LINEAR ALGEBRA WITH PROGRAMMING

Chapter 1
VECTORS AND MATRICES
 LINEAR ALGEBRA WITH PROGRAMMING

1.3 Linear Transformations

FUNCTIONS ON PYTHON
LINEAR ALGEBRA WITH PYTHON

Idea
Seeing matrices as “functions”…

The vector 𝒙 is transformed into the vector 𝐴𝒙

𝒙 𝑨𝒙

ℝ𝑛 - Domain ℝ𝑚 - Codomain
Linear transformation

A map 𝑇: ℝ𝑛 ⟶ ℝ𝑚 is called a linear transformation if there is a 𝑚 × 𝑛 matrix 𝐴 𝑇 , such

that 𝑇 is defined by
𝑇 𝒙 = 𝐴𝑇 𝒙
for all 𝒙 ∈ ℝ𝑛 .
In that case, the matrix 𝐴 𝑇 is unique and it is called the transformation matrix of 𝑇.

4
Example −1 2
Consider the linear transformation 𝑇: ℝ2 ⟶ ℝ3 with transformation matrix 𝐴 𝑇 = 1 2 .
0 5

• What is the image of the vector 𝑢 = (3, 1) through 𝑇 ?

−1 2 −1
3
𝑇 𝑢 = 𝐴𝑇 𝑢 = 1 2 = 5 = (−1, 5, 5)
1
0 5 5

• What are the objects which image through 𝑇 is 𝑣 = (3, 1, 5) ?

−1 2 𝑥 3 −1 2 3 1 0 −1 𝑥1 = −1
1
𝑇 𝒙 = 𝑣 ⇔ 𝐴𝑇 𝒙 = 𝑣 ⇔ 1 2 𝑥 = 1 ⇝ 1 2 1 ⟶ ⋯ ⟶ 0 1 1 ⇝ ቐ 𝑥2 = 1
2
0 5 5 0 5 5 0 0 0 0=0

Only the vector 𝒙 = (−1, 1) has image equal to 𝑣 through 𝑇.

• What is the general expression that defines 𝑇 ?

−1 2 𝑥 −𝑥1 + 2𝑥2
1
𝑇 𝒙 = 𝐴 𝑇 𝒙 = 1 2 𝑥 = 𝑥1 + 2𝑥2 = (−𝑥1 + 2𝑥2 , 𝑥1 + 2𝑥2 , 5𝑥2 )
2
0 5 5𝑥2
5
Kernel and range
Let 𝑇: ℝ𝑛 ⟶ ℝ𝑚 be a linear transformation.

• The kernel of 𝑇, denoted by Ker(𝑇), is the set of vectors in ℝ𝑛 which are mapped
onto 0𝑚×1 through 𝑇,
Ker 𝑇 = 𝒙 ∈ ℝ𝑛 : 𝑇 𝒙 = 0𝑚×1 ⊆ ℝ𝑛 ;

• The range of 𝑇, denoted by Range(𝑇), is the set of images through 𝑇 of vectors in

ℝ𝑛 ,
Range 𝑇 = 𝑇 𝒙 : 𝒙 ∈ ℝ𝑛 ⊆ ℝ𝑚 .

6
Example
How can we find the transformation matrix of the linear transformation 𝑇: ℝ3 ⟶ ℝ2 defined by

𝑇(𝑥1 , 𝑥2 , 𝑥3 ) = (𝑥1 − 3𝑥2 , 2𝑥2 − 4𝑥3 )

for all (𝑥1 , 𝑥2 , 𝑥3 ) ∈ ℝ3 ?

𝑥 𝑥1
𝑥 − 3𝑥2 1 −3 0 𝑥1
𝑇 𝑥1 , 𝑥2 , 𝑥3 = 𝑥1 − 3𝑥2 , 2𝑥2 − 4𝑥3 = 1 = 2 = 𝐴 𝑇 𝑥2
2𝑥2 − 4𝑥3 0 2 −4 𝑥 𝑥3
3

1 −3 0
∴ 𝐴𝑇 =
0 2 −4

Notice that:

1 1 0 −3 0 0
𝑇 0 = , 𝑇 1 = , 𝑇 0 =
0 0 0 2 −4
1
7
Proposition

If 𝑇: ℝ𝑛 ⟶ ℝ𝑚 is a linear transformation, then the following conditions hold:

1. 𝑇 𝑢+𝑣 =𝑇 𝑢 +𝑇 𝑣 , ∀𝑢, 𝑣 ∈ ℝ𝑛 ;
2. 𝑇 𝑘𝑢 = 𝑘 𝑇 𝑢 , ∀𝑘 ∈ ℝ, ∀𝑢 ∈ ℝ𝑛 ;
3. 𝑇 0𝑛×1 = 0𝑚×1 .

8
Remark
If 𝑇: ℝ𝑛 ⟶ ℝ𝑚 is a linear transformation, notice that:

𝑥1 1 0 0 1 0 0
𝑥2 0 +𝑥 1 + ⋯+ 𝑥 0 0 +𝑥 𝐴 1 + ⋯+ 𝑥 𝐴 0
𝑇 𝒙 = 𝐴𝑇 ⋮ = 𝐴 𝑇 𝑥1 2 𝑛 = 𝑥1 𝐴 𝑇 2 𝑇 𝑛 𝑇
⋮ ⋮ ⋮ ⋮ ⋮ ⋮
𝑥𝑛 0 0 1 0 0 1

9
Proposition

A linear transformation 𝑇: ℝ𝑛 ⟶ ℝ𝑚 is completely determined by the images of the

columns of the identity matrix 𝐼𝑛 .

10
Example
Consider 𝑇: ℝ3 → ℝ3 , such that 𝑇 𝑥, 𝑦, 𝑧 = 2𝑦, 𝑥, −𝑧 , ∀(𝑥, 𝑦, 𝑧) ∈ ℝ3 .
We have
𝑇 1,0,0 = 0,1,0 ; 𝑇 0,1,0 = 2,0,0 ; 𝑇(0,0,1) = (0,0, −1).
Then
0 2 0
𝐴𝑇 = 1 0 0 .
0 0 −1
Notice that, for example,
0 2 0 2 8 4 0 2 0 1
𝑇 2,4, −6 = 1 0 0 4 = 2 =2 1 =2 1 0 0 2 = 2𝑇 1,2, −3
0 0 −1 −6 6 3 0 0 −1 −3
or, for example,
0 2 0 2 8 0 8 0 2 0 2 0 2 0 0
𝑇 2,4, −6 = 1 0 0 4 = 2 = 2 + 0 = 1 0 0 0 + 1 0 0 4
0 0 −1 −6 6 0 6 0 0 −1 0 0 0 −1 −6
= 𝑇 2,0,0 + 𝑇 0,4, −6

11
Proposition

A linear transformation 𝑇: ℝ𝑛 ⟶ ℝ𝑚 is invertible if and only if 𝑚 = 𝑛 and the matrix 𝐴 𝑇

is invertible.
In that case,
𝐴 𝑇 −1 = (𝐴 𝑇 )−1 .

12
Example
Let 𝑇: ℝ3 → ℝ3 defined by 𝑇(𝑥, 𝑦, 𝑧) = (𝑦 + 𝑧, 2𝑥, 4𝑧) for all (𝑥, 𝑦, 𝑧) ∈ ℝ3 .

1
0 0
−1 2
0 1 1 0 1 1 1
𝐴𝑇 = 2 0 0 ⟹ 𝐴 𝑇 −1 = 2 0 0 = 1 0 − .
0 0 4 0 0 4 4
1
0 0
4

Then 𝑇 is invertible and the linear transformation 𝑇 −1 : ℝ3 → ℝ3 is defined by

1
0 0
2 𝑥
−1 1 𝑦 𝑧 𝑧
𝑇 𝑥, 𝑦, 𝑧 = 1 0 − 𝑦 = ,𝑥 − ,
4 𝑧 2 4 4
1
0 0
4
for all (𝑥, 𝑦, 𝑧) ∈ ℝ .
3

13
Proposition

Let 𝑇1 : ℝ𝑛 ⟶ ℝ𝑝 and 𝑇2 : ℝ𝑝 ⟶ ℝ𝑚 be linear transformations.

The transformation matrix of the composition 𝑇2 ∘ 𝑇1 can be find by

𝐴 𝑇2 ∘ 𝑇1 = 𝐴 𝑇2 𝐴 𝑇1 .

14
 CHALLENGE
Define on Python the linear transformation 𝑇 of the
previous example and check the image through 𝑇 of
the vector (23,477,100).
 Exercises
1.4.1 Find the transformation matrices of the linear transformations defined below.
a) 𝑇: ℝ3 → ℝ2 , 𝑇 𝑥1 , 𝑥2 , 𝑥3 = 𝑥1 − 5𝑥2 + 4𝑥3 , 𝑥2 − 6𝑥3
b) 𝑇: ℝ2 → ℝ4 , 𝑇 𝑥1 , 𝑥2 = 2𝑥2 − 3𝑥1 , 𝑥1 − 4𝑥2 , 0, 𝑥2
c) 𝑇: ℝ4 → ℝ4 , 𝑇 𝑥1 , 𝑥2 , 𝑥3 , 𝑥4 = 0, 𝑥1 + 𝑥2 , 𝑥2 + 𝑥3 , 𝑥3 + 𝑥4

1.4.2. Show that the transformations below are not linear.

a) 𝑇: ℝ2 → ℝ3 , 𝑇 𝑥1 , 𝑥2 = 2𝑥1 − 3𝑥2 , 𝑥1 + 4, 5𝑥2
b) 𝑇: ℝ2 → ℝ2 , 𝑇 𝑥1 , 𝑥2 = 4𝑥1 − 2𝑥2 , 3 𝑥2
c) 𝑇: ℝ4 → ℝ2 , 𝑇 𝑥1 , 𝑥2 , 𝑥3 , 𝑥4 = 𝑥1 𝑥2 , 𝑥3 + 𝑥4 ,

16
 Exercises
1.4.3 Consider two linear transformations 𝑇: ℝ3 → ℝ3 and 𝐺: ℝ3 → ℝ3 . It is known that
𝑇 1,1,0 = 1,0,1 , 𝑇 1,0,0 = 1,0,2 , 𝑇 0,1,1 = 0,1,0 and 𝐺 𝑥, 𝑦, 𝑧 = 𝑥, 2𝑥 + 𝑦 − 𝑧, 𝑦 .
a) Find 𝑇 0,1,0 .
b) Find 𝐴 𝑇 .
c) Is the 𝐺 the inverse of 𝑇?

1.4.4. Consider two linear transformations 𝑇1 : ℝ → ℝ3 and 𝑇2 : ℝ3 → ℝ3 defined as:

𝑇1 𝑥 = 𝑥, 0,2𝑥 and 𝑇2 𝑥, 𝑦, 𝑧 = 𝑦 − 𝑧, 𝑥 + 2𝑦 − 𝑧, −𝑥 − 𝑦 .
a) Find (𝑇2 ∘ 𝑇1 ) 3 .
b) Find the transformation matrices of 𝑇1 , 𝑇2 and 𝑇2 ∘ 𝑇1 .
c) Show that 𝑇2 is not surjective, i.e., there are vectors of ℝ3 that are not images
of any vector of the domain of 𝑇2 .

17
 Exercises
1.4.5 Let 𝑇: ℝ3 → ℝ3 and 𝑆: ℝ2 → ℝ3 be linear transformations, 𝑆 ∘ 𝑇 their composition,
−2 −1 3
𝑇 𝑥1 , 𝑥2 , 𝑥3 = 𝑥1 , 𝑥2 − 𝑥3 , 𝑥3 − 𝑥1 and 𝐴𝑆∘𝑇 = 1 −1 1 .
1 0 −1
a) Find the transformation matrix of 𝑇.
b) Find the transformation matrix of 𝑆.

1.4.6. Consider 𝑇: ℝ4 → ℝ3 , defined by 𝑇 𝑥1 , 𝑥2 , 𝑥3 , 𝑥4 = 2𝑥1 + 𝑥3 + 3𝑥4 , 2𝑥1 + 2𝑥4 , 𝑥3 + 𝑥4 .

a) Find 𝐴𝑇 , the transformation matrix of 𝑇.
b) Find the general expression of the linear transformation 𝑆 whose
transformation matrix, 𝐴𝑆 , is the transpose of 𝐴𝑇 .
c) Verify if different objects have necessarily different images through 𝑇.
d) Consider 𝒗 = 𝑎, 𝑏, 𝑐 where 𝑎, 𝑏, 𝑐 are the last three digits of your student’s
number. Verify if there is an object which image through 𝑇 is 𝒗.

18
 Exercises
1.4.7 For 𝑘 ∈ ℝ, consider 𝑇: ℝ3 → ℝ3 , defined by 𝑇 𝑥, 𝑦, 𝑧 = −2𝑥 + 𝑦, 2𝑥 + 𝑦 + 𝑧𝑘, 𝑦 + 𝑘𝑧 .
a) Find the transformation matrix of 𝑇 as a function on 𝑘.
b) For which values of 𝑘 is 𝑇 invertible?

1.4.8. True or false?

a) There are linear transformations from ℝ4 to ℝ2 , such that to different objects
necessarily correspond different images.
b) There is no linear transformation 𝑇 that transforms the vectors 1,1,0 , 1,0,1
and (2,1,1) into the vectors 1,0,0 , 0,1,0 and (0,0,1), respectively.

19
1.4.9 Carefully explain the output of the following programs. Exercises
i. A

ii. B

iii. C

20