Lecture 2 : Linear Maps

ENSIA, 2024
 Contents

• Definition
• Endomorphism, isomorphism and automorphism
• Composition of linear maps
• Image and Kernel
• The inverse of a linear map
• Rank Theorem
 Definition of a linear map

Throughout this lecture, 𝐾 will denote an arbitrary field.

Definition 1
Let V and W be K- vector spaces. A function 𝑓: 𝑉 ⟶ 𝑊 is said to be a linear map if
for any two vectors 𝑢, 𝑣 ∈ 𝑉 and any scalar 𝛼 ∈ 𝐾, the following two conditions are
satisfied :
1) 𝑓 𝑢 + 𝑣 = 𝑓 𝑢 + 𝑓 𝑣 ;
2) 𝑓 𝛼𝑢 = 𝛼𝑓(𝑢).
 Definition of a linear map

Remark 1
A linear map is also called homomorphism of vector spaces.

Exercise 1
Show that the following statements are equivalent :
1) The map 𝑓 is linear ;
2) ∀𝑢, 𝑣 ∈ 𝑉 and ∀𝛼 ∈ 𝐾, 𝑓 𝛼𝑢 + 𝑣 = 𝛼𝑓 𝑢 + 𝑓 𝑣 ;
3) ∀𝑢, 𝑣 ∈ 𝑉 and ∀𝛼, 𝛽 ∈ 𝐾, 𝑓 𝛼𝑢 + 𝛽𝑣 = 𝛼𝑓 𝑢 + 𝛽𝑓 𝑣 .
 Endomorphism, isomorphism and
automorphism
Definition 2
A linear map from a vector space 𝑉 to itself is called an endomorphism
of 𝑽.
A bijective linear map is called an isomorphism. Two 𝐾- vector spaces
are isomorphic if there exists an isomorphism between them.
When the endomorphism is bijective, it is called an automorphism
 Examples
Example 1
The identity map
𝐼𝑑𝑉 ∶ 𝑉 ⟶ 𝑉
𝑥 ↦ 𝐼𝑑𝑉 𝑥 = 𝑥
is a linear map and also an automorphism of the vector space 𝑉.
Example 2
The zero function
𝑓: 𝑉 ⟶ 𝑊
𝑥 ↦ 𝑓 𝑥 = 0𝑊
is linear.
 Examples
Example 3
The map
𝑓: ℝ ⟶ ℝ
𝑥 ↦ 𝑓 𝑥 = 2𝑥
is an automorphism.
Example 4
The map
𝑓: ℝ2 ⟶ ℝ
(𝑥, 𝑦) ↦ 𝑓 𝑥, 𝑦 = 2𝑥 + 𝑦
is linear
 Examples
Example 5
The map
𝑓: ℝ ⟶ ℝ
𝑥 ↦ 𝑓 𝑥 = sin 𝑥
is not linear.

Example 6
The map
𝑓: ℝ2 ⟶ ℝ
(𝑥, 𝑦) ↦ 𝑓 𝑥, 𝑦 = 𝑥𝑦
is not linear.
 Examples
Example 7
Consider the subspace
𝒞 ∞ ( 𝑎, 𝑏 ,ℝ) = 𝑓 ∈ ℱ 𝑎, 𝑏 , ℝ : 𝑓 is infinitely differentiable ,
and define
𝐷: 𝒞 ∞ ( 𝑎, 𝑏 ,ℝ) → 𝒞 ∞ ( 𝑎, 𝑏 ,ℝ)
𝑓↦𝐷 𝑓 ,
where
𝐷 𝑓 𝑥 = 𝑓 ′ 𝑥 , ∀𝑥 ∈ 𝑎, 𝑏 .
Then 𝐷 is a linear map.
 Examples
Example 8
Consider the subspace
𝒞( 𝑎, 𝑏 ,ℝ) = 𝑓 ∈ ℱ 𝑎, 𝑏 , ℝ : 𝑓 is continuous ,
and define
𝐼: 𝒞( 𝑎, 𝑏 ,ℝ) → 𝒞( 𝑎, 𝑏 ,ℝ)
𝑓↦𝐼 𝑓 ,
where
𝑥
𝐼 𝑓 (𝑥) = 𝑓 𝑡 𝑑𝑡, ∀𝑥 ∈ 𝑎, 𝑏 .
𝑎
Then 𝐼 is a linear map.
 Composition of linear maps

Theorem 1
Let 𝑓 ∶ 𝑈 ⟶ 𝑉 and 𝑔 ∶ 𝑉 ⟶ 𝑊 be two linear maps. Then the
composed function 𝑔 ∘ 𝑓 ∶ 𝑈 ⟶ 𝑊 is a linear map.
 Image and Kernel

Definition 3
Let 𝑓: 𝑈 → 𝑉 be a linear map. We define the image of 𝑓 by
Im 𝑓 = {𝑦 ∈ 𝑉 : ∃𝑥 ∈ 𝑈 such that 𝑦 = 𝑓(𝑥)},
and we define the kernel of 𝑓 by
Ker 𝑓 = 𝑥 ∈ 𝑈: 𝑓 (𝑥) = 0𝑉 .
 Some properties
Proposition 1
Let 𝑓: 𝑉 → 𝑊 be a linear map of 𝐾- vector spaces. Then we have :
1) 𝑓 0𝑉 = 0𝑊 ;
2) Ker𝑓 is a subspace of 𝑉 ;
3) Ker𝑓 = 0𝑉 if, and only if 𝑓 is injective ;
4) Im𝑓 is a subspace of 𝑊 ;
5) If dim 𝑉 < +∞ and 𝑣1 , 𝑣2 , ⋯ , 𝑣𝑛 is a basis of 𝑉 then
Im𝑓 = 𝑓(𝑣1 ), 𝑓(𝑣2 ), ⋯ , 𝑓(𝑣𝑛 ) .
 Example
Example 9
Let
𝑓: 𝐾3 𝑋 → 𝐾2 𝑋
𝑓 𝑃 = 𝑃′ .
We have
Ker 𝑓 = 𝑃 ∈ 𝐾3 𝑋 : 𝑓 (𝑃) = 0𝑉 = 1 ,
then 1 is a basis of Ker 𝑓, so dim(ker𝑓) = 1.
Im 𝑓 = 𝑓 1 , 𝑓 𝑋 , 𝑓 𝑋 2 , 𝑓 𝑋 3 = 1,2𝑋, 3𝑋 2 ,
then 1,2𝑋, 3𝑋 2 is a basis of Im 𝑓, so dim Im𝑓 = 3.
 Example
Example 10
Let
𝜑: 𝒞 ∞ ( 𝑎, 𝑏 ,ℝ) → 𝒞 ∞ ( 𝑎, 𝑏 ,ℝ)
𝑓 ↦ 𝜑 𝑓 = 𝑓 ′′ + 𝑓 ′ − 2𝑓.
Then 𝜑 is a linear map, and we have
Ker 𝜑 = 𝛼𝑔 + 𝛽𝑕 ∶ 𝛼, 𝛽 ∈ ℝ ,
where 𝑔 𝑥 = 𝑒 𝑥 and 𝑕 𝑥 = 𝑒 −2𝑥 for all 𝑥 ∈ ℝ.
 Inverse of a linear map

Theorem 2
Let 𝑓: 𝑈 → 𝑉 be a linear map. If 𝑓 is an isomorphism, then
𝑓 −1 : 𝑉 ⟶ 𝑈
is also an isomorphism.
 Values on a basis

Proposition 2
Let 𝑉 be a finite-dimensional 𝐾- vector space and 𝑣1 , 𝑣2 , ⋯ , 𝑣𝑛 be a
basis of 𝑉. Let 𝑊 be a 𝐾- vector space and 𝑤1 , 𝑤2 , ⋯ , 𝑤𝑛 be vectors of
𝑊. Then, there exists a unique linear map 𝑓 from 𝑉 to 𝑊 such that
𝑓 𝑣𝑖 = 𝑤𝑖 , 𝑖 = 1, ⋯ , 𝑛.
Remark 2
The proposition means that a linear map is entirely defined by its
values on a given basis.
 Linear maps and bases

Theorem 3
Let 𝑉 and 𝑊 be 𝐾- vector spaces and let 𝑣1 , 𝑣2 , ⋯ , 𝑣𝑛 be a basis of
𝑉. Let 𝑤1 , 𝑤2 , ⋯ , 𝑤𝑛 be vectors of 𝑊 and let 𝑓 be the linear map from
𝑉 to 𝑊 given by 𝑓 𝑣𝑖 = 𝑤𝑖 , 𝑖 = 1, ⋯ , 𝑛. Then we have :
1) 𝑓 is injective if, and only if, 𝑤1 , 𝑤2 , ⋯ , 𝑤𝑛 are linearly independent ;
2) 𝑓 is surjective if, and only if, 𝑤1 , 𝑤2 , ⋯ , 𝑤𝑛 span 𝑊 ;
3) 𝑓 is bijective if, and only if, 𝑤1 , 𝑤2 , ⋯ , 𝑤𝑛 is a basis of 𝑊.
 Isomorphic spaces
Corollary 1
Let 𝑉, 𝑊 be finite-dimensional vector spaces over 𝐾 and let 𝑓 be a
linear map from 𝑉 to 𝑊. The map 𝑓 is an isomorphism from 𝑉 to 𝑊 if,
and only if, the image of a basis of 𝑉 is a basis of 𝑊.
Corollary 2
Two isomorphic finite-dimensional vector spaces over 𝐾 have the same
dimension.
Corollary 3
Any vector space of finite dimension 𝑛 over 𝐾 is isomorphic to 𝐾 𝑛 .
 Rank Theorem
Definition 3
Let 𝑉 and 𝑊 be finite-dimensional vector spaces and 𝑓: 𝑉 → 𝑊 be a
linear map. The rank of 𝒇, denoted by 𝑟 𝑓 , is defined as the
dimension of Im𝑓.
Theorem 3
Let 𝑓: 𝑉 → 𝑊 be a linear map between finite-dimensional 𝐾-vector
spaces. Then
𝑟(𝑓) = dim 𝑉 − dim(ker𝑓).
 Rank Properties

Theorem 4
Let 𝑓: 𝑉 → 𝑊 be a linear map between finite-dimensional 𝐾-vector
spaces. Then we have :
1) 𝑟(𝑓) ≤ dim(𝑉) and 𝑟(𝑓) ≤ dim 𝑊 ;
2) 𝑟(𝑓) = dim(𝑉) ⟺ 𝑓 is injective ;
3) 𝑟(𝑓) = dim(𝑊) ⟺ 𝑓 is surjective.