You are on page 1of 30

MATH 211 Winter 2013

Lecture Notes
(Adapted by permission of K. Seyarth)
Sections 2.6
Sections 2.6 Page 1/1
2.6 Matrix Transformations
Sections 2.6 Page 2/1
Linear Transformations
Denition
A transformation T : R
n
R
m
is a linear transformation if and only if,
for all X, Y R
n
and all scalars a,
T1. T(X + Y) = T(X) + T(Y) (preservation of addition)
T2. T(aX) = aT(X) (preservation of scalar multiplication)
As a consequence of T2, for any linear transformation T,
T(0X) = 0T(X), implying T(0) = 0,
and
T((1)X) = (1)T(X), implying T(X) = T(X),
i.e., T preserves the zero vector and T preserves the negative of a vector.
Sections 2.6 Page 3/1
Furthermore, if X
1
, X
2
, . . . , X
k
are vectors in R
n
and Y is a linear
combination of X
1
, X
2
, . . . , X
k
, i.e.,
Y = a
1
X
1
+ a
2
X
2
+ + a
k
X
k
for some a
1
, a
2
, . . . , a
k
R, then T1 and T2 used repeatedly give us
T(Y) = T(a
1
X
1
+ a
2
X
2
+ + a
k
X
k
)
= a
1
T(X
1
) + a
2
T(X
2
) + + a
k
T(X
k
),
i.e.,T preserves linear combinations.
Sections 2.6 Page 4/1
Example
Let T : R
4
R
3
be a linear transformation such that
T
_

_
1
1
0
2
_

_
=
_
_
2
3
1
_
_
and T
_

_
0
1
1
1
_

_
=
_
_
5
0
1
_
_
. Find T
_

_
1
3
2
4
_

_
.
The only way it is possible to solve this problem is if
_

_
1
3
2
4
_

_
is a linear combination of
_

_
1
1
0
2
_

_
and
_

_
0
1
1
1
_

_
,
i.e., if there exist a, b R so that
_

_
1
3
2
4
_

_
= a
_

_
1
1
0
2
_

_
+ b
_

_
0
1
1
1
_

_
Sections 2.6 Page 5/1
Example (continued)
Solve the system of four equations in two variables:
_

_
1 0 1
1 1 3
0 1 2
2 1 4
_

_

_

_
1 0 1
0 1 2
0 0 0
0 0 0
_

_
Thus a = 1, b = 2, and
_

_
1
3
2
4
_

_
=
_

_
1
1
0
2
_

_
2
_

_
0
1
1
1
_

_
.
Sections 2.6 Page 6/1
Example (continued)
It follows that
T
_

_
1
3
2
4
_

_
= T
_
_
_
_
_

_
1
1
0
2
_

_
2
_

_
0
1
1
1
_

_
_
_
_
_
= T
_

_
1
1
0
2
_

_
2T
_

_
0
1
1
1
_

_
=
_
_
2
3
1
_
_
2
_
_
5
0
1
_
_
=
_
_
8
3
3
_
_
Sections 2.6 Page 7/1
Example (2.6 Example 2)
Every matrix transformation is a linear transformation.
Proof.
Suppose T : R
n
R
m
is a matrix transformation induced by the m n
matrix A, i.e., T(X) = AX for each X R
n
.
Let X, Y R
n
and let a R. Then
T(X + Y) = A(X + Y) = AX + AY = T(X) + T(Y),
proving that T preserves addition. Also,
T(aX) = A(aX) = a(AX) = aT(X),
proving that T preserves scalar multiplication.
Since T1 and T2 are satised, T is, in fact, a linear transformation.
Sections 2.6 Page 8/1
It turns out that the converse of this statement is also true, i.e., every
linear transformation from R
n
to R
m
is a matrix transformation.
Theorem (2.6 Theorem 2)
Let T : R
n
R
m
be a transformation.
1
T is linear if and only if T is a matrix transformation.
2
If T is linear, then T is induced by the unique matrix
A =
_
T(E
1
) T(E
2
) T(E
n
)

,
where E
j
is the j
th
column of I
n
.
Why does this work for nding A?
Sections 2.6 Page 9/1
The uniqueness in Theorem 2 guarantees that there is exactly one matrix
for any linear transformation, so it makes sense to say the matrix of a
linear transformation.
Sections 2.6 Page 10/1
Examples
Q
0
: R
2
R
2
is reection in the x-axis, i.e.,
Q
0
_
x
y
_
=
_
x
y
_
.
We saw earlier that Q
0
is induced by the matrix A =
_
1 0
0 1
_
, so
Q
0
is a linear transformation.
By Theorem 2, Q
0
is induced by the matrix
_
Q
0
(E
1
) Q
0
(E
2
)

=
_
Q
0
_
1
0
_
Q
0
_
0
1
_ _
=
_
1 0
0 1
_
= A.
Sections 2.6 Page 11/1
Examples (continued)
R

2
: R
2
R
2
is (counterclockwise) rotation about the origin through
an angle of

2
, and
R

2
_
x
y
_
=
_
y
x
_
.
We saw earlier that R

2
is induced by the matrix B =
_
0 1
1 0
_
, so
R

2
is a linear transformation.
By Theorem 2, R

2
is induced by the matrix
_
R

2
(E
1
) R

2
(E
2
)

=
_
R

2
_
1
0
_
R

2
_
0
1
_ _
=
_
0 1
1 0
_
= B.
Sections 2.6 Page 12/1
Example (2.6 Example 4)
Q
1
: R
2
R
2
is reection in the line y = x, i.e.,
Q
1
_
a
b
_
=
_
b
a
_
.
Notice that
_
b
a
_
=
_
0 1
1 0
_ _
a
b
_
,
so Q
1
is a matrix transformation, and hence a linear transformation.
Again, illustrating Theorem 2, Q
1
is induced by the matrix
_
Q
1
(E
1
) Q
1
(E
2
)

=
_
Q
1
_
1
0
_
Q
1
_
0
1
_ _
=
_
0 1
1 0
_
.
Sections 2.6 Page 13/1
Example
Let T : R
2
R
3
be a transformation dened by
T
_
x
y
_
=
_
_
2x
y
x + 2y
_
_
.
Show that T is a linear transformation.
Solution. If T were a linear transformation, then T would be induced by
the matrix
A =
_
T(E
1
) T(E
2
)

=
_
T
_
1
0
_
T
_
0
1
_ _
=
_
_
2 0
0 1
1 2
_
_
.
Sections 2.6 Page 14/1
Example
Let T : R
2
R
3
be a transformation dened by
T
_
x
y
_
=
_
_
2x
y
x + 2y
_
_
.
Show that T is a linear transformation.
Solution. If T were a linear transformation, then T would be induced by
the matrix
A =
_
T(E
1
) T(E
2
)

=
_
T
_
1
0
_
T
_
0
1
_ _
=
_
_
2 0
0 1
1 2
_
_
.
Sections 2.6 Page 15/1
Example (continued)
Since
A
_
x
y
_
=
_
_
2 0
0 1
1 2
_
_
_
x
y
_
=
_
_
2x
y
x + 2y
_
_
= T
_
x
y
_
,
T is a matrix transformation, and therefore a linear transformation.
Sections 2.6 Page 16/1
Example
Let T : R
2
R
2
be a transformation dened by
T
_
x
y
_
=
_
xy
x + y
_
.
Is T is a linear transformation? Explain.
Solution. If T were a linear transformation, then T would be induced by
the matrix
A =
_
T(E
1
) T(E
2
)

=
_
T
_
1
0
_
T
_
0
1
_ _
=
_
0 0
1 1
_
.
Now
A
_
x
y
_
=
_
0 0
1 1
_ _
x
y
_
=
_
0
x + y
_
.
Sections 2.6 Page 17/1
Example (continued)
In particular,
A
_
1
1
_
=
_
0 0
1 1
_ _
1
1
_
=
_
0
2
_
,
while
T
_
1
1
_
=
_
1
2
_
.
Since A
_
1
1
_
= T
_
1
1
_
, T is not a linear transformation.
There is an alternate way to show that T is not a linear transformation.
Sections 2.6 Page 18/1
Example (continued)
Notice that
_
1
1
_
=
_
1
0
_
+
_
0
1
_
, and
T
_
1
1
_
=
_
1
2
_
, T
_
1
0
_
=
_
0
1
_
, T
_
0
1
_
=
_
0
1
_
.
From this we see that
T
_
1
1
_
= T
_
1
0
_
+ T
_
0
1
_
,
i.e., T does not preserve addition, and so T is not a linear transformation.
Sections 2.6 Page 19/1
Denition (Vector scalar multiplication)
Let X R
2
and let k R. Then kX is the vector in R
2
that is |k| times
the length of X; kX points the same directions as X if k > 0, and
opposite to X if k < 0.
Denition (Vector addition)
Let X, Y R
2
, and consider the parallelogram dened by 0, X and Y.
The vector X + Y corresponds to the fourth vertex of this parallelogram.
Sections 2.6 Page 20/1
Rotation in R
2
Let R

: R
2
R
2
denote counterclockwise rotation about the origin
through and angle of , and let X, Y R
2
and k R. Then
R

(kX) = kR

(X),
and
R

(X + Y) = R

(X) + R

(Y).
This means that R

is a linear transformation, and hence a matrix


transformation.
Therefore, by our earlier theorem, we can nd the matrix that induces R

by simply nding R

(E
1
) and R

(E
2
).
Sections 2.6 Page 21/1
R

(E
1
) = R

_
1
0
_
=
_
cos
sin
_
,
and
R

(E
2
) = R

_
0
1
_
=
_
sin
cos
_
Theorem (2.6 Theorem 4)
The transformation R

: R
2
R
2
is a linear transformation, and is
induced by the matrix
_
cos sin
sin cos
_
.
Sections 2.6 Page 22/1
Reection in R
2
Let Q
m
: R
2
R
2
denote reection in the line y = mx, and let
X, Y R
2
and k R. Then
Q
m
(kX) = kQ
m
(X),
and
Q
m
(X + Y) = Q
m
(X) + Q
m
(Y).
This means that Q
m
is a linear transformation and hence a matrix
transformation.
Therefore, by our earlier theorem, we can nd the matrix that induces Q
m
by simply nding Q
m
(E
1
) and Q
m
(E
2
).
Sections 2.6 Page 23/1
However, an easier way to obtain the matrix for Q
m
is to use the following
observation:
Q
m
= R

Q
0
R

,
where is the angle between y = mx and the positive x axis.
Using the standard trigonometric identities:
cos =
1

1 + m
2
and sin =
m

1 + m
2
,
the matrix for Q
m
can be found by computing the product
_
cos sin
sin cos
_ _
1 0
0 1
_ _
cos() sin()
sin() cos()
_
Sections 2.6 Page 24/1
Theorem (2.6 Theorem 5)
The transformation Q
m
: R
2
R
2
, denoting reection in the line y = mx,
is a linear transformation and is induced by the matrix
1
1 + m
2
_
1 m
2
2m
2m m
2
1
_
.
For reection in the x-axis, m = 0, and the theorem yields the expected
matrix
_
1 0
0 1
_
.
However, the y-axis has undened slope, so the theorem does not apply.
We use Q
Y
to denote reection in the y-axis, and as weve already seen,
Q
Y
is induced by the matrix
_
1 0
0 1
_
,
and thus is a linear transformation.
Sections 2.6 Page 25/1
Example
Find the rotation or reection that equals reection in the x-axis followed
by rotation through an angle of

2
.
We must nd the matrix for the transformation R

2
Q
0
.
Q
0
is induced by A =
_
1 0
0 1
_
, and R

2
is induced by
B =
_
cos

2
sin

2
sin

2
cos

2
_
=
_
0 1
1 0
_
R

2
Q
0
is induced by
BA =
_
0 1
1 0
_ _
1 0
0 1
_
=
_
0 1
1 0
_
.
Sections 2.6 Page 26/1
Example (continued)
Notice that BA =
_
0 1
1 0
_
is a reection matrix.
How do we know this?
Now, since 1 m
2
= 0, we know that m = 1 or m = 1. But
2m
1+m
2
= 1 > 0, so m > 0, implying m = 1.
Therefore,
R

2
Q
0
= Q
1
,
reection in the line y = x.
Sections 2.6 Page 27/1
In general,
The composite of two reections is a rotation.
The composite of two rotations is a rotation.
The composite of a reection and a rotation is a reection.
Sections 2.6 Page 28/1
Example
Find the rotation or reection that equals reection in the line y = x
followed by reection in the y-axis.
We must nd the matrix for the transformation Q
Y
Q
1
.
Q
1
is induced by
A =
1
2
_
0 2
2 0
_
=
_
0 1
1 0
_
,
and Q
Y
is induced by
B =
_
1 0
0 1
_
.
Therefore, Q
Y
Q
1
is induced by BA.
Sections 2.6 Page 29/1
Example (continued)
BA =
_
1 0
0 1
_ _
0 1
1 0
_
=
_
0 1
1 0
_
.
What transformation does BA induce?
Rotation through an angle such that
cos = 0 and sin = 1.
Therefore, Q
Y
Q
1
= R

2
= R3
2
.
Sections 2.6 Page 30/1