Announcements

1. Tutorial classes on Tuesday next week.

2. Lab quiz duration: 50 minutes
3. Please bring along your student (matriculation)
card.
4. Please be reminded to check your group and
arrive on time.
5. Open ‘everything’ quiz.
 MA1101R Lab Quiz
• Once you are inside the lab, find a workstation and get seated.
• If this is NOT your official lab timeslot, you are NOT allowed to
take the quiz (unless you have obtained permission).
• Once at your seat, you may take out your belongings and log in to
the PC using your NUSNET ID and password.
• Take out your student card and place it on your desk.
• Read the instructions on the cover page of your question paper.
• Once your PC is ready, you may start MATLAB. Remember to
change the directory using the command cd c:/ma1101r
• You may flip over (after reading the instructions) and read the
question but you cannot start writing until you are told to do so.
• This is an open book quiz, but no communication (e.g. email,
MSN) is allowed.
• Read the question and enter the matrices CAREFULLY!
Lecture 20 Recap
1) Definition of a linear transformation.
2) The standard matrix and formula for a linear
transformation.
3) Abstract definition of a linear transformation.
4) Two properties that linear transformations have.

5) What do we need to know to completely determine

a linear transformation T : n
 m
.
6) Composition of linear transformations.
Lecture 21
Ranges and Kernels
Definition
Let T : n
 m
be a linear transformation.

R(T )  {T (u)| u  } 
n m
T
R(T )  { } 'set of all

n m images'

The range, denoted by R(T ) of T is the set of images of T .

Example
Let I : n
 n
be the transformation defined by
'what you put into I ,
I (u)  u, u  n
you get back the same thing'

R(I )  n

n n
Example
Let O : n
 m
be the transformation defined by
'whatever you put into O,
O(u)  0, u  n
you get back the zero vector'
m
(in )

R(O)  {0}  m

0
n m
Example
Let T : 2
 3
be the transformation defined by

 x y 'all the images are

 x     x
T    y    2
.  x y
  y    x   y 
  of the form  y 
 
 x 
 x  y    
 
R(T )   y  x , y 
for some real numbers x , y '
  
 x  
  
Example
 x  y    1 1 
       
R(T )   y  x , y     x  0   y  1  x, y  
 
 x    1  0 
        

Equation of the plane?  1   1  

    
 span  0 , 1  (a plane in 2
)
  
ax  by  cz  0  1   0  
    

x yz 0
 Discussion
Let T : n
 m
be a linear transformation and
{u1 , u2 ,..., un } be a basis for n .
v For any v  n , we have already
observed that T (v) is some
linear combination of
T (u1) T (u2 ) T (un ) T (v) T (u1),T (u2 ),...,T (un ).
Each T (v) is a linear
So R(T )  {T (v)|v  n
}
combination
 span{T (u1),...,T (un )}
of T (u1),T (u2 ),...,T (un ).
Discussion
Let T : n
 m
be a linear transformation and
{u1 , u2 ,..., un } be a basis for n
. Conversely,
w  span{T (u1),T (u2 )...,T (un )} since u1 , u2 ,..., un
n
 w  a1T (u1),  a2T (u2 )  ...  anT (un ) is a basis for
 T (a1u1  a2 u2  ...  an un )  T (v) for some v  n

 w  R(T )
 span{T (u1),T (u2 ),...,T (un )}  R(T )
So R(T )  {T (v)|v  n
} R(T )
 span{T (u1),...,T (un )}  span{T (u ),T (u ),...,T (u )}
1 2 n
What does this mean?
Let T : n
 m
be a linear transformation and
n
{u1 , u2 ,..., un } be a basis for .

R(T )
 span{T (u1),T (u2 ),...,T (un )}

1) R(T ) is always a subspace. {e1 , e2 ,..., en }

2) R(T )  span{T (e1),T (e2 )...,T (en )}  standard basis for n
 column space of A Remember
standard matrix for T  A  T (e1) T (e2 ) T (en )
Theorem
Let T : n  m be a linear transformation and A the
standard matrix for T . Then

R(T )  column space of A

m
and is a subspace of .
Definition
Let T : n  m be a linear transformation and A the
standard matrix for T . Then

R(T )  column space of A

m
and is a subspace of .
dim(R(T )) is called the rank of T and is denoted
by rank(T ).
By theorem above, rank(A)  rank(T ) where A is the
standard matrix for T .
Example
Let T : 4
 4
be a linear transformation defined by

 w  x  2y  z   w
  x    x  3y   x
T         4
.
 y   x  4y  z   y
    z
  
z y  z   
Find a basis for R(T ) and determine rank(T ).
If A is the standard matrix for T ...
Find a basis for the column space of A and
determine rank(A).
 Example
Let T : 4
 4
be a linear transformation defined by

 w  x  2y  z   w
  x    x  3y   x
T         4
.
 y   x  4y  z   y
    z
  
z y  z   
0 1 2 1  w 
We will find a basis 0 1 3 0  x 
for the column space    
0 1 4 1 y 
of and rank of 0
 0 1 1
 
z 
Example
0 1 2 1 0 1 2 1
0 1 3 0  Gaussian 0 0 1 1
  Elimination  
0 1 4 1 0 0 0 0
0 1 1 0 0 0 
 0  0

 1   2  
 1   3  
    
So a basis for R(T ) is  ,  and rank(T )  2.
 1   4  
 0   1  
Definition
Let T : n
 m
be a linear transformation.

Ker(T )  { u  n
|T (u)  0}
T
0  n

n m

{ }
The kernel, denoted by Ker(T ) of T is the set of vectors
n m
in whose image is the zero vector in .
Example
Let I : n
 n
be the transformation defined by
'what you put into I ,
I (u)  u, u  n
you get back the same thing'

Ker(I )  ?

n n
Example
Let O : n
 m
be the transformation defined by
'whatever you put into O,
O(u)  0, u  n
you get back the zero vector'
m
(in )
Ker(O)  ?

0
n m
Example
Let T : 2
 3
be the transformation defined by

 x y
 x     x
T    y    2
.
  y    x   y 
 

Ker(T )  ?
 Example
Let T : 3
 4
be the transformation defined by
 2x  y 
 x     x
    x  y  3z
T y    y   3.
     5 x  y   
 z  z
   x  z   
  Solve!
To find Ker(T ), we need to
0
 x  x   
      0
find all vectors y such that T   y     
   z  0
z
    0
 
Example  2 1 0 0
 1 1 0
 2x  y  0 
3

 x  y  3z   5 1 0 0
 0
 1 0 1 
5 x  y  0  0
 x  z  0

 x  0, y  0, z  0 1 0 0 0
0 1 0 0
So Ker(T ) contains only the zero vector,  
that is, Ker(T )  {0}. 0 0 1 0
0 0 
 0 0
Example
Let T : 3
 3
be the transformation defined by

 x   x  z   x 
T   y     0   y   3
.
     

 z     z
   y   

To find Ker(T ), we need to

 x  x  0
find all vectors  y  such that T   y     0 
     
z  z  0
     
Example
Let T : 3
 3
be the transformation defined by
 x   x  z   x 
T   y     0   y   3.
     

 z     z
   y     x  
  
 x   x  z   0 Ker(T )   0 x  
 
         x  
T y  0  0   
     
 z   y   0  1     1  
     
      
x  z  0 x  z span  0    x 0 x  
     
 1     1  
 y  0  y  0       
Theorem
Since Ker(T ) is simply the solution space of Ax  0 (where
A is the standard matrix for T ),
the following theorem is obvious. nullspace of A

Let T : n  m be a linear transformation and A the

standard matrix for T . Then
Ker(T )  nullspace of A
n
and is a subspace of .
Definition
Let T : n  m be a linear transformation and A the
standard matrix for T . Then
Ker(T )  nullspace of A
n
and is a subspace of .
dim(Ker(T )) is called the nullity of T and is denoted
by nullity(T ).
By theorem above, nullity(A)  nullity(T ) where A is the
standard matrix for T .
Example
Let I : n
 n
be the transformation defined by

I (u)  u, u  n
Ker(I )  {0}  nullity(I )  0

Let O : n
 m
be the transformation defined by

O(u)  0, u  n
Ker(O)  n
 nullity(O)  n
 Example
Let T : 3
 3
be the transformation defined by

 x   x  z   x 
T   y     0   y   3.
     
 z   y   z 
       1  
  
 x    1    0   is a basis
       1  
Ker(T )   0 x    span  0    
   

 x    1  for Ker( T ) and
     
nullity(T )  1
Example
Let T : 4
 4
be the transformation defined by

 w  x  2y  z   w
  x    x  3y   x 
T         4
.
 y   x  4y  z   y 
     
 z   y  z   z 

Find a basis for Ker(T ) and determine its dimension.

Example
 w  x  2y  z   w
  x    x  3y   x 
T         4
.
 y   x  4y  z   y 
     
 z   y  z   z 

The standard matrix for T is A 

Let's find a basis for (and determine the dimension of)

the nullspace of A.
Example
Solving Ax  0:
0 1 2 1 0 0 1 0 3 0
0 1 3 0 0 0 0 1 1 0
   
0 1 4 1 0 0 0 0 0 0
0 1 1 0  0 0 
 0  0 0 0
w  s  s  
x  3t  3t  
   s ,t  
 So Ker(T )   
y  t  t  
 z  t , s ,t  .  t  
Example
 s    1  0  
 3t     0   3  
        
So Ker(T )   s ,t    s t s ,t  
 t    0  1  
 t     0   1  
 1   0  
 0   3     1   0 
    
 ,  is a basis for Ker(T )  0   3  
 0   1      
   span  , 
   
 0   1    0   1  
 0   1  
and dim(Ker(T ))  nullity(T )  2.
Theorem
If T : n
 m
is a linear transformation then

rank(T )  nullity(T )  n.
Proof:
Let A be the standard matrix for T . So A is a m  n matrix.
By dimension theorem for matrices:
rank( A)  nullity( A)  n
by earlier
observation
 rank(T )  nullity(T )  n
End of Lecture 21
Lecture 22:
Revision