Sei sulla pagina 1di 12

Chapter 2: Determinants

17

SECTION B Properties of a Determinant


By the end of this section you will be able to
prove that the determinant of a triangular or diagonal matrix is the product of the
leading diagonal entries
evaluate determinants of a triangular and diagonal matrix
prove properties of determinant of an elementary matrix
establish certain properties of determinants of other matrices

You will need to remember the definition of a determinant of a matrix and the
technique to evaluate it by using cofactors. This section is a difficult section to follow
because we need to prove a number of propositions and some of these rely on the
results of chapter 1.
B1 Revision of Properties of a Determinant
In the last section we established certain properties of the determinant of a matrix such
as:
Proposition (2.10). If a square n by n matrix A consists of two identical rows then
det ( A ) = 0 .
Further properties were also established in the associated Exercise 2(a) and we give
these reference numbers.
Proposition (2.14). Let A be a square matrix then det ( AT ) = det ( A ) . What does this
mean?
The determinant of the transposed matrix is the same as the determinant of the initial
matrix.
Proposition (2.15). Let B be a matrix obtained from matrix A by multiplying one row
(or column) of A by a non-zero scalar k then
det ( B ) = k det ( A )

We will restate this result later in this section.


Proposition (2.16). Let A be a square n by n matrix and k be a scalar then
det ( kA ) = k n det ( A )
What does this mean?
The determinant of the scalar multiplication kA is k n times the determinant of the
matrix A.
Note that since det ( AT ) = det ( A ) , expanding along a row or column is equivalent.

Generally propositions about the determinant of the matrix with the word row can be
swapped by the word column because det ( AT ) = det ( A ) .
B2 Determinant Properties of Particular Matrices
We can find determinants of particular matrices such as triangular matrices. What are
triangular matrices?
Definition (2.17). A triangular matrix is a n by n matrix where all entries to one side of
the leading diagonal are zero.
What is meant by the leading diagonal?
Leading diagonal of a matrix is the entries going from the top left hand corner to the
bottom right hand corner of the matrix,
For example, the following are triangular matrices:

Chapter 2: Determinants

18

1 2 3
1 0 0
Leading

Diagonal
(a) 0 4 5 and (b) 2 3 0
0 0 6
4 5 6

(a) is an example of an upper triangular matrix.


(b) is an example of a lower triangular matrix.
Another type of matrix is a diagonal matrix. Do you know what is meant by a diagonal
matrix?
Definition (2.18). A diagonal matrix is a n by n matrix where all entries to both sides of
the leading diagonal are zero.
Can you think of an example of a diagonal matrix?
1 0 0
1 0 0

The identity matrix I = 0 1 0 . Another example is 0 2 0 .


0 0 1
0 0 3

A diagonal matrix is both an upper and lower triangular matrix.


Triangular and diagonal matrices have the following property.
Proposition (2.19). The determinant of a triangular or diagonal matrix is a product of
the entries along the leading diagonal.
What does this proposition mean?
Let A be an upper triangle matrix such that
a11 a12 " a1n

0 a22 " a2 n
A=
# " % #

0 " ann
0
then det ( A ) = a11a22 a33 " ann .

We prove it for the upper triangular matrix and the proof for the lower triangular matrix
is similar. Since the diagonal matrix is a particular upper (or lower) triangular matrix
therefore the proof of the diagonal matrix follows from the upper triangular result.
We use proof by induction. Remember the method of proof by induction is threefold.
1. Prove it for a base n = 1 or n = 2 (or for some other base n = k0 ).
2. Assume it is true for n = k .
3. Prove it for n = k + 1 .
Proof.
For a 2 by 2 matrix we have
a
a
det 11 12 = a11a22 0 = a11a22
0 a22
Hence the determinant of a 2 by 2 upper triangular matrix is the product of entries in
the leading diagonal. Therefore the result is true for n = 2 .
Assume it is true for n = k , that is
a11 a12 " a1k

0 a22 " a2 k

()
= a a a " akk
det
# " % # 11 22 33

0 " akk
0

Chapter 2: Determinants

19

We need to prove the result for n = k + 1 , that is a matrix of size k + 1 by k + 1 . For


n = k + 1 we can find the determinant by expanding along the bottom row because all
entries in the bottom row are zero apart from a( k +1)( k +1) :
a11

0
det
#
0

a1( k +1)
a11 a12 " a1k

a22 " a2( k +1)


0 a22 " a2 k
( k +1) + ( k +1)

a( k +1)( k +1) det


= ( 1)
# " % #
" %
#

0
0 " akk
0 " a( k +1)( k +1)

a12 "

= a11a22 a33"akk by ()

= ( 1)

2k +2

a( k +1)( k +1) a11a22 a33 " akk

= a11a22 a33 " akk a( k +1)( k +1)


Remember ( 1)

2k +2

= ( 1)

2( k +1)

= 1 because the index 2 ( k + 1) is even.

Hence the determinant of an upper triangular matrix of size k + 1 by k + 1 is the


product of the entries in the leading diagonal. We have proven the case for n = k + 1 .
Therefore by induction we have shown that the determinant of an upper triangular
matrix is the product of the entries along the leading diagonal.

Example 10
Find the determinants of the following matrices:
1

0
(a) U =
0

2
5
0
0

3 4
0
1

6 7
3 2
(b) L =
11 8
8 9

0 10
89 9

0
0
5
3

0
0

2 0

0 6
(c) D = 0 0

0 0
0 0

0
0
4
0
0

0 0

0 0
0 0

3 0
0 10

Solution
We use the result of the above Proposition (2.19) because all 3 matrices are upper
triangular, lower triangular and diagonal respectively.
In each case the determinant is the product of the leading diagonal entries.
(a) det ( U ) = 1 5 8 10 = 400
(b) det ( L ) = 1 ( 2 ) 5 2 = 20
(c) det ( D ) = 2 ( 6 ) 4 3 10 = 1440
You may like to check these answers by using MATLAB. The MATLAB command for
determinant is det().
B3 Determinant Properties of Elementary Matrices
Do you remember what an elementary matrix is?
An elementary matrix is a matrix obtained by a single row operation on the identity
matrix I. Examples of 3 by 3 elementary matrices are
1 0 0 0 0 1
1 0 5

0 5 0 , 0 1 0 and 0 1 0
0 0 1 1 0 0
0 0 1

There are three different types of elementary matrices.

Chapter 2: Determinants

20

1. An elementary matrix E obtained from the identity matrix, I, by multiplying a


row by a non-zero scalar k .
2. An elementary matrix E obtained from the identity matrix, I, by adding (or
subtracting) one row to another.
3. An elementary matrix E obtained from the identity matrix, I, by interchanging
two rows (or columns).
We can evaluate the determinant of these three different elementary matrices by using
the following proposition.
Proposition (2.20). Let E be an elementary matrix.
(a) If the elementary matrix E is obtained from the identity matrix I by multiplying a
row by a non-zero scalar k then det ( E ) = k .
(b) If the elementary matrix E is obtained from the identity matrix I by adding (or
subtracting) a multiple of one row to another then det ( E ) = 1 .
(c) If the elementary matrix E is obtained from the identity matrix I by interchanging
two rows (or columns) then det ( E ) = 1 .
Proof.
(a) Remember the determinant of the identity is 1 therefore by the above Proposition
(2.15) with one row of the identity multiplied by non-zero k we have
det ( E ) = k det ( I ) = k 1 = k

(b) Remember the identity matrix is a diagonal matrix. If we add a multiple of one row
of the identity matrix to another then we have a triangular matrix and by the above
proposition (2.19) the determinant of a triangular matrix is the product of the elements
along the main diagonal, which is 111" 1 = 1 . Hence det ( E ) = 1 .
(c) We have to prove it for the case when two rows have been interchanged. This is a
more complex proof and is by induction.
We first prove the result for n = 2 :
Rows of I 2 have
0 1
det ( E ) = det

been interchanged
1 0

= 0 1 = 1
Hence if we interchange the two rows of the 2 by 2 identity matrix then the determinant
is 1 .
Assume the result is true for n = k , that is k by k elementary matrix E k with rows i

and j interchanged, det ( Ek ) = 1 .

We need to prove the result for n = k + 1 . Let Ek +1 be the k + 1 by k + 1 elementary


matrix with rows i and j of the identity matrix interchanged.
1 0 " " " " 0

ith row
# % # # # # #
0 " 0 " 1 0 #

Ek +1 = # # 0 1 0 # #
jth row
0 " 1 0 0 # #

# # # # # 1 #
0 0 " " " " 1

(2.15) det ( B ) = k det ( A ) where B is obtained by multiplying one row of A by k

21

Chapter 2: Determinants

To find the determinant of this matrix we can expand along the kth row where kth row
is not one of ith or jth row. Note that in the kth row all the entries should be zeros apart
from the diagonal element ekk which is equal to 1. Therefore the determinant of this
matrix Ek +1 is
det ( Ek +1 ) = ( 1)

k +k

det ( Ek )

(*)

Why?
Because if you delete the elements containing the row and column containing the entry
ekk then the remaining matrix is the k by k elementary matrix with rows i and j
interchanged, which is Ek . What is the determinant of Ek ?

By our induction hypothesis we have det ( Ek ) = 1 . Substituting this into (*) gives
det ( Ek +1 ) = ( 1)

( 1)
2k
= ( 1) ( 1) = 1
k +k

Because ( 1)2 k = 1

Hence we have proven that the determinant of an elementary matrix of size k + 1 by


k + 1 which has two rows interchanged is 1 .
Therefore by induction we have our result that if the elementary matrix E is obtained
from the identity matrix I by swapping two rows (or columns) then det ( E ) = 1 .

Note that we can use this Proposition (2.20) to find the determinants of elementary
matrices. Summarizing this Proposition (2.20) we have that the determinant of an
elementary matrix E is given by
1 if a multiple of one row is added to another

(2.21)
det ( E ) = 1 if two rows have been interchanged
k if a row has been multiplied by non-zero k

Example 11
Find the determinants of the following elementary matrices:
1 0 0
0 0 1
1 0 5

(a) A = 0 5 0
(b) B = 0 1 0
(c) C = 0 1 0
0 0 1
1 0 0
0 0 1

Solution
Each of these is an elementary matrix so we can use the above Result (2.21).
(a) Since this A is the elementary matrix obtained from the identity matrix by
multiplying the middle row by 5 we have, by Result (2.21) with k = 5 ,
det ( A ) = 5
(b) How is matrix B obtained from the identity matrix?
By interchanging top and bottom rows, therefore by the middle line of (2.21) we have
det ( B ) = 1
(c) How is matrix C obtained from the identity matrix?
By subtracting 5 times the bottom row from the top row. Hence by the top line of
Result (2.21) we have
det ( C ) = 1

22

Chapter 2: Determinants

B4 Determinant Properties of Other Matrices


We can extend Proposition (2.20) or result (2.21) to any square matrix A as the
following proposition states.
Proposition (2.22). Let B be a matrix obtained from the matrix A by
(a)
multiplying a row (or column) by a non-zero scalar k. In this case
det ( B ) = k det ( A ) .
(b)
adding (or subtracting) a multiple of one row to another. In this case
det ( B ) = det ( A ) .
(c)

interchanging two rows (or columns). In this case det ( B ) = det ( A ) .

Proof.
Part (a) was proven in Exercise 2(a) and is really Proposition (2.11) stated earlier in this
section. See Exercise 2(b) for proofs of parts (b) and (c).
We can summarize this into the following result.
det ( A ) if a multiple of one row is added to another

(2.23)
det ( B ) = det ( A ) if two rows have been interchanged

k det ( A ) if a row has been multiplied by non-zero k


In the next example we apply this result (2.23) to find the determinant of a matrix
which has fractional entries.
Example 12
Find the determinant of the following matrix:
2
1

1
23 23

1
1
5

A=
2
6
6

3
1
1

11
11
11
Solution
How can we find the determinant of this matrix?
We can find the determinant of another matrix which is matrix A with top row
multiplied by 23, second row multiplied by 6 and bottom row multiplied by 11.
How is the determinant of this new matrix, call it B say, related to the determinant of
matrix A?
23
1 2

()
( 23 6 11) det ( A ) = det ( B ) = det 3 1 5
1 3 1

We can find the determinant of matrix B as in the last section:


1 2 23
1 5
3 5
3 1

det ( B ) = det 3 1
5 = det

( 2 ) det
+ 23det
3 1
1 1
1 3
1 3 1

= ( 1 15 ) + 2 ( 3 5 ) + 23 ( 9 1)
= 16 16 + 23 ( 8 ) = 152

Substituting this, det ( B ) = 152 , into () gives

23

Chapter 2: Determinants

( 23 6 11) det ( A ) = 152


det ( A ) =

Hence det ( A ) =

152
152
76
=
=
( 23 6 11) 1518 759

76
.
759

For larger size matrices it is normally easier to convert these into triangular matrices by
applying row operations. Why carry out this conversion?
Because the determinant of a triangular matrix is just the product of the entries on the
leading diagonal. The following two examples shows how we use this approach.
Example 13
Find the determinant of the following matrix by using row operations:
1 2 2 4

7 8 3 0

A=
3 2 0 0

1 0 0 0
Solution
Can we convert this into a triangular matrix?
Yes by using row operations. Note that the given matrix A is not a triangular matrix
because both sides of the leading diagonal contain non-zero entries. You will need to
recall your work from chapter 1 on row operations. How can we convert the matrix A
into a triangular matrix?
First we label the rows of matrix A and then we apply row operations:
R1 1 2 2 4
R2 7 8 3 0
R3 3 2 0 0

R4 1 0 0 0
Interchanging rows R1 and R4 , R2 and R3 we have

R4 1 0 0 0
R3 3 2 0 0
R2 7 8 3 0

R1 1 2 2 4
What is the determinant of this matrix?
We have a lower triangular matrix so the determinant is the product of the entries on
the leading diagonal, that is 1 2 3 4 = 24 . What is the determinant of the given
matrix A?
The bottom matrix is obtained from matrix A by interchanging rows R1 and R4 , R2
and R3 . How does interchanging rows affect the determinant of the matrix?
By (2.23) interchanging rows multiplies the determinant by 1 . Since we have two
interchanges therefore the determinant of the matrix A is given by
( 1)( 1) det ( A ) = 24 which gives det ( A ) = 24
(2.23)

det ( B ) = det ( A )

if two rows have been interchanged

Chapter 2: Determinants

24

Example 14
Find the determinant of the following matrix:
5 1 1 1 1

1 5 1 1 1
A = 5 1 5 1 1

1 5 1 5 1
5 1 5 1 5

Solution
Labelling the rows of this matrix we have
R1 5 1 1 1 1
R2 1 5 1 1 1
R3 5 1 5 1 1

R4 1 5 1 5 1
R5 5 1 5 1 5
Executing the following row operations:
R1
5 1 1 1 1
R2 * = R2 R1 4 4 0 0 0
R3 * = R3 R1 0 0 4 0 0

R4 * = R4 R2 0 0 0 4 0
R5 * = R5 R3 0 0 0 0 4
Carrying out the row operation 5 R2 * gives
R1
5 1 1 1 1
R2 ** = 5R2 * 20 20 0 0 0
0 0 4 0 0
R3 *

R4 *
0 0 0 4 0
0 0 0 0 4
R5 *

Executing the row operation R2 ** + 4 R1 yields


R1
5 1 1 1 1
R2 *** = R2 ** + 4 R1 0 24 4 4 4
0 0 4 0 0
R3 *

R4 *
0 0 0 4 0
0 0 0 0 4
R5 *

What is the determinant of this matrix?


We have an upper triangular matrix so the determinant is the product of all the entries
on the leading diagonal, that is 5 24 4 4 4 = 7680 . What is the determinant of the
given matrix A?
All the above row operations apart from 5R2 * makes no difference to the determinant.
Why not?
Because by result (2.23) adding a multiple of one row to another has the same
determinant. How does the row operation 5R2 * change the determinant?
(2.23)

det ( B ) = det ( A )

if a multiple of one row is added to another

Chapter 2: Determinants

25

By (2.23) we have the above determinant 7680 is 5 det ( A ) , that is


5det ( A ) = 7680 which gives det ( A ) =

7680
= 1536
5

Proposition (2.24). Let E be an elementary matrix. For any square matrix A of the same
size as E we have
det ( EA ) = det ( E ) det ( A )

What does this proposition mean?


Means the determinant of matrix multiplication EA is equal to the determinant of the
elementary matrix E times the determinant of matrix A.
Proof.
We consider the three different cases of elementary matrices separately.
Case 1. Let E be the elementary matrix obtained from the identity matrix by adding a
multiple of one row to another. Then from Chapter 1 we have that the matrix
multiplication EA performs the same row operation of adding one row to another of
matrix A. By the first line of Result (2.23) we have
det ( EA ) = det ( A )
By result (2.21) we have det ( E ) = 1 therefore

det ( EA ) = det ( A )
= 1 det ( A ) = det ( E ) det ( A )
Hence for the first case we have det ( EA ) = det ( E ) det ( A ) .
Case 2. Let E be the elementary matrix obtained from the identity matrix by
multiplying a row (or column) by a non-zero scalar k. By result (2.21) we have
det ( E ) = k . The matrix multiplication EA performs the same row operation of
multiplying a row by a non-zero scalar k on matrix A. By result (2.23) we have
det ( EA ) = k det ( A ) .
det ( EA ) = kN det ( A ) = det ( E ) det ( A )
= det ( E )

We have proven det ( EA ) = det ( E ) det ( A ) for the second case.


Case 3. See Exercise 2(b).
Proposition (2.25). Let E1 , E2 , E3 , " and Ek be elementary matrices and B be a
square matrix of the same size. Then
det ( E1E2 E3 " Ek B ) = det ( E1 ) det ( E2 ) det ( E3 )" det ( Ek ) det ( B )
What does this proposition mean?
Means if we have a matrix multiplication E1E 2 E3 " Ek B then the determinant of this is
equal to determinant of each matrix multiplied together.
Proof.
Exercise 2(b).

(2.23)

det ( A ) if a multiple of one row is added to another

det ( B ) = det ( A ) if two rows have been interchanged

k det ( A ) if a row has been multiplied by non-zero k

Chapter 2: Determinants

26

Next we prove an important property of invertible (non-singular) matrices.


Theorem (2.26). A square matrix A is invertible (non-singular) if and only if
det ( A ) 0
What does this proposition mean?
Means if matrix A is invertible then the determinant of A does not equal zero. Also if
the determinant is not equal to zero then the matrix A is invertible (non-singular).
How do we prove this result?
Since it is an if and only if statement, we need to prove it both ways. That is we first
assume matrix A is invertible and show that this leads to det ( A ) 0 . This direction is

normally symbolized by . Then we assume det ( A ) 0 and prove that matrix A is


invertible. This direction is normally symbolized by .
Proof.
( ). Assume the matrix A is invertible. By Proposition (1.27) part d) of chapter 1 we
know the matrix A is a product of elementary matrices. We can write the matrix A as
A = E1E2 E3 " E k
where E1 , E2 , E3 , " and Ek are elementary matrices. By Proposition (2.25) we
have
det ( A ) = det ( E1E2 E3 " E k )
= det ( E1 ) det ( E2 ) det ( E3 )" det ( Ek )
Remember
(2.21)

det ( E ) = 1
k

if a multiple of one row is added to another


if two rows have been interchanged
if a row has been multiplied by non-zero k

The determinant of an elementary matrix can only be 1, 1 or the non-zero k.


Multiplying these non-zero numbers, det ( E1 ) det ( E2 ) det ( E3 )" det ( Ek ) , cannot give
0. Therefore det ( A ) 0 .
Now we go the other way ( ). Assume det ( A ) 0 then by Proposition (2.13) we
1
adj ( A ) which means that matrix A is invertible (non-singular).
det ( A )
Hence we have proven our result.

have A 1 =

We can extend Proposition (2.24) to any two square matrices of the same size as the
next proposition states.
Proposition (2.27). If A and B are square matrices of the same size then
det ( AB ) = det ( A ) det ( B )
Proof.
We consider the two cases of matrix A. Case 1 is where the matrix A is invertible (nonsingular) and case 2 is where matrix A is non-invertible (singular).
Case 1. Assume the matrix A is invertible. Then by Proposition (1.27) part d) of chapter
1 we know matrix A is a product of elementary matrices. We can write
A = E1E2 E3 " E k
(1.27) Part (d) A is invertible A is a product of elementary matrices

Chapter 2: Determinants

27

where E1 , E2 , E3 , " and Ek are elementary matrices. We have


det ( AB ) = det ( E1E 2 E3 " Ek B )

= det ( E1 ) det ( E2 ) det ( E3 )" det ( Ek ) det ( B )

[ By (2.25)]

= det E1E2 E3 " Ek det ( B )


=A
= det ( A ) det ( B )
Case 2. Assume matrix A is non-invertible (singular). By the above Proposition (2.26)
we conclude that det ( A ) = 0 . Since matrix A is non-invertible therefore matrix

multiplication AB is also non-invertible. Why?


Suppose AB is invertible then by Proposition (1.20) of the last chapter we have matrix
A (and B) is invertible. Since AB is also non-invertible therefore det ( AB ) = 0 . Hence
we have our result

det ( AB ) = det ( A ) det ( B )

because det ( A ) = 0 and det ( AB ) = 0 .

Again we can extend the result of Proposition (2.27) to n square matrices as the next
proposition states.
Proposition (2.28). If A1 , A 2 , A 3 , " and A n are square matrices of the same size
then
det ( A1A 2 A 3 " A n ) = det ( A1 ) det ( A 2 ) det ( A 3 )" det ( A n )

Proof.
Exercise 2(b).
Generally a function in mathematics which has these properties of Proposition (2.27) or
(2.28) is called a multiplicative function.
Generally in mathematics we say a function f is multiplicative if
f ( xy ) = f ( x ) f ( y )
The determinant is an example of a multiplicative function.
Proposition (2.29). If A is an invertible (non-singular) matrix then
1
det ( A 1 ) =
det ( A )
Proof. Exercise 2(b)

(1.20)
(2.25)
(2.26)

( AB ) = B 1A 1
det ( E1E2 E3 " Ek B ) = det ( E1 ) det ( E2 ) det ( E3 )" det ( Ek ) det ( B )
A square matrix A is invertible (non-singular) if and only if det ( A ) 0
1

Chapter 2: Determinants

28

SUMMARY
A triangular matrix is a square matrix where all entries to one side of the leading
diagonal are zero.
A diagonal matrix is a square matrix where all entries to both sides of the leading
diagonal are zero.
The determinant of a triangular or diagonal matrix is a product of the entries along the
leading diagonal.
det ( A ) if a multiple of one row is added to another

(2.23)
det ( B ) = det ( A ) if two rows have been interchanged

k det ( A ) if a row has been multiplied by non-zero k


A square matrix A is invertible if and only if the determinant of A does not equal to
zero.
If A and B are square matrices of the same size then
det ( AB ) = det ( A ) det ( B )

Potrebbero piacerti anche