Download as pdf or txt
Download as pdf or txt
You are on page 1of 27

Linear transformation

Phan Thi Khanh Van

E-mail: khanhvanphan@hcmut.edu.vn

May 13, 2021

(Phan Thi Khanh Van) Linear transformation May 13, 2021 1 / 27


Table of contents

1 Linear transformation

2 The transformation matrix

3 Change of basis matrix and transformation matrix

4 Kernel and Image of a linear transformation

(Phan Thi Khanh Van) Linear transformation May 13, 2021 2 / 27


Linear transformation
Let U, V be 2 vector spaces over the field K(R, C). f : U → V is called a
linear transformation if:
∀u1 , u2 ∈ U, f (u1 + u2 ) = f (u1 ) + f (u2 ).
∀u ∈ U, α ∈ K : f (αu) = αf (u).

Example 1
Let f : R2 → R2 be a linear transformation satisfying:
f (x1 , x2 ) = (2x1 + x2 , x1 − 3x2 ). Find f (3, 4).

f (3, 4) = (2.3 + 4, 3 − 3.4) = (10, −9).


Represent f as matrix
 multiplication:
  
2x1 + x2 2 1 x1
[f (x)] = = = A[x].
x1 − 3x2 1 −3 x2

(Phan Thi Khanh Van) Linear transformation May 13, 2021 3 / 27


Example 2
Let f : R3 → R2 be a linear transformation satisfying:
f (1, 1, 3) = (1, 0), f (1, −1, 0) = (2, −3), f (2, 0, 0) = (1, 1). Find
a) f (2, 0, 3)
b) f (−3, −1, 6) = f (2(1, 1, 3) + 3(1, −1, 0) − 4(2, 0, 0))
c) f (2, 3, −4).

a) (2, 0, 3) = (1, 1, 3) + (1, −1, 0)


⇒ f (2, 0, 3) = f (1, 1, 3) + f (1, −1, 0)
= (1, 0) + (2, −3) = (3, −3).
b) f (−3, −1, 6) = f (2(1, 1, 3) + 3(1, −1, 0) − 4(2, 0, 0))
= 2f (1, 1, 3) + 3f (1, −1, 0) − 4f (2, 0, 0) = 2(1, 0) + 3(2, −3) − 4(1, 1)
= (4, −13).

(Phan Thi Khanh Van) Linear transformation May 13, 2021 4 / 27


c) We have that E = {(1; 1; 3), (1; −1; 0), (2; 0; 0)} is a basis of R 3
(det(E ) 6= 0).
 −1    4 
1 1 2 2 −3
[(2, 3, −4)]E = 1 −1 0  3  = − 13 3 .

3 0 0 −4 23
6
Or (2, 3, −4) = − 43 (1, 1, 3) − 13 3 (1, −1, 0) + 236 (2, 0, 0).
Consequently,
[f (2, 3, −4)] = − 43 [f (1, 1, 3)] − 13 [f (1, −1, 0)] + 23
6 [f (2, 0, 0)]
      3  37 
1 2 1 −
= − 34 − 13
3 −3 + 6 1 =
23
101
6
0 6

(Phan Thi Khanh Van) Linear transformation May 13, 2021 5 / 27


Example 3
Let f : R3 → R2 be a linear transformation satisfying:
f (1, 1, 3) = (1, 0), f (1, −1, 0) = (2, −3), f (2, 0, 0) = (1, 1). Find
f (x1 , x2 , x3 ).

E = {(1; 1; 3), 3
 −1; 0), (2; 0; 0)} is a basis of R .
 (1;
α
Put [x]E = β  = E −1 [x].
γ
[x] = α[e1 ] + β[e2 ] + γ[e3 ]
⇒ [f (x)] = α[f (e1 )] + β[f(e2 )] + γ[f (e3 )]
 α
= f (e1 ) f (e2 ) f (e3 ) β  = f (e1 ) f (e2 ) f (e3 ) .E −1 [x]
  
γ
 1 1 2 −1 x1
   
  x1 −3x2 2x3 
1 2 1  + +
= . 1 −1 0 x2  = x21 7x22 −4x33
0 −3 1 + +
3 0 0 x3 2 2 3
x1 3x2 2x3 x1 7x2 4x3

Therefore, f (x1 , x2 , x3 ) = 2 − 2 + 3 , 2 + 2 − 3 .
[f (x)] = A[x] = fE .E −1 [x]
(Phan Thi Khanh Van) Linear transformation May 13, 2021 6 / 27
Example
Given a triangle ABC : A(0, 1), B(2, 3), C (4, 2). Find the image of the
triangle under the transformation:

a) Rotation about the origin 60o


counterclockwise.
b) Reflection in the line y = −3x.
c) Shear transformation with the
fixed line x−axis:
S(x, y ) = (x + 3y , y )

(Phan Thi Khanh Van) Linear transformation May 13, 2021 7 / 27


 
~ ~ ~ 0 2 4
The matrix of coordinates of the triangle: [OA OB OC ] = .
1 3 2
a) Rotation about the origin 60o counterclockwise: π3 . The standard
basis of R2 : E = {i = (1, 0), j = (0, 1)T }.

cos( π3 )
 " √
3
√ √ #
f (i) = − 1√− 3 2 3 2√− 3
sin( π3 ) = 1
2
2 3 + 32 2 3+1
− sin( π3 )
 
f (j) =
cos( π3 )
Rotation matrix: " √ #
1 3

R = f (E ).E −1 = √2
3 1
2
2 2
The coordinates of the image:
~ 0 OB
[OA ~ 0 OC
~ 0 ] = R.[OA
~ OB
~ OC
~ ]

(Phan Thi Khanh Van) Linear transformation May 13, 2021 8 / 27


b) Reflection about the line y = −3x.
Choose the directional vector: e1 = (1, −3), and the normal vector
e2 = (3, 1) to form a basis of R2 .

− 53 − 17 − 22
  
1 5 5
f (e1 ) = e1 == = 4 6
−3
  5 5 − 45
−3
f (e2 ) = −e2 =
−1
Reflection matrix:  4
− 5 − 53

T = fE .E = −1
− 35 4
5
The coordinates of the image:
[OA~ 0 OB
~ 0 OC
~ 0 ] = R.[OA~ OB~ OC
~ ]

(Phan Thi Khanh Van) Linear transformation May 13, 2021 9 / 27


b) Shear transformation
 S(x, y ) = (x + 3y , y )
1 3
Shear matrix: S =
0 1
 
3 11 10
The coordinates of the image: S(X ) = S.X =
1 3 2

(Phan Thi Khanh Van) Linear transformation May 13, 2021 10 / 27


Example 4
The projection f in Oxyz with the dot product onto the plane
P : x − y + 2z = 0 is a linear transformation. Find f (x, y , z)

The equation x − y + 2z = 0 has the general solution:


u = m(−2, 0, 1) + n(1, 1, 0).
Then, one basis of P: {e1 = (−2, 0, 1), e2 = (1, 1, 0)}.
f (e1 ) = e1 = (−2, 0, 1)
f (e2 ) = e2 = (−2, 0, 1).
Choose e3 to be the normal vector of the plane: e3 = (1, −1, 2). Vector
projection of e3 onto P is 0: f (e3 ) = 0.
  −1  
−2 1 0 −2 1 1 x
Then, [f (x, y , z)] = fE .E −1 [X ] =  0 1 0  0 1 −1 y 
1 0 0 1 0 2 z
 5x y −z

6 + 6 + 3
=  x6 + 5y z .
6 + 3
−x y z
3 + 3 + 3

(Phan Thi Khanh Van) Linear transformation May 13, 2021 11 / 27


Exercises
1 In R3 with the dot product, find the reflection mapping in the plane
x + 2y + z = 0.
2 In R3 with the dot product, find the rotation mapping about the z axis
by 45o counterclockwise.

(Phan Thi Khanh Van) Linear transformation May 13, 2021 12 / 27


The transformation matrix
Let f : U → V be a linear transformation. Let E = {e1 , e2 , ..., en } and
F = {f1 , f2 , ..., fm } be bases of U and V , respectively.
An m × n matrix whose the i-th column is the coordinate vector of f (ei )
with respect to the basis F is called  the transformation matrix of f with
respect to E , F . Denote: AEF = [f (e1 )]F [f (e2 )]F . . . [f (en )]F
[f (x)]F = AEF [x]E

Remark: If f : Rn → Rm : [f (ei )]F = F −1 [f (ei )], then,


AEF = F [f (e1 )] F −1 [f (e2 )] . . . F −1 [f (en )] = F −1 f (E ).
 −1

Remark:
If f : Rm → Rn , E , F are two corresponding bases. Then,
AEF = F −1 .f (E )
.
If f : Rn → Rn , E is a basis of R n . Then,
AE = E −1 .f (E )
.
(Phan Thi Khanh Van) Linear transformation May 13, 2021 13 / 27
Example
Let f be a linear transformation
f : R3 → R2 : f (x1 , x2 , x3 ) = (2x1 + x2 + x3 , x1 − x2 − x3 ). Find the
transformation matrix of f w.r.t E = {(1, 2, 1), (1, 1, 0), (1, 0, 0)},
F = {(1, 3), (2, 4)}
 
  x1
2 1 1  
[f (x)] = A[x] = x2
1 −1 −1
x3
AEF = F −1 .fE = F −1 .A.E  
 −1   1 1 1  
1 2 2 1 1 −12 −6 −3
= . . 2 1 0 = 17
  9 5
3 4 1 −1 −1 2 2 2
1 0 0

(Phan Thi Khanh Van) Linear transformation May 13, 2021 14 / 27


Theorem
Let E and F be bases of U and V , respectively.
For any linear transformation f : U → V , there exists a unique matrix
AEF satisfying: ∀x ∈ U : [f (x)]F = AEF [x]E .
For any m × n matrix AEF , there exists a unique linear transformation
f : U → V such that:
∀x ∈ U : [f (x)]F = AEF [x]E .

Conclusion: For any linear transformation f , we can find 1 and only 1


matrix A : f (v ) = Av . If f is invertible, then f −1 has the matrix A−1 . The
product of 2 transformations f1 : f1 (v ) = A1 v and f2 : f2 (v ) = A2 v
corresponds to A1 A2 . This is where the matrix multiplication came from!

(Phan Thi Khanh Van) Linear transformation May 13, 2021 15 / 27


Example
Let f be a linear transformation f : R3 → R2 . The transformation matrix
 E = {(1,
of f w.r.t.  1, −1), (2, 1, 0), (1, 0, 0)}, F = {(1, 2), (−2, 3)} is
2 −1 3
AEF = . Find f (x).
1 2 4

We have: [f (x)]F = AEF .[x]E = AEF .E −1 [x]


Then [f (x)] = F .[f (x)]F = F .AEF .E −1 [x]
 −1  
    1 2 1 x1
1 −2 2 −1 3 
= . . 1 1 0 . x2 
2 3 1 2 4
−1 0 0 x3
 
  x1  
−5 5 0 −5x1 + 5x2
[f (x)] = . x2 =
 
18 −32 −21 18x1 − 32x2 − 21x3
x3

(Phan Thi Khanh Van) Linear transformation May 13, 2021 16 / 27


Example
Let f be a linear transformation
f : R3 → R3 : f (x1 , x2 , x3 ) = (2x1 + x2 + x3 , x1 − x2 − x3 , 4x1 − x2 − x3 ).
Find the matrix of f in E = {(1, 1, 1), (1, 1, 0), (1, 0, 0)},
F = {(1, 1, 2), (1, −2, 1), (0, 0, 3)}
 
2 1 1
[f (x)] = 1 −1 −1 .[x] = A.[x] ⇒ f (E ) = A.E
4 −1 −1
We have: AEF = F −1 f (E ) = F −1 .A.E
 −1   
1 1 0 2 1 1 1 1 1
= 1 −2 0 1 −1 −1 1 1 0 = ...
2 1 3 4 −1 −1 1 0 0

(Phan Thi Khanh Van) Linear transformation May 13, 2021 17 / 27


Change of basis matrix and transformation matrix
Let f : U → V be a linear transformation; E and E 0 be 2 bases of U; F
and F 0 be 2 bases of V .
Transformation matrix of f w.r.t. E , F is AEF . Change of basis matrices:
PE 0 →E = E −1 E 0 , and QF 0 →F = F −1 F 0 .
Then, the transformation matrix of f w.r.t. E 0 , F 0 is:

AE 0 F 0 = QF−1
0 →F AEF PE 0 →E

Remark
Let f : U → U be a linear transformation; E and E 0 be 2 bases of U. The
change of basis matrix: PE 0 →E = E −1 E 0 .
Then, the transformation matrix of f with respect to E 0 :

AE 0 = PE−1
0 →E AE PE 0 →E

(AE 0 and AE are 2 similar matrices).

(Phan Thi Khanh Van) Linear transformation May 13, 2021 18 / 27


Example
Given a linear transformation f : R3 → R3 . Matrix of f w.r.t

2 −1 3
E = {(1, 1, −1), (2, 1, 0), (1, 0, 0)} is AE = 1 2 4 . Find the matrix
1 −3 −1
representation of f w.r.t the standard basis (standard matrix representation)
   
1 2 1 12 −24 −17
E =  1 1 0 ⇒ A = E .AE .E −1 =  7 −13 −9  .
−1 0 0 −3 7 6

(Phan Thi Khanh Van) Linear transformation May 13, 2021 19 / 27


Example
Let f be a linear transformation f : R3 → R3 . The matrix
 of f in
2 −1 3
E = {(1; 1; 1), (1; 2; −1), (0; 2; 1)} is AE = 1 2 0. Find the matrix
1 0 1
of f w.r.t F = {(1; −1; −1), (1; 1; 1), (0; 0; 1)}.

AF =F −1 E .AE .E −1
 .F , where:
 
1 1 0 1 1 0
E = 1 2 2, F = −1 1 0
1 −1 1 −1 1 1

(Phan Thi Khanh Van) Linear transformation May 13, 2021 20 / 27


Kernel and Image of a linear transformation
Let f be a linear transformation f : U → V .
Kernel of the linear transformation f : Kerf = {x ∈ U|f (x) = 0}
Image of the linear transformation f : Imf = {y ∈ V |∃x ∈ U : y = f (x)}

Theorem
Let f be a linear transformation f : U → V .
Kerf is a subspace of U.
Imf is a subspace of V .
dim(Imf ) + dim(Kerf ) = dim(U)

Proposition
Let f : U → V be a linear transformation.
If U = span{e1 , e2 , ...en }, then Imf = span{f (e1 ), f (e2 )...f (en )}.

(Phan Thi Khanh Van) Linear transformation May 13, 2021 21 / 27


Example
Let f be a linear transformation
f : R3 → R3 : f (x1 , x2 , x3 ) = (2x1 + x2 + x3 , x1 − x2 − x3 , 4x1 − x2 − x3 ).
Find the dimensions, bases of Kerf , Imf .
Find a basis of Kerf :
x ∈ Kerf ⇔ [f (x)] =0⇔  A[x] = 0 
2 1 1 0 0 3 3 0  
 1 −1 −1 0  ∼  1 −1 −1 0  ∼ 1 −1 −1 0
.
0 1 1 0
4 −1 −1 0 0 3 3 0
General solution: x = (0, −m, m).
One basis for Kerf : {(0, −1, 1)}, dim(Kerf ) = 1
Find a basis of Imf :
Choose a basis of R3 : {(1, 0, 0), (0, 1, 0), (0, 0, 1)}
⇒ A spanning set of Imf : {f (1, 0, 0), f (0, 1, 0), f (0, 0, 1)}
= {(2, 1, 4),
 (1, −1, −1),(1, −1,
 −1)} 
2 1 1 2 1 1
We have 1 −1 −1 ∼ 0 −3 −3
4 −1 −1 0 −3 −3
One basis of Imf : {(2, 1, 4), (1, −1, −1)}, dim(Imf ) = 2
(Phan Thi Khanh Van) Linear transformation May 13, 2021 22 / 27
Example
Given f : R3 → R3 : f (1, 1, 1) = (1, 2, 1), f (1, 1, 2) = (2, 1, −1), f (1, 2, 1) =
(5, 4, −1). Find the dimensions, bases of Kerf , Imf .

We have that E = {e1 = (1, 1, 1), e2 = (1, 1, 2), e3 = (1, 2, 1)} is a basis of
R3 .
Then {f (e1 ), 
 f (e2 ),f (e3 )} is a spanning
  set of Imf.
1 2 5 1 2 5 1 2 5
2 1 4  ∼ 0 −3 −6 ∼ 0 −3 −6
1 −1 −1 0 −3 −6 0 0 0
One basis for Imf : {(1, 2, 1), (2, 1, −1)}, dim(Imf ) = 2

(Phan Thi Khanh Van) Linear transformation May 13, 2021 23 / 27


Find a basis of Kerf :
We have: x ∈ Kerf ⇔ [f (x)] = A[x] = 0 ⇔ f (E )E −1 [x] = 0
  −1    
1 2 5 1 1 1 x1 −4 4 1 0
⇔ 2 1 4  1 1 2 x2  = 0 ⇔  1 2 −1 0 
1 −1 −1 1 2 1 x3 5 −2 −2 0
 
1 2 −1 0
∼  0 12 −3 0 
0 12 −3 0
 

1 2 −1 0

0 4 −1 0
General solution: x = (2m, m, 4m).
One basis for Kerf : {(2, 1, 4)}, dim(Ker (f )) = 1

(Phan Thi Khanh Van) Linear transformation May 13, 2021 24 / 27


Exercises
1 The projection f in Oxyz onto the plane P : x − y + 2z = 0 is a linear
transformation. Find the dimensions, bases of Imf , Kerf
2 The reflection f in Oxyz with respect to the plane x + y + z = 0 is a
linear transformation. Find the dimensions, bases of Imf , Kerf
3 The rotation f in Oxyz with respect to z-axis by 45o counterclockwise
is a linear transformation. Find the dimensions, bases of Imf , Kerf

(Phan Thi Khanh Van) Linear transformation May 13, 2021 25 / 27


Example
Given f : R3 → R3 . The matrix of f in the  basis 
2 −1 3
E = {(1, 1, 1), (1, 1, 0), (1, 0, 0)} is AE = 1 2 4 . Find the
1 −3 −1
dimensions, bases of Kerf , Imf .
 
6 −8 6
A = E .AE .E −1 = 7 −6 2.
3 −4 3
 
6x1 − 8x2 + 6x3
Hence, [f (x)] = A.[x] = 7x1 − 6x2 + 2x3  Continue!!!
3x1 − 4x2 + 3x3

(Phan Thi Khanh Van) Linear transformation May 13, 2021 26 / 27


Thank you for your attention!

(Phan Thi Khanh Van) Linear transformation May 13, 2021 27 / 27

You might also like