10.
2 The Kernel and Range
DEF ( p. 441, 443)
Let L : V W be a linear transformation. Then
(a) the kernel of L is the subset of V comprised
of all vectors whose image is the zero vector:
ker L = { v | L( v ) = 0 }
(b)
the range of L is the subset of W comprised
of all images of vectors in V:
range L = { w | L( v ) = w }
DEF ( p. 440, 443)
Let L : V W be a linear transformation. Then
(a) L is one-to-one if v 1 v 2 L(v 1 ) L(v 2 )
(b) L is onto W if range L = W.
MATH 316U (003) - 10.2 (The Kernel and Range) / 1
EXAMPLE 1
Let L : R 3 R 3 be defined by
L(x, y, z) = (x, y, 0). (Projection onto the xy-plane.)
ker L = {(x, y, z) | (x, y, 0) = (0, 0, 0)}
ker L consists of (x, y, z) that are solutions of
the system
x
=0
y =0
z is arbitrary, and x = y = 0.
ker L = span {(0, 0, 1)}.
range L = span {(1, 0, 0), (0, 1, 0)}.
L is not one-to-one (e.g.,
L(1, 2, 3) = L(1, 2, 5) = (1, 2, 0).)
L is not onto (range L R 3 ).
MATH 316U (003) - 10.2 (The Kernel and Range) / 2
TH ( Th. 10.4 p. 442, Th. 10.6 p. 443)
Let L : V W be a linear transformation. Then
ker L is a subspace of V and
range L is a subspace of W.
TH 10.5 p. 443
A linear transformation L is one-to-one if and
only if ker L = { 0 }.
MATH 316U (003) - 10.2 (The Kernel and Range) / 3
EXAMPLE 2
Let L : R 2 R 3 be defined by
x1
x1
L
=
.
x1 + x2
x2
x 1 + 2x 2
ker L =
{
x1
x2
x1
|
0
=
x1 + x2
x 1 + 2x 2
Solve the system of equations:
= 0
x1
x1
x2 = 0
x1
+ 2x 2 = 0
Coefficient matrix:
MATH 316U (003) - 10.2 (The Kernel and Range) / 4
1 0
1 0
has r.r.e.f.
1 1
0 1
1 2
0 0
ker L = {(0, 0)}. By Theorem 10.5, L is
one-to-one.
x1
range L = {
| for all x 1 , x 2 }
x1 + x2
x 1 + 2x 2
1
= {x 1
0
+ x2
| for all x 1 , x 2 }
= span {(1, 1, 1), (0, 1, 2)} R 3
basis for range L
L is not onto
MATH 316U (003) - 10.2 (The Kernel and Range) / 5
EXAMPLE 3 Let L : R 3 R 2 be defined by
x1
L
x1 + x2
x2
x2 + x3
x3
ker L =
x1
{
x2
x3
x1 + x2
x2 + x3
0
0
The homogeneous system coefficient matrix:
1 1 0
0 1 1
has r.r.e.f.
1 0 1
0 1 1
x 3 is arbitrary, x 1 = x 3 , x 2 = x 3 .
ker L = span { (1, 1, 1) }
basis for ker L
Th. 10.5
ker L {(0, 0, 0)}
L is not one-to-one.
MATH 316U (003) - 10.2 (The Kernel and Range) / 6
range L = {
= {x 1
x1 + x2
| for all x 1 , x 2 , x 3 }
x2 + x3
+ x2
+ x3
0
1
all x 1 , x 2 , x 3 }
Find a basis for range L = span
{(1, 0), (1, 1), (0, 1)}:
1 1 0
0 1 1
has r.r.e.f.
0
1
| for
1 0 1
0 1 1
range L = span {(1, 0), (1, 1)}.
basis for range L
range L = R 2 L is onto.
MATH 316U (003) - 10.2 (The Kernel and Range) / 7
Note that in EXAMPLE 3 we used r.r.e.f.
1 0 1
0 1 1
of the homogeneous system
coefficient matrix A =
1 1 0
to
0 1 1
determine both the kernel and the range of L. In
this case, we had:
ker L = null space of A
range L = column space of A
Recall Th. 6.12 p. 288:
If A is an m n matrix then
rank A + nullity A = n.
TH 10.7 p. 446
Let L : V W be a linear transformation. Then
dim(ker L) + dim(range L) = dim V
MATH 316U (003) - 10.2 (The Kernel and Range) / 8
EXAMPLE 4 ( EXAMPLE 1 from the previous
lecture)
L : P 2 P 3 is defined by
L(at 2 + bt + c) = ct 3 + (a + b)t.
ker L = {at 2 + bt + c | ct 3 + (a + b)t = 0}
Set up the homogeneous equation:
= 0
= 0
= 0
b
0
= 0
The coefficient matrix
0 0 1
1 1 0
0 0 0
0 0 1
1 1 0
0 0 0
has r.r.e.f.
0 0 0
0 0 0
b is arbitrary; c = 0; a = b.
ker L = span {t 2 + t}.
ker L {0} L is not one-to-one.
MATH 316U (003) - 10.2 (The Kernel and Range) / 9
range L = {ct 3 + (a + b)t | for all a, b, c}
= {a(t) + b(t) + c(t 3 ) | for all a, b, c}
}
t, t 3
= span {
basis for range L
range L P 3 L is not onto.
Verify Th. 10.7 for the four examples:
EX L : V W dim(ker L) dim(rangeL) dim V
1
L : R3 R3
L : R2 R3
L : R3 R2
L : P2 P3
dim(ker L) = nullity of L
dim(range L) = rank of L.
COROLLARY 10.2 p. 443
Let L : V W be a linear transformation and
dim V = dim W. L is one-to-one if and only if L is
onto.
MATH 316U (003) - 10.2 (The Kernel and Range) / 10
