You are on page 1of 10

Linear Transformation

 Mapping/Transformation :Let V and W be two vector spaces. A mapping T from V


to W is a function that assigns to each vector v  V a unique vector w W . In this case
we say that T maps V into W and is written as T : V  W . For each v  V the vector
w  T (v) W is the image of v under T .
 Linear Mapping/Linear Transformation : Let V and W be two vector spaces. A
mapping T : V  W is called linear transformation or linear mapping if
i) T (u  v)  T (u )  T (v), u, v V (Additivity)
ii) T ( u )  T (u ),   , u V (Homogeneity)
When V  W , T is called as linear operator.
Note That :1. Substituting , in condition (ii), we get T (0)  0 , thus every linear mapping
maps zero vector into zero vector.
2. If for transformation T : U  V , T (0)  0, then T may or may not be linear.
 Some Examples of Linear Transformation
1) Any matrix transformation is a linear transformation.
2) Derivative operator, integration operator are linear operators.
T : 3  2 is projection of mapping is a linear transformation from
3
3) to 2 .
 x
1 0 0  1 0 0     x 
A2  3     0 1 0   y    y  .
0 1 0    z  
 
4) Multiplication by a fixed polynomial is a linear transformation.
5) Rotation Matrix is a linear transformation –
 Rotation matrix in 2 dimensions for anticlockwise rotation through an angle  is

cos   sin 
A()   
 sin  cos  

 Rotation matrix in 3 dimension


a) Rotation about X - axis
1 0 0 

A( x)  0 cos   sin 
0 sin  cos  
 cos  0 sin  
b) Rotation about Y-axis A( y)   0 1 0 

  sin  0 cos 
cos   sin  0
c) Rotation about Z-axis A( z )   sin  cos  0
 
 0 0 1 
Illustrative Examples
Q 1) Let A be n  n matrix. Define T : n  m by T ( x)  Ax, x  n
.
i) Show that T is a linear transformation.

Prepared by Dr RSD
Linear Transformation
1  7 
 1 2 1 1 &  1 under the mapping T ( x)  Ax.
ii) Let A    . Find the image of
 1 3 2    
1  5 
Soln. i) Let u, v  n &  . Then T (u  v)  A(u  v)  Au  Av  T (u )  T (v) and
T ( u )  A( u )   Au  T (u ) . Hence T is linear.
 1  1  7  7
     1 2 1    2       1 2 1   0 
ii) T  1     1    and T   1    1 3 2   1  0  .
 1   1 3 2  1  4   5    5   
      

Note That : Every matrix of order m n determines a linear transformation from n to m .


 x 
   x  2y  z 
Q 2) Define T : 3
 3
by T   y     
  z    x  5 y  z 
 
i) Show that T is a linear transformation. ii) Find all vectors that are mapped to 0 .
Sol . i) Let u, v  &  . Then
n 3

  u1   v1     u1  v1  
       (u1  v1 )  2(u2  v2 )  (u3  v3 ) 
a) T  u2   v2    T  u2  v2     
 u   v     u  v    (u1  v1 )  5(u2  v2 )  (u3  v3 ) 
 3  3  3 3
 (u  2u2  u3 )  (v1  2v2  v3 ) 
 1 
(u1  5u2  u3 )  (v1  5v2  v3 ) 
 (u  2u2  u3 )   (v1  2v2  v3 ) 
 1    T (u )  T (v) .
(u1  5u2  u3 )  (v1  5v2  v3 ) 
  u1     u1  
      ( u1  2 u2   u3 )    (u1  2u2  u3 ) 
b) T   u2    T   u2        (u  5u  u ) 
 u     u    (  u1  5 u 2   u3 )   1 2 3 
  3  3
  u1  
 (u1  2u2  u3 )   
     T  u2     T (u ) .
(u1  5u2  u3 )   u  
 3
 (u  2u2  u3 )  0
ii) To find u such that T (u )  0, i.e.,  1    ,
(u1  5u2  u3 )  0
i.e., u1  2u2  u3  0&  u1  5u2  u3  0 .
Solving the above homogeneous system of 2 equations in 3 unknowns , the
  11  
  
possible set of vectors is   2  t / t   .
 7  
  
Q 3) Determine whether the function T : 3  2 defined by
T ( x, y, z )  (2 x, x  y ) is alinear transformation.

Prepared by Dr RSD
Linear Transformation
Soln. T ( x1 , y1 , z1 )  (2 x1 , x1  y1 )  T (u)
T ( x2 , y2 , z2 )  (2 x2 , x2  y2 )  T (v)
T (u  v)  ( x1  x2 , y1  y2 , z1  z2 )
 {2( x1  x2 ), [( x1  x2 )  ( y1  y2 )]}  (2 x1 , x1  y1 )  (2 x2 , x2  y2 )  T (u )  T (v)
T (u )  T (x1 , y1 , z1 )  (2x1 , x1  y1 )  (2 x1 , x1  y1 )  T (u )
 It is linear transformation.
Examples of Non-linear Transformations
Q 4) Show that T :  defined by T ( x)  x  1 is not linear,i. e., translation is not a
linear transformation.
n
Sol . Let x1 , x2  &   then T ( x1  x2 )  ( x1  x2 )  1  x1  x2  1 while
T ( x1 )  T ( x2 )  x1  1  x2  1  ( x1  x2 )  2 .
Thus, T ( x1  x2 )  T ( x1 )  T ( x2 ) .
Also, T ( x)  ( x)  1 and  T ( x)   ( x  1) Thus, T ( x)   T ( x) .
T is not a linear operation.
Or equivalently T (0)  1  0 . Therefore T is not a linear transformation.
Q 5) Show that T :  defined by T ( x)  x 2 is not linear.
Soln. Let x1 , x2  . T ( x1  x2 )  ( x1  x2 ) 2 T ( x1 )  T ( x2 )  x12  x22
Thus, T ( x1  x2 )  T ( x1 )  T ( x2 ) .  T is not additive.
T ( x)  ( x)2   2 x 2 while  T ( x)   x 2  T ( x)
 T is not satisfying homogeneity also.  T is not linear.
Q 6) Let T : M nn  be the transformation that maps an n  n matrix to a number set by
T ( A)  det( A). Show that the transformation is not linear.
Sol . T ( A  B)  det( A  B)  det( A)  det( B) and det(A)  n det( A)   det( A) .
n

Therefore T is not linear transformation.


 Matrices For Linear Transformation
Consider the transformation T : 3  3 defined by
T ( x, y, z)   x  2 y  3z,  2x  3 y  2z, x  y  z  .
 1 2 3   x 
This can be expressed as T ( x, y, z )   2 3 2   y   AX .
  
 1 1 1  z 
 Method of Finding Standard Matrix for Linear Transformation
Consider the linear transformation T : n  m such that
 a11   a12   a1n 
a  a  a 
 21   22   2n 
 .   .   . 
T(e1 )    , T(e 2 )    , ..., T(e n )    , where
 .   .   . 
 .   .   . 
     
 am1   am 2   amn 
e1 , e2 , , en  is a standard basis for n . Then the m n matrix whose n correspond to
images of e1 , e2 , , en under T , i.e., T (e1 ),T(e2 ), ... T(en ) is called the standard

Prepared by Dr RSD
Linear Transformation
matrix of T . Thus
 a11 a12 ... a1n 
a a 
 21 22 ... a2 n 
 . . . . 
A   T(e1 ) T(e 2 ) ... T(e n )    .
 . . . . 
 . . . . 
 
 am1 am 2 ... amn 
Hence for every v  n , T (v)  Av.
 Kernel and image of a matrix transformation
Consider the linear transformation T : n  m such that TX  Amn X . Then
Image of T  Y  m
: such that AX  Y   col ( A) .
Kernel of T   X  n
: AX  0  Null ( A) .
Illustrative Example:
Q 1) Find the standard matrix of the linear transformation T : 3
 2
defined by
T (x, y, z)   x  3 y, 2 x  y  .
 1    0    0  
          0 
Soln. T (e1 )  T  0      , T (e2 )  T  1      and T (e1 )  T  0      .
   
1 3
 0    2   0    1  1    0 
     
1 3 0
Thus A  T (e1 ) T(e2 ) T (e3 )   .
 2 1 0
Q 2) Let T : 4  3 be the linear transformation defined by
T ( x1 , x2 , x3 , x4 )   x1  x2  x3  x4 , 2x1  2 x2  3x3  4x4 , 3x1  3x2  4x3  5x4  . Find
a) Basis and dimension of the image of T .
b) Basis and dimension of the kernel of T .
Soln. T 1, 0,0, 0  1, 2, 3 , T  0, 1,0, 0   1,  2,  3 , T  0, 0,1, 0  1, 3, 4 and
T  0, 0,0, 1  1, 4, 5 . Therefore the matrix of the transformation is
1 1 1 1 
A   2 2 3 4  .
 3 3 4 5 
a) To find the basis of image of T which is nothing but column space of A , we reduce A to
row Echelon form
1 1 1 1  1 1 1 1  1 1 1 1 
A   2 2 3 4 0 0 1 2 0 0 1 2  .
 3 3 4 5  0 0 1 2 0 0 0 0 
 1  1  
 
Therefore thebasis of image of T is   2  ,  3   and the dimension of the space is 2.
3 4 
    
b) To find the basis for Kernel of T which is the null space of A , consider AX  0 .

Prepared by Dr RSD
Linear Transformation
Solving this homogeneous system of linear equations the solution is

 x1  l  k  1   1   1   1  
 x   l  1   0      
 2        1   0  
x  l  . Therefore the basis for kernel is  ,  and
 x3   2k  0  2   0   2  
            
 4 
x k     
0 1     
0 1
the dimension of the kernel is 2 .
Q 3) Let B  v1 , v2  be basis of 2 , where v1  (1, 1) and v2  (1, 2) . Let T : 2  3 is
such that T (v1 )  (1, 2, 0) and T (v2 )  (0,  2, 4). Find T (5,  1).
Soln.Since B  v1 , v2  is basis of 2
, every vector in 2
can be expressed as a linear
combination of v1 and v2 .
 (5,  1)  c1 (1, 1)  c2 (1, 2)
 c1  c2  5 and c1  2c2  1
 c1  11 and c2  6
Thus (5,  1)  11 (1, 1)  6  (1, 2) .
 T (5, 1)  T 11 (1, 1)  6  (1, 2) 
 11 T (1, 1)  6  T (1, 2)
 11 (1, 2, 0)  6  (0,  2, 4)
 (11, 34,  24).
 Regular / Non Singular Transformation: A transformation Y  AX is said to be
regular / non-singular if A is non singular matrix. The corresponding inverse
transformation is X  A1Y .
 Orthogonal Transformation : A transformation Y  AX is said to orthogonal if the
matrix A is orthogonal matrix.
 Orthogonal Matrix : Matrix A is said to be orthogonal matrix if AAT  AT A  I .
Note That :1) Every rotation matrix is an orthogonal matrix & vice versa.
2) Orthogonal transformation preserves the length.
If A is an orthogonal matrix, then || AX ||  || X || .
( || AX || 2  AX , AX  | A | 2 X , X  X , X  || X || 2 )
3) If the matrix is orthogonal, then its rows and columns are orthonormal.
4) If A is orthogonal then | A |  1.
5) If for certain matrix A, | A |  1thenAmay or may not be orthogonal.
6) If for a matrix A, we observe that its rows (or columns) are orthogonal but
Columns (rows) are not, then it can be converted into an orthogonal matrix by
simply normalizing each row (column) of the matrix.
7) Orthogonal matrixA is always invertible and A1  AT .
Illustrative Examples
0 a a
Q 1) Is A   2b b b  orthogonal?
 
 c c c 
Soln. No, because rows are orthogonal but columns are not.

Prepared by Dr RSD
Linear Transformation
Note That: 1) If rows are orthogonal, normalize the rows to obtain an orthogonal matrix.
2) Similarly, if columns are orthogonal, then normalize columns to reduce it to
orthogonal matrix.
0 2m n 
Q 2) Find l , m, n so that the matrix A   l m n  is orthogonal.
 
 l m n 
Soln. A is anOrthogonal matrix, therefore every column androw must be orthonormal
1
 o2  l 2  l 2  1  2l 2  1  l   ,
2
1
(2m) 2  m2  m2  1  6m2  1  m   and
6
1
n 2  (n)2  n2  1  3n2  1  n   .
3
Q 3) Is T : 2  2 defined by T ( x, y )  ( x  y, 2 x  y ) regular? If regular, find the inverse
transformation.
1 1  x  1 1
Soln. T ( x, y)  ( x  y, 2 x  y)      . Thus TU  AU , where A  .
2 1   y  2 1 
Further det( A)  0 . Therefore T is a regular transformation. The inverse transformation
1  1 1
is S : 2  2 is given by S ( x, y)  A1 ( x, y)T , where A1  
3  2 1
.

 Geometry of Linear Operators on 2


a b 
If T : 2  2 is the matrix operator whose standard matrix is A    , then
c d 
 a b   x   ax  by 
X      .
 c d   y   cx  dy 
It is natural question that Geometrically how can we view the obove transformation ?
We may view entries in the matrices as components of vectors or as co-ordinates of points.
 Compressions and Expansions
If T: 2  2 is a compression or expansion in the x  direction with the factor k , then
 1    k   0  0
T  e1   T        and T  e2   T        .
 0   0   1    1 
k 0
So the standard matrix for T is  .
0 1
Similarly, the standard matrix for compression or expansion in the y direction is
1 0 
0 k  .
 

Prepared by Dr RSD
Linear Transformation
 Special operators in 2

Sr. Operator Matrix


No. Representation
1 0
1. Reflection about X -axis 0 1

 1 0
2. Reflection about Y -axis 0 1 

0 1
Reflection about the line y  x . 1 0
3.

0 0
4. Projection on x-axis 0 1 

1 0 
5. Projection on y-axis 0 0
 
Counterclockwise Rotation through an cos   sin  
6.
angle  .  sin  cos  
 
Compression  0  k  1 or k 0
7. 0 1
Expansion (k  1) in the x1 - direction  
Compression  0  k  1 or 1 0 
8. 0 k 
Expansion (k  1) in the x2 - direction  
Contraction  0  k  1 or k 0 
9. 0 k 
Dilation (k  1)  
1 k
Shear in the x1 -direction with factor k. 0 1 
10.

1 0
Shear in the x2 -direction with factor k. k 1 
11.

Prepared by Dr RSD
Linear Transformation
 Special operators in 3

Sr. Operator Matrix Representation


No.
1 0 0
0 1 0 
1. Reflection about XY -plane 
0 0 1
1 0 0
0 1 0 
2. Reflection about XZ -plane 
0 0 1 
 1 0 0
0 1 0 
3. Reflection about YZ -plane 
 0 0 1 
1 0 0
0 1 0 
4. Projection on XY -plane 
0 0 0 
0 0 0 
0 1 0 
5. Projection on YZ -plane  
0 0 1 
1 0 0 
0 0 0 
6. Projection on XZ -plane  
0 0 1 
1 0 0 
0 cos   sin 
7. Rotation about X -axis  
0 sin  cos  
 cos  0 sin  
 0 1 0 
8. Rotation about Y -axis 
  sin  0 cos 
cos   sin  0
 sin  cos  0
9. Rotation about Z -axis 
 0 0 1 

Illustrative Examples

2 2
Q 1) Find a transformation from to that first shears in x1 direction by a factor of 3 and
then reflects about y  x.
1 3
Soln.The standard shear matrix in x1 direction by a factor of 3 is A1 =  .
 0 1
0 1 
The standard matrix of reflection about y  x is A2   .
1 0 

Prepared by Dr RSD
Linear Transformation
0 1  1 3 0 1
Hence the required matrix is A2 A1     .
1 0 0 1 1 3
Q 2) Find a transformation from 2 to 2 that first reflects about y  x and then shears by
a factor of 3 in x1 direction.
1 3
Soln.The standard shear matrix in x1 direction by a factor of 3 is A1 =  .
 0 1
0 1 
The standard matrix of reflection about y  x is A2   .
1 0 
1 3 0 1  3 1 
The required transformation is A1 A2     .
0 1 1 0 1 0
Q 3) Find the standard matrix for T: 3  3 , that first reflects about xy - plane , then
rotates the resulting vector in counterclockwise direction through an angle θ, about
z - axis and then finally resultant vector is projected on xz - plane.
Soln. The standard matrix for reflection about xy - plane is

1 0 0 
A1  0 1 0  .
0 0 1
cos   sin  0
The standard matrix for rotation about z - axis A2   sin  cos  0 .

 0 0 1 

1 0 0 
The standard matrix for projection onto xz - plane A3  0 0 0 .
 
0 0 1 
Therefore the required transformation is

1 0 0 cos   sin  0 1 0 0  cos   sin  0


A3 A2 A1  0 0 0  sin  cos  0 0 1 0    0 0 0  .
0 0 1   0 0 1  0 0 1  0 0 1

Note that elementary matrix represents shear, compression, expansion or reflection.


Thus every 2 by 2 invertible matrix geometrically represents the appropriate succession of shear,
compression, expansion or reflection.
 Geometric properties of Linear Operator on 2 .
Theorem 2.3 : If T : 2  2 is multiplication by an invertible matrix A, then the geometric
effect of T is the appropriate succession of shears, compressions, expansions and reflections.
Proof : Since A is invertible, it can be reduced to identity matrix by a finite sequence of
elementary row transformation. An elementary row operations can be performed by
multiplying on the left by elementary matrix and so there exist elementary matrices

Prepared by Dr RSD
Linear Transformation
E1 , E2 , ..., Ek such that Ek  E2 E1 A  I . Therefore
A  E11E21  En1I  E11E21  En1 .
Illustrative Example:

Q 1) Express the following matrix as a product of elementary matrices.Describe the


effect of multiplication by the given matrix in terms of compression, expression,
reflection and shear.
1 4 
A=   .
2 9
Soln. A can be reduced to identity as follows :
1 4  R2  2 R1 1 4  R1 4 R2 1 0 
 2 9   0 1   0 1 
     
Two row operations can be performed on the left successively by
 1 0 1 4
E1    , E  0 1  . Therefore E2 E1 A  I .
 2 1 
2
 
 1 0  1 4 
A  E11E21     .
 2 1  0 1 
It follows that the effect of multiplication by A
i) Shearing by a factor 4 in the x1  direction.
ii) Followed by shearing by a factor 2 in the x2  direction.
Q 2) Express the following matrix as a product of elementary matrices. Describe the effect of
multiplication by the given matrix in terms of compression, expression, reflection and
shear.
 4 3
A=   .
 3 2
Soln. A can be reduce to identity as follows :
 4 3 R1  R2 1 1 R2 3 R1 1 1 R1  R2 1 0
 3 2     0 1    
  3 2   0 1 
Three row operations can be performed on the left successively by

1 1  1 0 1 1
E1    , E2    , E3    . Therefore E3 E2 E1 A  I .
0 1   3 1  0 1

1 1 1 0 1 1
A  E11E21E31     .
0 1 3 1  0 1 
Hence the effect of multiplication is
i) Shearing by a factor 1 in the negative x1  direction.
ii) Shearing by a factor 3 in the negative x2  direction.
iii) Shearing by a factor 1 in the x1  direction.

Prepared by Dr RSD

You might also like