 𝐴 = {𝑥 ∈ 𝑈 ∣ (𝑓𝑛 (𝑥))𝑛⩾0 converges in 𝑌 }, which is non-empty by
hypothesis. One easily obtains that 𝐴 is open and closed, hence 𝐴 = 𝑈 since 𝑈 is
connected, that is, lim𝑛→∞ 𝑓𝑛 (𝑥) = 𝑓 (𝑥) for any 𝑥 ∈ 𝑈 .
It remains to prove that, for any 𝑎 ∈ 𝑈 , 𝑓 is differentiable at 𝑎 and 𝑓 ′ (𝑎) = 𝑔(𝑎). We

write, for any ℎ ∈ 𝑋 small enough and 𝑛 ∈ N,
‖𝑓 (𝑎 + ℎ) − 𝑓 (𝑎) − 𝑔(𝑎)(ℎ)‖ ⩽ ‖𝑓 (𝑎 + ℎ) − 𝑓 (𝑎) − (𝑓𝑛 (𝑎 + ℎ) − 𝑓𝑛 (𝑎))‖

+ ‖𝑓𝑛 (𝑎 + ℎ) − 𝑓𝑛 (𝑎) − 𝑓𝑛′ (𝑎)(ℎ)‖
+ ‖𝑓𝑛′ (𝑎)(ℎ) − 𝑔(𝑎)(ℎ)‖.
Let 𝜀 > 0, then there is 𝑁0 ∈ N such that for any integer 𝑛, 𝑚 ⩾ 𝑁0 one has
𝜀
sup ‖𝑓𝑛′ (𝑦) − 𝑓𝑚′ (𝑦)‖ < ,
𝑦∈𝐵𝑋 (𝑎,𝑟) 3
hence using (∗) one gets, for any ℎ ∈ 𝐵𝑋 (0, 𝑟),

𝜀
‖𝑓𝑛 (𝑎 + ℎ) − 𝑓𝑚 (𝑎 + ℎ) − (𝑓𝑛 (𝑎) − 𝑓𝑚 (𝑎))‖ < ‖ℎ‖,
3
and by letting 𝑚 → ∞ one concludes that
𝜀
‖𝑓 (𝑎 + ℎ) − 𝑓𝑛 (𝑎 + ℎ) − (𝑓 (𝑎) − 𝑓𝑛 (𝑎))‖ ⩽ ‖ℎ‖.
3
Furthermore, there is 𝑁1 ∈ N such that for any integer 𝑛 ⩾ 𝑁1 one has
𝜀
‖𝑓𝑛′ (𝑎) − 𝑔(𝑎)‖ < .
3
Let us fix some 𝑛 ⩾ max(𝑁0 , 𝑁1 ). Since 𝑓𝑛 is differentiable at 𝑎, there is 𝑟 ′ ∈ (0, 𝑟] such
that for any ℎ ∈ 𝐵𝑋 (0, 𝑟 ′ ) there holds
𝜀
‖𝑓𝑛 (𝑎 + ℎ) − 𝑓𝑛 (𝑎 + ℎ) − 𝑓𝑛′ (𝑎)(ℎ)‖ ⩽ ‖ℎ‖.
3
Finally one obtains
‖𝑓 (𝑎 + ℎ) − 𝑓 (𝑎) − 𝑔(𝑎)(ℎ)‖ ⩽ 𝜀‖ℎ‖.
10.4 High-order derivatives

10.4.1 Second-order derivatives
Let 𝑋 , 𝑌 be Banach spaces and 𝑈 be an open subset of 𝑋 .
Definition 10.5. A map 𝑓 ∶ 𝑈 → 𝑌 is twice differentiable at 𝑎 ∈ 𝑈 if 𝑓 is differentiable

on some open neighborhood 𝑉 of 𝑎, and if the map 𝑓 ′ ∶ 𝑉 → ℒ (𝑋 , 𝑌 ) is differentiable
at 𝑎. We then denote
d2 𝑓 (𝑎) or 𝑓 ″ (𝑎) = (𝑓 ′ )′ (𝑎).
Recall from Proposition 8.21 that the spaces ℒ (𝑋 , ℒ (𝑋 , 𝑌 )) and ℒ2 (𝑋 , 𝑋 ; 𝑌 ) are

isometrically isomorphic. We will therefore always identify 𝑓 ″ (𝑎) ∈ ℒ (𝑋 , ℒ (𝑋 , 𝑌 )) to
the corresponding continuous bilinear map in ℒ2 (𝑋 , 𝑋 ; 𝑌 ), more precisely to the map
𝑋 ×𝑋 → 𝑌
(ℎ, 𝑘) ↦ 𝑓 ″ (𝑎)(ℎ)(𝑘).
10.4 High-order derivatives 67

(i) One says that 𝑓 is twice differentiable on 𝑈 if 𝑓 is twice differentiable at every point
of 𝑈 or, equivalently, if 𝑓 is differentiable on 𝑈 and the map 𝑓 ′ ∶ 𝑈 → ℒ (𝑋 , 𝑌 ) is
differentiable on 𝑈 .
(ii) One says that 𝑓 is of class 𝒞 2 on 𝑈 if it is twice differentiable on 𝑈 and the map
𝑓 ″ ∶ 𝑈 → ℒ (𝑋 , ℒ (𝑋 , 𝑌 )) is continuous.
Proposition 10.7. If 𝑓 ∶ 𝑈 → 𝑌 is twice differentiable at 𝑎 ∈ 𝑈 , then 𝑓 ″ (𝑎) is a symmetric
bilinear continuous map, that is, 𝑓 ″ (𝑎)(ℎ)(𝑘) = 𝑓 ″ (𝑎)(𝑘)(ℎ) for any (ℎ, 𝑘) ∈ 𝑋 × 𝑋 .
Proof. Let 𝜀 > 0, then there is 𝑟 > 0 such that 𝐵(𝑎, 𝑟) ⊆ 𝑈 and for any ℎ, 𝑘 ∈ 𝐸 with
‖ℎ‖, ‖𝑘‖ < 𝑟/2, and any 𝑡 ∈ [0, 1], one has
‖𝑓 ′ (𝑎 + 𝑡ℎ + 𝑘) − 𝑓 ′ (𝑎) − 𝑓 ″ (𝑎)(𝑡ℎ + 𝑘)‖ ⩽ 𝜀(‖𝑡ℎ + 𝑘‖) ⩽ 𝜀(‖ℎ‖ + ‖𝑘‖).
Define the map 𝑔 ∶ 𝑡 ↦ 𝑓 ′ (𝑎+𝑡ℎ+𝑘)−𝑓 ′ (𝑎+𝑡ℎ) from [0, 1] into 𝑌 , then 𝑔 is differentiable
(by composition of differentiable maps) and
𝑔 ′ (𝑡) = 𝑓 ′ (𝑎 + 𝑡ℎ + 𝑘)(ℎ) − 𝑓 ′ (𝑎 + 𝑡ℎ)(ℎ)

= (𝑓 ′ (𝑎 + 𝑡ℎ + 𝑘) − 𝑓 ′ (𝑎)) (ℎ) − (𝑓 ′ (𝑎 + 𝑡ℎ) − 𝑓 ′ (𝑎)) (ℎ).
Hence one obtains

‖𝑔 ′ (𝑡) − 𝑓 ″ (𝑎)(𝑘, ℎ)‖ = ‖𝑔 ′ (𝑡) − 𝑓 ″ (𝑎)(𝑡ℎ + 𝑘, ℎ) + 𝑓 ″ (𝑎)(𝑡ℎ, ℎ)‖
⩽ ‖ (𝑓 ′ (𝑎 + 𝑡ℎ + 𝑘) − 𝑓 ′ (𝑎) − 𝑓 ″ (𝑎)(𝑡ℎ + 𝑘)) (ℎ)‖
+ ‖ (𝑓 ′ (𝑎 + 𝑡ℎ) − 𝑓 ′ (𝑎) − 𝑓 ″ (𝑎)(𝑡ℎ)) (ℎ)‖
⩽ 2𝜀(‖ℎ‖ + ‖𝑘‖)‖ℎ‖ ⩽ 2𝜀(‖ℎ‖ + ‖𝑘‖)2 .
Applying Theorem 10.2 to the function 𝑡 ↦ 𝑔(𝑡) − 𝑡𝑓 ″ (𝑎)(𝑘, ℎ) one gets
‖𝑔(1) − 𝑔(0) − 𝑓 ″ (𝑎)(𝑘, ℎ)‖ ⩽ sup ‖𝑔 ′ (𝑡) − 𝑓 ″ (𝑎)(𝑘, ℎ)‖ ⩽ 2𝜀(‖ℎ‖ + ‖𝑘‖)2 ,
0⩽𝑡⩽1
and 𝑆(ℎ, 𝑘) ∶= 𝑔(1) − 𝑔(0) = 𝑓 (𝑎 + ℎ + 𝑘) − 𝑓 (𝑎 + ℎ) − 𝑓 (𝑎 + 𝑘) + 𝑓 (𝑎) is symmetric on

(ℎ, 𝑘). Therefore, for any ℎ, 𝑘 ∈ 𝐵(0, 𝑟/2), it follows that
‖𝑓 ″ (𝑎)(ℎ, 𝑘) − 𝑓 ″ (𝑎)(𝑘, ℎ)‖ ⩽ ‖𝑓 ″ (𝑎)(ℎ, 𝑘) − 𝑆(ℎ, 𝑘)‖ + ‖𝑆(𝑘, ℎ) − 𝑓 ″ (𝑎)(𝑘, ℎ)‖

⩽ 4𝜀(‖ℎ‖ + ‖𝑘‖)2 .
By homogeneity, this last inequality holds for any (ℎ, 𝑘) ∈ 𝑋 × 𝑋 , and one concludes that
𝑓 ″ (𝑎)(ℎ, 𝑘) = 𝑓 ″ (𝑎)(𝑘, ℎ) since 𝜀 > 0 is arbitrary.
10.4.2 Second-order partial derivatives

Let 𝑋 = 𝑋1 × ⋯ × 𝑋𝑛 be a finite product of Banach spaces, 𝑌 be a Banach space, and 𝑈
an open subset of 𝑋 .
Consider a map 𝑓 ∶ 𝑈 → 𝑌 twice differentiable at 𝑎 ∈ 𝑈 . Then there exists some
open neighborhood 𝑉 of 𝑎 such that, for any 𝑥 ∈ 𝑉 , one has
𝑛
𝑓 ′ (𝑥)(ℎ1 , … , ℎ𝑛 ) = ∑ 𝜕𝑗 𝑓 (𝑥)(ℎ𝑗 ), ∀ℎ = (ℎ1 , … , ℎ𝑛 ) ∈ 𝑋 .
𝑗=1
Moreover, one also has

𝑛
𝑓 ″ (𝑎)(𝑘 1 , … , 𝑘𝑛 ) = ∑ 𝜕𝑖 𝑓 ′ (𝑎)(𝑘𝑖 ), ∀𝑘 = (𝑘1 , … , 𝑘𝑛 ) ∈ 𝑋 .
𝑖=1
Combining both identities one finally gets, for any (ℎ, 𝑘) ∈ 𝑋 × 𝑋 ,

𝑛
𝑓 ″ (𝑎)(𝑘1 , … , 𝑘𝑛 )(ℎ1 , … , ℎ𝑛 ) = ∑ 𝜕𝑖 (𝜕𝑗 𝑓 )(𝑎)(𝑘𝑖 )(ℎ𝑗 ),
𝑖,𝑗=1
which relates the second-order derivative of 𝑓 at 𝑎 with the partial derivatives 𝜕𝑖 (𝜕𝑗 𝑓 )(𝑎).
Applying twice (the proof of) Theorem 10.3, one obtains:
Proposition 10.8. A map 𝑓 ∶ 𝑈 → 𝑌 is twice differentiable at 𝑎 ∈ 𝑈 if and only if
(i) the partial derivatives 𝜕𝑗 𝑓 exist on 𝑈 for any 𝑗 = 1, … , 𝑛 and they are continuous on
𝑈;
(ii) for any 𝑗 = 1, … , 𝑛, the partial derivatives 𝜕𝑖 (𝜕𝑗 𝑓 ) exist on 𝑈 for any 𝑖 = 1, … , 𝑛 and
they are continuous at 𝑎.
10.4.3 Derivatives of higher order

Let 𝑋 , 𝑌 be Banach spaces and 𝑈 be an open subset of 𝑌 . We denote 𝑓 (1) = 𝑓 ′ , ℒ1 (𝑋 , 𝑌 ) =
ℒ (𝑋 , 𝑌 ) and ℒ𝑛 (𝑋 , 𝑌 ) = ℒ (𝑋 , ℒ𝑛−1 (𝑋 , 𝑌 )) for any 𝑛 ∈ N∗ .
(i) One says that 𝑓 is 𝑛-times differentiable at 𝑎 ∈ 𝑈 if 𝑓 is (𝑛 − 1)-times differentiable
on some open neighborhood 𝑉 of 𝑎, and if the map 𝑓 (𝑛−1) ∶ 𝑉 → ℒ𝑛−1 (𝑋 , 𝑌 ) is
differentiable at 𝑎. We then denote
d𝑛 𝑓 (𝑎) or 𝑓 (𝑛) (𝑎) = (𝑓 (𝑛−1) )′ (𝑎).
(ii) One says that 𝑓 is 𝑛-times differentiable on 𝑈 if it is 𝑛-times differentiable at every

point of 𝑈 .
(iii) One says that 𝑓 is of class 𝒞 𝑛 on 𝑈 if it is 𝑛-times differentiable on 𝑈 and the map
𝑓 (𝑛) ∶ 𝑈 → ℒ𝑛 (𝑋 , 𝑌 ) is continuous.
(iv) One says that 𝑓 is of class 𝒞 ∞ on 𝑈 if it is of class 𝒞 𝑛 on 𝑈 for all 𝑛 ∈ N.
As for the case of second-order derivatives, we often identify 𝑓 (𝑛) (𝑎) ∈ ℒ𝑛 (𝑋 , 𝑌 ) to
the corresponding element of ℒ𝑛 (𝑋 , … , 𝑋 ; 𝑌 ), the vector space of 𝑛-multilinear contin-
uous maps, through the canonical isomorphism
ℒ𝑛 (𝑋 , 𝑌 ) → ℒ𝑛 (𝑋 , … , 𝑋 ; 𝑌 )
𝑢 ↦ {(ℎ1 , … , ℎ𝑛 ) ↦ 𝑢(ℎ1 )(ℎ2 ) ⋯ (ℎ𝑛 )}.
One also obtains an analogue result of Proposition 10.7, which states that if 𝑓 ∶ 𝑈 →
𝑌 is 𝑛-times differentiable at some point 𝑎 ∈ 𝑈 , then its derivative 𝑓 (𝑛) (𝑎) ∈ ℒ𝑛 (𝑋 , 𝑌 ) is
a 𝑛-multilinear symmetric map, that is, for any (ℎ1 , … , ℎ𝑛 ) ∈ 𝑋 𝑛 and any permutation 𝜎
on {1, … , 𝑛}, one has
𝑓 (𝑛) (𝑎)(ℎ1 , … , ℎ𝑛 ) = 𝑓 (𝑛) (𝑎)(ℎ𝜎 (1) , … , ℎ𝜎 (𝑛) ).

10.5 Taylor formulas 69
10.5 Taylor formulas

Consider a map 𝑣 ∶ 𝐼 → 𝑌 where 𝐼 = (𝑎, 𝑏) is an open interval of R and 𝑌 a Banach
space. If 𝑣 is (𝑛 + 1)-times differentiable on 𝐼 , then one has for any 𝑡 ∈ 𝐼
d (1 − 𝑡)𝑛 (𝑛) (1 − 𝑡)𝑛 (𝑛+1)

[𝑣(𝑡) + (1 − 𝑡)𝑣 ′ (𝑡) + ⋯ + 𝑣 (𝑡)] = 𝑣 (𝑡).
d𝑡 𝑛! 𝑛!
As a consequence on obtains the following particular case of Taylor’s formula:
Proposition 10.10. Let 𝑣 ∶ 𝐼 → 𝑌 be (𝑛 + 1)-times differentiable on 𝐼 ⊇ [0, 1].

(i) Assume that 𝑣 (𝑛+1) is continuous on 𝐼 , then
1
1 ″ 1 (𝑛) (1 − 𝑡)𝑛 (𝑛+1)
𝑣(1) − 𝑣(0) − 𝑣 ′ (0) − 𝑣 (0) − ⋯ − 𝑣 (0) = ∫ 𝑣 d𝑡.
2! 𝑛! 0 𝑛!
(ii) Assume that there is 𝑀 > 0 such that ‖𝑣 (𝑛+1) (𝑡)‖ ⩽ 𝑀 for any 𝑡 ∈ [0, 1], then
1 ″ 1 𝑀
‖𝑣(1) − 𝑣(0) − 𝑣 ′ (0) − 𝑣 (0) − ⋯ − 𝑣 (𝑛) (0)‖ ⩽ .
2! 𝑛! (𝑛 + 1)!
1
Proof. (i) Recall that if 𝑓 ∶ 𝐼 → 𝑌 is of class 𝒞 1 then 𝑓 (1) − 𝑓 (0) = ∫0 𝑓 ′ (𝑡) d𝑡. The result
follows applying this to the function
(1 − 𝑡)2 ″ (1 − 𝑡)𝑛 (𝑛)

𝑓 (𝑡) = 𝑣(𝑡) + (1 − 𝑡)𝑣 ′ (𝑡) + 𝑣 (𝑡) + ⋯ + 𝑣 (𝑡).
2! 𝑛!
(1−𝑡)𝑛+1
(ii) Consider the function 𝑓 (𝑡) defined above, and define 𝑔(𝑡) = −𝑀 (𝑛+1)!
so that 𝑔 ′ (𝑡) =
(1−𝑡)𝑛
𝑀 𝑛!
. One has
(1 − 𝑡)𝑛 (𝑛+1) (1 − 𝑡)𝑛 (𝑛+1) (1 − 𝑡)𝑛

‖𝑓 ′ (𝑡)‖ = ‖ 𝑣 (𝑡)‖ ⩽ ‖𝑣 (𝑡)‖ ⩽ 𝑀 .
𝑛! 𝑛! 𝑛!
Thanks to Theorem 10.1 one deduces ‖𝑓 (1) − 𝑓 (0)‖ ⩽ 𝑔(1) − 𝑔(0), which concludes the
proof.
Consider now 𝑋 and 𝑌 Banach spaces, 𝑈 an open set in 𝑋 , and a map 𝑓 ∶ 𝑈 → 𝑌 .

Let 𝑎, 𝑎 + ℎ ∈ 𝑈 be points such that the segment [𝑎, 𝑎 + ℎ] = {𝑎 + 𝑡ℎ ∶ 𝑡 ∈ [0, 1]} is included
in 𝑈 . Define the function 𝑣(𝑡) = 𝑓 (𝑎 + 𝑡ℎ) for 𝑡 ∈ [0, 1] ⊆ 𝐼 (where 𝐼 in an open interval
of R). If 𝑓 is (𝑛 + 1)-times differentiable, then 𝑣 is also (𝑛 + 1)-times differentiable by
composition and, moreover, one has for any 𝑡 ∈ [0, 1]
𝑣 ′ (𝑡) = 𝑓 ′ (𝑎 + 𝑡ℎ)(ℎ)
𝑣 ″ (𝑡) = 𝑓 ″ (𝑎 + 𝑡ℎ)(ℎ, ℎ)
⋮
𝑣 (𝑛) (𝑡) = 𝑓 (𝑛) (𝑎 + 𝑡ℎ)(ℎ, … , ℎ).
Therefore, applying Proposition 10.10 we obtain the following two Taylor formulas.
Theorem 10.11 (Taylor formula with integral remainder). Let 𝑓 ∶ 𝑈 → 𝑌 be a map of

class 𝒞 𝑛+1 on 𝑈 . If [𝑎, 𝑎 + ℎ] ⊆ 𝑈 then
1 ″
𝑓 (𝑎 + ℎ) = 𝑓 (𝑎) + 𝑓 ′ (𝑎)(ℎ) +
𝑓 (𝑎)(ℎ, ℎ)
2!
1
1 (𝑛) (1 − 𝑡)𝑛 (𝑛+1)
+ ⋯ + 𝑓 (𝑎)(ℎ, … , ℎ) + ∫ 𝑓 (𝑎 + 𝑡ℎ)(ℎ, … , ℎ) d𝑡.
𝑛! 0 𝑛!
Theorem 10.12 (Taylor formula with Lagrange remainder). Let 𝑓 ∶ 𝑈 → 𝑌 be (𝑛 + 1)-
times differentiable on 𝑈 and assume that there is 𝑀 > 0 such that ‖𝑓 (𝑛+1) (𝑥)‖ ⩽ 𝑀 for
any 𝑥 ∈ 𝑈 . Then
1 ″ 1
‖𝑓 (𝑎 + ℎ) − 𝑓 (𝑎) − 𝑓 ′ (𝑎)(ℎ) − 𝑓 (𝑎)(ℎ, ℎ) − ⋯ − 𝑓 (𝑛) (𝑎)(ℎ, … , ℎ)‖
2! 𝑛!
𝑀
⩽ ‖ℎ‖𝑛+1 .
(𝑛 + 1)!
One can actually obtain a third formula that is similar to that obtained in Theo-
rem 10.12, but under weaker hypothesis.
Theorem 10.13. Let 𝑓 ∶ 𝑈 → 𝑌 be (𝑛 − 1)-times differentiable on 𝑈 , and assume that 𝑓

is 𝑛-times differentiable at 𝑎 ∈ 𝑈 . Then for any 𝜀 > 0, there is 𝛿 > 0 such that for any ℎ ∈ 𝐸,
‖ℎ‖ < 𝛿 implies
1 ″ 1
‖𝑓 (𝑎 + ℎ) − 𝑓 (𝑎) − 𝑓 ′ (𝑎)(ℎ) − 𝑓 (𝑎)(ℎ, ℎ) − ⋯ − 𝑓 (𝑛) (𝑎)(ℎ, … , ℎ)‖ < 𝜀‖ℎ‖𝑛 .
2! 𝑛!
Proof. The case 𝑛 = 1 follows from the definition of differentiability. We shall prove the
general case by induction.
Assume that the result holds for some 𝑛 − 1 with 𝑛 ⩾ 2. Define the map
1 ″ 1
𝜑(ℎ) = 𝑓 (𝑎 + ℎ) − 𝑓 (𝑎) − 𝑓 ′ (𝑎)(ℎ) − 𝑓 (𝑎)(ℎ, ℎ) − ⋯ − 𝑓 (𝑛) (𝑎)(ℎ, … , ℎ)
2! 𝑛!
which is differentiable. For any ℎ, 𝑘 ∈ 𝑋 one has
1 ″
𝜑 ′ (ℎ)(𝑘) = 𝑓 ′ (𝑎 + ℎ)(𝑘) − 𝑓 ′ (𝑎)(𝑘) − {𝑓 (𝑎)(ℎ, 𝑘) + 𝑓 ″ (𝑎)(𝑘, ℎ)} − ⋯
2!
1 (𝑛)
− {𝑓 (𝑎)(ℎ, … , ℎ, 𝑘) + 𝑓 (𝑛) (𝑎)(ℎ, … , ℎ, 𝑘, ℎ) + ⋯ + 𝑓 (𝑛) (𝑎)(𝑘, ℎ, … , ℎ)} .
𝑛!
Using the symmetry of the derivatives 𝑓 (𝑘) (𝑎) and denoting (ℎ)𝑘 = (ℎ, … , ℎ) ∈ 𝐸 𝑘 , one
gets
1 (3) 2 1
𝜑 ′ (ℎ) = 𝑓 ′ (𝑎 + ℎ) − 𝑓 ′ (𝑎) − 𝑓 ″ (𝑎)(ℎ) − 𝑓 (ℎ) − ⋯ − 𝑓 (𝑛) (𝑎)(ℎ)𝑛−1 .
2! (𝑛 − 1)!
Therefore, by the induction hypothesis applied to 𝑓 ′ , one obtains that for any 𝜀 > 0,
there is 𝛿 > 0 such that for any ℎ ∈ 𝐸, ‖ℎ‖ ⩽ 𝛿 implies ‖𝜑 ′ (ℎ)‖ ⩽ 𝜀‖ℎ‖𝑛−1 . Hence, thanks
to Theorem 10.2, on deduces that ‖𝜑(1) − 𝜑(0)‖ ⩽ 𝜀‖ℎ‖𝑛 for any ‖ℎ‖ ⩽ 𝛿, which concludes
the proof since 𝜑(0) = 0.
Chapter 11
Inverse function and implicit

function theorems
11.1 Diffeomorphisms
Let 𝑋 and 𝑌 be Banach spaces.
Definition 11.1. Let 𝑉 be an open subset of 𝑋 and 𝑊 an open subset of 𝑌 . A map

𝑓 ∶ 𝑉 → 𝑊 is a 𝒞 1 -diffeomorphism if it is bijective, of class 𝒞 1 on 𝑉 , and its inverse
map 𝑓 −1 ∶ 𝑊 → 𝑉 is of class 𝒞 1 on 𝑊 .
Remark that a map of class 𝒞 1 can be a homeomorphism without being a 𝒞 1 -diffeomorphism,

in other words the inverse homeomorphism need not be of class 𝒞 1 .
Proposition 11.2. Let 𝑉 be an open subset of 𝑋 and 𝑊 an open subset of 𝑌 . Let 𝑓 ∶ 𝑉 → 𝑊

be a homeomorphism of class 𝒞 1 . Then 𝑓 is a 𝒞 1 -diffeomorphism if and only if 𝑓 ′ (𝑎) is
an isomorphism from 𝑋 onto 𝑌 for all 𝑎 ∈ 𝑉 .
Proof. The “only if” part is trivial. Conversely, let 𝑓 ∶ 𝑉 → 𝑊 be a homeomorphism of

class 𝒞 1 and assume that 𝑓 ′ (𝑎) is an isomorphism from 𝑋 into 𝑌 for all 𝑎 ∈ 𝑉 . Thanks
to Proposition 9.9 the inverse map 𝑔 = 𝑓 −1 ∶ 𝑊 → 𝑉 is differentiable on 𝑊 with
𝑔 ′ (𝑏) = (𝑓 ′ (𝑔(𝑏)))−1 , ∀𝑏 ∈ 𝑊 .
It remains to check that 𝑔 is of class 𝒞 1 on 𝑊 , that is, the map 𝑔 ′ ∶ 𝑊 → ℒ (𝑌 , 𝑋 ) is

continuous. But the map 𝑔 can be written as the composition of three continuous maps,
more precisely, 𝑔 ′ = inv ∘ 𝑓 ′ ∘ 𝑔, where, denoting by Isom(𝑋 , 𝑌 ) the set of isomorphisms
from 𝑋 into 𝑌 , we denote
inv ∶ Isom(𝑋 , 𝑌 ) → Isom(𝑌 , 𝑋 )

𝑢 ↦ 𝑢 −1
which is continuous thanks to Proposition 8.17. Therefore 𝑔 is of class 𝒞 1 on 𝑊 , and

thus 𝑓 ∶ 𝑉 → 𝑊 is a 𝒞 1 -diffeomorphism.
11.2 Inverse function theorem

Let 𝑋 and 𝑌 be Banach spaces, and 𝑈 be an open subset of 𝑋 .
71
72 Chapter 11. Inverse function and implicit function theorems
Theorem 11.3 (Inverse function theorem). Let 𝑓 ∶ 𝑈 → 𝑌 be a map of class 𝒞 1 on 𝑈

and assume that there is 𝑎 ∈ 𝑈 such that 𝑓 ′ (𝑎) is an isomorphism from 𝑋 into 𝑌 . Then
there is an open neighborhood 𝑉 ⊆ 𝑈 of 𝑎 in 𝑋 and an open neighborhood 𝑊 of 𝑓 (𝑎) in 𝑌
such that 𝑓 is a 𝒞 1 -diffeomorphism from 𝑉 onto 𝑊 .
Proof. Consider the translation maps 𝜏 ∶ 𝑥 ↦ 𝑥 + 𝑎 from 𝑋 into 𝑋 and 𝜏 ̄ ∶ 𝑦 ↦ 𝑦 − 𝑓 (𝑎)
from 𝑌 into 𝑌 , and the set 𝑈𝑎 ∶= 𝑈 − 𝑎 = {𝑥 − 𝑎 ∶ 𝑥 ∈ 𝑈 } which is open in 𝑋 and contains
0. Define the map
𝑔 ∶ 𝑈𝑎 → 𝑋
𝑥 ↦ (𝑓 ′ (𝑎))−1 (𝑓 (𝑥 + 𝑎) − 𝑓 (𝑎))
in other words 𝑔 can be written as the following composition
𝑔 = (𝑓 ′ (𝑎))−1 ∘ 𝜏 ̄ ∘ 𝑓 ∘ 𝜏 .
The map 𝑔 is of class 𝒞 1 on 𝑈𝑎 with 𝑔 ′ (𝑥) = (𝑓 ′ (𝑎))−1 ∘ 𝑓 ′ (𝑥 + 𝑎), 𝑔(0) = 0 and
𝑔 ′ (0) = id𝑋 ∈ Isom(𝑋 , 𝑋 ). We will prove that 𝑔 is a 𝒞 1 -diffeomorphism from some
open neighborhood of 0 onto some open neighborhood of 0, which will imply the result
for 𝑓 since 𝑓 will be the composition of 𝒞 1 -diffeomorphisms, more precisely one has
𝑓 = (𝜏 ̄)−1 ∘ 𝑓 ′ (𝑎) ∘ 𝑔 ∘ 𝜏 −1 .
Define the map 𝜙 = 𝑔 − id𝑋 from 𝑈𝑎 into 𝑋 . One has 𝜙(0) = 0 and 𝜙 is of class 𝒞 1
on 𝑈𝑎 . Since 𝜙 ′ (𝑥) = 𝑔 ′ (𝑥) − id𝑋 = 𝑔 ′ (𝑥) − 𝑔 ′ (0), Isom(𝑋 , 𝑋 ) is open in ℒ (𝑋 , 𝑋 ) and 𝑔 ′
is continuous, there there is 𝑟 > 0 such that
1
𝑔 ′ (𝑥) ∈ Isom(𝑋 , 𝑋 ) and ‖𝜙 ′ (𝑥)‖ ⩽ , ∀𝑥 ∈ 𝐵𝑋 (0, 𝑟) ⊆ 𝑈𝑎 . (∗)
2
Applying Theorem 10.2 one deduces
1
‖𝜙(𝑥) − 𝜙(𝑦)‖ ⩽ ‖𝑥 − 𝑦‖ ∀𝑥, 𝑦 ∈ 𝐵𝑋 (0, 𝑟),
2
and in particular 𝜙(𝐵𝑋 (0, 𝑟)) ⊆ 𝐵𝑋 (0, 𝑟/2).
We now check that 𝑔 is injective on 𝐵𝑋 (0, 𝑟). Let 𝑥, 𝑦 ∈ 𝐵𝑋 (0, 𝑟) be such that 𝑔(𝑥) =
𝑔(𝑦), then one gets ‖𝑥 − 𝑦‖ = ‖𝜙(𝑥) − 𝜙(𝑦)‖ ⩽ 12 ‖𝑥 − 𝑦‖, thus 𝑥 = 𝑦.
Define the open set 𝑊 ′ = 𝐵𝑋 (0, 𝑟/2) and the closed set
𝑍 = {ℎ ∈ 𝒞 (𝑊 ′ , 𝑋 ) ∣ ℎ(𝑊 ′ ) ⊆ 𝐵𝑋 (0, 𝑟)} ,
thus 𝑍 is a complete metric space endowed with the supremum norm, since it is a closed
subspace of (𝒞 (𝑊 ′ , 𝑋 ), ‖ ⋅ ‖∞ ). Observe that the identity map id𝑊 ′ ∶ 𝑊 ′ → 𝑋 belongs
to 𝑍 and define the map
𝑇 ∶ 𝑍 → 𝒞 (𝑊 ′ , 𝑋 )
ℎ ↦ id𝑊 ′ − 𝜙 ∘ ℎ.
One easily verifies that 𝑇 (𝑍 ) ⊆ 𝑍 and that 𝑇 is 21 -Lipschitz, therefore by the Banach-
Picard fixed-point theorem (Theorem 4.10) the map 𝑇 has a unique fixed point, that is,
there is a unique ℎ ∈ 𝑍 such that id𝑊 ′ − 𝜙 ∘ ℎ = ℎ, thus 𝑔 ∘ ℎ = id𝑊 ′ .
Set 𝑉 ′ ∶= 𝑔 −1 (𝑊 ′ ) ∩ 𝐵𝑋 (0, 𝑟) which is an open set in 𝑋 . Then 𝑔 ∶ 𝑉 ′ → 𝑊 ′ is
continuous, injective since 𝑉 ′ ⊆ 𝐵𝑋 (0, 𝑟), and surjective since 𝑔 ∘ℎ = id𝑊 ′ . Moreover the
inverse map of 𝑔 is ℎ = 𝑔 −1 which is continuous. Hence 𝑔 is a homeomorphism of class
𝒞 1 from 𝑉 ′ onto 𝑊 ′ and 𝑔 ′ (𝑥) ∈ Isom(𝑋 , 𝑋 ) for any 𝑥 ∈ 𝑉 ′ ⊆ 𝐵𝑋 (0, 𝑟) from (∗), therefore
we deduce from Proposition 11.2 that 𝑔 ∶ 𝑉 ′ → 𝑊 ′ is a 𝒞 1 -diffeomorphism.
11.3 Implicit function theorem 73
11.3 Implicit function theorem

Let 𝑋 , 𝑌 , 𝑍 be Banach spaces and 𝑈 ⊆ 𝑋 × 𝑌 be open.
Theorem 11.4 (Implicit function theorem). Consider a map 𝑓 ∶ 𝑈 → 𝑍 that is of class
𝒞 1 on 𝑈 . Let (𝑎, 𝑏) ∈ 𝑈 be such that 𝑓 (𝑎, 𝑏) = 0, and assume that 𝜕2 𝑓 (𝑎, 𝑏) ∈ ℒ (𝑌 , 𝑍 ) is
an isomorphism from 𝑌 onto 𝑍 .
Then there are an open neighborhood 𝑉 ⊆ 𝑈 of (𝑎, 𝑏), an open neighborhood 𝑊 of 𝑎,
and a map 𝑔 ∶ 𝑊 → 𝑌 of class 𝒞 1 on 𝑊 such that
(𝑥, 𝑦) ∈ 𝑉 and 𝑓 (𝑥, 𝑦) = 0 ⟺ 𝑥∈𝑊 and 𝑔(𝑥) = 𝑦.
In particular 𝑔(𝑎) = 𝑏 and
𝑔 ′ (𝑎) = −(𝜕2 𝑓 (𝑎, 𝑏))−1 ∘ 𝜕1 𝑓 (𝑎, 𝑏).
Proof. Define the map
𝜑∶ 𝑈 → 𝑋 ×𝑍
(𝑥, 𝑦) ↦ (𝑥, 𝑓 (𝑥, 𝑦))
which is of class 𝒞 1 on 𝑈 since each component is 𝒞 1 on 𝑈 . Therefore one has
𝜑 ′ (𝑎, 𝑏) ∶ 𝑋 × 𝑌 → 𝑋 × 𝑍
(ℎ, 𝑘) ↦ (ℎ, 𝜕1 𝑓 (𝑎, 𝑏)(ℎ) + 𝜕2 𝑓 (𝑎, 𝑏)(𝑘)) .
Since 𝜕2 𝑓 (𝑎, 𝑏) is an isomorphism from 𝑌 onto 𝑍 , one gets that 𝜑 ′ (𝑎, 𝑏) is an isomorphism
from 𝑋 × 𝑌 onto 𝑋 × 𝑍 with inverse map given by
𝑋 ×𝑍 → 𝑋 ×𝑌
(ℎ′ , 𝑘 ′ ) ↦ (ℎ′ , (𝜕2 𝑓 (𝑎, 𝑏))−1 (𝑘 ′ ) − (𝜕2 𝑓 (𝑎, 𝑏))−1 ∘ 𝜕1 𝑓 (𝑎, 𝑏)(ℎ′ )) .
Applying Theorem 11.3 to 𝜑, one gets that there is an open neighborhood 𝑉 ⊆ 𝑈 of (𝑎, 𝑏)
in 𝑋 × 𝑌 and an open neighborhood 𝑊1 of 𝜑(𝑎, 𝑏) = (𝑎, 0) in 𝑋 × 𝑍 such that 𝜑 ∶ 𝑉 → 𝑊1
is a 𝒞 1 -diffeomorphism from 𝑉 to 𝑊1 .
The inverse map of 𝜑, that we denote by 𝜓 = 𝜑 −1 , is of the form
𝜓 (𝑥, 𝑧) = (𝑥, 𝑔(𝑥,
̄ 𝑧)) ∀(𝑥, 𝑧) ∈ 𝑊1 ,
for some function 𝑔̄ ∶ 𝑊1 → 𝑌 of class 𝒞 1 on 𝑊1 . Therefore one has
(𝑥, 𝑦) ∈ 𝑉 and 𝑓 (𝑥, 𝑦) = 𝑧 ⟺ (𝑥, 𝑧) ∈ 𝑊1 and 𝑔(𝑥,
̄ 𝑧) = 𝑦.
Taking 𝑧 = 0 above one gets
(𝑥, 𝑦) ∈ 𝑉 and 𝑓 (𝑥, 𝑦) = 0 ⟺ (𝑥, 0) ∈ 𝑊1 and 𝑔(𝑥,
̄ 0) = 𝑦.
Define 𝑊 = {𝑥 ∈ 𝐸 ∣ (𝑥, 0) ∈ 𝑊1 } which is open in 𝑋 and contains 𝑎. Moreover, define
̄ 0) from 𝑊 to 𝑌 so that 𝑔 is of class 𝒞 1 on 𝑊 . One finally obtains
the function 𝑔(𝑥) = 𝑔(𝑥,
(𝑥, 𝑦) ∈ 𝑉 and 𝑓 (𝑥, 𝑦) = 0 ⇔ 𝑥∈𝑊 and 𝑔(𝑥) = 𝑦.
Let us now check the last statement. The map 𝑢 ∶ 𝑥 ↦ 𝑓 (𝑥, 𝑔(𝑥)) is constant on 𝑊 ,
thus 𝑢 ′ (𝑥) = 0 for any 𝑥 ∈ 𝑊 , which implies that
𝜕1 𝑓 (𝑥, 𝑔(𝑥)) + 𝜕2 𝑓 (𝑥, 𝑔(𝑥)) ∘ 𝑔 ′ (𝑥) = 0 ∀𝑥 ∈ 𝑊 .
We complete the proof by taking 𝑥 = 𝑎.
Chapter 12
Optimization
In this chapter we shall consider maps 𝑓 ∶ 𝑈 → R with 𝑈 an open subset of a Banach

space 𝑋 .
12.1 Extremum points

Definition 12.1. Let 𝑓 ∶ 𝑈 → R be a map and 𝑎 ∈ 𝑈 . One says that:
(i) 𝑓 has a (global) minimum at 𝑎 if 𝑓 (𝑎) ⩽ 𝑓 (𝑥) for any 𝑥 ∈ 𝑈 ;
(ii) 𝑓 has a strict (global) minimum at 𝑎 if 𝑓 (𝑎) < 𝑓 (𝑥) for any 𝑥 ∈ 𝑈 with 𝑥 ≠ 𝑎;
(iii) 𝑓 has a local minimum at 𝑎 if there is some neighborhood 𝑉 of 𝑎 such that 𝑓 (𝑎) ⩽
𝑓 (𝑥) for any 𝑥 ∈ 𝑉 ;
(iv) 𝑓 has a strict local minimum at 𝑎 if there is some neighborhood 𝑉 of 𝑎 such that
𝑓 (𝑎) < 𝑓 (𝑥) for any 𝑥 ∈ 𝑉 with 𝑥 ≠ 𝑎;
(v) 𝑓 has a (global) maximum at 𝑎 if 𝑓 (𝑎) ⩾ 𝑓 (𝑥) for any 𝑥 ∈ 𝑈 ;
(vi) 𝑓 has a strict (global) maximum at 𝑎 if 𝑓 (𝑎) > 𝑓 (𝑥) for any 𝑥 ∈ 𝑈 with 𝑥 ≠ 𝑎;
(vii) 𝑓 has a local maximum at 𝑎 if there is some neighborhood 𝑉 of 𝑎 such that 𝑓 (𝑎) ⩾
𝑓 (𝑥) for any 𝑥 ∈ 𝑉 ;
(viii) 𝑓 has a strict local maximum at 𝑎 if there is some neighborhood 𝑉 of 𝑎 such that
𝑓 (𝑎) > 𝑓 (𝑥) for any 𝑥 ∈ 𝑉 with 𝑥 ≠ 𝑎.
One says that 𝑓 has an (strict local/local/strict) extremum at 𝑎 if 𝑓 has a (strict lo-
cal/local/strict) minimum or maximum at 𝑎. Moreover, if 𝑉 is a subset of 𝑈 , one says that
𝑓 has an extremum on 𝑉 if the restriction map 𝑓|𝑉 ∶ 𝑉 → R of 𝑓 to 𝑉 has an extremum.
12.2 First-order conditions

Our first result is a necessary condition for a differentiable function to have a extremum
at some point.
Proposition 12.2. Let 𝑓 ∶ 𝑈 → R be differentiable at 𝑎 ∈ 𝑈 . If 𝑓 has a local extremum

at 𝑎 then 𝑓 ′ (𝑎) = 0.
Proof. Assume that 𝑓 has a local minimum at 𝑎 (the proof of the other case being similar).
Then there is some neighborhood 𝑉1 of 0 in 𝑋 such that 𝑓 (𝑎 +ℎ)−𝑓 (𝑎) ⩾ 0 for any ℎ ∈ 𝑉1 .
75
76 Chapter 12. Optimization
Moreover, since 𝑓 is differentiable at 𝑎, there is a neighborhood 𝑉2 of 0 in 𝑋 such that

𝑓 (𝑎 + ℎ) − 𝑓 (𝑎) = 𝑓 ′ (𝑎)(ℎ) + ‖ℎ‖𝜀(ℎ), where 𝜀 ∶ 𝑉2 → R satisfies limℎ→0 𝜀(ℎ) = 0.
For any ℎ ∈ 𝑋 and 𝑡 > 0 small enough so that 𝑡ℎ ∈ 𝑉 ∶= 𝑉1 ∩ 𝑉2 , one obtains
𝑓 (𝑎 + 𝑡ℎ) − 𝑓 (𝑎) = 𝑓 ′ (𝑎)(𝑡ℎ) + ‖𝑡ℎ‖𝜀(𝑡ℎ) ⩾ 0
hence
𝑓 ′ (𝑎)(ℎ) + ‖ℎ‖𝜀(𝑡ℎ) ⩾ 0,
and taking the limit 𝑡 → 0 one gets 𝑓 ′ (𝑎)(ℎ) ⩾ 0. Applying the same argument with −ℎ
instead of ℎ one obtains 𝑓 ′ (𝑎)(ℎ) ⩽ 0, which concludes the proof.
Definition 12.3. Let 𝑓 ∶ 𝑈 → R be a differentiable function. We call critical points (or

stationary points) of 𝑓 all points 𝑥 ∈ 𝑈 such that 𝑓 ′ (𝑥) = 0.
Theorem 12.4. Let 𝑓 ∶ 𝑈 → R be a differentiable map. Assume that 𝑈 is a convex open

subset of 𝑋 , and that 𝑓 is a convex map. Let 𝑎 ∈ 𝑈 be such that 𝑓 ′ (𝑎) = 0. Then 𝑓 has a
minimum at 𝑎.
Proof. By way of contradiction, assume that there is some 𝑏 ∈ 𝑈 ⧵ {𝑎} such that 𝑓 (𝑏) <
𝑓 (𝑎). Since 𝑈 is convex one has 𝑡𝑏 + (1 − 𝑡)𝑎 ∈ 𝑈 for all 𝑡 ∈ [0, 1] and
𝑓 (𝑡𝑏 + (1 − 𝑡)𝑎) ⩽ 𝑡𝑓 (𝑏) + (1 − 𝑡)𝑓 (𝑎) = 𝑓 (𝑎) + 𝑡(𝑓 (𝑏) − 𝑓 (𝑎)).
Define the function 𝑔 ∶ 𝑡 ↦ 𝑓 (𝑡𝑏 + (1 − 𝑡)𝑎) from 𝐽 into R, where 𝐽 is some open interval
containing [0, 1]. It is clear that 𝑔 is differentiable and
𝑔(𝑡) ⩽ 𝑔(0) + 𝑡(𝑓 (𝑏) − 𝑓 (𝑎)).
On the one hand, one computes
𝑔(𝑡) − 𝑔(0)
𝑔 ′ (0) = lim ⩽ 𝑓 (𝑏) − 𝑓 (𝑎) < 0.
𝑡→0, 𝑡>0 𝑡
On the other hand one has 𝑔 ′ (0) = 𝑓 ′ (𝑎)(𝑏 − 𝑎), thus 𝑓 ′ (𝑎)(𝑏 − 𝑎) < 0, which is a
contradiction with 𝑓 ′ (𝑎) = 0.
12.3 Optimization with constraints

Theorem 12.5 (Lagrange multipliers’ theorem). Let 𝑓 ∶ 𝑈 → R be a map of class 𝒞 1
on 𝑈 . Consider a constraint function 𝜑 = (𝜑1 , … , 𝜑𝑝 ) ∶ 𝑈 → R𝑝 of class 𝒞 1 on 𝑈 , and
define the set 𝑆 ∶= {𝑥 ∈ 𝑈 ∣ 𝜑(𝑥) = 0}. Assume that 𝑓 has a local extremum on 𝑆 at a point
𝑎 ∈ 𝑆 and that 𝜑 ′ (𝑎) is a surjective map. Then there exist 𝜆1 , ⋯ , 𝜆𝑝 ∈ R (called Lagrange
multipliers) such that
𝑝
𝑓 ′ (𝑎) = ∑ 𝜆𝑖 𝜑𝑗′ (𝑎).
𝑖=1
Proof. Let (𝑒1 , … , 𝑒𝑝 ) be a basis of R𝑝 . Since 𝜑 ′ (𝑎)(𝑋 ) = R𝑝 , there is 𝑢1 , … , 𝑢𝑝 ∈ 𝑋 such

that 𝜑 ′ (𝑎)(𝑢𝑖 ) = 𝑒𝑖 for any 𝑖 = 1, … , 𝑝. Define 𝑉 = span(𝑢1 , … , 𝑢𝑝 ) a vector subspace of
𝑋.
12.4 Second-order conditions 77
Consider the set Ω = {(𝑥, 𝑡) ∈ 𝑋 × R𝑝 ∣ 𝑥 + 𝑡1 𝑢1 + ⋯ + 𝑡𝑝 𝑢𝑝 ∈ 𝑈 } where we denote

𝑡 = (𝑡1 , … , 𝑡𝑝 ) ∈ R𝑝 , and define the function Φ ∶ Ω → R𝑝 By
Φ(𝑥, 𝑡) = 𝜑(𝑥 + 𝑡1 𝑢1 + ⋯ + 𝑡𝑝 𝑢𝑝 ).
One has (𝑎, 0) ∈ Ω, Φ(𝑎, 0) = 𝜑(𝑎) = 0 and
𝜕𝑡𝑖 Φ(𝑎, 0) = 𝜑 ′ (𝑎)𝑢𝑖 = 𝑒𝑖 ∀𝑖 = 1, … , 𝑝.
In particular, one has 𝜕𝑡 Φ(𝑎, 0) = idR𝑝 ∈ ℒ (R𝑝 , R𝑝 ). Hence by the Implicit Function
Theorem 11.4 there is some open neighborhood 𝑊 of 𝑎 in 𝑋 and some function 𝑔 =
(𝑔1 , … , 𝑔𝑝 ) = 𝑊 → R𝑝 of class 𝒞 1 such that 𝑔(𝑎) = 0 and Φ(𝑥, 𝑔(𝑥)) = 0 for any 𝑥 ∈ 𝑊 ,
that is
𝑥 + 𝑔1 (𝑥)𝑢1 + ⋯ 𝑔𝑝 (𝑥)𝑢𝑝 ∈ 𝑆 = Φ−1 ({0}),
and moreover
𝑔 ′ (𝑎)(ℎ) = −(𝜕𝑡 Φ(𝑎, 0))−1 ∘ 𝜕𝑥 Φ(𝑎, 0)(ℎ),
hence
𝑔𝑖′ (𝑎)(ℎ) = −𝜑𝑖′ (𝑎)(ℎ) ∀𝑖 = 1, … , 𝑝.
Assume now that 𝑓 has a local maximum on 𝑆 at the point 𝑎 (the other case can be
treated in a similar fashion). For 𝑥 close enough to 𝑎 in 𝑊 one has
𝑓 (𝑥 + 𝑔1 (𝑥)𝑢1 + ⋯ 𝑔𝑝 (𝑥)𝑢𝑝 ) ⩽ 𝑓 (𝑎) = 𝑓 (𝑎 + 𝑔1 (𝑎)(𝑢1 ) + ⋯ + 𝑔𝑝 (𝑎)𝑢𝑝 ).
Define the map 𝜓 ∶ 𝑥 ↦ 𝑓 (𝑥 + 𝑔1 (𝑥)𝑢1 + ⋯ 𝑔𝑝 (𝑥)𝑢𝑝 ) from 𝑊 into R, so that 𝜓 has a

local maximum at 𝑎 by above, therefore for any ℎ ∈ 𝑋 one has
𝜓 ′ (𝑎)(ℎ) = 𝑓 ′ (𝑎) (ℎ + 𝑔1′ (𝑎)(ℎ)𝑢1 + ⋯ + 𝑔𝑝′ (𝑎)(ℎ)𝑢𝑝 ) = 0.
It follows that
𝑝
𝑓 ′ (𝑎)(ℎ) = ∑(−𝑔𝑖′ (𝑎)(ℎ))𝑓 ′ (𝑎)(𝑢𝑖 )
𝑖=1
which implies
𝑝
𝑓 ′ (𝑎)(ℎ) = ∑(𝑓 ′ (𝑎)(𝑢𝑖 ))𝜑𝑖′ (𝑎)(ℎ),
𝑖=1
which concludes the proof by taking 𝜆𝑖 = 𝑓 ′ (𝑎)(𝑢𝑖 ) for any 𝑖 = 1, … , 𝑝.
12.4 Second-order conditions

Proposition 12.6. Let 𝑓 ∶ 𝑈 → R be of class 𝒞 2 on 𝑈 . If 𝑓 has a local minimum (resp.
local maximum) at 𝑎 ∈ 𝑈 , then 𝑓 ″ (𝑎)(ℎ, ℎ) ⩾ 0 (resp. 𝑓 ″ (𝑎)(ℎ, ℎ) ⩽ 0) for any ℎ ∈ 𝑋 .
Proof. If 𝑓 has a local minimum at 𝑎 one knows that 𝑓 ′ (𝑎) = 0 and that there is some
neighborhood 𝑉 of 0 in 𝑋 such that 𝑓 (𝑎 + ℎ) ⩾ 𝑓 (𝑎) for any ℎ ∈ 𝑉 .
Let ℎ ∈ 𝑋 . For 𝑡 ∈ R small enough one has 𝑓 (𝑎 + 𝑡ℎ) ⩾ 𝑓 (𝑎). Define the function
𝑔 ∶ 𝑡 ↦ 𝑓 (𝑎 + 𝑡ℎ) which also is of class 𝒞 2 . Applying Taylor formula (Theorem 10.13)
one obtains
𝑡2
𝑔(𝑡) = 𝑔(0) + 𝑡𝑔 ′ (0) + 𝑔 ″ (0) + 𝑜(𝑡 2 ).
2
78 Chapter 12. Optimization
This implies that

2
0⩽ [𝑓 (𝑎 + 𝑡ℎ) − 𝑓 (𝑎)] = 𝑓 ″ (𝑎)(ℎ, ℎ) + 𝑜(1)
𝑡2
and taking the limit 𝑡 → 0 one concludes that 𝑓 ″ (𝑎)(ℎ, ℎ) ⩾ 0.
Definition 12.7. Let 𝑓 ∶ 𝑈 → R be of class 𝒞 2 on 𝑈 . A point 𝑎 ∈ 𝑈 is a saddle-point

if it is a critical point and if for every neighborhood of 𝑎 the function 𝑓 attains values
strictly greater and strictly smaller than 𝑓 (𝑎).
Theorem 12.8. Let 𝑓 ∶ 𝑈 → R be of class 𝒞 2 on 𝑈 . Suppose that 𝑎 ∈ 𝑈 is a critical point

of 𝑓 and that there is 𝛼 > 0 such that 𝑓 ″ (𝑎)(ℎ, ℎ) ⩾ 𝛼‖ℎ‖2 for any ℎ ∈ 𝑋 . Then 𝑓 has a
strict local minimum at 𝑎.
Proof. Since 𝑓 ″ is continuous there is 𝑟 > 0 such that ‖𝑓 ″ (𝑎 + ℎ) − 𝑓 ″ (𝑎)‖ ⩽ 𝛼2 for any
ℎ ∈ 𝐵𝑋 (0, 𝑟).
Define the function 𝑔 ∶ 𝑥 ↦ 𝑓 (𝑎 + 𝑥) − 12 𝑓 ″ (𝑎)(𝑥, 𝑥) from 𝐵𝑋 (0, 𝑟) into R. The map
𝑔 is also of class 𝒞 2 and one has
𝑔 ′ (𝑥)(ℎ) = 𝑓 ′ (𝑎 + 𝑥)(ℎ) − 𝑓 ″ (𝑎)(𝑥, ℎ), 𝑔 ″ (𝑥)(ℎ, 𝑘) = 𝑓 ′ (𝑎 + 𝑥)(ℎ, 𝑘) − 𝑓 ″ (𝑎)(ℎ, 𝑘)

𝛼
and 𝑔 ′ (0) = 𝑓 ′ (𝑎) = 0. Hence ‖𝑔 ″ (𝑥)‖ ⩽ 2
for any 𝑥 ∈ 𝐵𝑋 (0, 𝑟) and by Taylor’s formula
(Theorem 10.12) one gets
1 𝛼
|𝑔(𝑥) − 𝑔(0) − 𝑔 ′ (0)(𝑥)| ⩽ sup ‖𝑔 ″ (𝑦)(𝑥, 𝑥)‖ ⩽ ‖𝑥‖2 ,
2 𝑦∈𝐵𝑋 (0,𝑟) 4
which implies that

1 ″ 𝛼
|𝑓 (𝑎 + 𝑥) − 𝑓 (𝑎) − 𝑓 (𝑎)(𝑥, 𝑥)| ⩽ ‖𝑥‖2 .
2 4
Therefore, for any 𝑥 ∈ 𝐵𝑋 (0, 𝑟) such that 𝑥 ≠ 0 ,
1 ″ 𝛼 𝛼
𝑓 (𝑎 + 𝑥) ⩾ 𝑓 (𝑎) + 𝑓 (𝑎)(𝑥, 𝑥) − ‖𝑥‖2 ⩾ 𝑓 (𝑎) + ‖𝑥‖2 > 𝑓 (𝑎),
2 4 4
Appendix A
Reminders: Sets and maps
A.1 Sets
Let 𝑋 be a set. We denote by ∅ the empty set, and by 𝔓(𝑋 ) the powerset of 𝑋 , that is,
𝔓(𝑋 ) = {𝐴 ∶ 𝐴 ⊆ 𝑋 } is the set of all subsets of 𝑋 .
A.1.1 Intersection, union, and complements

If 𝐴 and 𝐵 are subsets of 𝑋 , we define the intersection of 𝐴 and 𝐵 by
𝐴 ∩ 𝐵 = {𝑥 ∈ 𝑋 ∣ 𝑥 ∈ 𝐴 and 𝑥 ∈ 𝐵};
the union of 𝐴 and 𝐵 by
𝐴 ∪ 𝐵 = {𝑥 ∈ 𝑋 ∣ 𝑥 ∈ 𝐴 or 𝑥 ∈ 𝐵};
and the set of elements of 𝑋 that belong to 𝐵 but do not belong to 𝐴 by
𝐵 ⧵ 𝐴 = {𝑥 ∈ 𝑋 ∣ 𝑥 ∈ 𝐵, 𝑥 ∉ 𝐴}.
In particular, 𝑋 ⧵ 𝐴 = {𝑥 ∈ 𝑋 ∣ 𝑥 ∉ 𝐴} is called the complement set of 𝐴.
A.1.2 Family of subsets

Let {𝐴𝑖 }𝑖∈𝐼 be a family of subsets of 𝑋 indexed by a set 𝐼 . We define the intersection of
the family {𝐴𝑖 }𝑖∈𝐼 by
⋂ 𝐴𝑖 = {𝑥 ∈ 𝑋 ∣ ∀𝑖 ∈ 𝐼 , 𝑥 ∈ 𝐴𝑖 }
𝑖∈𝐼
and the union of the family {𝐴𝑖 }𝑖∈𝐼 by
⋃ 𝐴𝑖 = {𝑥 ∈ 𝑋 ∣ ∃𝑖 ∈ 𝐼 , 𝑥 ∈ 𝐴𝑖 }.
𝑖∈𝐼
Similarly, if 𝒜 is a (non indexed) family of subsets of 𝑋 , then the intersection ⋂ 𝒜 and

the union ⋃ 𝒜 are defined by, respectively,
⋂𝒜 = ⋂ 𝐴 and ⋃ 𝒜 = ⋃ 𝐴.
𝐴∈𝒜 𝐴∈𝒜
We shall adopt the convention that the union of an empty family of subsets of 𝑋 is equal
to the empty set ∅, and that the intersection of an empty family of subsets of 𝑋 is equal
to 𝑋 .
79
80 Chapter A. Reminders: Sets and maps
Proposition A.1. Let 𝑋 be a set, {𝐴𝑖 }𝑖∈𝐼 be a family of subsets of 𝑋 . Then

𝑋 ⧵ ⋂ 𝐴𝑖 = ⋃(𝑋 ⧵ 𝐴𝑖 ) and 𝑋 ⧵ ⋃ 𝐴𝑖 = ⋂(𝑋 ⧵ 𝐴𝑖 ).
𝑖∈𝐼 𝑖∈𝐼 𝑖∈𝐼 𝑖∈𝐼
Proof. Let 𝑥 ∈ 𝑋 , then
𝑥 ∈ 𝑋 ⧵ ⋂ 𝐴𝑖 ⟺ 𝑥 ∉ ⋂ 𝐴𝑖 ⟺ ∃𝑖 ∈ 𝐼 , 𝑥 ∉ 𝐴𝑖
𝑖∈𝐼 𝑖∈𝐼
⟺ ∃𝑖 ∈ 𝐼 , 𝑥 ∈ 𝑋 ⧵ 𝐴𝑖 ⟺ 𝑥 ∈ ⋃(𝑋 ⧵ 𝐴𝑖 ).
𝑖∈𝐼
Let 𝑥 ∈ 𝑋 , then
𝑥 ∈ 𝑋 ⧵ ⋃ 𝐴𝑖 ⟺ 𝑥 ∉ ⋃ 𝐴𝑖 ⟺ ∀𝑖 ∈ 𝐼 , 𝑥 ∉ 𝐴𝑖
⟺ ∀𝑖 ∈ 𝐼 , 𝑥 ∈ 𝑋 ⧵ 𝐴𝑖 ⟺ 𝑥 ∈ ⋂(𝑋 ⧵ 𝐴𝑖 ).
𝑖∈𝐼
Proposition A.2. Let 𝑋 be a set, {𝐴𝑖 }𝑖∈𝐼 be a family of subsets of 𝑋 , and 𝐵 be a subset of
𝑋 . Then
(⋃ 𝐴𝑖 ) ∩ 𝐵 = ⋃(𝐴𝑖 ∩ 𝐵) and (⋂ 𝐴𝑖 ) ∪ 𝐵 = ⋂(𝐴𝑖 ∪ 𝐵).
𝑥 ∈ (⋃ 𝐴𝑖 ) ∩ 𝐵 ⟺ 𝑥 ∈ ⋃ 𝐴𝑖 and 𝑥 ∈ 𝐵 ⟺ (∃𝑖 ∈ 𝐼 , 𝑥 ∈ 𝐴𝑖 ) and 𝑥 ∈ 𝐵

⟺ ∃𝑖 ∈ 𝐼 , 𝑥 ∈ 𝐴𝑖 ∩ 𝐵 ⟺ 𝑥 ∈ ⋃(𝐴𝑖 ∩ 𝐵).
𝑖∈𝐼
𝑥 ∈ (⋂ 𝐴𝑖 ) ∪ 𝐵 ⟺ 𝑥 ∈ ⋂ 𝐴𝑖 or 𝑥 ∈ 𝐵 ⟺ (∀𝑖 ∈ 𝐼 , 𝑥 ∈ 𝐴𝑖 ) or 𝑥 ∈ 𝐵
⟺ ∀𝑖 ∈ 𝐼 , 𝑥 ∈ 𝐴𝑖 ∪ 𝐵 ⟺ 𝑥 ∈ ⋂(𝐴𝑖 ∪ 𝐵).
𝑖∈𝐼
A.1.3 Product of sets

Consider two sets 𝑋 and 𝑌 . We define the (Cartesian) product 𝑋 × 𝑌 of the sets 𝑋 and 𝑌
as the set of all ordered pairs (𝑥, 𝑦) with 𝑥 ∈ 𝑋 and 𝑦 ∈ 𝑌 , more precisely
𝑋 × 𝑌 = {(𝑥, 𝑦) ∣ 𝑥 ∈ 𝑋 , 𝑦 ∈ 𝑌 }.
Proposition A.3. If 𝐴 and 𝐵 are subsets of 𝑋 and 𝐶 and 𝐷 are subsets of 𝑌 , then we have
𝐴 × (𝐶 ∩ 𝐷) = (𝐴 × 𝐶) ∩ (𝐴 × 𝐷),
𝐴 × (𝐶 ∪ 𝐷) = (𝐴 × 𝐶) ∪ (𝐴 × 𝐷),
𝐴 × (𝐶 ⧵ 𝐷) = (𝐴 × 𝐶) ⧵ (𝐴 × 𝐷),
(𝐴 ∩ 𝐵) × (𝐶 ∩ 𝐷) = (𝐴 × 𝐶) ∩ (𝐵 × 𝐷).
A.1 Sets 81
Proof. Let (𝑥, 𝑦) ∈ 𝑋 × 𝑌 , then
(𝑥, 𝑦) ∈ 𝐴 × (𝐶 ∩ 𝐷) ⟺ 𝑥 ∈ 𝐴, 𝑦 ∈ 𝐶 ∩ 𝐷 ⟺ 𝑥 ∈ 𝐴, 𝑦 ∈ 𝐶, 𝑦 ∈ 𝐷
⟺ (𝑥, 𝑦) ∈ 𝐴 × 𝐶, (𝑥, 𝑦) ∈ 𝐴 × 𝐷
⟺ (𝑥, 𝑦) ∈ (𝐴 × 𝐶) ∩ (𝐴 × 𝐷).
Let (𝑥, 𝑦) ∈ 𝑋 × 𝑌 , then
(𝑥, 𝑦) ∈ 𝐴 × (𝐶 ∪ 𝐷) ⟺ 𝑥 ∈ 𝐴, 𝑦 ∈ 𝐶 ∪ 𝐷 ⟺ 𝑥 ∈ 𝐴, 𝑦 ∈ 𝐶 or 𝑦 ∈ 𝐷
⟺ (𝑥, 𝑦) ∈ 𝐴 × 𝐶 or (𝑥, 𝑦) ∈ 𝐴 × 𝐷
⟺ (𝑥, 𝑦) ∈ (𝐴 × 𝐶) ∪ (𝐴 × 𝐷).
(𝑥, 𝑦) ∈ 𝐴 × (𝐶 ⧵ 𝐷) ⟺ 𝑥 ∈ 𝐴, 𝑦 ∈ 𝐶 ⧵ 𝐷 ⟺ 𝑥 ∈ 𝐴, 𝑦 ∈ 𝐶, 𝑦 ∉ 𝐷
⟺ (𝑥, 𝑦) ∈ 𝐴 × 𝐶, (𝑥, 𝑦) ∉ 𝐴 × 𝐷
⟺ (𝑥, 𝑦) ∈ (𝐴 × 𝐶) ⧵ (𝐴 × 𝐷).
(𝑥, 𝑦) ∈ (𝐴 ∩ 𝐵) × (𝐶 ∩ 𝐷) ⟺ 𝑥 ∈ 𝐴 ∩ 𝐵, 𝑦 ∈ 𝐶 ∩ 𝐷
⟺ 𝑥 ∈ 𝐴, 𝑥 ∈ 𝐵, 𝑦 ∈ 𝐶, 𝑦 ∈ 𝐷
⟺ (𝑥, 𝑦) ∈ 𝐴 × 𝐶, (𝑥, 𝑦) ∈ 𝐵 × 𝐷
⟺ (𝑥, 𝑦) ∈ (𝐴 × 𝐶) ∩ (𝐵 × 𝐷).
Consider now a family {𝑋𝑖 }𝑖∈𝐼 of sets indexed by 𝐼 , we then define the (Cartesian)
product ∏𝑖∈𝐼 𝑋𝑖 of the sets {𝑋𝑖 }𝑖∈𝐼 by
∏ 𝑋𝑖 = {𝑥 ∶ 𝐼 → ⋃ 𝑋𝑖 | ∀𝑖 ∈ 𝐼 , 𝑥(𝑖) ∈ 𝑋𝑖 } .
We often denote 𝑥 ∈ ∏𝑖∈𝐼 𝑋𝑖 by 𝑥 = (𝑥𝑖 )𝑖∈𝐼 . Moreover, for each 𝑗 ∈ 𝐼 we define the
canonical projection pr𝑗 ∶ ∏𝑖∈𝐼 𝑋𝑖 → 𝑋𝑗 given by pr𝑗 (𝑥) = 𝑥(𝑗).
A.1.4 Axiom of choice and Zorn’s lemma

Axiom of Choice. For every set 𝑋 there is a map 𝑓 ∶ 𝔓(𝑋 )⧵{∅} → 𝑋 such that 𝑓 (𝐴) ∈ 𝐴
for all non-empty subset 𝐴 of 𝑋 .
The Axiom of Choice is actually equivalent (under the Zermelo-Fraenkel axioms of
set theory) to the Zorn’s Lemma described below.
A partially ordered set (poset) is a pair (𝑃, ⪯) where 𝑃 is a set and ⪯ is a reflexive,
antisymmetric, and transitive binary relation over 𝑃, that is, for any 𝑥, 𝑦, 𝑧 ∈ 𝑃 there
holds:
(i) 𝑥 ⪯ 𝑥 ;
(ii) 𝑥 ⪯ 𝑦 and 𝑦 ⪯ 𝑥 imply 𝑥 = 𝑦 ;
(iii) 𝑥 ⪯ 𝑦 and 𝑦 ⪯ 𝑧 imply 𝑥 ⪯ 𝑧.
Consider a subset 𝑆 of a given poset 𝑃. An element 𝑚 ∈ 𝑆 is a maximal element of 𝑆 if

for any 𝑥 ∈ 𝑆, 𝑚 ⪯ 𝑥 implies 𝑚 = 𝑥. An element 𝑢 ∈ 𝑃 is an upper bound of 𝑆 if for any
𝑥 ∈ 𝑆, 𝑥 ⪯ 𝑢. One says that 𝑆 is a totally ordered subset or a chain if the restriction of ⪯
to 𝑆 is total, i.e. for any 𝑥, 𝑦 ∈ 𝑆 one has 𝑥 ⪯ 𝑦 or 𝑦 ⪯ 𝑥.
Lemma A.4 (Zorn’s Lemma). If 𝑃 is a partially ordered set with the property that every
chain has an upper bound, then 𝑃 contains a maximal element.
A.1.5 Countable sets

A set 𝑋 is said to be countable if it is in bijection with some subset of N. In other words,
𝑋 is countable if either 𝑋 is a finite set, or 𝑋 is an infinite set and there is a bijective
map from 𝑋 into N. A set is said to be uncountable if it is not countable.
Proposition A.5. There is no surjective map from a non-empty set 𝑋 onto 𝔓(𝑋 )
As a consequence of this result, we deduce that there is no surjection from N onto

𝔓(N), which is equivalent to say that there is no injection from 𝔓(N) to N, and thus
𝔓(N) is uncountable.
Proof of Proposition A.5. By way of contradiction, assume that 𝑓 ∶ 𝑋 → 𝔓(𝑋 ) is surjec-

tive. Consider the set 𝐴 = {𝑥 ∈ 𝑋 ∣ 𝑥 ∉ 𝑓 (𝑥)}. Since 𝑓 is surjective, there is 𝑎 ∈ 𝑋
such that 𝑓 (𝑎) = 𝐴. On the other hand, by construction we have 𝑎 ∈ 𝐴 if and only if
𝑎 ∉ 𝑓 (𝑎) = 𝐴, thus a contradiction.
Proposition A.6. Every finite product of countable sets is countable.
Proof. We only show that the product of two countable sets is countable, the general
case being deduced from this by induction. Let 𝑋 and 𝑌 be countable sets. We know
that there is an injection 𝑋 → N and an injection 𝑌 → N, thus there exists an injection
𝜙 ∶ 𝑋 × 𝑌 → N × N. But N × N is countable, since the map 𝑓 ∶ N × N → N given by
𝑓 (𝑛, 𝑚) = 2𝑛 3𝑚 is injective. Therefore the map 𝑓 ∘ 𝜙 ∶ 𝑋 × 𝑌 → N is injective, thus 𝑋 × 𝑌
is countable.
Proposition A.7. Every countable union of countable sets is countable
Proof. Let {𝑋𝑛 }𝑛∈N be a countable family of countable sets and consider their union 𝑋 =
⋃𝑛∈N 𝑋𝑛 . Define, for any 𝑛 ∈ N, the set ℐ𝑛 of all injections from 𝑋𝑛 to N. We remark
that ℐ𝑛 ≠ ∅ since 𝑋𝑛 is countable.
Thanks to the Axiom of Choice (actually we need only a weaker version of it, called
the “Axiom of Countable Choice”) there is a sequence (𝑖𝑛 )𝑛∈N such that 𝑖𝑛 ∈ ℐ𝑛 for all
𝑛 ∈ N. Define the map
𝜙 ∶ 𝑋 → N×N
𝑥 ↦ (𝑛, 𝑖𝑛 (𝑥))
where 𝑛 is the smallest natural number verifying 𝑥 ∈ 𝑋𝑛 . By construction 𝜙 is an injec-
tion, since every 𝑖𝑛 is an injection. But there is an injection 𝑓 ∶ N × N → N since N × N
is countable. Therefore 𝑓 ∘ 𝜙 ∶ 𝑋 → N is a injective map and thus 𝑋 is countable.
A.2 Maps 83
A.1.6 Equivalence relation and quotient set

A partition of a set 𝑋 is a collection {𝐴𝑖 }𝑖∈𝐼 of non-empty subsets of 𝑋 such that 𝑋 is the
disjoint union of 𝐴𝑖 , that is 𝑋 = ⋃𝑖∈𝐼 𝐴𝑖 and 𝐴𝑖 ∩ 𝐴𝑗 = ∅ for all 𝑖 ≠ 𝑗.
A binary relation ℛ on a set 𝑋 is a subset of 𝑋 × 𝑋 . For all 𝑥, 𝑦 ∈ 𝑋 , if (𝑥, 𝑦) ∈ ℛ
we then denote 𝑥 ℛ 𝑦.
An equivalence relation on 𝑋 is a binary relation ℛ that is reflexive, symmetric and
transitive, that is, if for any 𝑥, 𝑦, 𝑧 ∈ 𝑋 there holds
(i) 𝑥 ℛ 𝑥;
(ii) 𝑥 ℛ 𝑦 if and only if 𝑦 ℛ 𝑥;
(iii) 𝑥 ℛ 𝑦 and 𝑦 ℛ 𝑧 imply 𝑥 ℛ 𝑧.
We denote by [𝑥] the equivalence class of 𝑥 ∈ 𝑋 under ℛ, which is defined by [𝑥] ∶=
{𝑦 ∈ 𝑋 ∣ 𝑥 ℛ 𝑦}. Any element of [𝑥] is called a representative of [𝑥].
One obtains that
• 𝑥 ℛ 𝑦 if and only if [𝑥] = [𝑦],
• the set of equivalence classes forms a partition of 𝑋 .
The set 𝑋 /ℛ = {[𝑥] ∶ 𝑥 ∈ 𝑋 } of equivalence classes of 𝑋 is called the quotient set
of 𝑋 by ℛ. There is a canonical projection 𝜋 ∶ 𝑋 → 𝑋 /ℛ, 𝑥 ↦ [𝑥], that maps each
element into its equivalence class.
A.2 Maps
A.2.1 Injectivity and surjectivity
Let 𝑋 and 𝑌 be sets and 𝑓 ∶ 𝑋 → 𝑌 a map.
The map 𝑓 is said to be injective (or one-to-one) provided for any 𝑥, 𝑥 ′ ∈ 𝑋 , if 𝑓 (𝑥) =
𝑓 (𝑥 ′ )
then 𝑥 = 𝑥 ′ .
The map 𝑓 is is said to be surjective (or onto) if for any 𝑦 ∈ 𝑌 there is 𝑥 ∈ 𝑋 such that
𝑦 = 𝑓 (𝑥).
The map 𝑓 is is said to be bijective (or to be a one-to-one correspondence) if it is both
injective and surjective.
Proposition A.8. Let 𝑓 ∶ 𝑋 → 𝑌 and 𝑔 ∶ 𝑌 → 𝑍 be two maps between sets, then there
holds:
(i) If 𝑓 and 𝑔 are injective, then 𝑔 ∘ 𝑓 is injective.
(ii) If 𝑓 and 𝑔 are surjective, then 𝑔 ∘ 𝑓 is surjective.
(iii) If 𝑔 ∘ 𝑓 is injective, then 𝑓 is injective.
(iv) If 𝑔 ∘ 𝑓 is surjective, then 𝑔 is surjective.
As a consequence, taking 𝑍 = 𝑋 above, we deduce that if 𝑔 ∘ 𝑓 = id𝑋 and 𝑓 ∘ 𝑔 = id𝑌 ,

then 𝑔 and 𝑓 are bijective and 𝑔 = 𝑓 −1 is the inverse of 𝑓 .
Proof of Proposition A.8. (i) Let 𝑥, 𝑥 ′ ∈ 𝑋 such that 𝑔 ∘ 𝑓 (𝑥) = 𝑔 ∘ 𝑓 (𝑥 ′ ). This means that
𝑔(𝑓 (𝑥)) = 𝑔(𝑓 (𝑥 ′ )) and hence, since 𝑔 is injective, one obtains 𝑓 (𝑥) = 𝑓 (𝑥 ′ ). Since 𝑓 is
injective one deduces that 𝑥 = 𝑥 ′ .
(ii) Let 𝑧 ∈ 𝑍 . Since 𝑔 is surjective there is 𝑦 ∈ 𝑌 such that 𝑔(𝑦) = 𝑧. Since 𝑓 is surjective
there is 𝑥 ∈ 𝑋 such that 𝑦 = 𝑓 (𝑥), therefore 𝑔 ∘ 𝑓 (𝑥) = 𝑔(𝑦) = 𝑧.
(iii) Let 𝑥, 𝑥 ′ ∈ 𝑋 such that 𝑓 (𝑥) = 𝑓 (𝑥 ′ ). Applying the map 𝑔 to last equality, one
obtains that 𝑔 ∘ 𝑓 (𝑥) = 𝑔 ∘ 𝑓 (𝑥 ′ ) and hence, since 𝑔 ∘ 𝑓 is injective, one deduces 𝑥 = 𝑥 ′ .
(iv) Let 𝑧 ∈ 𝑍 . Since 𝑔 ∘ 𝑓 is surjective there is 𝑥 ∈ 𝑋 such that 𝑔 ∘ 𝑓 (𝑥) = 𝑧. Therefore
𝑦 = 𝑓 (𝑥) ∈ 𝑌 and 𝑔(𝑦) = 𝑧.
A.2.2 Image and preimage

Let 𝑋 and 𝑌 be sets and 𝑓 ∶ 𝑋 → 𝑌 a map.
For any subset 𝐴 of 𝑋 , one defines the image of 𝐴 under 𝑓 as the subset of 𝑌 given
by
𝑓 (𝐴) = {𝑦 ∈ 𝑌 ∣ ∃𝑥 ∈ 𝐴, 𝑓 (𝑥) = 𝑦} = {𝑓 (𝑥) ∶ 𝑥 ∈ 𝐴}.
For any subset 𝐵 of 𝑌 , one defines the preimage or inverse image of 𝐵 under 𝑓 as the
subset of 𝑋 given by
𝑓 −1 (𝐵) = {𝑥 ∈ 𝑋 ∣ 𝑓 (𝑥) ∈ 𝐵}.
We easily obtain the following properties.
Proposition A.9. For any subset 𝐴 ⊆ 𝑋 one has 𝐴 ⊆ 𝑓 −1 (𝑓 (𝐴)). Moreover, the map 𝑓 is
injective if and only if 𝐴 = 𝑓 −1 (𝑓 (𝐴)) for any subset 𝐴 ⊆ 𝑋 .
Proof. Let 𝐴 ⊆ 𝑋 be a subset and 𝑥 ∈ 𝐴, then 𝑓 (𝑥) ∈ 𝑓 (𝐴) and hence 𝑥 ∈ 𝑓 −1 (𝑓 (𝐴)).
This proves 𝐴 ⊆ 𝑓 −1 (𝑓 (𝐴)).
We give an example in which the previous inclusion is proper: Consider the map
𝑓 ∶ {0, 1} → {0, 1} defined by 0 ↦ 0 and 1 ↦ 0. Then {0} ⊊ {0, 1} = 𝑓 −1 (𝑓 ({0})).
We now prove the second assertion. Suppose 𝑓 injective and let 𝐴 ⊆ 𝑋 . Let 𝑥 ∈
−1
𝑓 (𝑓 (𝐴)), then 𝑓 (𝑥) ∈ 𝑓 (𝐴). Therefore there is 𝑎 ∈ 𝐴 such that 𝑓 (𝑥) = 𝑓 (𝑎), which
implies that 𝑥 = 𝑎 ∈ 𝐴, whence 𝑓 −1 (𝑓 (𝐴)) ⊆ 𝐴. From the first assertion one deduces
𝐴 = 𝑓 −1 (𝑓 (𝐴)).
Conversely, assume that 𝐴 = 𝑓 −1 (𝑓 (𝐴)) for any subset 𝐴 ⊆ 𝑋 . Let 𝑥, 𝑦 ∈ 𝑋 such that
𝑓 (𝑥) = 𝑓 (𝑦). Then, by hypothesis, {𝑥} = 𝑓 −1 (𝑓 ({𝑥})) = 𝑓 −1 ({𝑓 (𝑥)}) = 𝑓 −1 ({𝑓 (𝑦)}) =
𝑓 −1 (𝑓 ({𝑦})) = {𝑦}, therefore 𝑥 = 𝑦 and thus 𝑓 is injective.
Proposition A.10. For any subset 𝐵 ⊆ 𝑌 one has 𝑓 (𝑓 −1 (𝐵)) ⊆ 𝐵. Moreover, the map 𝑓 is
surjective if and only if 𝑓 (𝑓 −1 (𝐵)) = 𝐵 for any subset 𝐵 ⊆ 𝑌 .
Proof. Let 𝐵 ⊆ 𝑌 and 𝑦 ∈ 𝑓 (𝑓 −1 (𝐵)), then there is 𝑥 ∈ 𝑓 −1 (𝐵) such that 𝑦 = 𝑓 (𝑥). Since
𝑓 (𝑥) ∈ 𝐵 we conclude that 𝑦 ∈ 𝐵. This proves 𝑓 (𝑓 −1 (𝐵)).
We give an example in which the inclusion is proper: Consider the map 𝑓 ∶ {0, 1} →
{0, 1} defined by 0 ↦ 0 and 1 ↦ 0. Then 𝑓 (𝑓 −1 ({0, 1})) = {0} ⊊ {0, 1}.
We now prove the second assertion. Suppose 𝑓 surjective and let 𝐵 ⊆ 𝑌 . Let 𝑦 ∈ 𝐵,
then there is 𝑥 ∈ 𝑋 such that 𝑓 (𝑥) = 𝑦, that is 𝑥 ∈ 𝑓 −1 (𝐵), whence 𝑦 ∈ 𝑓 (𝑓 −1 (𝐵)). From
the first assertion one deduces that 𝑓 (𝑓 −1 (𝐵)) = 𝐵.
Conversely, assume that 𝑓 (𝑓 −1 (𝐵)) = 𝐵 for any subset 𝐵 ⊆ 𝑌 . Let 𝑦 ∈ 𝑌 , then
𝑓 (𝑓 −1 ({𝑦})) = {𝑦} is non-empty, therefore there is 𝑥 ∈ 𝑋 such that 𝑓 (𝑥) = 𝑦, and thus 𝑓
is surjective.
Proposition A.11. For any subset 𝐵 ⊆ 𝑌 one has 𝑋 ⧵ 𝑓 −1 (𝐵) = 𝑓 −1 (𝑌 ⧵ 𝐵).

A.2 Maps 85
Proof. Let 𝐵 ⊆ 𝑌 and 𝑥 ∈ 𝑋 . Then 𝑥 ∉ 𝑓 −1 (𝐵) if and only if 𝑓 (𝑥) ∉ 𝐵, which means
𝑥 ∈ 𝑓 −1 (𝑌 ⧵ 𝐵).
Consider a non-empty family {𝐴𝑖 }𝑖∈𝐼 of subsets of 𝑋 . We obtain the following prop-
erties concerning the image of unions and intersections.
Proposition A.12. There holds
𝑓 (⋃ 𝐴𝑖 ) = ⋃ 𝑓 (𝐴𝑖 ) and 𝑓 (⋂ 𝐴𝑖 ) ⊆ ⋂ 𝑓 (𝐴𝑖 ).
Moreover, if 𝑓 is injective then the above inclusion is an equality.
Proof. Let 𝑦 ∈ 𝑌 , then
𝑦 ∈ 𝑓 (⋃ 𝐴𝑖 ) ⟺ ∃𝑥 ∈ ⋃ 𝐴𝑖 , 𝑦 = 𝑓 (𝑥) ⟺ ∃𝑖 ∈ 𝐼 , ∃𝑥 ∈ 𝐴𝑖 , 𝑦 = 𝑓 (𝑥)
⟺ ∃𝑖 ∈ 𝐼 , 𝑦 ∈ 𝑓 (𝐴𝑖 ) ⟺ 𝑦 ∈ ⋃ 𝑓 (𝐴𝑖 ).
𝑖∈𝐼
This proves the first assertion.
Let 𝑦 ∈ 𝑓 (⋂𝑖∈𝐼 𝐴𝑖 ), then there is 𝑥 ∈ ⋂𝑖∈𝐼 𝐴𝑖 such that 𝑦 = 𝑓 (𝑥), which means that
𝑥 ∈ 𝐴𝑖 for any 𝑖 ∈ 𝐼 and 𝑦 = 𝑓 (𝑥). This implies that for any 𝑖 ∈ 𝐼 there is a 𝑥 ∈ 𝐴𝑖 such
that 𝑦 = 𝑓 (𝑥), thus 𝑦 ∈ 𝑓 (𝐴𝑖 ) for any 𝑖 ∈ 𝐼 . Therefore 𝑦 ∈ ⋂𝑖∈𝐼 𝑓 (𝐴𝑖 ). This proves the
second assertion.
Let us give an example of a proper inclusion: Take 𝑓 ∶ 𝑥 ↦ 𝑥 2 from R to R, 𝐴1 = {1}
and 𝐴2 = {−1}, then 𝑓 (𝐴1 ∩ 𝐴2 ) = ∅ and 𝑓 (𝐴1 ) ∩ 𝑓 (𝐴2 ) = {1}.
Finally, we assume that 𝑓 is injective and we prove the inclusion ⋂𝑖∈𝐼 𝑓 (𝐴𝑖 ) ⊆ 𝑓 (⋂𝑖∈𝐼 𝐴𝑖 ).
Let 𝑦 ∈ ⋂𝑖∈𝐼 𝑓 (𝐴𝑖 ), then 𝑦 ∈ 𝑓 (𝐴𝑖 ) for any 𝑖 ∈ 𝐼 . This means that, for any 𝑖 ∈ 𝐼 , there is
𝑥𝑖 ∈ 𝐴𝑖 such that 𝑦 = 𝑓 (𝑥𝑖 ). Since 𝑓 is injective we deduce that all the 𝑥𝑖 are the same,
that is, there is 𝑥 ∈ 𝑋 such that 𝑥 = 𝑥𝑖 ∈ 𝐴𝑖 for any 𝑖 ∈ 𝐼 . Therefore we have that
𝑥 ∈ ⋂𝑖∈𝐼 𝐴𝑖 and 𝑦 = 𝑓 (𝑥), hence 𝑦 ∈ 𝑓 (⋂𝑖∈𝐼 𝐴𝑖 ).
Finally, we consider a non-empty family {𝐵𝑗 }𝑗∈𝐽 of subsets of 𝑌 and obtain the fol-
lowing properties concerning the preimage of unions and intersections.
Proposition A.13. There holds
𝑓 −1 (⋃ 𝐵𝑗 ) = ⋃ 𝑓 −1 (𝐵𝑗 ) and 𝑓 −1 (⋂ 𝐵𝑗 ) = ⋂ 𝑓 −1 (𝐵𝑗 ).
𝑗∈𝐽 𝑗∈𝐽 𝑗∈𝐽 𝑗∈𝐽

𝑥 ∈ 𝑓 −1 (⋃ 𝐵𝑗 ) ⟺ 𝑓 (𝑥) ∈ ⋃ 𝐵𝑗 ⟺ ∃𝑗 ∈ 𝐽 , 𝑓 (𝑥) ∈ 𝐵𝑗
𝑗∈𝐽 𝑗∈𝐽
⟺ ∃𝑗 ∈ 𝐽 , 𝑥 ∈ 𝑓 −1 (𝐵𝑗 ) ⟺ 𝑥 ∈ ⋃ 𝑓 −1 (𝐵𝑗 ).
𝑗∈𝐽
This proves the first assertion.

𝑥 ∈ 𝑓 −1 (⋂ 𝐵𝑗 ) ⟺ 𝑓 (𝑥) ∈ ⋂ 𝐵𝑗 ⟺ ∀𝑗 ∈ 𝐽 , 𝑓 (𝑥) ∈ 𝐵𝑗
𝑗∈𝐽 𝑗∈𝐽
⟺ ∀𝑗 ∈ 𝐽 , 𝑥 ∈ 𝑓 −1 (𝐵𝑗 ) ⟺ 𝑥 ∈ ⋂ 𝑓 −1 (𝐵𝑗 ).
𝑗∈𝐽
This proves the second assertion.

Topology Polytechnique

Uploaded by

Document Information

Original Title

Copyright

Available Formats

Share this document

Share or Embed Document

Sharing Options

Did you find this document useful?

Is this content inappropriate?

Copyright:

Available Formats

Topology Polytechnique

Uploaded by

Copyright:

Available Formats

MAA302

Topology and differential calculus

1 Topological and metric spaces 1

2 Neighborhoods, basis, interior and closure 5

3 Continuity and limits 13

8 Banach spaces and continuous linear maps 43

10 Mean-value theorem and applications 63

11 Inverse function and implicit function theorems 71

A Reminders: Sets and maps 79

Topological and metric spaces

1.1 Topological spaces

Definition 1.4. A subset 𝑈 of a topological space 𝑋 is closed if its complement 𝑋 ⧵ 𝑈 is

It easily follows from Definition 1.1 that:

1.2 Metric spaces

𝐵(𝑥, 𝑟) ∶= {𝑦 ∈ 𝑋 ∣ 𝑑(𝑥, 𝑦) ⩽ 𝑟},

and the sphere centred at 𝑥 of radius 𝑟 by

𝑆(𝑥, 𝑟) ∶= {𝑦 ∈ 𝑋 ∣ 𝑑(𝑥, 𝑦) = 𝑟}.

Definition 1.10. A subset 𝑌 of a metric space (𝑋 , 𝑑) is said to be bounded if there are

Equivalently, 𝑌 is bounded if and only if sup𝑥,𝑦∈𝑌 𝑑(𝑥, 𝑦) < ∞.

𝑑1 ((𝑥, 𝑦), (𝑥 ′ , 𝑦 ′ )) ∶= 𝑑𝑋 (𝑥, 𝑥 ′ ) + 𝑑𝑌 (𝑦, 𝑦 ′ ),

1.2.1 Topology associated to a metric

called the topology associated to or induced by the metric 𝑑.

Let us verify that 𝒪𝑑 is indeed a topology on 𝑋 . It is clear that ∅ and 𝑋 belong to

Definition 1.12. A topological space (𝑋 , 𝒪) is metrizable if there is a metric 𝑑 on 𝑋

1.2.2 Equivalent metrics

𝑐1 𝑑1 (𝑥, 𝑦) ⩽ 𝑑2 (𝑥, 𝑦) ⩽ 𝑐2 𝑑1 (𝑥, 𝑦).

It is easy to verify that “equivalent” implies “topologically equivalent”.

Neighborhoods, basis, interior and

Definition 2.1. A neighborhood of 𝑥 (resp. 𝐴) is a subset 𝑉 ⊆ 𝑋 that contains an open

In other words, denoting by 𝒪 the topology of 𝑋 , a subset 𝑉 of 𝑋 is a neighborhood

{(𝑥 − 𝑟, 𝑥 + 𝑟) ∶ 𝑟 > 0}, {[𝑥 − 𝑟, 𝑥 + 𝑟] ∶ 𝑟 > 0},

Proposition 2.3. A subset of a topological space 𝑋 is open if and only if it is a neighbor-

Proof. Let 𝑈 be a subset of 𝑋 . If 𝑈 is open then it is clear by definition that 𝑈 is a

(i) any subset of 𝑋 containing an element of 𝒱 (𝑥) belongs to 𝒱 (𝑥);

We easily verify that 𝒪 is a topology on 𝑋 : ∅, 𝑋 ∈ 𝒪 is trivial; the fact that finite

(i) 𝑥 belongs to all elements of 𝒮 (𝑥) ;

Equivalently, we can caractherise a basis for a topology in the following way.

2.2.1 Sub-basis and topology generated by a family of subsets

Definition 2.8. Let 𝒮 be a family of subsets of a set 𝑋 . The topology generated by 𝒮

It is easy to verify that the topology 𝒪𝒮 generated by 𝒮 corresponds to the collection

Proposition 2.9. Let ℬ be a family of subsets of a set 𝑋 satisfying

2.3 Interior and closure

Definition 2.10. One says that 𝑥 is an interior point of 𝐴 if 𝐴 is a neighborhood of 𝑥.

Definition 2.13. One says that 𝑥 is a boundary point of 𝐴 if every neighborhood of 𝑥

It is easy to verify that 𝜕𝐴 = 𝐴 ⧵ 𝐴̊ = 𝐴 ∩ 𝑋 ⧵ 𝐴.

Equivalently, 𝑥 is an isolated point of 𝐴 if and only it is not a limit point of 𝐴. Another

Definition 2.15. One says that 𝐴 is dense in 𝑋 if 𝐴 = 𝑋 .

Equivalently, a subset 𝐴 ⊆ 𝑋 is dense if and only if every non-empty open set in 𝑋

Proposition 2.16. Let 𝐴 and 𝐵 be subsets of a topological space 𝑋 , then:

2.4 Separation and countability

Definition 2.17. We say that a topological space 𝑋 is

2.4.2 Axioms of countability

Definition 2.18. We say that a topological space 𝑋 is

Proposition 2.19. Every second-countable space is first-countable and separable.

Proposition 2.20. Every metric space is first-countable.

Continuity and limits

3.1 Continuous maps

Remark that 𝑓 (𝑈 ) ⊆ 𝑉 is equivalent to 𝑈 ⊆ 𝑓 −1 (𝑉 ), and thus we can reformulate

∀𝜀 > 0, ∃𝛿 > 0, ∀𝑥 ∈ 𝑋 , 𝑑𝑋 (𝑥, 𝑎) < 𝛿 ⟹ 𝑑𝑌 (𝑓 (𝑥), 𝑓 (𝑎)) < 𝜀.

Example 3.1. A constant map between topological spaces is continuous.

It is easy to obtain that the composition of continuous maps is continuous.

Proposition 3.2. Let 𝑋 , 𝑌 and 𝑍 be topological spaces. Suppose that 𝑓 ∶ 𝑋 → 𝑌 is

Proof. Let 𝑉 be a neighborhood of (𝑔 ∘ 𝑓 )(𝑎) = 𝑔(𝑓 (𝑎)) in 𝑍 . Since 𝑔 is continuous at