Linear Algebra(Bretscher) Chapter 3 Notes

1 Image and Kernel of a Linear Transformation

Denition The image of a function consist of all the values the function takes in its target space. If f is
a function from X to Y , then
image(f) = {f(x) : x X}
= {b Y : b = f(x), x X}
Theorem 1.1 (Image of a linear transformation). The image of a linear transformation T(x) = Ax is the
span of the column vectors of A.
Theorem 1.2 (Properties of the image). The image is a linear space.
1.1 The kernel of a Linear Transformation
Denition The kernel of a linear transformation T(x) = Ax from R
m
to R
n
consists of all zeros of the
transformation. In other words. Solutions to T(x) = Ax =

0.
Note:
im(T) = {T(x) : x R
m
} is a subset of the target space R
n
of T.
ker(T) = {x R
m
: T(x) = 0} is a subset of the domain R
m
of T.
Theorem 1.3. The kernel is a linear space.
Theorem 1.4. n m matrix A. Square matrix B
1. ker(A) = {

0} rank(A) = m
2. ker(A) = {

0} m n
3. ker(B){

0} B is invertible
1.2 Characterization of invertible matrices
For n n matrix A to be invertible:
1. A is invertible.
2. linear system Ax =

b has a unique solution x for all

b in R
n
.
3. rref(A) = I
n
4. rank(A) = n
5. im(A) = R
n
6. ker(A) = {

0}
1
2 Subspace of R
n
; Bases and Linear Independence
Denition A subset W of the vector space R
n
is called a (linear) subspace of R
n
if it has the following
three properties:
1. W contains the zero vector in R
n
2. W is closed under addition.
3. W is closed under scalar multiplication
Theorem 2.1 (Image and the kernel are subspaces). If T(x) = Ax is a linear transformation from R
m
to
R
n
, then
ker(T) = ker(A) is a subspace of R
m
image(T) = im(A) is a subspace of R
n
Denition Vectors v
1
, . . . , v
m
in R
n
1. v
i
is redundant if v
i
is a linear combination of v
1
, . . . , v
i1
2. v
1
, . . . , v
m
are linearly independent if non of them are redundant, else they are linearly dependent.
3. v
1
, . . . , v
m
form a basis of subspace V of R
n
if they span V and are linearly independent and in V .
Theorem 2.2 (Basis of the image). To construct a basis of the image of a matrix A. List all the column
vectors of A, and omit the redundant vectors from the list.
Theorem 2.3 (Linear independence and zero components). Vectors v
1
, . . . , v
m
in R
n
, If v
1
is nonzero, and
if each of the vectors v
i
(i 2) has a nonzero entry in a component where all the preceding vectors v
1
, . . . , v
i1
have a 0, then the vectors v
1
, . . . , v
m
are linearly independent.
Denition Vectors v
1
, . . . , v
m
in R
n
. An equation of the form
c
1
v
1
+ . . . + c
m
v
m
=

0
is called a (linear) relation among the vectors v
1
, . . . , v
m
. There is always the trivial relation, with
c
1
= . . . = c
m
= 0. Nontrivial relations may or may not exist among the vectors v
1
, . . . , v
m
.
Theorem 2.4 (Relations and linear dependence). v
1
, . . . , v
m
in R
n
are linear dependent if there exist a
trivial relation.
Theorem 2.5 (Kernel and relations). The vectors in the kernel of an n m matrix A correspond to the
linear relations among the column vectors v
1
, . . . , v
m
of A: The equation
Ax =

0
means that
x
1
v
1
+ . . . + x
m
v
m
=

0
The column vectors of A are linearly independent iif ker(A) = {

0} or rank(A) = m. Thus there are at most

n linearly independent vectors in R
n
.
2.1 Various characterizations of linear independence
For a list of v
1
, . . . , v
m
vectors in R
n
, the following statements are equivalent:
1. Vectors v
1
, . . . , v
m
are linearly independent.
2. None of the vectors v
1
, . . . , v
m
is redundant.
3. Non of the vectors v
i
is a linear combination of other vectors.
4. c
1
v
1
+ . . . + c
m
v
m
=

0 c
1
= . . . = c
m
= 0.
5. ker[v
1
. . . v
m
] =

0.
6. rank[v
1
. . . v
m
] = m.
Theorem 2.6 (Basis and unique representation). v
1
, . . . , v
m
in a subspace V of R
n
. It forms a basis of V
iif every vector v in V can be expressed uniquely as a linear combination
v = c
1
v
1
+ . . . + c
m
v
m
3 The Dimension of a Subspace of R
n
Theorem 3.1. Consider vectors v
1
, . . . , v
p
and w
1
, . . . , w
q
in a subspace V of R
n
. If the vectors v
1
, . . . , v
p
are linearly independent, and the vectors w
1
, . . . , w
q
span V , then q p.
Denition The number of vectors in a basis of V is called the dimension of V , denoted by dim(V )
Theorem 3.2 (Independent vectors and spanning vectors in a subspace of R
n
). Consider a subspace V of
R
n
with dim(V ) = m.
1. We can nd at most m linearly independent vectors in V .
2. We need at least m vectors to span V
3. If m vectors in V are linearly independent, then they form a basis of V
4. If m vectors in V span V , then they form a basis of V .
Theorem 3.3 (using rref to construct a basis of the image). To construct a basis of the image of A, pick
the column vectors of A that correspond to the columns of rref(A) containing the leading 1s.
Theorem 3.4 (Dimension of the image). For any matrix A,
dim(imA) = rank(A)
Theorem 3.5 (Rank-Nullity Theorem). For any n m matrix A, the equation
dim(ker A) + dim(imA) = m
Theorem 3.6 (Finding bases of the kernel and image by inspection). Each redundant column as a linear
combination of the preceding columns, v
i
= c
1
v
1
+ . . . + c
i1
v
i1
, write a corresponding relation, c
1
v
1

. . . c
i1
v
i1
+v
i
=

0, and generate the vector

_
_
_
_
_
_
_
_
_
_
_
_
c
1
.
.
.
c
i1
1
0
.
.
.
0
_
_
_
_
_
_
_
_
_
_
_
_
in the kernel of A. The vectors so constructed form a basis of the kernel of A. The nonredundant columns
form a basis of the image of A.
Theorem 3.7. The vectors v
1
, . . . , v
n
in R
n
form a basis of R
n
iif the matrix
_
_
| |
v
1
. . . v
n
| |
_
_
is invertible
4 Coordinates
Denition Consider a basis B= (v
1
, v
2
, . . . , v
m
) of a subspace V of R
n
. Any vector x in V can be written
uniquely as
x = c
1
v
1
+ . . . + c
m
v
m
The scalars c
1
, c
2
, . . . , c
m
are called the B-coordinates of x, and the vector
_

_
c
1
c
2
.
.
.
c
m
_

_
is the B-coordinate vector of x, denoted by [x]
B
.
x = S[x]
B
, where S =
_
v
1
v
2
. . . v
m

Theorem 4.1 (Linearity of Coordinates). If B is a basis of a subspace V of R

n
, then
1. [x + y]
B
= [x]
B
+ [y]
B
2. [kx]
B
= k[x]
B
where x and y are in V .
Denition The linear transformation T from R
n
to R
n
and a basis B of R
n
. The n n matrix B that
transforms [x]
B
into [T(x)]
B
is called the B-matrix of T:
[T(x)]
B
= B[x]
B
for all x in R
n
. We can construct B column by column as follows: If B= (v
1
, . . . , v
n
), then
B =
_
[T(v
1
)]
B
. . . [T(v
n
)]
B

Theorem 4.2 (Standard matrix versus B-matrix). Linear transformation T from R

n
to R
n
and a basis
B = (v
1
, . . . , v
n
) of R
n
. Let B be the B-matrix of T, and let A be the standard matrix of T (such that
T(x) = Ax for all x in R
n
). Then
AS = SB, B = S
1
AS, and A = SBS
1
, where S =
_
v
1
v
2
. . . v
m

Denition Consider two nn matrices A and B. We say that A is similar to B if there exists an invertible
matrix S such that
AS = SB, or B = S
1
AS
Theorem 4.3 (Similarity is an equivalence relation). 1. An nn matrix A is similar to A itself (reex-
ivity).
2. If A is similar to B, then B is similar to A (symmetry).
3. If A is similar to B and B is similar to C, then A is similar to C (transitivity).