Math1120 - Mathematics II

Linear Algebra (I): Systems of Linear Equations

§1 Introduction

An equation such as
a1 x1  a2 x2   an xn  b

(where the ai and b are numbers and the xi are variables) is a linear equation in n variables,
(or n unknowns). In the special case in which b  0 the equation is called homogeneous. (Of
course when there are only a small number of variables, 2 or 3 say, it is common to use
completely different labels (such as x, y and z) for the variables rather than to use subscripts
(such as x1 , x2 and x3 ).)

A system of linear equations is a set of m simultaneous linear equations in n variables,

i.e.
a11 x1  a12 x2   a1n xn  b1
a21 x1  a22 x2   a2 n xn  b2

am1 x1  am 2 x2   amn xn  bm

A solution to a system of linear equations is a set of values for the variables x1 , x2 , , xn that
satisfy all of the equations in the system. The solution set to the system is the set of all
possible solutions.

Example 1: The following is a homogeneous system of three linear equations in four

unknowns.
3x1  x2  2 x3  x4  0
 x1  6 x2  11x4  0 .
x1  x2  7 x3  4 x4  0

A solution to this system of equations is x1  1 , x2  2 , x3  1 , x4  1 which can be

written as
 x1, x2 , x3 , x4   (1, 2, 1,1) .
The solution set for this system is the set of points
 x1, x2 , x3 , x4   (t, 2t, t, t ) , where t  .
☼

You should already be familiar with solving systems of linear equations involving two
equations in two variables and interpreting the solution in terms of lines in the plane.

Example 2: Solve the following system of linear equations and interpret the answer
geometrically
2 x  3 y  13
x y 5 .

Page 1
 Linear Algebra (I) - Lecture Notes

Multiplying the second equation by 2 and then subtracting this from the first
equation gives y  3 . Substituting this result into either of the original equations gives
x  2 . Thus the solution is ( x, y)  (2, 3) .

As shown in Figure 1, we can interpret the solution ( x, y)  (2, 3) as the point in
the plane at which the straight lines represented by the equations 2 x  3 y  13 and
x  y  5 intersect.

Figure 1
☼

Example 3: Solve the following system of linear equations and interpret the answer
geometrically
2x  3y  4
.
4x  6 y  8

Multiplying the first equation by 2 and then subtracting this from the second
equation gives 0  0 . This tells us that any values of x and y that satisfy the first
equation will also satisfy the second equation. Thus there are infinitely many solutions.
We may choose x (say) to be anything we like, and then find a corresponding value for
y (from either equation). So if we set x  t , then
4 2
y  t,
3 3
and the solution set is
 x, y    t ,  t 
4 2
 3 3 
Thus, the parameter t labels the infinite family of solutions.

The geometric interpretation of this system of linear equations is that both

equations represent the same line (as shown in Figure 2).

Page 2
 Math1120 - Mathematics II

Figure 2

Note that the solution written in terms of the parameter t can be interpreted as a vector
equation for the line. By writing the solution as
 x, y    0,   t 1,  
4 2
 3  3
 4
we can see that the line passes through the point  0,  and is parallel to the vector
 3
 2
1,   .
 3
☼

Example 4: Solve the following system of linear equations and interpret the answer
geometrically
2 x  3 y  13
.
4x  6 y  5

Multiplying the first equation by 2 and then subtracting this from the second equation
gives 0  21 . This tells us that no values of x and y will satisfy these equations. Thus there
are no solutions. A system of equations with no solution is said to be inconsistent.

Geometrically, this system of linear equations can be interpreted as two parallel lines in
the plane, as shown in Figure 3.

Figure 3 ☼

Page 3
 Linear Algebra (I) - Lecture Notes

The above (simple) examples illustrate two important facts that hold for all systems of
linear equations. Firstly:

A system of m linear equations in n unknowns has either

• A unique solution,
• An infinite number of solutions, or
• No solution

Thus it is impossible, for example, for a system of linear equations to have exactly two or
exactly three solutions. (We will see why this is later.) Secondly, we cannot tell which of
these cases (i.e. a unique solution, infinitely many solutions or no solution) applies simply by
counting the number of equations and unknowns.

For systems of two linear equations in two unknowns we have seen that we can solve
the system via the “substitution method” (viz. making one variable the subject of one
equation and then substituting this into the second equation) or the “elimination method” (viz.
adding a multiple of one equation to other). For larger system of equations this basic idea can
still be used but the working can be become quite messy. In the following sections we will
look at a systematic way of solving a system of m linear equations in n unknowns using
matrices.

§2 Matrices and Systems of Linear Equations

An m  n matrix is a rectangular array of numbers with m rows and n columns, i.e.

 a11 a12 a1n 

 
 a21 a22   aij 
a2 n 
  mn
 
 am1 am 2 amn 

We say that the number aij is the  i, j  entry (i.e. the entry in row i, column j).
th

Example 5: The following are examples of matrices.

 1  2
1 2 3   0 4  
     1 
 4 5 10   1 3   3 
   
 2 
2  3 matrix 2 2 matrix 3 1 matrix
☼

Matrices have proved useful in many areas of mathematics and we will study them in
more detail throughout this strand. With respect to the following system of linear equations

Page 4
 Math1120 - Mathematics II

a11 x1  a12 x2   a1n xn  b1

a21 x1  a22 x2   a2 n xn  b2

am1 x1  am 2 x2   amn xn  bm

so long as the variables are always written in the same order and the constants are always put
on to the right hand side we can represent this system via the m  (n  1) matrix

 a11 a12 a1n b1 

 
 a21 a22 a2 n b2 
 
 
 am1 am 2 amn bn 

This matrix is called the augmented matrix for the system of linear equations. Each row of
this matrix corresponds to an equation in the system. Each of the first n columns corresponds
to the coefficients of a variable in the equations and the last column gives the constants on the
right hand sides of the equations.

Example 6: Write down the augmented matrix for the following non-homogenous
system of linear equations
3x1  x2  2 x3  x4  2
 x1  6 x2  11x4  22
x1  x2  7 x3  4 x4  8

The augmented matrix will be the 3  5 matrix

 3 1 2 1 2 
 
 1 6 0 11 22 
1 1 7 4 8 

☼

Note that sometimes a partition line is put before the last column in an augmented
matrix to emphasise that the last column represents the right hand side of the equations while
the remaining columns represent the coefficients of the variables in the system. The matrix
without the last column is called the coefficient matrix for the system of equations.

Example 6: (cont.) For this system of linear equations the augmented matrix is
sometimes written as
 3 1 2 1 2 
 
 1 6 0 11 22 
1 1 7 4 8  .


The coefficient matrix for this system is

Page 5
 Linear Algebra (I) - Lecture Notes

 3 1 2 1 
 
 1 6 0 11
1 1 7 4 
 
☼

Example 7: Write down the system of linear equations that correspond to the following
augmented matrices.

(a) (b)
 3 2 2 3  1 0 0 1 
   
 1 6 4 7  0 1 0 2 
 1 1 5 2   0 0 1 1
   

(a) Calling the variables as x, y and z the system of equations is

3x  2 y  2 z  3
x  6 y  4z  7 .
x  y  5 z  2

(b) Again calling the variables x, y and z this augmented matrix represents the very
simple system of equations
x 1
y2
z  1
☼

§3 Solving Systems of Linear Equations

Two systems of linear equations are said to be equivalent if they have the same number
of variables and have the same solution sets.

Example 7: (cont.) It turns out that the systems of linear equations in Example 7 are
equivalent, i.e.

3x  2 y  2 z  3 x 1
x  6 y  4z  7 is equivalent to y2
x  y  5 z  2 z  1

The solution to these equivalent systems can be seen from the system on the right
whereas they are not at all obvious from the system on the left.
☼

Notice that, given a system of linear equations, we can produce an equivalent system
by doing any one of the following:
• Multiplying any one equation by a non-zero constant
• Re-writing the equations in a different order
• Adding any multiple of one equation to another

Page 6
 Math1120 - Mathematics II

Example 8: The following systems of linear equations are equivalent because we have
just written the equations in a different order.

3x  2 y  2 z  3 3x  2 y  2 z  3
x  6 y  4z  7 is equivalent to x  y  5 z  2
x  y  5 z  2 x  6 y  4z  7

The following systems of linear equations are equivalent because we have just
multiplied the second equation by 3.

3x  2 y  2 z  3 3x  2 y  2 z  3
x  y  5 z  2 is equivalent to 3x  3 y  15 z  6
x  6 y  4z  7 x  6 y  4z  7

The following systems of linear equations are equivalent because we have just
subtracted the first equation from the second equation, (i.e. added -1 times the first
equation to the second equation.)

3x  2 y  2 z  3 3x  2 y  2 z  3
3x  3 y  15 z  6 is equivalent to 5 y  13z  3
x  6 y  4z  7 x  6 y  4z  7
☼

Thus the strategy for solving a system of linear equations will be to apply the above
operations in a systematic way to produce an equivalent system from which we can determine
the solution (such as that given in Example 7b). This strategy, in the form described below, is
called Gauss-Jordon Elimination. In order to describe this strategy we need some
terminology.

Given a matrix an equivalent matrix can be produced by applying any one of the
following three elementary row operations:
• Swapping two rows
• Multiplying a row by a non-zero constant
• Adding a multiple of one row to another row

Note: i. We use the symbol to denote that two matrices are equivalent.
ii. The process of applying elementary row operations to a matrix is called row
reduction

Example 8: (cont.) The following matrices are equivalent because we have just swapped
two rows.
 3 2 2 3   3 2 2 3 
    
 1 6 4 7   1 1 5 2  R2  R3
 1 1 5 2   1 6 4 7  R  R
    3 2

The following matrices are equivalent because we have just multiplied Row 2 by 3.

Page 7
 Linear Algebra (I) - Lecture Notes

 3 2 2 3   3 2 2 3 
    
 1 1 5 2   3 3 15 6  R2  3R2
 1 6 4 7   1 6 4 7 
   

The following matrices are equivalent because we have just subtracted Row 1 from
Row 2, (i.e. added -1 times Row 1 to Row 2)

 3 2 2 3   3 2 2 3 
    
 3 3 15 6   0 5 13 3  R2  R2  R1
 1 6 4 7   1 6 4 7 
   
☼

A matrix is said to be in row-echelon form if the first non-zero entry (i.e. the leading
entry) in each row occurs further to the right than in the row above it.

A matrix is said to be in reduced row-echelon form if it is in row-echelon form and each

leading entry is 1 with no non-zero entries above it.

Example 9: For the following matrices decide if it is row-echelon form, reduced row-
echelon form or neither. For those matrices in reduced row-echelon form what can you
say about the system of linear equations represented by those matrices?
(a) (b)
1 0 3 2   3 1 2 1 2 
   
 0 1 1 3   0 7 7 7 16 
0 0 0 0  1 1 7 4 8 
  
(c) (d)
 1 4 1 2 
  1 0 0 1 
 0 2 3 2   
0 0 4 5  0 1 0 2 
 0 0 1 1
   
 0 0 0 6 

(a) This matrix is in reduced row-echelon form. The system of equations represented
by this matrix is
x  3z  2
y  z  3 .
00
The third equation in this system is satisfied no matter what values we choose for
x, y and z. Once we choose a value for z (say) then from the first two equations
we can determine the values for x and y as
x  2  3z
.
y  3  z
Thus this system of equations has an infinite number of solutions given by
 x, y, z   (2  3t, 3  t, t )

Page 8
 Math1120 - Mathematics II

or
( x, y, z)  (2, 3,0)  t (3,1,1)

(b) This matrix isn’t in row-echelon form (and hence not in reduced row-echelon
form either) since the leading entry in row 3 occurs to the left of the leading entry
is row 2.

(c) This matrix is in row-echelon form but not reduced row-echelon form. Notice that
in this case we can tell that the associated system of linear equations is
inconsistent even though the matrix is not in reduced row-echelon form. The
fourth equation in this system is
06
and this cannot be satisfied no matter which values of x, y and z we choose.

(d) This matrix is in reduced row-echelon form and the solution of the associated
systems of linear equations is
( x, y, z )  (1, 2, 1)
☼

In Gauss-Jordon Elimination, the augmented matrix associated with the system of

linear equations is systematically row reduced to reduced row-echelon form by:

• Obtaining a 1 in entry a11

• Obtaining 0’s in the remainder of column 1 by using the elementary row
operation of adding multiples of row 1 to another row.
• Obtaining a 1 in entry a22
• Obtaining 0’s in the remainder of column 2 by using the elementary row
operation of adding multiples of row 2 to another row.
• And so on across the columns.

The solution is then read off from the equivalent system represented by the matrix in
reduced row-echelon form

Example 10: Use Gauss Jordon Elimination to solve the following system of linear
equations:
2x  3y  1
3x  4 y  10

The augmented matrix for this system is

2 3 1 
 .
 3 4 10 
We now want to row reduce this matrix to reduced row-echelon form. Following the
steps as outlined:
2 3 1  1 3 2 1 2  R  R1 2
   
1

 3 4 10   0 17 2 17 2  R  R  3R  2 2 1

 1 0 2  R  R  3R2 2
 
1 1

 0 1 1 R   2 R2 17
2

Page 9
 Linear Algebra (I) - Lecture Notes

We can now read off the solution as

( x, y)  (2, 1) .

Of course we can check that this solution is correct by substituting these values into the
original equations.

In doing the row reduction here were had to use fractions. It is possible to apply the
elementary row operations in a manner that avoids the fractions until the last steps.
However you have to be very careful in doing this that you don’t lose some of the 0’s
that you have already created and hence start going around in circles!
☼

Example 11: Use Gauss Jordon Elimination to solve the following system of linear
equations.
2x  6 y  2z  4
3x  2 y  z  0 .
2x  3y  1

The augmented matrix for this system is

 2 6 2 4
 
3 2 1 0 .
2 3 0 1
 

Now using the steps as outlined to reduce this to reduced row-echelon form:

 2 6 2 4 1 3 1 2  R  R 2
   
1 1

3 2 1 0 0 7 2 6  R   R  3R 
2 2 1

2 3 0 1 0 3 2 3  R   R  2 R 
   3 3 1

1 0 1 7  4 7  R  R  3R
 
1 1 2

0 1 27 6 7  R   R 7 2 2

0 0  8 7  3 7  R   R  3R 
 3 3 2

1 0 0 5 8 R1  R1  R3 7
 
R2  R2  2 R3 7
0 1 0 3 4 
0 0 1 3 8  R3  7 R3 8
 

Thus the solution is

 x, y, z     5 8,3 4,3 8
☼

Page 10
 Math1120 - Mathematics II

Example Tasks

ET 1: For the following augmented matrices write down the general solution of the
associated system of linear equations, (i.e. write down the set of all solutions).
(a) (b)
1 0 0 2
0 2 2 2  
   0 1 0 3
0 0 3 3 0 0 0 0
 
(c) (d)
 1 6 0 0 3 2 
 1 1 2 3   
  0 0 1 0 4 7 
 0 0 0 0 
0 0 0 0 0 0 0 1 5 8 
   
0 0 0 0 0 0 

ET 2: Solve the following system of linear equations by Gauss-Jordan elimination.

5 x  2 y  z  11
x  y  z 1
4 x  2 y  3z  5

ET 3: The general equation of a circle is x2  y 2  ax  by  c  0 . Find the equation of

the circle that passes through the points 1, 0  ,  2,1 and  2, 1 .

Page 11
 Linear Algebra (I) - Lecture Notes

Page 12