5.

4 Eigenvectors and
Linear Transformations

4/21/2017 1
Eigenvectors and Linear transformations.

How is the factorization A = PDP −1 related to linear transformations?

Any linear transformation T : n →  m can be implemented by

We want to extend this kind of representation to any linear

4/21/2017 2
Suppose V is an n-dimensional vector space and W is an m-dimensional vector space.
Let T :V → W be a linear transformation. Suppose further that B and C are ordered bases
for V and W respectively.

 
For any x ∈V , the coordinate vector  x  B is in  n and the coordinate vector of its image,

T ( x )  , is in  m .
 C 

V  W 
x T ( x)


T ( x ) 

x   C
 B

n m

4/21/2017 3
 
     
Let  b1 , , bn  be the basis B for V . If x = r1b1 +  + rnbn , then


r 
  1
 x  =   
B  
r 
 n 

and, since T is linear,

    
T ( x=) T (r1b1 +  + rnb=n ) r1T (b1 ) +  + rnT (bn ) (*)

Now using the basis C in W , we can write (*) in terms of

  
T=

( x ) r1 T (b1) + + rn T (bn ) (**)
C  C  C

4/21/2017 4
 But since C -coordinate vectors are in  m , the vector equation

 
  
 
T=
( x ) r1 T (b1) + + rn T (bn )
 C  C  C

can be written as a matrix equation

 
T ( x ) 
 C
= M  x  B

   
  
  
with M =  T (b1) T (b2 )   T (bn )  
  C C C 

The matrix M is a matrix representation of T called the matrix

4/21/2017 5
  
T ( x ) 
 C
= M  x  B

   
  
  
M =  T (b1 ) T (b2 )   T (bn )  
  C C C 

4/21/2017 6
  
Example: Suppose B { }
b=1 , b2 is a basis for V and C
  
{c1, c2 , c3} is
a basis for W . Let T : V → W be a linear transformation with the
       
property that T (b1 ) =3c1 − 2c2 + 5c3 and T (b2 ) = 4c1 + 7c2 − c3
Find the matrix M for T relative to B and C.
 
The C -coordinate vectors for the images of b1 and b2 are

  3   4
       
T (b )
1  =−2
  and T (b )
2  =
 7
C  5 C  −1
   

 3 4
 
Then the matrix M =  −2 7
 5 −1


4/21/2017 7
 Linear transformations on  n .

Theorem. Diagonal matrix representation.

If B is a basis for  n formed from the columns of P,

4/21/2017 8
Linear transformations T :V → V , f rom a vector space V into itself.

When the domain and codomain of T are the same, that is when
T :V → W and W is the same as V and basis C is the same as B,
the matrix

   
  
  
M =  T (b1 ) T (b )
2   T (bn )  
  C C C 

is called the B-matrix for T , written T  B .

The B-matrix for T :V → V satisfies the equation

  
T ( x ) 
 B
= T  B  x  B , ∀x ∈V .

4/21/2017 9
 Matrix similarity

Let A and B be n × n matrices. A is said to be similar to B if

=P −1 AP B=
or equivalently, A PBP −1.

If we write Q for P −1, we have Q −1BQ = A.

So B is similar to A as well. Thus A and B are simply said to

Theorem

If n × n matrices A and B are similar, then they have the

4/21/2017 10
 End presentation

4/21/2017 11