Solution by

Gauss Elimination

8.3

Introduction
Engineers often need to solve large systems of linear equations; for example in determining the forces
in a large framework or nding currents in a complicated electrical circuit. The method of Gauss
elimination provides a systematic approach to their solution.

Prerequisites

be familiar with matrix algebra

Before starting this Section you should . . .


'

identify the row operations which allow the

reduction of a system of linear equations to
upper triangular form

Learning Outcomes
On completion you should be able to . . .

&

22


$

use back-substitution to solve a system of

equations in echelon form
understand and use the method of Gauss
elimination to solve a system of three
simultaneous linear equations
HELM (2005):
Workbook 8: Matrix Solution of Equations

1. Solving three equations in three unknowns

The easiest set of three simultaneous linear equations to solve is of the following type:
3x1 = 6,
2x2 = 5,
4x3 = 7

T
which obviously has solution [x1 , x2 , x3 ]T = 2, 52 , 74 or x1 = 2, x2 = 52 , x3 = 74 .
In matrix form AX = B the equations are

x1
3 0 0
6
0 2 0 x2 = 5
0 0 4
7
x3
where the matrix of coecients, A, is clearly diagonal.
Task

Solve the equations

x1
2
0 0
8
0 1 0 x2 = 2 .
0
0 3
6
x3
Your solution

Answer
[x1 , x2 , x3 ]T = [4, 2, 2]T .
The next easiest system of equations to solve is of the following kind:
3x1 + x2 x3 = 0
2x2 + x3 = 12
3x3 = 6.
The last equation can be solved immediately to give x3 = 2.
Substituting this value of x3 into the second equation gives
2x2 + 2 = 12

from which

2x2 = 10

so that

x2 = 5

Substituting these values of x2 and x3 into the rst equation gives

3x1 + 5 2 = 0

from which

3x1 = 3

so that

x1 = 1

Hence the solution is [x1 , x2 , x3 ]T = [1, 5, 2]T .

This process of solution is called back-substitution.
In matrix form the system of equations is

x1
3 1 1
0
0 2
1 x2 = 12 .
0 0
3
6
x3
HELM (2005):
Section 8.3: Solution by Gauss Elimination

23

The matrix of coecients is said to be upper triangular because all elements below the leading
diagonal are zero. Any system of equations in which the coecient matrix is triangular (whether
upper or lower) will be particularly easy to solve.

Task

Solve the following system of equations by back-substitution.

x1
7
2 1
3

0
3 1
x2 = 5 .
2
0
0
2
x3
Write the equations in expanded form:
Your solution

Answer
2x1 x2 + 3x3 = 7
3x2 x3 = 5
2x3 = 2
Now nd the solution for x3 :
Your solution
x3 =
Answer
The last equation can be solved immediately to give x3 = 1.
Using this value for x3 , obtain x2 and x1 :
Your solution
x2 =

x1 =

Answer
x2 = 2, x1 = 3. Therefore the solution is x1 = 3, x2 = 2 and x3 = 1.
Although we have worked so far with integers this will not always be the case and fractions will enter
the solution process. We must then take care and it is always wise to check that the equations
balance using the calculated solution.

24

HELM (2005):
Workbook 8: Matrix Solution of Equations

2. The general system of three simultaneous linear equations

In the previous subsection we met systems of equations which could be solved by back-substitution
alone. In this Section we meet systems which are not so amenable and where preliminary work must
be done before back-substitution can be used.
Consider the system
x1 + 3x2 + 5x3 = 14
2x1 x2 3x3 = 3
4x1 + 5x2 x3 = 7
We will use the solution method known as Gauss elimination, which has three stages. In the rst
stage the equations are written in matrix form. In the second stage the matrix equations are replaced
by a system of equations having the same solution but which are in triangular form. In the nal
stage the new system is solved by back-substitution.
Stage 1: Matrix Formulation
The rst step is to write the equations in matrix form:

14
x1
1
3
5
2 1 3 x2 = 3 .
7
x3
4
5 1
Then, for conciseness, we combine the matrix of coecients with the column vector of right-hand
sides to produce the augmented matrix:

1
3
5 14
2 1 3 3
4
5 1 7
If the general system of equations is written AX = B then the augmented matrix is written [A|B].
Hence the rst equation
x1 + 3x2 + 5x3 = 14
is replaced by the rst row of the augmented matrix,
1

14

and so on.

Stage 1 has now been completed. We will next triangularise the matrix of coecients by means of
row operations. There are three possible row operations:
interchange two rows;
multiply or divide a row by a non-zero constant factor;
add to, or subtract from, one row a multiple of another row.
Note that interchanging two rows of the augmented matrix is equivalent to interchanging the two
corresponding equations. The shorthand notation we use is introduced by example. To interchange
row 1 and row 3 we write R1 R3. To divide row 2 by 5 we write R2 5. To add three times row
1 to row 2, we write R2 + 3R1. In the Task which follows you will see where these annotations are
placed.
Note that these operations neither create nor destroy solutions so that at every step the system of
equations has the same solution as the original system.
HELM (2005):
Section 8.3: Solution by Gauss Elimination

25

Stage 2: Triangularisation
The second stage proceeds by rst eliminating x1 from the second and third equations using row
operations.

1
3
5 14
1
3
5
14
2 1 3 3 R2 2 R1 0 7 13 25
4
5 1 7
0 7 21 49
R3 4 R1
In the above we have subtracted twice row (equation) 1 from row (equation) 2.
In full these operations would be written, respectively, as
(2x1 x2 3x3 ) 2(x1 + 3x2 + 5x3 ) = 3 2 14 or

7x2 13x3 = 25

(4x1 + 5x2 x3 ) 4(x1 + 3x2 + 5x3 ) = 7 4 14 or

7x2 21x3 = 49.

and

Now since all the elements in rows 2 and 3 are negative

1
3
5
1
14
0 7 13 25 R2 (1) 0
0 7 21 49
0
R3 (1)

we multiply throughout by 1:

3 5 14
7 13 25
7 21 49

Finally, we eliminate x3 from the third equation by subtracting equation 2 from equation 3
i.e. R3 R2:

1 3 5 14
1 3 5 14
0 7 13 25
0 7 13 25
0 7 21 49
0 0 8 24
R3 R2
The system is now in triangular form.
Stage 3: Back Substitution
Here we solve the equations from bottom to top. At each step of the back substitution process we
encounter equations which only have a single unknown and so can be easily solved.

Task

Now complete the solution to the above system by back-substitution.

Your solution

26

HELM (2005):
Workbook 8: Matrix Solution of Equations

Answer
In full the equations are
x1 + 3x2 + 5x3 = 14
7x2 + 13x3 = 25
8x3 = 24
From the last equation we see that x3 = 3.
Substituting this value into the second equation gives
7x2 + 39 = 25 or 7x2 = 14 so that x2 = 2.
Finally, using these values for x2 and x3 in equation 1 gives x1 6 + 15 = 14. Hence x1 = 5. The
solution is therefore [x1 , x2 , x3 ]T = [5, 2, 3]T
Check that these values satisfy the original system of equations.
Task

Solve
2x1 3x2 + 4x3 = 2
4x1 + x2 + 2x3 = 2
x1 x2 + 3x3 = 3
Write down the augmented matrix for this system and then interchange rows 1 and 3:
Your solution

Answer
Augmented
matrix

2 3 4 2
R1 R3
1 1 3 3
4
1 2 2
1 2 2
4
1 1 3 3
2 3 4 2
Now subtract suitable multiples of row 1 from row 2 and from row 3 to eliminate the x1 coecient
from rows 2 and 3:
Your solution

Answer

1 1 3 3
4
1 2 2
2 3 4 2

R2 4R1
R3 2R1

HELM (2005):
Section 8.3: Solution by Gauss Elimination

1 1
3
3
5 10 10
0
0 1 2 4
27

Now divide row 2 by 5 and add a suitable multiple of the result to row 3:
Your solution

Answer

1 1
3
1 1
3
1 1
3
3
3
3
0
5 10 10 R2 5 0
1 2 2
1 2 2
0
0 1 2 4
0 1 2 4
0
0 4 6
R3 + R2
Now complete the solution using back-substitution:
Your solution

Answer
The equations in full are
x1 x2 + 3x3 = 3
x2 2x3 = 2
4x3 = 6.
The last equation reduces to x3 = 32 .
Using this value in the second equation gives x2 3 = 2 so that x2 = 1.
Finally, x1 1 +

= 3 so that x1 = 12 .


T
The solution is therefore [x1 , x2 , x3 ]T = 12 , 1, 32 .
9
2

You should check these values in the original equations to ensure that the equations balance.
Again we emphasise that we chose a particular set of procedures in Stage 1. This was chosen mainly
to keep the arithmetic simple by delaying the introduction of fractions. Sometimes we are courageous
and take fewer, harder steps.
An important point to note is that when in Stage 2 we wrote R2 2R1 against row 2; what we
meant is that row 2 is replaced by the combination (row 2) 2(row 1).
In general, the operation
row i row j
means replace row i by the combination
row i row j.

28

HELM (2005):
Workbook 8: Matrix Solution of Equations

3. Equations which have an infinite number of solutions

Consider the following system of equations
x1 + x2 3x3 = 3
2x1 3x2 + 4x3 = 4
x1 x2 + x3 = 1
In augmented form we have:

1
1 3
3
2 3
4 4
1 1
1 1
Now performing the usual Gauss elimination operations we have

1
1 3
1
1 3
3
3
2 3
4 4 R2 2 R1 0 5 10 10
1 1
1 1
0 2
4 4
R3 R1
Now applying

1 1
0 1
0 1

R2 5 and R3 2 gives

3 3
2 2
2 2

Then R2 R3 gives

1 1 3 3
0 1 2 2
0 0
0 0
We see that all the elements in the last row are zero. This means that the variable x3 can take any
value whatsoever, so let x3 = t then using back substitution the second row now implies
x2 = 2 + 2x3 = 2 + 2t
and then the rst row implies
x1 = 3 x2 + 3x3 = 3 (2 + 2t) + 3(t) = 1 + t
In this example the system of equations has an innite number of solutions:
x1 = 1 + t,

x2 = 2 + 2t,

x3 = t or [x1 , x2 , x3 ]T = [1 + t, 2 + 2t, t]T

where t can be assigned any value. For every value of t these expressions for x1 , x2 and x3 will
simultaneously satisfy each of the three given equations.
Systems of linear equations arise in the modelling of electrical circuits or networks. By breaking
down a complicated system into simple loops, Kirchhos law can be applied. This leads to a set of
linear equations in the unknown quantities (usually currents) which can easily be solved by one of
the methods described in this Workbook.

HELM (2005):
Section 8.3: Solution by Gauss Elimination

29

Engineering Example 3
Currents in three loops
In the circuit shown nd the currents (i1 , i2 , i3 ) in the loops.
4v
2

3
i1

i2
6

5v

i3
6v

Figure 2
Solution
Loop 1 gives
2(i1 ) + 3(i1 i2 ) = 5

5i1 3i2 = 5

Loop 2 gives
6(i2 i3 ) + 3(i2 i1 ) = 4

3i1 + 9i2 6i3 = 4

6i2 + 10i3 = 1

Loop 3 gives
6(i3 i2 ) + 4(i3 ) = 6 5

Note that in loop 3, the current generated by the 6v cell is positive and for the 5v cell negative in
the direction of the arrow.
In matrix form

5 3
0 i1
5

9 6 i2 = 4
0 6 10 i3
1
Solving gives
i1 =

30

34
,
15

i2 =

19
,
9

i3 =

41
30

HELM (2005):
Workbook 8: Matrix Solution of Equations

Engineering Example 4
Velocity of a rocket
The upward velocity of a rocket, measured at 3 dierent times, is shown in the following table
Time, t
(seconds)
5
8
12

Velocity, v
(metres/second)
106.8
177.2
279.2

The velocity over the time interval 5 t 12 is approximated by a quadratic expression as

v(t) = a1 t2 + a2 t + a3
Find the values of a1 , a2 and a3 .
Solution
Substituting the values from the table into the quadratic

25 5
106.8 = 25a1 + 5a2 + a3

64 8
177.2 = 64a1 + 8a2 + a3
or
144 12
279.2 = 144a1 + 12a2 + a3

equation for v(t) gives:

1 a1
106.8
1 a2 = 177.2
1 a3
279.2

Applying one of the methods from this Workbook gives the solution as
a1 = 0.2905

a2 = 19.6905

a3 = 1.0857 to 4 d.p.

As the original values were all experimental observations then the values of the unknowns are all
approximations. The relation v(t) = 0.2905t2 + 19.6905t + 1.0857 can now be used to predict the
approximate position of the rocket for any time within the interval 5 t 12.

HELM (2005):
Section 8.3: Solution by Gauss Elimination

31

Exercises
Solve the following using Gauss elimination:
1.

2.

2x1 + x2 x3 = 0
x1
+ x3 = 4
x1 + x2 + x3 = 0
x1 x2 + x3 = 1
x1
+ x3 = 1
x1 + x2 x3 = 0

3.
x1 + x2 + x3 = 2
2x1 + 3x2 + 4x3 = 3
x1 2x2 x3 = 1
4.

x1 2x2 3x3 = 1
3x1 + x2 + x3 =
4
11x1 x2 3x3 = 10

You may need to think carefully about this system.

Answers
8
4
(1) x1 = , x2 = 4, x3 =
3
3
1
3
(2) x1 = , x2 = 1, x3 =
2
2
(3) x1 = 2, x2 = 1, x3 = 1
(4) innite number of solutions: x1 = t, x2 = 11 10t, x3 = 7t 7

32

HELM (2005):
Workbook 8: Matrix Solution of Equations