Module 2

Vectors and Matrices

Prof. Serena Brianzoni

Mathematics
Digital Economics and Business - Bachelor’s Degree
Marche Polytechnic University
Preliminary concepts
The space R2

• Any point in the plane is represented in a coordinate system

defined by two orthogonal lines
• These two lines, that are orthogonal to each other and
intersect at the point O, are called coordinate axes and
denoted by x−axis and y−axis
• Each axis represents the real line (with the arrow indicating
positive direction), hence R2 is the result of the Cartesian
product R × R (Cartesian product)

The two-dimensional space R2 is identified by the Cartesian plane

1
Cartesian plane

2
The space R3

• Any point in the space is represented in a coordinate system

defined by three orthogonal lines
• These three lines, that are orthogonal to each other and
intersect at the point O, are called coordinate axes and
denoted by x−axis, y−axis and z−axis
• Each axis represents the real line (with the arrow indicating
positive direction), hence R3 is the result of the Cartesian
product R × R × R

The three-dimensional space R3 is identified by the Cartesian space

3
Cartesian space

4
IMPORTANT

1. Both the pairs of real numbers (elements of R2 ) and the

triples of real numbers (elements of R3 ) are ORDERED
√ √
For example (−2, 10) ̸= (10, −2), (1, 3 −2, 0) ̸= (1, 0, 3 −2)
2. Given a = (a1 , a2 ) , b = (b1 , b2 ) ∈ R2
a < b ⇔ a1 < b1 and a2 < b2
Contrary to R the ordering is not total because there are
non-comparable elements. For example (1, 5) and (2, 3)
(the same holds in R3 )
5
The n-dimensional space Rn

By generalizing, we arrive at:

• Rn = R × R × · · · × R ( n times)
• Elements a ∈ Rn are ordered n-tuple of real numbers

a = (a1 , a2 , . . . , an ) : ai ∈ R, ∀i = 1, . . . , n
It is point of coordinates ai

6
Vectors
Vectors

A vector is an ordered n-tuple of real numbers:

a = (a1 , . . . , an ) ∈ Rn (ai is called vector’s component)

• example: unit prices vector of n goods p = (p1 , p2 , . . . , pn )

• null vector: 0 = (0, . . . , 0) ∈ Rn
| {z }
n times
• equality between two vectors of Rn :
a = (a1 , . . . , an ) and b = (b1 , . . . , bn )
a = b ⇔ a1 = b1 ∧ a2 = b2 ∧ · · · ∧ an = bn
7
Operations on vectors

• Sum (difference) a ± b = (a1 ± b1 , . . . , an ± bn )

with a = (a1 , . . . , an ) and b = (b1 , . . . , bn )
• Scalar multiplication λ · a = (λ · a1 , . . . , λ · an )
with a = (a1 , . . . , an ) and λ ∈ R
• Given p vectors x1 , . . . , xp ∈ Rn and p scalars k1 , . . . , kp ∈ R,
then the vector x of n components:
y = k1 · x1 + · · · + kp · xp = pi=1 ki xi
P

is said to be a linear combination of the vectors x1 , . . . , xp

8
Operations on vectors

By expanding the linear combination, one obtains a system:



 x1 = (x11 , . . . , x1n )

 x = (x , . . . , x )
2 21 2n
y = (y1 , . . . , yn )

 ...

xp = (xp1 , . . . , xpn )



 y1 = k1 · x11 + · · · + kp · xp1

 y = k · x + ··· + k · x X p
2 1 12 p p2
⇔ yi = kj · xji (i = 1, . . . , n)

 · · · j=1

yn = k1 · x1n + · · · + kp · xpn

9
Operations on vectors

• The opposite of a vector a ∈ Rn is

−a = (−1) · a = (−a1 , . . . , −an ) (⇒ a − a = 0 = (0, . . . , 0))
• Given two vectors x = (x1 , . . . , xn ) , y = (y1 , . . . , yn ) ∈ Rn ,
the scalar product or inner product or dot product is:
x · y = x1 y1 + · · · + xn yn = ni=1 xi yi ∈ R
P

NOTE
If the product of two vectors is zero ⇏ at least one factor is zero

In fact, x = (1, 1), y = (1, −1) ⇒ x · y = 1 − 1 = 0

10
Operations on vectors

• The NORM of a vector is the length of the vector

√ pPn
2
∥x∥ = x · x = i=1 xi
• Distance between two vectors
∀x = (x1 , x2 , . . . , xn ) , y = (y1 , y2 , . . . , yn ) ∈ Rn , the distance
between x and y is defined as the non-negative real number
v
u n
uX
d(x, y) = ∥x − y∥ = t (xi − yi )2
i=1

11
Matrices
Matrices

• An m-by-n (m × n) matrix A is a table of numbers with m

rows and n columns
 
a11 a12 · · · a1n
 21 a22 · · · a2n 
 a 
A =  .. .. . . . .. 
 . . . 
am1 am2 · · · amn

• aij is a generic elements identified by two indexes (i, j) (ith

row and jth column)
• A contains m · n elements: the dimension (or order) is m × n 12
Vectors are particular matrices

• A matrix with either only one row or only one column is a

vector
• It is usual to distinguish between a row vector (with only one
row) and a column vector (with only one column)
• The vector is the same, because it has the same components,
but the corresponding matrices are different
 
  2
x= 2 4 6 x=4
 
1×3
6
3×1
13
Square matrix (m = n)

For the particular case of square matrix (same number of columns

as rows)
 
a11 a12 · · · a1n
 a a ··· a2n 
 21 22
A =  ..

 . .. . . . .. 
. . 
an1 an2 · · · ann

It is defined the main diagonal as the elements

{aij : i = j} = {a11 , a22 , ..., ann }
14
Square matrix (m = n)
Square matrix

n=m=3⇒ matrix 3 × 3
 
1 2 1
A=7 9 2
 
0 1 3
1
Main diagonal 9
3
15
Rectangular matrix (m ̸= n)

Rectangular matrix
n = 2, m = 3 ⇒ matrix 2 × 3
!
1 −1 0
A=
3 −1 2

16
Particular matrices

• Null matrix (zero matrix): all zero elements

• Identity matrix: a diagonal matrix with all its diagonal
elements equal to 1 and zero everywhere else
 
1 0 0
I=0 1 0
 
0 0 1

17
Operations on matrices

• Sum of two matrices with same sizes:

C = A + B → cij = aij + bij ∀i = 1, ..., m j = 1, ..., n
! ! !
1 4 5 2 −2 0 3 2 5
+ =
−2 0 3 3 −1 −3 1 −1 0
• Scalar multiplication (multiplication of a matrix by a scalar)
! !
2 1 −3 4 2 −6
λ · A = B → bij = λ · aij → 2 · =
0 4 5 0 8 10 18
Operations on matrices

• Transposition (by turning rows into columns):

AT = B → bij = aji

Example
 
! 1 4
1 2 3 T
A= →A =2 5
 
4 5 6
3 6

19
Symmetric matrix

• A square matrix is said to be symmetric if it is equal to its

own transpose: AT = A
A = (aij ) with i = 1, . . . , n; j = 1, . . . , n
aij = aji ∀i, j (i ̸= j)

Example
   
1 2 1 1 2 1
T
A =  2 7 3 , A =2 7 3
   
1 3 9 1 3 9
20
Matrix Multiplication

Multiplication of two matrices is defined if and only if the number

of columns of the left matrix is the same as the number of rows of
the right matrix (CONFORMABLE matrices)
A(m × n) · B(n × p) = C(m × p)

where:
C = [cij ] , cij = ai,∗ · b∗,j = ai,1 · b1,j + ai,2 · b2,j + · · · + an,1 · bn,j
Entries cij are given by dot product of the corresponding row i of
A and the corresponding column j of B.
21
Matrix multiplication
Example
 
! 1 0 1
1 0 0
A= ,B =  0 2 0 
 
2 1 0
1 3 0
A(2 × 3) · B(3 × 3) = C(2 × 3)
! !
1+0+0 0+0+0 1+0+0 1 0 1
C= =
2+0+0 0+2+0 2+0+0 2 2 2

22
Matrix multiplication
NOTE
• Generally, if A and B are conformable, B and A are not
conformable (thus, A · B is possible, but not B · A)
• If A e B are squared with same sizes, then A · B and B · A are
possible but with different result
(matrix multiplication is not commutative)
• A · B = 0 (null matrix) does not imply that A or B are null
! ! !
1 0 0 0 0 0
A= ,B = ,A · B = =0
1 0 0 1 0 0
23
Multiplication of a matrix by a vector

Let A be an m × n matrix and consider the vector x as a matrix:

• the product x × A is possible if and only if x is a row vector of
m components !
  1 0 −1  
2 4 · = 2 −8 2 ∈ R3
0 −2 1
• the product A × x is possible if and only if x is a column
vector of n components  
! 1 !
1 0 −1 2
· 0 = ∈ R2
 
0 −2 1 −1
−1 24
Determinant (only for square matrices)

• The determinant of a square matrix is a real number and it is

denoted as follows

 
a11 . . . a1n a11 . . . a1n
A =  ... . . . ..  ⇒ det A = |A| =
.
.. . . .
.
..
.

an1 . . . ann an1 . . . ann

• If n = 1 ⇒ the matrix has order 1 ⇒ the determinant is the

element itself: A = (−2) ⇒ |A| = det(A) = −2
25
Determinant of order 2

The determinant of a 2 × 2 matrix is calculated by subtracting the

product of the off-diagonal elements from the product of the main
diagonal elements:
a11 a12
|A| = = a11 · a22 − a12 · a21
a21 a22

Example 2 × 2
!
1 2
A= ⇒ det A = 1 · 3 − 2 · 9 = −15
9 3
26
Determinant of order 3

• Rule of Sarrus applies only to 3 × 3 matrices

 
a11 a12 a13
A =  a21 a22 a23 
 
a31 a31 a33
a11 a12 a13 a11 a12
det A = |A| = a21 a22 a23 a21 a22 =
a31 a31 a33 a31 a31
= a11 · a22 · a33 + a12 · a23 · a31 + a13 · a21 · a32
−a31 · a22 · a13 − a32 · a23 · a11 − a33 · a21 · a12 27
Determinant of order 3

Example of the Sarrus rule

1 2 3 1 2 3 1 2
1 0 1 = 1 0 1 1 0 =
2 2 2 2 2 2 2 2

=1·0·2+2·1·2+3·1·2−2·0·3−2·1·1−2·1·2=4

28
Minor

• Let A be an m × n matrix and k an integer with

k ≤ min(m, n)
• A minor of order k of A is the determinant of a k × k
submatrix obtained from A by deleting m − k rows and n − k
columns
• Minors are defined for both square and rectangular matrices

29
Minor - example
 
1 2 3
A=4 5 6
 
7 8 9

Some 2 × 2 minors (obtained by deleting 1 row and 1 column):

1 2 4 6 2 3
= −3, = −6, = −3
4 5 7 9 5 6
1 3 1 3
= −12, = −6
7 9 4 6
30
Minor - example
!
1 −1 0
A=
3 −1 2

Minors of order 2 :
1 −1 1 0 −1 0
= 2; = 2; = −2
3 −1 3 2 −1 2
Minors of order 1 :
|1| = 1; |0| = 0; | − 1| = −1
| − 1| = −1; |3| = 3; |2| = 2
31
Cofactor

• Given an n × n matrix A, the cofactor of the element aij is

defined as Cij = (−1)i+j · Mij
where Mij is the minor obtained deleting the i-th row and j-th
column of aij
• Be careful, the cofactor sign depends on the position of aij
• Procedure for calculating the cofactor Cij : first delete row i
and column j to arrive at a minor of order n − 1, second
multiply the minor by the factor (−1)i+j

32
Determinant of any order

We can calculate the determinant of a matrix of order 3 using

cofactors, through the so-called Laplace expansion:

1. Choose an arbitrary row or column

2. Multiply each element aij of this row (or column) by the
corresponding cofactor Cij
3. Add the obtained products ⇒ this is the determinant

33
Laplace expansion

• Laplace expansion along the i-th row:

X n
det A = aik · Cik
k=1

• Laplace expansion along the j-th column:

X n
det A = akj · Ckj
k=1

• The result is independent on the selected row or column

34
Laplace expansion

Example

1 2 3
0 1 1 1 1 0
1 0 1 = 1· −2· +3· = −2 − 0 + 6 = 4
2 2 2 2 2 2
2 2 2

• This rule also applies to higher order determinants, by

iterating it until 2 × 2 determinant are obtained
35
Properties of determinants

These properties are useful in the calculation of the determinant:

• The matrix has a row or column of zeros ⇒ the det is zero

• The matrix has two equal rows or columns ⇒ the det is zero
• A row (column) is a linear combination of other rows
(columns) ⇒ the det is zero
• The determinant does not change if we add a linear
combination of rows (columns) to a row (column)

36
Properties of determinants

Example
1 4 5
2 5 7 =0
3 6 9

• The third column is the sum of the first two columns

NOTE A matrix with zero determinant is called singular matrix

37
Rank

The rank of a matrix A, rank(A) or rk(A) is the maximum order of

the non-zero minors that we can extract from the matrix

Example Square matrix of order three

1. The rank is three (maximum possible rank) if the determinant
of the matrix is different from zero, otherwise, the rank is at
most equal to two
2. The rank is two if we find at least one minor of order two
different from zero
3. If all minors of order two are null, the rank will be equal to one 38
Rank
Example: verify if rk(A) = 3
 
1 0 −3
A =  −3 5 9 
 
2 22 −6
The only minor of order three is:

1 0 −3
1 0
−3 5 9 = 0 ⇒ rk(A) < 3, = 5 ̸= 0
−3 5
2 22 −6
39
⇒ rk(A) = 2
Rank

Example
!
1 2 3
C=
4 5 6
The maximum possible rank is 2

1 2
= −3 ̸= 0 ⇒ rk(C) = 2
4 5

40
Rank
Example
 
1 4 3 2
A =  −2 1 0 −3 
 
−1 5 3 −1
• The maximum possible rank is 3
• All minors of order three are zero, so rk(A) < 3 (the third row
is the sum of the first two)

1 4
= 9 ̸= 0 ⇒ rk(A) = 2
−2 1 41
Rank
Rule of thumb
• Consider the maximum possible rank rmax (equal to the
smallest of the number of rows and columns)
• If we find a non-zero minor of order rmax , the rank will be rmax
• Otherwise, we consider a smaller order minor (one unit smaller)
and apply the same procedure until we find a non-null minor
• In the special case of the null matrix which has rank zero

NOTE Differently from the determinant, the rank is defined for

both square and rectangular matrices 42
Rank dependent on a parameter

• When a matrix contains a real parameter k, we discuss the

rank of the matrix depending on the parameter’s values
Example
 
1 3 2
A =  −3 1 5 
 
−2 4 k

• If matrix A is non-singular (det(A) ̸= 0) ⇒ rk(A) = 3

43
Rank dependent on a parameter

Example (cont’d)
• Compute the determinant:
1 3 2
−3 1 5 = 10k − 70 ̸= 0 ⇒ k ̸= 7
−2 4 k

• If k ̸= 7 ⇒ rk(A) = 3

44
Rank dependent on a parameter
Example (cont’d)
• If k = 7
 
1 3 2
A =  −3 1 5  ⇒ det(A) = 0 ⇒ rk(A) < 3
 
−2 4 7

• There is a non-zero minor of order two, so the rank is two

1 3
= 10 ̸= 0 ⇒ rk(A) = 2
−3 1
45
Rank dependent on a parameter
Example
 
1 3 −2
A =  −3 −9 k 
 
2 k −4

• If the matrix A is non-singular ⇒ rk(A) = 3

• Compute the determinant:
det A = −k 2 + 12k − 36 = −(k − 6)2 ̸= 0 ⇒ k ̸= 6

46
Rank dependent on a parameter
Example (cont’d)
• If k ̸= 6 ⇒ rk(A) = 3
• If k =6 
1 3 −2
A =  −3 −9 6  ⇒ det(A) = 0 ⇒ rk(A) < 3
 
2 6 −4
• Observe that all minors of order 2 are null (last two columns
are derived from the first) ⇒ the rank cannot be equal to two
⇒ rk(A) = 1
47
Rank dependent on a parameter

Example
!
4 1 2
A=
k 3 k

• The maximum possible rank of the 2 × 3 matrix A is 2

• Examine all possible minors of order two

48
Rank dependent on a parameter
Example (cont’d)
Minors of order 2:

4 1
= 12 − k ̸= 0 ⇒ k ̸= 12
k 3

4 2
= 2k ̸= 0 ⇒ k ̸= 0
k k

1 2
= k − 6 ̸= 0 ⇒ k ̸= 6
3 k
49
Rank dependent on a parameter

Example (cont’d)
Conclusion:

• If k ̸= 0 the second minor is non-zero

• If k = 0 the first minor is non-zero

Therefore, we can find a non-zero minor of order two for any

value of the parameter k so the rank of the matrix is always equal
to two

50
Systems of linear equations
Systems of linear equations

• A linear equation is a first order equation with n variables

a1 x1 + a2 x2 + · · · + an xn = b
2x1 + 3x2 + 4x3 + 9x4 = −3
xi are the variables (unknowns)

• A linear system is composed of linear equations

(
2x1 − 3x2 + 5x3 = 2
4x1 − 2x2 − 6x3 = 8
51
Systems of linear equations

• Generally: m equations with n unknowns (x1 , x2 , ..., xn ),

where m may be greater than, equal to, or less than n


 a11 x1 + a12 x2 + · · · + a1n xn = b1

 a x + a x + ··· + a x = b
21 1 22 2 2n n 2
 .
..


am1 x1 + am2 x2 + · · · + amn xn = bm


or in matrix form:
Ax = b
52
Systems of linear equations
 
a11 a12 . . . a1n
 a21 a22 . . . a2n 
Coefficient matrix A(m × n) = 
 
.. ... . . . ... 
 . 
am1 am2 . . . amn
   
b1 x1
 b2  x
 2

Constant vector b =   ∈ Rm Unknows x =  ..  ∈ Rn
  
..
 .   . 
bm xn

53
Solving a system

• A solution of an m × n system is an n-tuple of numbers that

satisfies all the equations of the system simultaneously
• A linear system may behave in 3 possible ways:
• The system has solutions (consistent)
• infinitely many solutions (undetermined)
• unique solution
• The system has no solution (inconsistent)
• Solving a system means finding all the solutions:
• Determine consistency (Cramer and Rouché-Capelli theorem)
• Find all possible solutions (Cramer’s rule)
54
Square system m = n

• Cramer theorem
det A ̸= 0 is a necessary and sufficient condition for a system
of n linear equations in n unknowns to have a unique solution
Example

 x+y+z =1

x − 2y − z = 2
 2x + y − 6z = 3

2

55
Square system m = n
Example (cont’d)
   
1 1 1 1
A =  1 −2 −1  , b= 2 
   
2 1 −6 3/2
1 1 1
det A = 1 −2 −1 =
2 1 −6
= 12 + (−2) + 1 − (−1) − (−4) − (−6) = 22 ̸= 0
⇒ unique solution
56
Solutions

• Cramer’s rule (solution)

If det A ̸= 0 ⇒

b1 a12 . . . a1m a11 b1 . . . a1m

b2 a22 . . . a2m a21 b2 . . . a2m
.. . . . . . . ... .. . . . . . . ...
. .
bn an2 . . . anm an1 bn . . . anm
x1 = , x2 =
det A det A

57
Cramer’s rule

• Generally:
det Ai
xi = , i = 1, . . . , n,
det A
where Ai is the matrix obtained from A by replacing the i-th
column with the vector of constant terms (or right-hand sides)

58
Cramer’s rule
Example
( !
x+y =3 1 1
, A= , det A = −2
x−y =1 1 −1
3 1 1 3
1 −1 −4 1 1 1−3
x= = = 2, y= = =1
1 1 −2 1 1 −2
1 −1 1 −1
x = 2, y=1
59
General case (m is not necessarily equal to n)

• Rouché-Capelli theorem
A system of linear equations with m equations and n
unknowns has solutions if the rank of the coefficient matrix is
equal to the rank of the augmented matrix:
If rk(A) = rk(A|b) ⇒ the system is consistent (has solutions)
 
a11 a12 . . . a1n
• Coefficient matrix: A =  ... . . . . . . ... 


am1 am2 . . . amn

60
General case

• Augmented matrix: coefficients matrix augmented with the

column vector of constant terms.
 
a11 a12 . . . a1n b1
 a
 21 a22 . . . a2n b2 

A|b =  .. .. . . . .. .. 
 . . . . 
am1 am2 . . . amn bm

61
Solutions of a system

We can have:

1. rk(A) < rk(A|b) ⇒ the system has no solutions (inconsistent)

2. rk(A) = rk(A|b) = r ⇒ at least one solution (consistent)
2.1 r = n → unique solution
2.2 r < n → infinite solutions

The case r > n is impossible because the rank of A(m × n) is less

than or equal to the minimum between m and n

62
Solutions of a system

Unique solution (r = n)
1. Consider the r-order minor used to identify the rank of A
2. The rows of this r-order minor correspond to the main
equations of the system, the other equations must be
eliminated
3. We arrive at a square system of order r that can be solved
with Cramer’s rule

63
Solutions of a system

Infinite solutions (r < n)

1. Consider the r-order minor used to identify the rank of A
2. The rows of this r-order minor correspond to the main
equations of the system, the other equations must be
eliminated
3. The columns of this r-order minor correspond to the main
unknowns, the other (secondary ) unknowns are moved to the
right of the equal sign and assume arbitrary constant values
4. Finally, we solve with Cramer’s rule
64
Examples

Example 1
  
 x+y =3
 1 1
x + 2y = 3, A =  1 2 , rk(A) = 2
 

 2x + y = 5 2 1
1 1 1 2 1 1
= 1, = −3, = −1
1 2 2 1 2 1

65
Examples

Example 1 (cont’d)
 
1 1 3
A|b =  1 2 3  , det A|b = −1 ̸= 0
 
2 1 5
rk(A|b) = 3 > rk(A) = 2
⇒ NO SOLUTIONS

66
Examples

Example 2
  

 x + y = 1 1 1
x + 2y = 2, , A =  1 2  , rk(A) = 2
 

 2x + 3y = 3 2 3
 
1 1 1
A|b =  1 2 2  , det A|b = 0 ⇒ rk(A|b) = 2
 
2 3 3

67
Examples

Example 2 (cont’d)
• Because rk(A) = rk(A|b) = 2, the system has a unique
solution
• We can use any two equations to find the solution, as all
minors of order 2 are different from zero

68
Examples
Example 2 (cont’d)


 1 1


 2 2

 x= =0
1 1




 
x+y =1  1 2
⇒
x + 2y = 2 
 1 1

1 2


y= =1






 1 1
1 2


69
Examples
Example 3

 x + 2y + z = 0

2x + 2y + 2z = 4

 x−y−z =1

rk(A) = rk(A|b) = 3 ⇒ Unique solution:

3 5
x = , y = −2, z =
2 2

70
Examples
Example 4 (m = 3 equations, n = 2 unknowns)

 3x − 2y = 0

x − y = −1

 2x − 3y = −4
 
3 −2
Coefficient matrix A =  1 −1 
 
2 −3
 
3 −2 0
Augmented matrix A|b =  1 −1 −1 
 
71
2 −3 −4
Examples

Example 4 (cont’d)
 
3 −2
A =  1 −1 
 
2 −3

3 −2
= −3 + 2 = −1 ̸= 0 ⇒ rk(A) = 2
1 −1

72
Examples

Example 4 (cont’d)
 
3 −2 0 3 −2 0
A|b =  1 −1 −1  , 1 −1 −1 = −1 ̸= 0
 
2 −3 −4 2 −3 −4
⇒ rk(A|b) = 3
rk(A|b) = 3 > rk(A) = 2
⇒ NO SOLUTIONS

73
Examples

Example 5

 3x − 2y = 0

x − y = −1

 4x − 3y = −1
 
3 −2
3 −2
A =  1 −1  , = −1 ̸= 0 ⇒ rk(A) = 2
 
1 −1
4 −3

74
Examples

Example 5 (cont’d)
 
3 −2 0 3 −2 0
A|b =  1 −1 −1  , 1 −1 −1 = 0
 
4 −3 −1 4 −3 −1
⇒ rk(A|b) < 3
rk(A|b) = rk(A) = 2
⇒ UNIQUE SOLUTIONS

75
Examples
Example 5 (cont’d). Delete the third equation.


 0 −2


 −1 −1
x = =2



 
 3 −2
3x − 2y = 0 
⇒ 1 −1
x − y = −1 

 3 0
y = =3


−2




 3
1 −1


76
Examples

Example 6
(
x + 2y − 3z = 2
m = 2 equations, n = 3 unknowns
4x − y + 2z = 3
!
1 2 −3
Coefficient matrix A =
4 −1 2
!
1 2 −3 2
Augmented matrix A|b =
4 −1 2 3

77
Examples
Example 6 (cont’d)
• The rank of matrix A is 2 :
1 2
= −9 ̸= 0 ⇒ rk(A) = 2
4 −1

• Rouché-Capelli theorem:
rk(A|b) = rk(A) = 2 < m = 3
⇒ INFINITE SOLUTIONS
Secondary variable: z 78
Examples
Example 6 (cont’d)
• Solution:
(
x + 2y = 3z + 2
4x − y = −2z + 3


 3z + 2 2
 − −2z + 3 −1 = −3z−2+4z−6 = 8−z



⇒ −9 −9 9

 1 3z + 2

4 −2z + 3


x= = −2z+3−12z−8 = 14z+5

−9 −9 9
79
Examples

Example 7 - Discuss depending on the real parameter k

    
 x+y =4
 1 1 1 1 4
4x − ky = 8 A =  4 −k  A|b =  4 −k 8 
   

 x−y =0 1 −1 1 −1 0

1 1
• = −2 ̸= 0 ⇒ rk(A) = 2
4 −1
• det A|b = 4k ⇒ The rank of A|b depends on k

80
Examples

Example 7 (cont’d)
• if k ̸= 0 ⇒ 4k ̸= 0 ⇒ rk(A|b) = 3 > rk(A) = 2
⇒ NO SOLUTIONS
• if k = 0 ⇒ 4k = 0 ⇒ rk(A|b) = 2 = rk(A)
⇒ UNIQUE SOLUTION

x=y=2

81
Homogeneous systems

• A linear system is homogeneous if all constant terms are zeros

⇒ it always has the trivial null solution (all unknowns are
equal to zero)
• In a square system, the unique solution is the null solution if
and only if the coefficient matrix A is non-singular
• A rectangular system is always consistent (because there exists
the null solution) ⇒ rk(A) = rk(A|b) HENCE:
1. If the rk(A|b) = n ⇒ unique solution (the null one)
2. If rk(A|b) < n ⇒ infinite solutions
82
Linearly dependent vectors

Vectors x1 , . . . , xp ∈ Rn are linearly dependent if at least one of

them is a linear combination of the others

p−1
X
xp = ki xi
i=1
p−1
X Xp
⇒ ki xi − xp = 0 ⇒ ki xi = 0, with kp = −1
i=1 i=1
⇒ k1 · x1 + · · · + kp · xp = 0
with at least one ki ̸= 0 (kp = −1) 83
Linearly dependent vectors
Example
• Is the vector (6, 2, 3) dependent on {(1, 2, 3), (2, −1, 4)} ?
• The following relation should hold:
(6, 2, 3) = α · (1, 2, 3) + β · (2, −1, 4)

 2α − β = 2 → α = 2

⇔ α + 2β = 6 → β = 2

 3α + 4β = 3 → not possible

The system has no solutions ⇒ the vector (6, 2, 3) cannot be

linear combination of the others: 84
Linearly independent vectors

Vectors x1 , . . . , xp ∈ Rn are linearly independent if no vector xi can

be represented as a linear combination of the remaining vectors:

p
X
ki xi = 0
i=1
⇒ k1 · x1 + · · · + kp · xp = 0 ⇔ k1 = . . . = kp = 0
This means that the only representation of 0 as a linear
combination of the vectors x1 , . . . , xp is the trivial representation in
which all the scalars ki are zero 85
Unit vectors Rn

Set of vectors, each of whose components are all zero, except one
that equals 1 (vector length = 1 )

e1 = (1, 0, 0, 0, . . . , 0) e2 = (0, 1, 0, 0, . . . , 0)
e3 = (0, 0, 1, 0, . . . , 0) en = (0, . . . , 0, 1)

Unit vectors are linearly independent and they generate all

vectors of Rn by linear combinations ⇒ they are said to be a
BASIS
86
Canonical basis (Standard basis)

Unit vectors e1 , e2 , . . . , en are a canonical basis of Rn

They are a privileged basis since the coordinates of the generic
vector v = (v1 , . . . , vn ) ∈ Rn with respect to this basis are exactly
v1 , . . . , vn :

v = (v1 , . . . , vn ) = v1 · (1, 0, . . . , 0) + · · · + vn · (0, . . . , 0, 1)

87
Examples
2 dimesions

(4, −2) = (4, 0) + (0, −2) = 4 · (1, 0) − 2 · (0, 1)

Basis: {(1, 0), (0, 1)}

3 dimesions
(4, −2, 5) = (4, 0, 0) + (0, −2, 0) + (0, 0, 5) =
4 · (1, 0, 0) − 2 · (0, 1, 0) − 5 · (0, 0, 1)
Basis: {(1, 0, 0), (0, 1, 0), (0, 0, 1)}
88