Topic 2: Matrices and Determinants [AR 1.31.7, 2.12.

4]
In studying linear systems we introduced the idea of a matrix. Next we
see that matrices are not only useful tools for solving systems of
equations but that they have their own algebraic structure and have
many other interesting properties.
Much of the material will have been seen already in your previous
studies.
2.1 Some notation
2.2 Matrix operations
2.3 Matrix inverses
2.4 Rank of a matrix
2.5 Solutions of non-homogeneous linear equations
2.6 Determinants
33

2.1 Some notation

We denote by Aij the entry in the i -th row and j-th column of A

A11

A21
A=
...

Am1

A12
A22
..
.

...
...
..
.

A1n
A2n
..
.

Am2

. . . Amn

or

A = [Aij ]

A matrix that has m rows and n columns has size m n

Example
The matrix A =

1 2
3
e 27.1

Some entries are: A12 = 2,

has size 2 3

A21 = , A23 = 27.1

34

Some special matrices

A matrix with the same number of rows as columns is a square matrix

A matrix with only one row is called a row matrix

A matrix with only one column is called a column matrix

A matrix with all elements equal to zero is a zero matrix

eg., [ 00 00 ], [ 00 00 00 ]

A square matrix with Aij = 0 for i 6= j is called a diagonal matrix

h2 0 0i
eg., 0 1 0
0 0 5
(
1 if i = j
is called an
A square matrix A satisfying Aij =
0 if i 6= j
identity matrix. The identity matrix of size n n is denoted In .
h1 0 0i
eg., I2 = [ 10 01 ], I3 = 0 1 0
001

35

2.2 Matrix operations

Some operations on matrices that we will look at are:
addition, scalar multiplication, multiplication and transpose.

Definition (Scalar multiple)

Let A be a matrix and c R. The product cA is the matrix obtained by
multiplying all entries of A by c. It is called a scalar multiple of A.
(cA)ij = c Aij

Definition (Addition of matrices)

Let A and B be matrices of the same size. The sum A + B is the matrix
obtained by adding corresponding entries of A and B. It has the same
size as A and B.
(A + B)ij = Aij + Bij
Be Careful: Matrices of different sizes can not be added.
Notation: We write A B in place of A + (1)B

36

Example
Let


2
0 3
A=
,
1 1
3

B=

1 1 1
0
1 2

and

C=

1 1
2 0

Calculate (where possible) A + B and (A + B) + C

37

Properties of Matrix Addition

For matrices A, B and C , all of the same size, the following properties
hold:
1. A + B = B + A
2. A + (B + C ) = (A + B) + C

(commutativity)
(associativity)

3. A A = 0
4. A + 0 = A

Here 0 denotes the zero matrix of the same size as A, B and C .

All these properties follow from the corresponding properties of the
scalars (real or complex numbers).

38

Definition (Matrix multiplication)

Let A be an m n matrix and B be a n q matrix. The product AB of
A and B is a matrix of size m q. The entry in position ij of the
matrix product is obtained by taking row i of A, and column j of B,
then multiplying together the entries in order and adding.
(AB)ij =

n
X

Aik Bkj

k=1

Note: The matrix product AB is only defined if the number of columns

of A is equal to the number of rows of B.

39

Example



1 1
4 0

Let A = 3
0 and B =
7 1
0
1
Calculate AB and BA (if they exist).

40

Example
Let A =

1 0
2 3

and B =

Calculate AB and BA.

4 3
2 1

Notice: In this example AB 6= BA, even though both are defined.

Matrix multiplication is not commutative (in general).

41

Properties of matrix multiplication

The following properties hold whenever the matrix products and sums
are defined:
1. A(B + C ) = AB + AC

(left distributivity)

2. (A + B)C = AC + BC
3. A(BC ) = (AB)C

(right distributivity)
(associativity)

4. A(B) = (AB)
5. AIn = Im A = A (where A has size m n)
6. A0 = 0 and 0A = 0

Here is a scalar and 0 denotes zero matrix of the appropriate size.

Matrix powers
If A is a square matrix and n > 1 is an integer we define An = AA A
as the product of n copies of A.
42

Matrix transpose

Definition (Transpose of a matrix)

Let A be an m n matrix. The transpose of A, denoted by AT , is
defined to be the n m matrix whose entries are given by interchanging
the rows and columns of A.
(AT )ij = Aji
Example 
Let A =

1 2 3
4 5 6

. Then AT =

43

Properties of the transpose

1. AT

T

=A

2. (A + B)T = AT + B T
3.

(A)T

4.

(AB)T

(whenever A + B is defined)

AT

(where is a scalar)

B T AT

(whenever AB is defined)

Parts 1, 2 and 3 follow easily from the definition.

To prove part 4, we also use the definition of matrix multiplication given
previously.
Exercise
Prove these!

44

A non-property of matrix multiplication

If two real numbers a and b satisfy ab = 0, then at least one of a and b
is equal to 0. This is not always true for matrices.
Example





1
1
1 1
Let A =
and B =
1 1
1
1
Then AB = 0, but A =
6 0 and B 6= 0.

What is going on here? The point is that the matrices A and B in the
above example do not have inverses.
45

2.3 Matrix inverses

Definition (Matrix inverse)
A matrix A is called invertible if there exists a matrix B such that
AB = I and BA = I where I is an identity matrix.
The matrix B is called the inverse of A and is denoted A1 . If A is not
invertible, we say that A is singular.
We can prove from the definition that:

For A to be invertible it must be square

(If it exists) A1 has the same size as A

(If it exists) A1 is unique

If A is invertible, then A1 is invertible and (A1 )1 = A

I 1 = I ,

A matrix that has a row consisting entirely of zeros is singular

0 has no inverse

46

Inverse of a 2 2 matrix
In general, for a 2 2 matrix A=

a b
c d

1. A is invertible iff (ad bc) 6= 0

2. If (ad bc) 6= 0, then
Example
Find the inverse of A =

A1

2 1
1
1

1
adbc

d b
c a

47

Finding the inverse of a square matrix

Calculating the inverse of a matrix

Given an n n matrix A, we can find A1 as follows:
1. Construct the (grand augmented) matrix [A | I ],
where I is the n n identity matrix.
2. Apply row operations to [A | I ] to get the block corresponding to
A into reduced row-echelon form. This gives
[A | I ] [R | B]
where R is in reduced row-echelon form.
3. If R = I , then A is invertible and A1 = B.
If R =
6 I , then A is singular , i.e., A1 does not exist.
We will see shortly why this method works.
48

Example

1
2 1
Find the inverse of A = 1 1 1
0
1 3

49

Properties of the matrix inverse

If A and B are invertible matrices of the same size, and a non-zero
scalar, then
1. (A)1 =

A1

2. (AB)1 = B 1 A1

n
3. (An )1 = A1
 1

T
= A1 .
4. AT

(for all n N)

Exercise
Prove these!

50

Row operations and matrix multiplication

The effect of a row operation can be achieved by multiplication on the
left by a suitable matrix.

Definition (Elementary matrix)

An n n matrix is an elementary matrix if it can be obtained from In by
performing a single elementary row operation.
Examples

1 0 0
0 0 1
0 1 0

2 0 0
0 1 0
0 0 1

1 0 0
0 1 0
3 0 1

51

Let Ep be the elementary matrix obtained by applying a row operation p

to the identity matrix In .
If A is a matrix such that the product Ep A is defined,
then the product Ep A is equal to the result of performing p on A.
We can perform a sequence of elementary row operations using a
corresponding sequence of elementary matrices.
(Be careful about the order!)
Example






1 2 p1 1 2 p2 1 2

=B

0 1
0 2
3 4




1 0
1 0
E2 =
E1 =
3 1
0 1
2

 



1 0
1 2
1 0 1 2
=B
=
E2 E1 A =
0 1
3 1 3 4
0 1
2
A=

52

If A I , then there is a sequence of elementary matrices E1 , E2 , . . . , En

such that En En1 E2 E1 A = I
This can be used to prove the following:

Theorem (AR 1.5.3)

1. Let A be an n n matrix. Then
A is invertible A In
2. If A and B are n n matrices such that AB = In ,
then A is invertible and B = A1 .
3. Every invertible matrix can be written as a product of elementary
matrices.
This theorem justifies why our procedure for finding inverses using row
operations actually works.... Can you see why?
53

Linear systems revisited

Any linear system in the variables x1 , . . . , xn can be written in the form
Ax = b where A is an m n matrix, b is an m 1 matrix and
T
x = x1 xn .
In the special case that A is invertible, the solution can be found using
the inverse of A.

Theorem
If A is an invertible matrix, then a linear system of the form Ax = b has
a unique solution. It is given by x = A1 b
Proof: Ax = b = A1 Ax = A1 b = x = A1 b

54

Example
Use a matrix inverse to solve the linear system
x + 2y + z = 3
x y + z = 11
y + 3z = 21

1
2 1
(Noting that weve calculated previously 1 1 1 )
the inverse of the matrix
0
1 3

55

2.4 Rank of a matrix

Definition (Rank of a matrix)

The rank of a matrix A is the number of non-zero rows in the reduced
row-echelon form of A.
Note

This is the same as the number of non-zero rows in a row-echelon

form of A.

If A has size m n, then clearly rank(A) 6 m.

We will see later that rank(AT ) = rank(A), from which it follows
that rank(A) 6 n.

56

Example
Find the rank of

1
2
1 1
0
1

each of the following matrices:

1
1 1 2
1
0
1
1 1 2
3
1 3 0
5

57

Theorem
The linear system Ax = b, where A is an m n matrix, has:
1. No solution if rank(A) < rank([A | b])

2. A unique solution if rank(A) = rank([A | b]) and rank(A) = n

3. Infinitely many solutions if rank(A) = rank([A | b]) and rank(A) < n
Proof: This is just a restatement of the results in section 1.3 using the
idea of rank.
Note
It is always the case that rank(A) 6 rank([A | b]) and rank(A) 6 n

58

Theorem
If A is an n n matrix, the following conditions are equivalent:
1. A is invertible

2. Ax = b has a unique solution for any b

3. The rank of A is n
4. The reduced row-echelon form of A is In
Proof:
1 2 Weve seen before (slide 54).
2 3 Follows from the previous theorem (or what we already knew about
linear systems).
3 4 Immediate from the definition of rank, and that fact that A is
square.
4 1 Let R be the RREF of A. Then R = EA, where E = Ek Ek1 . . . E1
is a product of elementary matrices. So we have I = EA. We have
already noted that this implies that A is invertible (slide 53 ).
59

2.5 Determinants

[AR 2.12.3]



a b
When we calculate the inverse of the 2 2 matrix A =
we see
c d
that the number ad bc is important:
If ad bc 6= 0 then we can find the inverse of A; if not, then A is not
invertible.
So this number plays an important role when we study A. We call it the
determinant of A and write it as det(A).
The determinant also has an important geometrical meaning: the area
of the parallelogram spanned by the vectors (a, b) and (c, d) is the
absolute value of det(A).
We now examine how the determinant extends to a function that
associates a real number to any square matrix (not just 2 2 matrices).
60

Defining the determinant

Definition (Determinant)
Let A be an n n matrix. The determinant of A, denoted det(A) or |A|,
can be defined as the signed sum of all the ways to multiply together n
entries of the matrix, with all chosen from different rows and columns.
To determine the sign of the products, imagine all but the elements in
the product in question are set to zero in the matrix. Now swap
columns until a diagonal matrix results. If the number of swaps required
is even, then the product has a + sign, while if it is odd, it is to be
given a sign.

61

Determinant of a 3 3 matrix
Suppose

a11 a12 a13

A = a21 a22 a23
a31 a32 a33
We can form a table of all the products of 3 entries taken from different
rows and columns, together with the signs:

Hence

product sign
a11 a22 a33 +
a11 a23 a32
a12 a21 a33
a12 a23 a31 +
a13 a21 a32 +
a13 a22 a31
det(A) = a11 a22 a33 a11 a23 a32 a12 a21 a33
+a12 a23 a31 + a13 a21 a32 a13 a22 a31

62

The formula for 3 3 matrices is complicated and it becomes quickly

worse as the size of the matrix increases (there are n! terms for an n n
matrix). But there are better ways to calculate determinants.
Cofactors
Here is one way for practical calculation of (small) determinants.

Definition
Let A be a square matrix. The (i , j)-cofactor of A, denoted by Cij , is
the number given by
Cij = (1)i +j det (A(i , j))
where A(i , j) is the matrix obtained from A by deleting the i th row and
jth column.
Example



1
2 1
1 2

If A = 1 1 1 , then A(2, 3) =
and C23 =
0 1
0
1 3

63

Cofactor Expansion
We can write the determinant of a 3 3 matrix, A, in terms of
cofactors.
det(A) = a11 C11 + a12 C12 + a13 C13
This is called the cofactor expansion along the first row of A.
Example

1
2 1
Calculate det 1 1 1
0
1 3

64

Theorem (Cofactor expansion)

The determinant of an n n matrix A can be computed by choosing
any row (or column) of A and multiplying the entries in that row (or
column) by their cofactors and then adding the resulting products.
That is, for each 1 6 i 6 n and 1 6 j 6 n,
det(A) = ai 1 Ci 1 + ai 2 Ci 2 + + ain Cin
(this is called cofactor expansion along the ith row)
and
det(A) = a1j C1j + a2j C2j + + anj Cnj
(this is called cofactor expansion along the jth column)
Proof: We can prove this in the n = 3 case using the formula on slide
62. The proof for the general case is essentially the same, but gets more
technical...
65

How do you remember the sign of the cofactor?

The (1, 1)-cofactor always has sign +. Starting from there, imagine
walking to the square you want using either horizontal or vertical steps.
The appropriate sign will change at each step.
We can visualize this arrangement

+
+

..
..
.
.

with the following matrix:

+ ...
+ ...

+ ...

+ ...

..
.. . .
.
.
.

So, for example, C13 is assigned + but C32 is assigned

66

Example




Calculate 




1 2 0 1 
3
1 2 0 
1
0 1 0 
2 2 1 2 

67

Some properties of determinants:

Suppose A and B are square matrices. Then


1. det AT = det(A)

2. If A has a row or column of zeros, then det(A) = 0

3. det(AB) = det(A) det(B)


4. If A is invertible, then det(A) 6= 0 and det A1 =

1
det(A)

5. If A is singular, then det(A) = 0



C
6. If A =
with C and D both square, then
0 D
det(A) = det(C ) det(D)

Idea of proof: 2: Cofactor expansion. 4: Follows from 3. 5: Follows

from the theorem on slide 59, 2 and 3. 1, 3 6: Need to use the
definition...
68

Row operations and determinants

Calculating determinants via cofactors becomes a very large calculation
as the size of the matrix increases.
We need a better way to calculate larger determinants.
First observe that for some types of matrix it is easy to write down their
determinant.

Definition (Triangular Matrix)

A matrix is said to be upper triangular (respectively lower triangular) if
all the elements below (respectively above) the main diagonal are zero.
Examples

2 1 9
0
3 2
0
0 2

2
0 0
1
3 0
2 3 2

2 0 0
0 3 0
0 0 2
69

Theorem
If A is an n n triangular matrix, then
det(A) is the product of the entries on the main diagonal of A.
Idea of proof: (Repeated) cofactor expansion along the first column.
Example

2 10 92 117
0
3 28 31

Let A =
0
0 1
27
0
0
0
2
What is det(A)?

70

We can use row operations to transform a matrix into triangular form.

The effect of each of the three types of elementary row operations are
given by the following.

Theorem
Let A be a square matrix.
1. If B is obtained from A by swapping two rows (or two columns) of
A, then det(B) = det(A)
2. If B is obtained from A by multiplying a row (or column) of A by
the scalar , then det(B) = det(A)
3. If B is obtained from A by replacing a row (or column) of A by
itself plus a multiple of another row (column), then
det(B) = det(A)
Proof: The corresponding elementary matrices have determinants 1,
and 1 (respectively). Then use det(B) = det(E ) det(A).
71

Example

 1
2 1

Calculate  1 1 1
 0
1 3





 2 4 1 



 and  3 6 3 




 2
1 4 

72

Example




Calculate 




1 2 0 1 
3
1 2 0 
1
0 1 0 
2 2 1 2 

73

We collect here some properties of the determinant function, most of

which weve already noted.

Theorem
Let A be an n n matrix. Then,


1. det AT = det(A)
2. det(AB) = det(A) det(B)
3. det(A) = n det(A)
4. If A is a triangular matrix, then its determinant is the product of
the elements on the main diagonal
5. If A has a row (or column) of zeros, then det(A) = 0
6. If A has a row (or column) which is a scalar multiple of another
row (or column) then det(A) = 0
7. A is singular iff det(A) = 0 (and A is invertible iff det(A) 6= 0)
74