Module 2

Systems of Linear Equations

The Inverse of a Matrix

Learning Objectives:
At the end of this unit, you are expected to:
✓ Write an augmented matrix from a system of linear
equations, or translate a matrix into a system of linear
equations.
✓ Solve a system of linear equations using Gauss-Jordan
Elimination Method.
✓ Solve a homogeneous system of linear equations.
✓ Explain the inverse of a matrix.

2.1 Systems of Linear Equations

2.1.1 Writing a System of Linear Equations in Matrix Form
2.1.2 Solving System of Linear Equations Using the Gauss – Jordan
Elimination Method

2.2 The Inverse of a Matrix

2.2.1 Definition and Examples of Inverse
2.2.2 Properties of the Inverse
2.2.3 Linear Systems and Inverses

1|Module 2
2.1 Systems of Linear Equations

We have learned from Module 1 that Systems of Linear Equations (Linear Systems, for short)
are either:
1. independent/consistent (unique solution);
2. inconsistent (no solution); or
3. dependent (many solutions).

Also, in Algebra, we can either solve Linear Systems using the following methods:
1. elimination
2. substitution
3. a combination of elimination and substitution methods

In this module, we will be able to solve linear systems using a new method named the
Gauss-Jordan Elimination Method. But before we do this, let us first learn how to transform the
linear system in matrix form.

Let us consider the system of linear equations:

𝑥 + 2𝑦 + 3𝑧 = 6
{ − 3𝑦 + 2𝑧 = 14
2𝑥
3𝑥 + 𝑦 − 𝑧 = −2

Now define the following matrices:

1 2 3 𝑥 6
𝑨 = [2 −3 2 ] 𝒙 = [𝑦] 𝑎𝑛𝑑 𝒃 = [ 14 ]
3 1 −1 𝑧 −2

Observe that the 3x3 matrix A consists of the numerical coefficients of the linear system
above. The entries in the first row of A coincide with the numerical coefficients (in red fonts
below) of the first linear equation. (Of course, variables with no numerical coefficients indicate a
numerical coefficient value equal to 1.)
1 2 3 𝑥 + 2𝑦 + 3𝑧 = 6
𝑨 = [2 −3 2 ] 2𝑥 − 3𝑦 + 2𝑧 = 14
3 1 −1 3𝑥 + 𝑦 − 𝑧 = −2

2|Module 2
 The second row of A consist of the numerical coefficients (in blue fonts) of the second
linear equation and the entries in third row of A are the numerical coefficients of the third linear
equation (in green fonts). Matrix A is called the COEFFICIENT MATRIX. Do you know why?
Indeed! It is called the coefficient matrix since this matrix consists of the numerical coefficients
of the given linear system.

What about the vector x?

𝑥 𝑥 + 2𝑦 + 3𝑧 = 6
𝒙 = [𝑦 ] 2𝑥 − 3𝑦 + 2𝑧 = 14
𝑧 3𝑥 + 𝑦 − 𝑧 = −2

Yes, x is formed from the variables (or the literal coefficients) used in the given system.
These are the unknowns, right? The first variable (𝑥) is written on the first row, second variable
(𝑦) on the second row of the vector x and the third variable (𝑧) on the third row of x.

Finally, the entries in vector b are the values on the right-hand side of each of the three
equations:

6 𝑥 + 2𝑦 + 3𝑧 = 6
𝒃 = [ 14 ] 2𝑥 − 3𝑦 + 2𝑧 = 14
−2 3𝑥 + 𝑦 − 𝑧 = −2

Unlike matrix A, vectors x and b have no names. But definitely, we know where they came
from right?

To summarize, the system of linear equations can be written in three matrices (two of which
are actually vectors), Ax = b. Observe that the left–hand side of the equation, Ax, involves matrix
multiplication. If we perform this, we will have

1 2 3 𝑥 𝑥 + 2𝑦 + 3𝑧
𝑨𝒙 = [2 −3 2 ] [𝑦] = [2𝑥 − 3𝑦 + 2𝑧]
3 1 −1 𝑧 3𝑥 + 𝑦 − 𝑧

What do you observe from the product Ax? Correct, it is actually the expressions on the
left–hand side of the system of linear equations!

3|Module 2
 1 2 3 𝑥 𝑥 + 2𝑦 + 3𝑧 𝑥 + 2𝑦 + 3𝑧 = 6
𝑨𝒙 = [2 −3 2 ] [𝑦] = [2𝑥 − 3𝑦 + 2𝑧] 2𝑥 − 3𝑦 + 2𝑧 = 14
3 1 −1 𝑧 3𝑥 + 𝑦 − 𝑧 3𝑥 + 𝑦 − 𝑧 = −2

Certainly, Ax = b is the matrix representation of the given system of linear equations.

1 2 3 𝑥 6 𝑥 + 2𝑦 + 3𝑧 = 6
𝑨𝒙 = 𝒃 ∶ [2 −3 2 ] [𝑦] = [ 14 ] 2𝑥 − 3𝑦 + 2𝑧 = 14
3 1 −1 𝑧 −2 3𝑥 + 𝑦 − 𝑧 = −2

We were now able to transform system of linear equations to matrix form and vice versa.
Let us now formally define a system of linear equations in its general form.

DEFINITION: SYSTEM OF LINEAR EQUATIONS

A system of linear equations (or simply linear system) is a set of equations with m
equations and n unknowns, is of the form
𝑎11 𝑥1 + 𝑎12 𝑥2 + 𝑎13 𝑥3 + ⋯ + 𝑎1𝑛 𝑥𝑛 = 𝑏1
𝑎21 𝑥1 + 𝑎22 𝑥2 + 𝑎23 𝑥3 + ⋯ + 𝑎2𝑛 𝑥𝑛 = 𝑏2
{
⋮
𝑎𝑚1 𝑥1 + 𝑎𝑚2 𝑥2 + 𝑎𝑚3 𝑥3 + ⋯ + 𝑎𝑚𝑛 𝑥𝑛 = 𝑏𝑛
The unknowns are denoted by the variables 𝑥𝑗 , 1 ≤ 𝑗 ≤ 𝑛, the numerical coefficients by
𝑎𝑖𝑗 , 1 ≤ 𝑖 ≤ 𝑚, 1 ≤ 𝑗 ≤ 𝑛 and the constant terms by 𝑏𝑗 , 1 ≤ 𝑗 ≤ 𝑛. In matrix form, the system
of equations above can be written as:
𝑎11 𝑎12 𝑎13 … 𝑎1𝑛 𝑥1 𝑏1
𝑎21 𝑎22 𝑎23 … 𝑎2𝑛 𝑥2 𝑏2
𝑎31 𝑎32 𝑎33 … 𝑎3𝑛 𝑥3 = 𝑏3
⋮ ⋮ ⋮ … ⋮ ⋮ ⋮
𝑎
[ 𝑚1 𝑎𝑚2 𝑎𝑚3 … 𝑎𝑚𝑛 ] [𝑥𝑛 ] [𝑏𝑛 ]
A simplified way of writing above is like this: Ax = b.

Let’s try these…

Can you transform this linear system to matrix form?

𝑥 − 2𝑦 + 3𝑧 = 9 1 −2 3 𝑥 9
{ −𝑥 + 3𝑦 = −4 𝐴𝑛𝑠𝑤𝑒𝑟: 𝑨 = [−1 3 0] 𝒙 = [𝑦] 𝒃 = [−4]
2𝑥 − 5𝑦 + 5𝑧 = 17 2 −5 5 𝑧 17

4|Module 2
What about transforming this matrix to a linear system?

0 1 −1 𝑥1 0 𝑥 2 − 𝑥3 = 0
[ 1 0 −3] [𝑥2 ] = [−1] Answer: { 𝑥1 − 3𝑥3 = −1
−1 3 0 𝑥3 1 −𝑥1 + 3𝑥2 = 1

Did you know that aside from the matrix representation Ax = b, we can rewrite our linear
system using only one matrix instead of three? This is done by adjoining matrices A and b together
forming what we call the AUGMENTED MATRIX, [𝑨 ⋮ 𝒃]. How does this matrix look like?

If we consider the previous example, let us adjoin matrices A and b:

1 2 3 6
𝑨 = [2 −3 2 ] 𝒃 = [ 14 ]
3 1 −1 −2

1 2 3 ⋮ 6
The augmented matrix is: [𝑨 ⋮ 𝒃] = [2 −3 2 ⋮ 14 ].
3 1 −1 ⋮ −2

What did we just do? Yes, matrix A is augmented with matrix b. (Just like when ladies
wanted their breasts to be enhanced, they have to undergo breast augmentation!)

But what happens to the vector x? If you will recall, we have said that the vector x is
comprised of the variables. Since these are just variables, then we can have many other
representations of vector x, right? We can use the variables 𝒙𝟏 , 𝒙𝟐 , and 𝒙𝟑 instead of the
𝑥1 𝑥
𝑥
variables x, y and z, making vector 𝒙 = [ 2 ] instead of 𝒙 = [𝑦]. Therefore, it does not matter
𝑥3 𝑧
what vector x holds for as long as we have A and b.

Thus, the given linear system can be represented by:

𝑥 + 2𝑦 + 3𝑧 = 6 1 2 3 ⋮ 6
{ 2𝑥 − 3𝑦 + 2𝑧 = 14 [𝑨 ⋮ 𝒃] = [2 −3 2 ⋮ 14 ]
3𝑥 + 𝑦 − 𝑧 = −2 3 1 −1 ⋮ −2

Therefore, aside from Ax = b, we can also represent our linear systems by the augmented
matrix [𝑨 ⋮ 𝒃].

5|Module 2
 𝑥 + 2𝑦 + 3𝑧 = 6 1 2 3 𝑥 6 1 2 3 ⋮ 6
{ 2𝑥 − 3𝑦 + 2𝑧 = 14 𝑨𝒙 = 𝒃 ∶ [2 −3 2 ] [𝑦] = [ 14 ] ; [𝑨 ⋮ 𝒃] = [2 −3 2 ⋮ 14 ]
3𝑥 + 𝑦 − 𝑧 = −2 3 1 −1 𝑧 −2 3 1 −1 ⋮ −2

Now, can you transform this linear system in augmented matrix?

𝑥 − 2𝑦 + 3𝑧 = 9 1 −2 3 ⋮ 9
{ −𝑥 + 3𝑦 = −4 𝐴𝑛𝑠𝑤𝑒𝑟: [𝑨 ⋮ 𝒃] = [−1 3 0 ⋮ −4]
2𝑥 − 5𝑦 + 5𝑧 = 17 2 −5 5 ⋮ 17

That was easy, isn’t it? Let us now use these ideas along with the previous lessons on
elementary row operations and matrices in row reduced echelon form to solve system of linear
equations using the Gauss – Jordan Reduction Method.

Let us try to solve the given system of linear equations from the previous example by
transforming the augmented matrix in reduced row echelon form.

𝑥 + 2𝑦 + 3𝑧 = 6
Example 1: { 2𝑥 − 3𝑦 + 2𝑧 = 14
3𝑥 + 𝑦 − 𝑧 = −2

1 2 3 ⋮ 6
Using the augmented matrix, [𝑨 ⋮ 𝒃] = [ 2 −3 2 ⋮ 14 ], we consider the entry in the
3 1 −1 ⋮ −2
first row, first column as the pivot. Since it is the leading one (in red font), let us make the entries
below it to be equal to zero by performing the third elementary row operation on the second and
third rows:

1 2 3 ⋮ 6 𝒑𝒊𝒗𝒐𝒕: 1 1 2 3 ⋮ 6
[𝑨 ⋮ 𝒃] = [2 −3 2 ⋮ 14 ] −𝟐𝑹𝟏 + 𝑹𝟐 ≈ [ 0 −7 −4 ⋮ 2 ]
3 1 −1 ⋮ −2 −𝟑 𝑹𝟏 + 𝑹𝟑 0 −5 −10 ⋮ −20

Solution:

−𝟐𝑹𝟏 : − 𝟐 [1 2 3 ⋮ 6] → [−2 −4 −6 ⋮ −12]

+ 𝑹𝟐 : + [2 −3 2 ⋮ 14] + [ 2 −3 2 ⋮ 14]
[ 0 −7 −4 ⋮ 2]

−𝟑𝑹𝟏 : − 𝟑 [1 2 3 ⋮ 6] → [−3 −6 −9 ⋮ −18]

+ 𝑹𝟑 : + [3 1 −1 ⋮ −2] +[ 3 1 −1 ⋮ − 2]
[ 0 −5 −10 ⋮ −20]

6|Module 2
 Now that the first column is in place, we consider the entry on the second row, second
column (−7) to be the pivot. We make it the leading one by performing the second elementary
row operation on the second row: (− 𝟏⁄𝟕)𝑹𝟐.

1 2 3 ⋮ 6 1 2 3 ⋮ 6
[ 0 −7 −4 ⋮ 2 ] (− 𝟏⁄𝟕)𝑹𝟐 ≈ [0 1 4⁄7 ⋮ − 2⁄7]
0 −5 −10 ⋮ −20 0 −5 −10 ⋮ −20

Solution:

0 −7 −4 2
1 1 ⋮ 1
∗ − ⁄7 − ⁄7 − 1⁄7 ⋮ − ⁄7
0 1 4⁄ ⋮ − 2⁄
7 7

We now have a leading one on the second column (in red font). Let us make the other
entries on the second column equal to zero. What elementary row operations will be used? We
perform the third elementary row operation on the first and third rows to make the entries on the
pivotal column equal to zero.

1 0 13⁄ 46⁄
1 2 3 ⋮ 6 −𝟐𝑹𝟐 + 𝑹𝟏 7 ⋮ 7
[0 1 4 2
⁄7 ⋮ − ⁄7] 𝒑𝒊𝒗𝒐𝒕: 1 ≈ 0 1 4 2
⁄7 ⋮ − ⁄7
0 −5 −10 ⋮ −20 𝟓𝑹𝟐 + 𝑹𝟑 50 ⋮ −150
[0 0 − ⁄7 ⁄7]

Solution:

−𝟐𝑹𝟐 : − 𝟐 [0 1 4⁄7 ⋮ − 2⁄7] → [0 −2 − 8⁄7 ⋮ 4⁄7]

+ 𝑹𝟏 : + [ 1 2 3 ⋮ 6 ] + [1 2 3 ⋮ 6 ]
[1 0 13⁄7 ⋮ 46⁄7]

𝟓𝑹𝟐 : 𝟓 [0 1 4⁄7 ⋮ − 2⁄7] → [0 5 20⁄7 ⋮ − 10⁄7 ]

+ 𝑹𝟑 : + [0 −5 −10 ⋮ − 20 ] + [0 −5 −10 ⋮ − 20 ]
[ 0 0 − ⁄7 ⋮ − 150⁄7]
50

The next pivot is − 50⁄7 . What shall we do with it? Yes, let us make it the leading one
and then make the entries above it all equal to zero.

7|Module 2
 We use the second elementary row operation to make the pivot as the leading one.

1 0 13⁄ 46⁄
7 ⋮ 7 1 0 13⁄7 ⋮ 46⁄7
0 1 4⁄ 2
7 ⋮ − ⁄7 ≈ [0 1 4⁄7 ⋮ − 2⁄7]
⋮ (− 𝟕⁄𝟓𝟎)𝑹𝟑 ⋮
[0 0 − 50⁄7 −150⁄7] 0 0 1 3

Solution:

[0 0 − 50⁄7 ⋮ − 150⁄7]
∗ [− 7⁄50 − 7⁄50 − 7⁄50 ⋮ − 7⁄50 ]
[0 0 1 ⋮ 3]

We now have a leading one on the third column (in red font). Let us make the other entries
above it equal to zero. We perform the third elementary row operation on the first and second
rows to make the entries on the pivotal column equal to zero.

1 0 13⁄7 ⋮ 46⁄7 (−𝟏𝟑/𝟕)𝑹𝟑 + 𝑹𝟏 1 0 0 ⋮ 1

[0 1 4⁄7 ⋮ − 2⁄7] (−𝟒/𝟕)𝑹𝟑 + 𝑹𝟐 ≈ [0 1 0 ⋮ −2]
⋮ 𝒑𝒊𝒗𝒐𝒕: 1 0 0 1 ⋮ 3
0 0 1 3

Solution:

(− 𝟏𝟑⁄𝟕)𝑹𝟑 : (− 𝟏𝟑⁄𝟕) [0 0 1 ⋮ 3 ] → [0 0 − 13⁄7 ⋮ − 39⁄7]

+ 𝑹𝟏 : + [1 0 13⁄7 ⋮ 46⁄7] + [1 0 13⁄ ⋮
7
46⁄ ]
7
[1 0 0 ⋮ 1 ]

(− 𝟒⁄𝟕)𝑹𝟑 : (− 𝟒⁄𝟕) [0 0 1 ⋮ 3 ] → [0 0 − 4⁄7 ⋮ − 12⁄7]

+ 𝑹𝟐 : + [0 1 4⁄7 ⋮ − 2⁄7] + [0 1 4⁄
7 ⋮ − 2⁄7 ]
[0 1 0 ⋮ −2 ]

1 0 0 ⋮ 1
The matrix [0 1 0 ⋮ −2] is now in its reduced row echelon form. If we transform this
0 0 1 ⋮ 3
matrix in a system of linear equations with variables x, y and z, we have:

𝑥 + 0𝑦 + 0𝑧 = 1 𝑥=1
{ + 𝑦 + 0𝑧 = −2
0𝑥 or simply {𝑦 = −2
0𝑥 + 0𝑦 + 𝑧 = 3 𝑧= 3

8|Module 2
which gives us the values of our variables, the solution to the system of linear equations. Let us
try to check if this is really a solution to the linear system by substituting the values in the three
equations.

𝑥 + 2𝑦 + 3𝑧 = 6 𝑥=1
{ 2𝑥 − 3𝑦 + 2𝑧 = 14 with {𝑦 = −2
3𝑥 + 𝑦 − 𝑧 = −2 𝑧= 3

x y z

Equation 1: 𝑥 + 2𝑦 + 3𝑧 = 6 1 + 2(−2) + 3(3) = 1 − 4 + 9 = 6

Equation 2: 2𝑥 − 3𝑦 + 2𝑧 = 14 2(1) − 3(−2) + 2(3) = 2 + 6 + 6 = 14

Equation 3: 3𝑥 + 𝑦 − 𝑧 = −2 3(1) + (−2) − 3 = 3 − 2 − 3 = −2

satisfying all the equations simultaneously. Hence, (1, −2, 3) is a solution to the given system of
linear equations. It is a unique solution making the linear system consistent. Isn’t it amazing?

What did we just do? We were able to solve the given linear system by reducing the
augmented matrix in reduced row echelon form. This process is called the Gauss – Jordan
Reduction Method. Let us summarize the procedure:

1. Form the augmented matrix [𝑨 ⋮ 𝒃].

2. Transform [𝑨 ⋮ 𝒃] to reduced row echelon form [𝑪 ⋮ 𝒅] by using a series of elementary
row operations.
3. For each nonzero row of the matrix [𝑪 ⋮ 𝒅], solve the corresponding equation for the
unknown associated with the leading one in that row. The rows consisting entirely of
zeros can be ignored, because the corresponding equation will be satisfied for any
values of the unknowns.

You can review the example before proceeding to our next example. On the next page, let
us look at another linear system and try to solve using the Gauss – Jordan Reduction Method.

9|Module 2
 𝑥 − 2𝑦 + 3𝑧 = 9
Example 2: { −𝑥 + 3𝑦 = −4
2𝑥 − 5𝑦 + 5𝑧 = 17

1. Let us form the augmented matrix [𝑨 ⋮ 𝒃]. How is the augmented matrix formed? Yes, it
is formed by augmenting the vector b with the coefficient matrix A. Hence,
1 −2 3 ⋮ 9
[𝑨 ⋮ 𝒃] = [−1 3 0 ⋮ −4]
2 −5 5 ⋮ 17

2. We are now ready to transform [𝑨 ⋮ 𝒃] to reduced row echelon form [𝑪 ⋮ 𝒅] by using a

series of elementary row operations.

𝟏 −2 3 ⋮ 9 𝒑𝒊𝒗𝒐𝒕: 𝟏 𝟏 −2 3 ⋮ 9
[𝑨 ⋮ 𝒃] = [−1 3 0 ⋮ −4] 𝑹𝟏 + 𝑹𝟐 ≈ [0 1 3 ⋮ 5]
2 −5 5 ⋮ 17 (−𝟐)𝑹𝟏 + 𝑹𝟑 0 −1 −1 ⋮ −1

Solution:

𝑹𝟏 : [1 −2 3 ⋮ 9 ]
+𝑹𝟐 : + [−1 3 0 ⋮ −4]
[0 1 3 ⋮ 5]

(−𝟐)𝑹𝟏 : (−𝟐)[1 −2 3 ⋮ 9 ] → [−2 4 −6 ⋮ −18 ]

+𝑹𝟑 : + [2 −5 5 ⋮ 17] → + [ 2 −5 5 ⋮ 17 ]
[ 0 −1 −1 ⋮ − 1 ]

Fortunately, for the second column, we have a pivot that is also a leading one.

1 −2 3 ⋮ 9 (𝟐)𝑹𝟐 + 𝑹𝟏 1 0 9 ⋮ 19
[0 𝟏 3 ⋮ 5 ] 𝒑𝒊𝒗𝒐𝒕: 𝟏 ≈ [0 1 3 ⋮ 5]
0 −1 −1 ⋮ −1 𝑹𝟐 + 𝑹𝟑 0 0 2 ⋮ 4

Can you show the solution?

(𝟐)𝑹𝟐 : (𝟐)[___ ___ ___ ⋮ ___] → [___ ___ ___ ⋮ ___ ]

+𝑹𝟏 : + [___ ___ ___ ⋮ ___] → + [ ___ ___ ___ ⋮ ___ ]
[ 1 0 9 ⋮ 19 ]
𝑹𝟐 : [___ ___ ___ ⋮ ___]
+𝑹𝟑 : + [___ ___ ___ ⋮ ___]
[0 0 2 ⋮ 4]

10 | M o d u l e 2
 Next, observe that we can perform first elementary row operation on the pivot of the third
column to make it a leading one. What must be multiplied to the third row?

1 0 9 ⋮ 19 1 0 9 ⋮ 19
[0 1 3 ⋮ 5] ≈ [0 1 3 ⋮ 5]
0 0 2 ⋮ 4 0 0 1 ⋮ 2

Solution:

0 0 2 ⋮ 4
∗ ___ ___ ___ ⋮ ___
0 0 1 ⋮ 2

Finally, we make the other entries on the third column above the leading one equal to zero.
How?

1 0 9 ⋮ 19 (___)𝑹𝟑 + 𝑹𝟏 1 0 0 ⋮ 1
[0 1 3 ⋮ 5 ] (___)𝑹𝟑 + 𝑹𝟐 ≈ [0 1 0 ⋮ −1]
0 0 1 ⋮ 2 𝒑𝒊𝒗𝒐𝒕: 𝟏 0 0 1 ⋮ 2

Solution:
(___)𝑹𝟑 : (___)[0 0 1 ⋮ 2 ] → [___ ___ ___ ⋮ ___ ]
+𝑹𝟏 : + [1 0 9 ⋮ 19] → + [1 0 9 ⋮ 19 ]
[1 0 0 ⋮ 1]

(___)𝑹𝟑 : (___) [0 0 1 ⋮ 2] → [___ ___ ___ ⋮ ___ ]

+𝑹𝟐 : + [0 1 3 ⋮ 5] → +[ 0 1 3 ⋮ 5 ]
[ 0 1 0 ⋮ −1]

3. We now have a matrix in reduced row echelon form. The given system has a unique
solution. Why? It is therefore consistent.

1 0 0 ⋮ 1 𝑥 + 0𝑦 + 0𝑧 = 1 𝑥=1
[0 1 0 ⋮ −1] 0𝑥 + 𝑦 + 0𝑧 = −1 𝑦 = −1
0 0 1 ⋮ 2 0𝑥 + 0𝑦 + 𝑧 = 2 𝑧= 2

𝑥 − 2𝑦 + 3𝑧 = 9 1 − 2(−1) + 3(2) = 1 + 2 + 6 =9
{ −𝑥 + 3𝑦 = −4 −(1) + 3(−1) + 0(2) = −1 + (−3) + 0 = −4
2𝑥 − 5𝑦 + 5𝑧 = 17 2(1) − 5(−1) + 5(2) = 2 + 5 + 10 = 17

It can be seen that (1,– 1, 2) is the only solution to the given linear system.

11 | M o d u l e 2
 Before we proceed to the next topic, let us have last two examples for the Gauss–Jordan
Elimination Method. This time, the solution is not detailed.

𝑥 2 − 𝑥3 = 0
Example 3: { 𝑥1 − 3𝑥3 = −1
−𝑥1 + 3𝑥2 = 1

0 1 −1 ⋮ 0
1. We form the augmented matrix [𝑨 ⋮ 𝒃] = [ 1 0 −3 ⋮ −1]
−1 3 0 ⋮ 1
2. We perform a series of elementary row operations on [𝑨 ⋮ 𝒃] until it is in its reduced row
echelon form [𝑪 ⋮ 𝒅] .

0 1 −1 ⋮ 0 1 0 −3 ⋮ −1 1 0 −3 ⋮ −1
[ 1 0 −3 ⋮ −1] 𝑹𝟏 ↔ 𝑹𝟐 ≈ [ 0 1 −1 ⋮ 0 ] ≈ [0 1 −1 ⋮ 0 ]
−1 3 0 ⋮ 1 −1 3 0 ⋮ 1 𝑹𝟏 + 𝑹𝟑 0 3 −3 ⋮ 0

1 0 −3 ⋮ −1 1 0 −3 ⋮ −1
[0 1 −1 ⋮ 0 ] ≈ [0 1 −1 ⋮ 0 ]
0 3 −2 ⋮ 0 (−𝟑)𝑹𝟐 + 𝑹𝟑 0 0 0 ⋮ 0

3. We now have a matrix in reduced row echelon form. But observe that there is a row
consisting entirely of zeros. What does this mean? If we transform the augmented into a
linear system, we will be able to come up with two equations in three unknowns:

1 0 −3 ⋮ −1 𝑥1 + 0𝑥2 − 3𝑥3 = −1 𝑥1 − 3𝑥3 = −1 𝑥1 = 3𝑥3 − 1

[0 1 −1 ⋮ 0 ]
0𝑥1 + 𝑥2 − 𝑥3 = 0 𝑥2 − 𝑥3 = 0 𝑥2 = 𝑥3
0 0 0 ⋮ 0

The values of the two variables 𝑥1 and 𝑥2 are dependent on the value of the variable z.
Hence, we have a dependent solution. This means that there are many solutions to the given linear
system. Why is this again? Suppose 𝑥3 = 0. Then 𝑥1 = 3𝑥3 − 1 = 3(0) − 1 = −1. Also, 𝑥2 =
𝑥3 = 0. Then, (−1, 0, 0) is a solution to the system.

But what if 𝑥3 = 1? Then 𝑥1 = 3(1) − 1 = 2. Also, 𝑥2 = 𝑥3 = 1. Then, (2, 1, 1) is

another solution to the system. We have two solutions so far. Can you give another solution to
the given linear system?

12 | M o d u l e 2
 3𝒂 − 2𝒃 + 4𝒄 = 1
Example 4: { 𝒂 + 𝒃 − 2𝒄 = 3
2𝒂 − 3𝒃 + 6𝒄 = 8

3 −2 4 ⋮ 1
1. We form the augmented matrix [𝑨 ⋮ 𝒃] = [1 1 −2 ⋮ 3]
2 −3 6 ⋮ 8
2. We perform a series of elementary row operations on [𝑨 ⋮ 𝒃] until it is in its reduced row
echelon form [𝑪 ⋮ 𝒅] .

3 −2 4 ⋮ 1 1 1 −2 ⋮ 3 1 1 −2 ⋮ 3
[1 1 −2 ⋮ 3] 𝑹𝟏 ↔ 𝑹𝟐 ≈ [3 −2 4 ⋮ 1] (−𝟑)𝑹 𝟏 + 𝑹 𝟐 ≈ [0 −5 10 ⋮ −8]
2 −3 6 ⋮ 8 2 −3 6 ⋮ 8 (−𝟐)𝑹𝟏 + 𝑹𝟑 0 −5 10 ⋮ 2

We have not transformed the given matrix in its reduced row echelon form yet. But look
at the second and third rows. What do you observe? If we are going to transform the second and
−5𝒃 + 10𝒄 = −8
third rows into linear equations, we will be having: {
−5𝒃 + 10𝒄 = 2

What is wrong with these two equations? Yes, if we are going to find values for the
variables b and c, and substitute it to these equations simultaneously, then we might only satisfy
only one of these two equations. Let us try to continue with the elementary row operations and
see what happens.

7
1 1 −2 ⋮ 3 1 1 −2 ⋮ 3 (−𝟏)𝑹𝟐 + 𝑹𝟏 1 0 0 ⋮ ⁄5
[0 −5 10 ⋮ −8] (− 𝟏⁄𝟓)𝑹𝟐 ≈ [0 1 −2 ⋮ 8⁄5] ≈ [0 1 −2 ⋮ 8⁄ ]
0 −5 10 ⋮ 2 0 −5 10 ⋮ 2 (𝟓)𝑹𝟐 + 𝑹𝟑 𝟎 𝟎 𝟎 ⋮ 5
𝟏𝟎

3. We now have a matrix in reduced row echelon form. Look at the entries in the last row of
the reduced matrix [𝑪 ⋮ 𝒅]. Transforming this to an equation, this would look like: 0𝒂 +
0𝒃 + 0𝒄 = 10 or simply, 0 = 10. But 0  10. This is inconsistent, right? Hence, this
system of linear equations has no solution.

We have seen examples of the three types of the system of linear equations. The Gauss –
Jordan Elimination Method still applies to any linear system with m equations in n unknowns. It
also applies to whatever type a linear system may be. A consistent system has an augmented
matrix that is row equivalent to an identity matrix (on the coefficient matrix side of the augmented
matrix). A system of linear equations usually has a dependent solution if the augmented matrix is

13 | M o d u l e 2
row equivalent to a matrix whose entries on the last row/s is/are all equal to zero. On the other
hand, if there is a row with entries that is all zero in the coefficient matrix A part but not on the
vector b part, then we have an inconsistent solution.

The following theorem and corollary may be of help when solving systems of linear
equations:

Theorem: Let 𝑨𝒙 = 𝒃 and 𝑪𝒙 = 𝒅 be two linear systems each of m equations in n unknowns. If

the augmented matrices [𝑨 ⋮ 𝒃] and [𝑪 ⋮ 𝒅] of these systems are row equivalent, then
both linear systems have exactly the same solutions.

Corollary: If A and C are row equivalent mxn matrices, then the linear system Ax = 0 and Cx = 0
have exactly the same solutions.

Look at the linear systems Ax = 0 and Cx = 0. These types of linear systems are called

HOMOGENEOUS SYSTEM of LINEAR EQUATIONS

DEFINITION: HOMOGENEOUS SYSTEMS

A linear system of the form

𝑎11 𝑥1 + 𝑎12 𝑥2 + 𝑎13 𝑥3 + ⋯ + 𝑎1𝑛 𝑥𝑛 = 0

𝑎 𝑥 + 𝑎22 𝑥2 + 𝑎23 𝑥3 + ⋯ + 𝑎2𝑛 𝑥𝑛 = 0
{ 21 1
⋮
𝑎𝑚1 𝑥1 + 𝑎𝑚2 𝑥2 + 𝑎𝑚3 𝑥3 + ⋯ + 𝑎𝑚𝑛 𝑥𝑛 = 0
is called a homogeneous system. A homogeneous system may be written in matrix form
as Ax = 0.

Observe that even without solving the system, we know that 𝑥1 = 𝑥2 = 𝑥3 = ⋯ = 𝑥𝑛 = 0 is

a solution. (Why?) Such a solution is called the trivial solution. Hence, any homogenous system
is always consistent because of the trivial solution. A solution to a homogeneous system in which
not all the 𝑥𝑖 ’s are zero is called a nontrivial solution. Therefore, a homogeneous system may
either have a consistent solution (trivial solution) or dependent solution (nontrivial solution).

14 | M o d u l e 2
 𝑥 − 2𝑦 + 3𝑧 = 0
Example 1: Solve the homogeneous system: { −𝑥 + 3𝑦 =0
2𝑥 − 5𝑦 + 5𝑧 = 0

Since the trivial solution applies to any homogeneous system, then p = q = r = s = 0 is a

solution to the given system. Let us see if there are other solutions aside from the trivial solution
using the Gauss – Jordan Elimination Method.

1 −2 3 ⋮ 0
1. We form the augmented matrix [𝑨 ⋮ 𝟎] = [−1 3 0 ⋮ 0]
2 −5 5 ⋮ 0
2. We perform a series of elementary row operations on [𝑨 ⋮ 𝒃] until it is in its reduced row
echelon form [𝑪 ⋮ 𝟎] .

𝟏 −2 3 ⋮ 0 𝒑𝒊𝒗𝒐𝒕: 𝟏 𝟏 −2 3 ⋮ 0
[𝑨 ⋮ 𝒃] = [−1 3 0 ⋮ 0] 𝑹𝟏 + 𝑹𝟐 ≈ [0 1 3 ⋮ 0]
2 −5 5 ⋮ 0 (−𝟐)𝑹𝟏 + 𝑹𝟑 0 −1 −1 ⋮ 0

Solution:

𝑹𝟏 : [1 −2 3 ⋮ 0]
+𝑹𝟐 : + [−1 3 0 ⋮ 0]
[ 0 1 3 ⋮ 0]

(−𝟐)𝑹𝟏 : (−𝟐)[1 −2 3 ⋮ 0] → [−2 4 −6 ⋮ 0 ]

+𝑹𝟑 : + + [2 −5 5 ⋮ 0] → + [ 2 −5 5 ⋮ 0]
[ 0 −1 −1 ⋮ 0 ]

Fortunately, for the second column, we have a pivot that is also a leading one.

1 −2 3 ⋮ 0 (𝟐)𝑹𝟐 + 𝑹𝟏 1 0 9 ⋮ 0
[0 𝟏 3 ⋮ 0] 𝒑𝒊𝒗𝒐𝒕: 𝟏 ≈ [0 1 3 ⋮ 0]
0 −1 −1 ⋮ 0 𝑹𝟐 + 𝑹𝟑 0 0 2 ⋮ 0

Can you show the solution?

(𝟐)𝑹𝟐 : (𝟐)[0 1 3 ⋮ 0] → [0 2 6 ⋮ 0]
+𝑹𝟏 : + [0 −2 3 ⋮ 0] → + [0 −2 3 ⋮ 0 ]
[1 0 9 ⋮ 0]
𝑹𝟐 : [0 1 3 ⋮ 0]
+𝑹𝟑 : + [0 −1 −1 ⋮ 0]
[ 0 0 2 ⋮ 0]
15 | M o d u l e 2
 Next, observe that we can perform first elementary row operation on the pivot of the third
column to make it a leading one.

1 0 9 ⋮ 0 1 0 9 ⋮ 0
[0 1 3 ⋮ 0] ≈ [0 1 3 ⋮ 0]
0 0 2 ⋮ 0 (𝟏⁄𝟐)𝑹𝟑 0 0 1 ⋮ 0

Solution:

0 0 2 ⋮ 0
𝟏
∗ ⁄𝟐 𝟏⁄ 𝟏⁄ 𝟏⁄𝟐
𝟐 𝟐 ⋮
0 0 1 ⋮ 0

Finally, we make the other entries on the third column above the leading one equal to zero.
How?

1 0 9 ⋮ 0 (−𝟗)𝑹𝟑 + 𝑹𝟏 1 0 0 ⋮ 0
[0 1 3 ⋮ 0] (−𝟑)𝑹𝟑 + 𝑹𝟐 ≈ [0 1 0 ⋮ 0]
0 0 1 ⋮ 0 𝒑𝒊𝒗𝒐𝒕: 𝟏 0 0 1 ⋮ 0

Solution:
(−𝟗)𝑹𝟑 : (−𝟗) [0 0 1 ⋮ 0] → [0 0 −9 ⋮ 0 ]
+𝑹𝟏 : + [1 0 9 ⋮ 0] → + [1 0 9 ⋮ 0]
[1 0 0 ⋮ 0 ]

(−𝟑)𝑹𝟑 : (−𝟑) [0 0 1 ⋮ 0] → [𝟎 𝟎 −𝟑 ⋮ 𝟎]
+𝑹𝟐 : + [0 1 3 ⋮ 0] → + [0 1 3 ⋮ 0]
[0 1 0 ⋮ 0]

3. We now have a matrix in reduced row echelon form. The given system has a unique
solution. Why? It is therefore consistent.

1 0 0 ⋮ 0 𝑥 + 0𝑦 + 0𝑧 = 0 𝑥=0
[0 1 0 ⋮ 0] 0𝑥 + 𝑦 + 0𝑧 = 0 𝑦 =0
0 0 1 ⋮ 0 0𝑥 + 0𝑦 + 𝑧 = 0 𝑧= 0

𝑥 − 2𝑦 + 3𝑧 = 0 0 − 2(0) + 3(0) = 0 + 0 + 0 =0
{ −𝑥 + 3𝑦 =0 −(0) + 3(0) + 0(0) = 0 + 0 + 0 = 0
2𝑥 − 5𝑦 + 5𝑧 = 0 2(0) − 5(0) + 5(0) = 0 + 0+ = 0

16 | M o d u l e 2
 It can be seen that the trivial solution, (0, 0, 0) is the only solution to the given linear
system.

𝑝+ 𝑞+𝑟+𝑠 =0
Example 2: Solve the homogeneous system: {𝑝 +𝑠 =0
𝑝 + 2𝑞 + 𝑟 = 0

Since the trivial solution applies to any homogeneous system, then p = q = r = s = 0 is a

solution to the given system. Let us see if there are other solutions aside from the trivial solution
using the Gauss – Jordan Elimination Method.

1 1 1 1 ⋮ 0
3. We form the augmented matrix [𝑨 ⋮ 𝟎] = [1 0 0 1 ⋮ 0]
1 2 1 0 ⋮ 0
4. We perform a series of elementary row operations on [𝑨 ⋮ 𝒃] until it is in its reduced row
echelon form [𝑪 ⋮ 𝟎] .
1 1 1 1 ⋮ 0 1 1 1 1 ⋮ 0
[𝑨 ⋮ 𝟎] = [1 0 0 1 ⋮ 0] (−𝟏)𝑹𝟏 + 𝑹𝟐 ≈ [0 −1 −1 0 ⋮ 0]
1 2 1 0 ⋮ 0 (−𝟏)𝑹𝟏 + 𝑹𝟑 0 1 0 −1 ⋮ 0

1 1 1 1 ⋮ 0 1 1 1 1 ⋮ 0 (−𝟏)𝑹𝟐 + 𝑹𝟏
[0 −1 −1 0 ⋮ 0 ] 𝑹 𝟐 ↔ 𝑹 𝟑 ≈ [0 1 0 −1 ⋮ 0]
0 1 0 −1 ⋮ 0 0 −1 −1 0 ⋮ 0 𝑹𝟐 + 𝑹𝟑

1 0 1 2 ⋮ 0 1 0 1 2 ⋮ 0 (−𝟏)𝑹𝟑 + 𝑹𝟏
≈ [0 1 0 −1 ⋮ 0] ≈ [0 1 0 −1 ⋮ 0]
0 0 −1 −1 ⋮ 0 (−𝟏)𝑹𝟑 0 0 1 1 ⋮ 0

1 0 0 1 ⋮ 0
≈ [0 1 0 −1 ⋮ 0]
0 0 1 1 ⋮ 0

5. We now have a matrix in reduced row echelon form. Rewriting the augmented matrix in
a system of linear equations, we will be having dependent variables on a particular
1 0 0 1 ⋮ 0 𝑝 + 0𝑞 + 0𝑟 + 𝑠 = 0 𝑝 +𝑠 =0
variable:[0 1 0 −1 ⋮ 0] { 0𝑝 + 𝑞 + 0𝑟 − 𝑠 = 0 {𝑞 −𝑠 =0
0 0 1 1 ⋮ 0 0𝑝 + 0𝑞 + 𝑟 + 𝑠 = 0 𝑟 +𝑠 = 0

17 | M o d u l e 2
 𝑝 +𝑠 =0 𝑝 +1=0 𝑝 = −1
If s = 1, then {𝑞 −𝑠 =0 𝑞−1=0 𝑞= 1
𝑟 +𝑠 = 0 𝑟+1=0 𝑟 = −1

Hence, a solution to the given system is (– 1, 1, –1, 1), which is different from the trivial
solution, (0, 0, 0, 0).

So when does a homogeneous system have a nontrivial solution? The following theorem
will tell us so:

Theorem: A homogeneous system of m equations in n unknowns always has a nontrivial solution

if m < n, that is, if the number of unknowns exceeds the number of equations.

References:

Bernard Kolman/David R. Hill. Introductory Linear Algebra with Applications .7th edition. New Jersey:
Prentice-Hall,Inc.,2002.

Larson, Ron. Linear Algebra. Singapore: Cengage Learning Asia Pte Ltd.,2011.

System of Linear Equations:

• https://www.youtube.com/watch?v=csgNflj69-Y
• https://www.youtube.com/watch?v=csgNflj69-
Y&list=PLybg94GvOJ9En46TNCXL2n6SiqRc_iMB8 &index=1

18 | M o d u l e 2
2.2 The Inverse of a Matrix

This section deals with an algebraic property which is similar to the multiplicative inverse,
or the reciprocal, of a real number. Here, we will find the inverse of a particular matrix. For
inverses, we will only be dealing square matrices.

Definition: Inverse of a Matrix

An n x n matrix A is called nonsingular (or invertible or nondegenerate) if there exists

an n x n matrix B such that 𝑨𝑩 = 𝑩𝑨 = 𝑰𝒏 , where 𝑰𝒏 denotes the n x n identity matrix and the
multiplication used is ordinary matrix multiplication. If this is the case, then the matrix B is
uniquely determined by A and is called the inverse of A, denoted by A−1.

If there exists no such matrix B, then A is not invertible (singular or degenerate).

Non-square matrices (m x n matrices for which m ≠ n) do not have an inverse.

From the definition above, matrix inversion is the process of finding the matrix B that
satisfies the prior equation for a given invertible matrix A.

NOTE: Not all square matrices have inverses. A square matrix which has an inverse is
called invertible or nonsingular or nondegenerate, and a square matrix without an inverse is
called noninvertible or singular or degenerate.

1 3 −2 3⁄2
Example 1: For matrix 𝑨 = [ ], its inverse is 𝑩 = [ ] since
2 4 1 − 1⁄2

3⁄ 3 1
1 3 −2 2 ] = [1(−2) + 3(1) 1( ⁄2) + 3(− ⁄2)] = [−2 + 3 3⁄ + (− 3⁄ )
2 2 ] = [1 0
𝑨𝑩 = [ ][ ] = 𝑰𝒏
2 4 1 − 1⁄2 2(−2) + 4(1) 2(3⁄2) + 4(− 1⁄2) −4 + 4 3−2 0 1

Also,
−2 3⁄ (−2)1 + (3⁄2)(2) (−2)(3) + (3⁄2)(4)
2 ] [1 3 −2 + 3 −6 + (6) 1 0
𝑩𝑨 = [ ] =[ ]= [ ]= [ ] = 𝑰𝒏
1 − 1⁄2 2 4 (1)1 + (− 1⁄2)(2) (1)(3) + (− 1⁄2)(4) 1−1 3−2 0 1

Hence, 𝑨𝑩 = 𝑩𝑨 = 𝑰𝒏 , and , B = A–1.

19 | M o d u l e 2
 Observe too that, BA = AB = In. What does this mean? Yes, this means that A = B–1.

4 3
Example 2. For the same matrix A as in example 1, and the matrix 𝑪 = [ ]. Then,
3 2

1 3 4 3 1(4) + 3(3) 1(3) + 3(2) 4+9 3+6 13 9 1 0

𝑨𝑪 = [ ][ ]= [ ]=[ ]=[ ] ≠[ ]
2 4 3 2 2(4) + 4(3) 2(3) + 4(2) 8 + 12 6 + 8 20 14 0 1

4 3 1 3 (4)1 + (3)2 (4)3 + (3)4 4+6 12 + 12 10 24 1 0

𝑪𝑨 = [ ][ ]= [ ]=[ ]=[ ] ≠[ ]
3 2 2 4 (3)1 + (2)2 (3)3 + (2)4 3+4 9+8 7 17 0 1

Since AC ≠ 𝑪𝑨 ≠ In, then C ≠ A–1. Also, A ≠ C–1

Then, how can we find the inverse of C, if it exists?

From the definition, we can find the inverse of C by looking for a 2x2 matrix, say D, such that:

CD = DC = I2.

𝑎 𝑏 4 3 𝑎 𝑏
If we represent D by 𝑫 = [ ], then CD = [ ][ ]. We want this product to be
𝑐 𝑑 3 2 𝑐 𝑑
equal to the product DC and the 2x2 identity matrix so that D will be the inverse of C. That is,

4 3 𝑎 𝑏 𝑎 𝑏 4 3 1 0
[ ][ ]= [ ][ ]=[ ]
3 2 𝑐 𝑑 𝑐 𝑑 3 2 0 1
Let us first solve for the values of a, b, c, d by using matrix multiplication CD and force
the product to be equal to the identity matrix.

4 3 𝑎 𝑏 1 0
[ ][ ]= [ ]
3 2 𝑐 𝑑 0 1
4𝑎 + 3𝑐 = 1
{ + 3𝑑 = 0
4𝑏
3𝑎 + 2𝑐 = 0
3𝑏 + 2𝑑 = 1
4 3 ⋮ 1 0
Then from Module 2.1, we can rewrite this as an augmented matrix: [ ]
3 2 ⋮ 0 1
and use the Gauss – Jordan Elimination method to solve for a, b, c, d.

4 3 ⋮ 1 0 (1⁄4)𝑅1 3⁄ ⋮ 1⁄ 0 1 3⁄4 ⋮ 1⁄4 0

[ ] ≈ [1 4 4 ] ≈ [ ]
3 2 ⋮ 0 1 3 2 ⋮ 0 1 (−3)𝑅1 + 𝑅2 0 − 1⁄4 ⋮ − 3⁄4 1

3⁄ 1⁄ 0
1 ⋮ 3⁄ ⋮ 1⁄ 0 (− 3⁄4)𝑅2 + 𝑅1 1 0 ⋮ −2 3
[ 4 4 ] ≈ [1 4 4 ] ≈ [ ]
0 − 1⁄4 ⋮ − 3⁄4 1 (−4)𝑅2 0 1 ⋮ 3 −4 0 1 ⋮ 3 −4

20 | M o d u l e 2
Then, we can rewrite the augmented matrix in reduced row echelon form as a system of linear
equations:
𝑎 + 0𝑐 = – 2 𝑎= – 2
−2 3
{ 𝑏 + 0𝑑 = 3 or simply {𝑏 = 3 with matrix D = [ ].
0𝑎 + 𝑐 = 3 𝑐 = 3 3 −4
0𝑏 + 𝑑 = – 4 𝑑 =– 4

Let us check if CD = DC = I2.

4(−2) + 3(3) 4(3) + 3(−4) −8 + 9 12 − 12
CD = [4 3] [−2 3
]= [ ]= [ ]=[
1 0
] = 𝑰𝟐
3 2 3 −4 3(−2) + 2(3) 3(3) + 2(−4) −6 + 6 9−8 0 1

(−2)4 + (3)3 (−2)3 + (3)2

DC = [−2 3 4 3
][ ]= [ ]= [
−8 + 9 −6 + 6
]=[
1 0
] = 𝑰𝟐
3 −4 3 2 (3)4 + (−4)3 (3)3 + (−4)2 12 − 12 9 − 8 0 1

So, CD = DC = I2. Therefore, D = C–1.

We have just found an inverse for the matrix C using the Gauss – Jordan Reduction
Method! Let us try another matrix.

3 −2 4
Example 3. Find A–1 if it exists where 𝑨 = [1 1 −2].
2 −3 6

Solution: We have to find A–1 such that AA–1=A–1A=I3.

𝑎 𝑏 𝑐 3 −2 4 𝑎 𝑏 𝑐 1 0 0
If we let 𝑨 −𝟏
= [ 𝑑 𝑒 𝑓], then AA–1=[1 1 −2] [𝑑 𝑒 𝑓] = [0 1 0] or
𝑔 ℎ 𝑖 2 −3 6 𝑔 ℎ 𝑖 0 0 1

3 −2 4 ⋮ 1 0 0
[1 1 −2 ⋮ 0 1 0]
2 −3 6 ⋮ 0 0 1
Let us use the Gauss – Jordan Reduction Method to identify A–1.

3 −2 4 ⋮ 1 0 0 1 1 −2 ⋮ 0 1 0
[1 1 −2 ⋮ 0 1 0] 𝑅1 ↔ 𝑅2 ≈ [3 −2 4 ⋮ 1 0 0] (−3)𝑅1 + 𝑅2
2 −3 6 ⋮ 0 0 1 2 −3 6 ⋮ 0 0 1 (−2)𝑅1 + 𝑅3

1 1 −2 ⋮ 0 1 0 1 1 −2 ⋮ 0 1 0 (−1)𝑅2 + 𝑅1
[0 −5 10 ⋮ 1 −3 0] (− 1⁄5)𝑅2 1 3
≈ [0 1 −2 ⋮ − ⁄5 ⁄5 0]
0 −5 10 ⋮ 0 −2 1 0 −5 10 ⋮ 0 −2 1 (5)𝑅2 + 𝑅3

21 | M o d u l e 2
 1⁄ 2⁄ 0
1 0 0 ⋮ 5 5
[0 1 −2 ⋮ − ⁄5 ⁄5 0] But observe that the left – hand side of the augmented matrix is
1 3
0 0 0 ⋮
−1 1 1
not row equivalent to the identity matrix I3. This will create linear equations that are inconsistent,
right. Hence, we cannot find values for the inverse of the given matrix A.

The practical procedure for computing the inverse of matrix A is summarized as follows:

1. Form the n x 2n matrix [𝐴 ⋮ 𝐼𝑛 ] obtained by augmenting the identity matrix In to the

given matrix A.
2. Transform the matrix obtained in step 1 to reduced row echelon form by using elementary
row operations. Remember that whatever we do to a row of A, we also do to the
corresponding row of In.
3. Suppose that step 2 has produced the matrix [𝐶 ⋮ 𝐷] in reduced row echelon form.
(a) If C = In, then D = A–1.
(b) If C  In, then C has a row of zeros. In this case A is singular and A–1does not exist.

The examples given above provide illustrations of some properties of the inverse of a given matrix.
These properties will be further discussed in detail through the following theorems:

PROPERTIES OF THE INVERSE OF A MATRIX:

1. If a matrix A has an inverse, then the inverse is unique.

Proof: Given: Matrix A has an inverse.

Need to show: The inverse of matrix A is unique.

Suppose, by contradiction, matrix A has two distinct inverses, say B and C.

Then, AB = BA = In and AC = CA = In by definition of inverse.

By transitivity, since BA = In and AC = In then BA = AC = In

Let us then consider matrix B:

B = BIn by definition of identity matrix

= B(AC) by substituting the value of In by AC (since AC = In)

22 | M o d u l e 2
 = (BA)C by Associative Property for Multiplication

= InC by substituting the value of BA by In (since BA = In)

=C by definition of identity matrix.

But this is a contradiction since we have said earlier that B and C are distinct. Therefore,
if the inverse exists, then it must be unique.

2. If A is a nonsingular matrix, then A–1 is nonsingular and (A–1)–1 = A.

Proof: Given: Matrix A is nonsingular.

Need to show: A–1 is nonsingular and (A–1)–1 = A.

Since A is nonsingular, then A–1 exists and

A–1A = A A–1 = In by definition of nonsingular matrix.

Now, if A–1 is nonsingular, then, there exists a matrix B such that

A–1B = BA–1 = In by definition of nonsingular matrix.

But we know that the inverse of A is A–1. This means

B=A since the inverse is unique.

Hence, A–1 is nonsingular and (A–1)–1 = A.

3. If A and B are nonsingular matrices, then AB is nonsingular and (AB)–1 = B–1A–1.

Proof: Given: A and B are nonsingular.

Need to show: AB is nonsingular and (AB)–1 = B–1A–1.

Since A and B are nonsingular, then A–1 and B–1 exist and

AA–1 = A–1A = In by definition of nonsingular matrix.

BB–1 = B–1B = In by definition of nonsingular matrix.

Let us multiply:

23 | M o d u l e 2
(AB) (B–1A–1) = A(BB–1) A–1 by Associative property for multiplication.

= A(In)A–1 by definition of inverse of B (also, B B–1 = B–1B = In)

= AA–1 by definition of identity matrix

= In by definition of inverse of A (also, AA–1 = A–1A = In)

Similarly, (B–1A–1)(AB) = In. Can you show why?

since (AB) (B–1A–1) = (B–1A–1)(AB) = In,

then (AB)–1 = B–1A–1 by definition of inverse of a matrix.

4. If A is a nonsingular matrix, then (AT)–1 = (A–1) T.

Proof: Given: A is nonsingular.

Need to show: (AT)–1 = (A–1) T.

Since A is nonsingular, then A–1 exists and AA–1 = A–1A = In (by definition of nonsingular
matrix).

Also, since AA–1 = In, then

(AA–1)T = (In)T taking transposes on both sides.

= In taking the transpose of the identity matrix.

Likewise, (A–1A)T = In and (AA–1)T = (A–1A)T = In. Therefore,

(A–1)T(A)T = (A T)(A–1)T = In by property of transpose: (AB)T = BTAT

Since (A T)(A–1)T = (A–1)T (A)T= In, then

(AT)–1 = (A–1) T by definition of inverse.

5. An n x n matrix is nonsingular if and only if it is row equivalent to In.

Linear Systems and Inverses

In this section, we will be using the inverse of a matrix to solve a given system of linear
equations with n equations in n unknowns.

24 | M o d u l e 2
 Let us observe the linear system Ax = b where A is an n x n matrix. Suppose too that A is
nonsingular. Then A–1 exists and

(A–1)Ax = (A–1)b multiplying A–1to both sides of the equation

(A–1A)x = (A–1)b LHS: Associative property for multiplication

(In)x = (A–1)b definition of inverse of A.

x = (A–1)b definition of identity matrix.

Therefore, x can be obtained by multiplying b by (A–1). Thus, if A is nonsingular, then we have a

unique solution to the linear system Ax = b. Let us use this concept to solve linear systems using
inverse.

4 3
Example 1: Let us consider the matrix 𝑪 = [ ] in example 2 above and use it as the coefficient
3 2
4𝑎 + 3𝑏 = 1
matrix of the linear system: { . Aside from the Gauss – Jordan
3𝑎 + 2𝑏 = −1
Reduction Method, let us use the inverse of C to solve for the values of a and b.

Solution:

−2 3
It has been shown in example 2 above that 𝑪−𝟏 = [ ].
3 −4
𝑎 −2 3 1
Then, x = (C–1)b or [ ] = [ ][ ]
𝑏 3 −4 −1
Using matrix multiplication, we have

𝑎 (−2)(1) + (3)(−1) −2 + (−3) −5

[ ]= [ ]= [ ]= [ ]
𝑏 (3)(1) + (−4)(−1) 3+4 7

So, (– 5, 7) is a solution to the linear system.

Example 2: Solve the homogeneous system (example 1 on homogeneous systems)

𝑥 − 2𝑦 + 3𝑧 = 0
{ −𝑥 + 3𝑦 = 0 using inverse, if it exist.
2𝑥 − 5𝑦 + 5𝑧 = 0

Solution: Let us first find the inverse of the coefficient matrix of the given homogeneous system.

1 −2 3 ⋮ 1 0 0 1 −2 3 ⋮ 1 0 0
[−1 3 0 ⋮ 0 1 0] 𝑅1 + 𝑅2 ≈ [0 1 3 ⋮ 1 1 0]
2 −5 5 ⋮ 0 0 1 (−2)𝑅1 + 𝑅3 0 −1 −1 ⋮ −2 0 1

25 | M o d u l e 2
 1 −2 3 ⋮ 1 0 0 (2)𝑅2 + 𝑅1 1 0 9 ⋮ 3 2 0
[0 1 3 ⋮ 1 1 0] ≈ [0 1 3 ⋮ 1 1 0]
0 −1 −1 ⋮ −2 0 1 𝑅2 + 𝑅3 0 0 2 ⋮ −1 1 1 (1⁄2)𝑅3

15⁄ 5 − 9⁄2
1 0 9 ⋮ 3 2 0 (−9)𝑅3 + 𝑅1 1 0 0 ⋮ 2 − ⁄2
[0 1 3 ⋮ 1 1 0 ] (−3)𝑅3 + 𝑅2 ≈ 0 ⋮ 5⁄ 1 − 3⁄2
1 1⁄ 1⁄
0 1 2 − ⁄2
0 0 1 ⋮ − ⁄2 0 0 1 ⋮
2 2 [ − 1⁄2 1⁄2 1⁄
2 ]
Therefore,

15⁄ 5 − 9⁄2 (15⁄2)(0) + (− 5⁄2)(0) + (− 9⁄2)(0)

2 − ⁄2 0 0
𝑥 = 5⁄2 − 1⁄2 − 3⁄2 [0] = (5⁄2)(0) + (− 1⁄2)(0) + (− 3⁄2)(0) = [0]
1 1⁄ 1⁄ 0 −1 1 1 0
[− ⁄2 2 2 ] [ ( ⁄2)(0) + ( ⁄2)(0) + ( ⁄2)(0) ]

the trivial solution.

The following theorems are very useful in this section:

Theorem: If A is an n x n matrix, the homogeneous system Ax = 0 has a nontrivial solution if

and only if A is singular.

Theorem: If A is an n x n matrix, then A is nonsingular if and only if the linear system Ax = b

has a unique solution for every n x 1 matrix b.

Summarized below are the results on homogeneous systems and nonsingular matrices:

The following statements are equivalent:

1. A is nonsingular.
2. Ax = 0 has only the trivial solution.
3. A is row equivalent to In.
4. The linear system Ax = b has a unique solution for every n x 1 matrix b.

References:

Bernard Kolman/David R. Hill. Introductory Linear Algebra with Applications .7th edition.
New Jersey: Prentice-Hall,Inc.,2002.

26 | M o d u l e 2
Larson, Ron. Linear Algebra. Singapore: Cengage Learning Asia Pte Ltd.,2011.

Inverse of a Matrix and its Properties

• https://www.youtube.com/watch?v=kWorj5BBy9k&list=PLybg94GvOJ9En46TNCXL2n6Si
qRc_iMB8&index=9&t=0s

27 | M o d u l e 2