2

Algebra of
systems of
linear
equations

© Copyright Universidad Europea. Todos los derechos reservados 1

 Introduction
Systems of linear equations appear as an essential element of
Linear Algebra and how to solve them has been one of the
fundamental problems in its history.

Swiss mathematician Gabriel Cramer

(1704-1752), was a very important
figure in this branch of mathematics,
publishing the book Introduction à
l'analyse des courbes algébriques.

© Copyright Universidad Europea. Todos los derechos reservados

 Systems of equations
A system of m equations with n unknowns is a set of m equalities
(equations) of the form:

ü The terms aij are called the coefficients of the system

ü The terms xj are the unknowns of the system

ü The terms bi are the independent terms of the system

© Copyright Universidad Europea. Todos los derechos reservados

 Systems of equations
Such systems can be expressed as a product of matrices:

Where

- A is the matrix of the

coefficients
- X is the matrix of
unknowns
- B is the matrix of the
A X = B independent terms

It can be
expressed as:

© Copyright Universidad Europea. Todos los derechos reservados

 Types of systems of linear equations
The first step when solving a system of linear equations is to know
whether it has a solution or not, and if so, how to calculate it.

Depending on the solutions that a system can have, we can classify

them into:

Determined or
independent:
One solution
Consistent:
It has a solution
Systems of Indetermined or
Linear dependent:
equations Infinite solutions
Inconsistent:
It has no solution

How do you know if a system has a solution? studying the RANK

of the system associated matrix
© Copyright Universidad Europea. Todos los derechos reservados
 Geometrical interpretation of the solutions
Let’s see what’s the geometrical view of a system of linear equations.
For the sake of simplicity, we’ll start with two dimensions.

An equation of the type ax+by=c represents a straight line in a 2-

dimensions space, so if we represent a system of 2 equations with 2
unknowns:
3x + 2y =11
x - 2y = 1

The solution is the set of

points of the 2-
dimensions space that
fulfils both equation. In
this case, this solution is
a single point:
x =3
y=1

© Copyright Universidad Europea. Todos los derechos reservados

 Geometrical interpretation of the solutions

Depending on the values of a, b and c, the solution can be a single

point, infinite number of points (a line) or no point at all.

Single solution. No solution. Infinite solutions.

Straight lines that Parallel lines Coincident lines
intersect at a point

Consistent Inconsistent Consistent

determined indetermined

© Copyright Universidad Europea. Todos los derechos reservados

 Resolution of systems of equations

Techniques for solving systems of equations:

ü Cramer’s Rule

ü Gauss Elimination Method

ü Inverse Matrix

Techniques for identifying solutions of systems of

equations:
ü Rouché-Frobenius Theorem

© Copyright Universidad Europea. Todos los derechos reservados

 Cramer’s Rule: preliminaries

To solve the system of equations, which we have expressed in the

form A X = B, we need to clear the matrix X.

! = #!" %

The inverse matrix of A must exist:

&
#!" = · (#)* #)#
#

In other words, the determinant of A cannot be 0 for the

system to be determined

© Copyright Universidad Europea. Todos los derechos reservados

 Cramer’s Rule
Let’s illustrate the technique with a system of 3 equations with 3
unknowns. The system has to be determined, i.e ! ≠ 0

The key idea is to define matrix Bi , that is the coefficient matrix

with the i-th column substituted by B (vector of independent
terms). For example:

If we take determinant of the three matrices to find x1, the result

is:

%&' ()
If applies to all unknowns, so, in general: xi = %&' *

© Copyright Universidad Europea. Todos los derechos reservados

 Cramer’s Rule

For a nxn system of equations it holds:

#"$"! + ##$#! + … + #$ $$!

!! =
$

Therefore, the solutions will be:

© Copyright Universidad Europea. Todos los derechos reservados

 Cramer’s Rule

Solution: First, we build the

coefficient matrix to see if it is square
and its determinant is non-zero

We calculate the
solutions using
Cramer's formulas

The system solution is x= -1 , y= 2.

To check that system is solved correctly, substitute the obtained

values of x and y in the equations and verify that both are true.

© Copyright Universidad Europea. Todos los derechos reservados

 Cramer’s Rule

Solution:

© Copyright Universidad Europea. Todos los derechos reservados

 Cramer’s Rule

ü Cramer's method can be applied as long as we have square

coefficients matrices with a non-zero determinant.

ü The solutions for each variable will be the quotient of two

determinants: in the denominator the determinant of the
coefficient matrix, and in the numerator the determinant of
the coefficient matrix in which we have replaced the column
corresponding to the variable being solved with the column of
independent terms.

ü The system has one single solution (system consistent

determinate).

© Copyright Universidad Europea. Todos los derechos reservados

 Linear dependency

We say that a row (or column) v is a linear combination of two other

rows u1 and u2 if there are two non-zero numbers λ, µ such that v= λu1+
µu2.

We say that three rows (or columns) are linearly independent if none of
them can be written as a linear combination of the others.

We say that three rows (or columns) are linearly dependent, if one of
them can be written as a linear combination of the other two.

© Copyright Universidad Europea. Todos los derechos reservados

 Linear dependency

EXAMPLE

The rows u=(1 2 3), v=(0 3 4) and w=(1 -1 -1) are linearly dependent since
(1 2 3)-(0 3 4)=(1 -1 -1)

The rows u=(1 2 3), v= (3 5 7) and w= (4 6 5) are linearly independent

since there are no pair of numbers such that
(1 2 3) = δ (3 5 7) + β (4 6 5)

© Copyright Universidad Europea. Todos los derechos reservados

 Rank of a matrix
The rank of a matrix A, denoted by rankA or r(A), is the number of
linearly independent rows (or columns) of that matrix.

ü The number of linearly independent rows matches the

number of linearly independent columns.

ü The maximum value that the rank of a matrix can have is the
smallest number of rows or columns. For example, in a 2x6
matrix, the maximum value that the rank of the matrix can
reach is 2

© Copyright Universidad Europea. Todos los derechos reservados

 How to calculate the rank of a matrix

ü Rank by determinant method: The rank is calculated from

the determinant of the largest square submatrix.

ü Rank by the Gaussian method: The rank is calculated by

removing all linearly dependent rows or columns (staggered
or echelon matrix). Linear transformations applied to the
system of equations do not change the range of the original
matrix.

© Copyright Universidad Europea. Todos los derechos reservados

 Example of the rank of a matrix

1 2 0 1
!= 2 4 1 3 Matrix 3x4. Max rank is 3.
3 6 2 5

1 2 0 1 1 2 0 1 1 2 0 1
!= 2 4 1 3 0 0 1 1 0 0 1 1
3 6 2 5 0 0 2 2 0 0 0 0

R2=R2-2R1 R3= R3-2R2

R3=R3-3R1

r(A) = 2

© Copyright Universidad Europea. Todos los derechos reservados

 Augmented matrix

An augmented matrix is the matrix formed by the coefficients and

independent terms of the system, separated by a vertical line.

%## %#$ %#% (#

%$# %$$ ⋯
%$% ($
!; # = !∗ =
⋮ ⋱ ⋮
%&# %&$ ⋯ %%& (&

To find out if a system has a solution, we must analyse the

range of both the coefficient matrix and the augmented
matrix.

© Copyright Universidad Europea. Todos los derechos reservados

 Rouché-Frobenius theorem
The Rouché-Frobenius Theorem allow us to analyze if the
system of equations has solutions and how many. This study
is called discussion of the system and it is based in the rank
of the matrices that are part of the system.

r(A) is the rank of the coefficient matrix and r(A*) of the

extended matrix. According to the theorem:

Consistent system Inconsistent system

r(A)=r(A*) rA<r(A*)

Determined Indetermined
r(A)=r(A*)=nº unknowns r(A)=r(A*)<nº unknowns

© Copyright Universidad Europea. Todos los derechos reservados

 Discussion of the system by the rank of A and A*

Coefficient
Matriz matrix A
de coeficientes Augmented matrix A*
Matriz ampliada
S.Consistent
Compatible
determinado
determined
rgA=rgA*=n.
rA=rA*=nº unknowns
incógnitas

S.Consistent
Compatible
indetermined
indeterminado
rA=rA*<nº unknowns
rgA=rgA*<n.
incógnitas

S. Incompatible
Inconsistent system
rgA≠rgA*
rA<rA*

© Copyright Universidad Europea. Todos los derechos reservados

 Geometrical interpretation in 3 dimensions

If the system is
consistent determined,
r(A)=r(A*)=3, it has a
single solution, i.e. the
three planes have a
unique intersection in
point P (red dot)

© Copyright Universidad Europea. Todos los derechos reservados

 Geometrical interpretation

If the system is
consistent
undetermined, and
r(A)=r(A*)=2 < n (being
unknowns 3), the three
planes cut in a line.

© Copyright Universidad Europea. Todos los derechos reservados

 Geometrical interpretation

If the system is
consistent
undetermined, and
r(A)=r(A*)=1 , the three
planes overlap (are the
same).

© Copyright Universidad Europea. Todos los derechos reservados

 Geometrical interpretation

If the system is
inconsistent, and the
three subsystems of two
equations are
inconsistent, r(A)=1,
r(A*)=2, the three planes
are parallel

© Copyright Universidad Europea. Todos los derechos reservados

 Geometrical interpretation

If the system is
inconsistent, and no
subsystem of two
equations is inconsistent,
r(A)=(1), r(A*)=3, two
planes cut in a line and
third is parallel to that
line

© Copyright Universidad Europea. Todos los derechos reservados

 Geometrical interpretation

If the system is
inconsistent, and just
one subsystem of two
equations is inconsistent,
r(A)=2,r(A*)=3, two planes
cut and third is parallel to
one of them

© Copyright Universidad Europea. Todos los derechos reservados

 Equivalent systems
Two systems of equations are equivalent if they have the same set of
solutions.

To transform a system into an equivalent, we can perform the

following operations:

ü Swap the position of two equations in the system

ü Multiply any of the equations in the system by a non-zero scalar

ü Substituting an equation in the system for the result of adding

another equation multiplied by a non-zero number

ü Add or delete an equation from the system that is a linear

combination of the existing equations.

ü Isolate an unknown in one of the equations and substitute in the

other equations
© Copyright Universidad Europea. Todos los derechos reservados
 Gaussian elimination method

ü One of the most important methods of solving systems of

linear equations is the Gaussian elimination method.

ü This method consists of transforming the system into an

equivalent system obtained by elementary operations.

Of Johann Carl Friedrich Gauss's many

contributions, we can highlight the
construction of a 17-sided regular polygon
with ruler and compass, as well as an
algebraic solution to the problem of
constructing any regular n-sided polygon
with a ruler and compass.

Called "the prince of mathematics", Gauss

legacy was decisive to the development of this
science.

© Copyright Universidad Europea. Todos los derechos reservados

 Gaussian elimination method
Gauss's method of solving consists of transforming the initial system
of m equations into an equivalent system that is simpler.

Given a system of m equations with n unknowns, the goal is to obtain

an equivalent system whose first equation maintains n unknowns, the
second n-1, the third n-2, and so on until the last equation has a
single unknown. Once this is done, we solve the last equation, then
the penultimate one, and so on until we reach the first one.

This method is also called the elimination method, as it involves

eliminating the unknowns in a staggered manner.

### ##$ ##% '#

It is applied to the augmented 0 #$$ ⋯
#$% '$
matrix of the system and the !∗ =
⋮ ⋱ ⋮
result is a staggered matrix
called echelon matrix
0 0 ⋯ #%& '&

© Copyright Universidad Europea. Todos los derechos reservados

 Gaussian elimination method. Procedure
1. Write the augmented matrix associated with the system of
equations, separating the solution column with a vertical bar.
2. The element chosen as the pivot (preferably a 1 or a -1) is
marked with a circle.
3. Once the pivot has been chosen, a new matrix is written, leaving
the pivot row in first place and without any modification.
4. Each element in the pivot row is multiplied by a number such
that adding or substracting the result to the element below the
pivot gives me zero (this process is called "zeroing").
5. This process is repeated for the following rows.
6. A new pivot is chosen that is not in the row we used in the
previous steps.
7. A new matrix is written in which the process of zeroing is
repeated.
8. The method ends when we have, under the main or secondary
diagonal, only zeros.

© Copyright Universidad Europea. Todos los derechos reservados

 Example

1. Write the coefficient matrix along with the independent terms

(Augmented Matrix)

2. By means of the transformations described, we transform the system

into an equivalent system that is staggered.

R2=R2-R1 R3=5R3 R3=R3-R2

R3=R3-R1

© Copyright Universidad Europea. Todos los derechos reservados

 Example
3. We analyze what type of system it is and, in consequence,
determine how many solutions it has.
Since r(A)=r(A)* and it matches the number of unknowns, the system
is Consistent Determinate (it has only one solution).

4. We rewrite the system from the echelon matrix obtained

5. To solve the system we can do it from the bottom up, that is, by
substituting the value of the unknowns in the previous equations.

The system's solution is therefore x=1, y=0, z=2.

© Copyright Universidad Europea. Todos los derechos reservados

 Example
Discuss the following system and solve
it using Gauss's method

Steps:

1. We write the augmented matrix

A
A*
2. Stagger the matrix using elementary operations.

R2=R2-2R1 R3=R3-2R2
R3=R3-3R1

3. Detemine the rank of the matrices.

Rank of A: 2
Rank of A*: 2 Consistent Undetermined
Nº unknowns: 3 System
© Copyright Universidad Europea. Todos los derechos reservados
 Undeterminated consistent systems

ü Undeterminated consistent systems have infinite solutions.

ü The term undetermined comes from the fact that some of the
unknowns remain not determined, that is, they depend on the value
of the others (through parameters). This is called the degree of
indeterminacy of the system.

ü The number of parameters we will use to write the solutions is the

difference

Nº parameters = Nº of unknowns - rank of the matrix

ü This process is known as system parameterization.

© Copyright Universidad Europea. Todos los derechos reservados

 Example
From the previous example:

We get:

This matrix corresponds to the system of equations:

© Copyright Universidad Europea. Todos los derechos reservados

 Example
Since the number of unknowns is 3, but the rank is 2, we need 1
parameter for its resolution (degree of indeterminacy 1).

We can take z as a parameter, and solve the other two unknowns

The infinite solutions appear from the different (real) values of

the parameter.

© Copyright Universidad Europea. Todos los derechos reservados

 Remarks about the Gaussian Method
ü The method is applicable to any system of linear equations
regardless of the number of equations and unknowns.

ü One of the advantages of this method is that allows the discussion

and resolution using the same calculations.

ü It's important to remember that elementary operations that can be

done in the matrix are only between rows and not between
columns, because the matrix is the representation of the system
and if we operate between columns we change the system to
another that is not equivalent.

ü Keep in mind that there are many ways of obtaining the echelon
matrix of a system. Depending on the operations you perform, you
might end up with different matrices. However, you must obtain the
same solutions for the system, regardless of the operations and the
particular echelon matrix.

© Copyright Universidad Europea. Todos los derechos reservados

 Inverse matrix
How do we solve the equation, how do we calculate the value of X?

AX=B
Remember that:

ü If A-1 is the inverse matrix of A, we have that

AA-1=A-1A=I

where I is the identity matrix.

ü The product of matrices is non-commutative.

THEREFORE: AX=B A-1AX=A-1B X= A-1B

© Copyright Universidad Europea. Todos los derechos reservados

 Inverse matrix method
Let's solve the following example using the inverse matrix method:

1. Calculate the matrix of coefficients and its determinant

2. Since the determinant is non-zero, the rank of A is 3 (the same as that of

the augmented matrix), so the system is consistent. Also, since it is equal to
the number of unknowns, the system is determinate (single solution).

rA=r(A*)=Nº unknowns: 3

© Copyright Universidad Europea. Todos los derechos reservados

 Example
3. To determine the unknowns, we solve the equation using the inverse matrix

By substituting and operating, we get the solution

x
y

z
The system's solution is x=1, y=0, z=2.
© Copyright Universidad Europea. Todos los derechos reservados
 Homogeneous systems

There is a class of systems of equations that, because of their

particularities, deserve to be mentioned separately: homogeneous
systems.

A system is said to be homogeneous if it has no independent terms,

that is, if its independent terms are all equal to zero

© Copyright Universidad Europea. Todos los derechos reservados

 Homogeneous systems

ü Homogeneous systems always have a solution (consistent

systems)

ü If its solution is unique (determined consistent system), it will be

(0,0,...,0) and it's called a trivial solution.

ü If the system is consistent undeterminate (infinite solutions), its

solutions always fullfil that:

ü The addition of two solutions is a homogeneous system

solution.
ü The product of a solution multiplied by a number (scalar) is a
homogeneous system solution.

© Copyright Universidad Europea. Todos los derechos reservados

 Thank you for
your
attention!

© Copyright Universidad Europea. Todos los derechos reservados 45