International Baccalaureate

MATHEMATICS
Applications and Interpretation HL
Lecture Notes

Christos Nikolaidis

TOPIC 1
NUMBER AND ALGEBRA

International Baccalaureate
1B. Complex numbers - Matrices

COMPLEX NUMBERS

1.8 COMPLEX NUMBERS – BASIC OPERATIONS …………………………………….. 1

1.9 THE COMPLEX PLANE …………………………………………..……………………………… 9

1.10 PRODUCTS, QUOTIENTS AND POWERS IN POLAR FORM ……………….. 16

1.11 ADDING SINUSOIDAL FUNCTIONS ………………………………………………………. 21

MATRICES

1.12 DEFINITION OF A MATRIX – MATRIX OPERATIONS ….……………………… 25

1.13 THE DETERMINANT detA – THE INVERSE A-1 ………………………………….. 35

1.14 MATRIX EQUATIONS – THE LINEAR SYSTEM AX=B ………...………………. 41

1.15 TRANSFORMATION MATRICES …………………………………………………………….. 48

1.16 EIGENVALUES AND EIGENVECTORS …………………………………………………… 55

January 2022
TOPIC 1: NUMBER AND ALGEBRA Christos Nikolaidis

1.8 COMPLEX NUMBERS – BASIC OPERATIONS

As we know, there are no real numbers of the form

1 4 , 9 , 5

However, we agree to accept an imaginary number i such that

i2 = -1

(so, in some way, the definition of i is: i =  1 )

The imaginary numbers mentioned above can be written as follows:

instead of  4 we write 2i
instead of  9 we write 3i
instead of  5 we write 5i

Consider now the equation

x 2  4x  13  0

Since Δ=-36, there are no real solutions. However, if we accept

that Δ  i 36  6i we obtain solutions of the following form

4  Δ 4  6i
x= = = 2  3i
2 2
These “new” numbers are known as complex numbers.

In general, if Δ<0 the complex roots of the quadratic are given by

-b  i Δ
x=
2a

NOTICE for the GDC - For Casio:

 use shift-0 for i
 for a quadratic with Δ<0 you may obtain the complex roots if
you use SET UP – Complex mode: a+bi

This gives rise to the following definition.

3
TOPIC 1: NUMBER AND ALGEBRA Christos Nikolaidis

 THE DEFINITION

A number z of the form z = x  yi

where x, y R, is called a complex number. We also say,

the real part of z is x: Re(z)=x

the imaginary part of z is y: Im(z)=y

The set of all complex numbers is denoted by C. A real number x is

also complex of the form x+0i (it has no imaginary part).

 THE CONJUGATE z

The conjugate complex number of z = x+yi is given by z =x-yi

(Sometimes, the conjugate number of z is denoted by z*)

EXAMPLE 1

For z=3+4i, we write Re(z)=3, Im(z)=4, z =3-4i.

Similarly

Complex number Real part Imaginary part Conjugate

z Re(z) Im(z) z
2+3i 2 3 2-3i
2-3i 2 -3 2+3i
-2+3i -2 3 -2-3i
-2-3i -2 -3 -2+3i
1+i 1 1 1-i
3i 0 3 -3i
2 2 0 2
0 0 0 0
i 0 1 -i
2  3i 1 3 2  3i
4 2 4 4

4
TOPIC 1: NUMBER AND ALGEBRA Christos Nikolaidis

 THE MODULUS |z|

The modulus of z=x+yi is defined by |z| = x 2  y2

For example, if z=2+3i, then |z|=|2+3i|= 2 2  3 2  13

Notice
z x+yi

z x-yi
-z -x-yi
-z -x+yi

all have the same modulus x 2  y2

Thus 3+4i, 3-4i, -3-4i, -3+4i have the same modulus

3 2  4 2  25  5

Finally, observe that |3|=3 and |-3|=3. That is, the modulus
generalizes the notion of the absolute value for real numbers.

 EQUALITY: z1 = z2

Two complex numbers are equal if they have equal real parts and
equal imaginary parts: Let z1=x1+y1i and z2=x2+y2i

x 1  x 2
z1 = z2  
y1  y 2

Thus, the equation of complex number must be thought as a system

of two simultaneous equations.

EXAMPLE 2

Let z1=3+4i and z2=a+(3b-2)i. Find a,b if z1 = z2.

3  a a  3
z1 = z2    
4  3b  2 b  2

5
TOPIC 1: NUMBER AND ALGEBRA Christos Nikolaidis

 ADDITION-SUBTRACTION-MULTIPLICATION-DIVISION

The four operations for complex numbers follow the usual laws of
algebra. We only have in mind that i2=-1.

EXAMPLE 3
Consider the two complex numbers z=7+4i and w=2+3i

 z+w =(7+4i)+(2+3i) =9+7i [add real parts; add imaginary parts]

 z-w =(7+4i)-(2+3i) = 5+i [similarly]

For multiplication we need some extra work:

 zw =(7+4i)(2+3i) = 14+21i+8i+12i2 = 14+21i+8i-12 =2+29i

What about the division?

z 7  4i
The fraction = is also a complex number!
w 2  3i
In order to obtain the usual form x+yi, we multiply both terms by
the conjugate of the denominator i.e. by w  2-3i

z 7  4i 7  4i (2  3i) 14  21i  8i  12 26  13i

 = = = = =2-i
w 2  3i 2  3i (2  3i) 13 13

Thus
7  4i
= 2-i
2  3i
(Confirm the result by multiplying (2+3i)(2-i); you must find 7+4i)

NOTICE

|z|2 = z. z

Indeed, both sides are equal to x2+y2: For z = x+yi

|z|2 = x2+y2
z z = (x+yi)(x-yi) = x2-y2i2 = x2+y2

6
TOPIC 1: NUMBER AND ALGEBRA Christos Nikolaidis

Get used to the multiplication by a conjugate:

EXAMPLE 4
(3+4i)(3-4i) = 9+16 = 25
(1+i)(1-i) = 1+1 = 2
(2-i)(2+i) = 4+1 =5
The result is always a real number (the square of the modulus).

EXAMPLE 5
Let us estimate the powers of i:

i0=1 i1=i i2=-1 i3=-i

i4=1 i5=i i6=-1 i7=-i
i8=1 i9=i i10=-1 i11=-i
… and so on

Thus, for example

i35=i32+3=i3=-i
(since 32 is a multiple of 4).

EXAMPLE 6
Calculate
(2  i) 3
(a) z = (2+i)3 (b) w=
1i
Solution
(a) z = (2+i)2(2+i) = (4+4i+i2)(2+i)

= (3+4i)(2+i)

= 6+3i+8i+4i2

= 2+11i

(2  i) 3 2  11i 2  11i 1  i 2  2i  11i  11i 2

(b) w= = = =
1i 1i 1i 1i 2
 9  13i  9 13
= = + i
2 2 2

7
TOPIC 1: NUMBER AND ALGEBRA Christos Nikolaidis

EXAMPLE 7

Find z if
z(1-i) = 2+11i

Solution

Method A (Analytical and safe but laborious):

Let z = x+yi. Then

z(1-i) = 2+11i  (x+yi)(1-i) = 2+11i

 x-xi+yi-yi2 = 2+11i

 (x+y)+(y-x)i = 2+11i

x  y  2
 
y  x  11
The solution of the system is x =-9/2 and y =13/2
 9 13
Hence, z = + i
2 2

Method B (quicker): think as in the equation ax=b

2  11i
z(1-i) = 2+11i  z =
1i
= …
 9 13
= + i [look at Exercise 6(b)]
2 2

8
TOPIC 1: NUMBER AND ALGEBRA Christos Nikolaidis

1.9 THE COMPLEX PLANE

The complex number z=x+yi can be represented on the Cartesian

plane as follows

y z = x+yi

O x

 z=x+yi is the point (x,y) *

 Real part = x-coordinate, Imaginary part = y-coordinate

 The modulus |z|= x 2  y 2 is in fact the distance from the origin.

EXAMPLE 1

4 3+4i
0+4i
3

2
1

-5 O 1 2 3
3+0i
-2
-5-2i

Notice

|3+4i|= 25 =5, |3|=3, |4i|=4, |-5|=5, |-5-2i|= 29

The modulus is always the distance from the origin.

*
We may also think of z as a vector from the origin to the point (x,y). Compare
x
with vectors   in paragraph 3.10
y
 

9
TOPIC 1: NUMBER AND ALGEBRA Christos Nikolaidis

NOTICE
We already know that the sets
N = natural numbers
Z = integers
Q = rational numbers
R = real numbers
can be represented on the real axis. We extend this representation
here to the complex plane (considering an imaginary y-axis).
It also holds
N Z Q R C

EXAMPLE 2
It is interesting to see the representations of z, z ,-z,- z . For
example,
z=3+4i z =3-4i z=-3+4i - z =-3-4i

- z =-3+4i z=3+4i

-z=-3-4i z =3-4i

 The modulus of all those is 5 (distance from the origin).

 z is symmetric to z about the x axis
 -z is symmetric to z about the origin

Think that these observations hold for real numbers as well:

 The absolute value of 5 and -5 is 5 (distance from origin)
 The conjugate of 5 is 5 itself (symmetric about x-axis)
 The opposite of 5 is -5 (symmetric about the origin)

10
TOPIC 1: NUMBER AND ALGEBRA Christos Nikolaidis

 THE POLAR FORM (MODULUS-ARGUMENT FORM)

We have just seen that a complex number z=x+yi is represented on

the plane by a pair of Cartesian coordinates (x,y).

An alternative way to describe a point on the plane (and thus the

position of z) is the so-called Polar coordinates (r,θ):

We draw a vector (an arrow) from O to the point and consider

r = the length of the vector

θ = the angle between the x-axis and the vector

y z = x+yi
r

θ
O x

Notice:

x y y
cosθ = , sinθ = , tanθ = (*)
r r x

For a complex number z=x+yi

r is in fact the modulus |z|

θ is called argument of z. We write arg(z)=θ

REMARK

Of course, the argument θ is not unique. For example, if θ=30ο is

an argument of z then

36oo+30ο= 390o, 720o+30o= 750o, etc

are also arguments of z.

For the principal argument θ, we agree

-180o<θ  180 o or -π<θ  π

but in our applications, we may consider any equivalent argument.

11
TOPIC 1: NUMBER AND ALGEBRA Christos Nikolaidis

The relations (*) above give

x=rcosθ and y=rsinθ

Thus, a complex number can be also written as

z = x+yi = (rcosθ)+(rsinθ)i = r (cosθ+isinθ)

The form
z = r (cosθ+isinθ)

is known as the polar form of the complex number z

(or otherwise modulus-argument form or trigonometric form)

 TRANSFORMATION FROM z = x+yi TO z = r (cosθ+isinθ)

Given: z = x+yi. We find r and θ by

 r = |z| = x 2  y2

y
 tanθ = , having in mind the quadrant of x+yi
x

EXAMPLE 3

Find the polar form of z =1+ 3 i and w =3+4i.

Solution
 For z =1+ 3 i :

r = 1  3 =2,
3 π
tanθ = = 3 , [1st quadrant]  θ=
1 3
π π
Therefore, z = 2(cos +isin ) [or 2 (cos 60 +isin 60 )]
3 3
 For w =3+4i :
r = 32  4 2 = 5 ,
4
tanθ = , [1st quadrant]  θ = 0.927 (by GDC)
3
Therefore, w = 5 [cos(0.927)+isin(0.927)]

12
TOPIC 1: NUMBER AND ALGEBRA Christos Nikolaidis

EXAMPLE 4

Find the polar form of z1 =1+i, z2 =-1-i, z3 = 1-i, z4 =-1+i

Solution

For all of them the modulus is r = 11 = 2

1 π
 For z1 =1+i: tanθ = = 1, [1st quadrant]  θ =
1 4
π π
Therefore, z1 = 2 (cos +isin )
4 4
1 5π
 For z2 =-1-i: tanθ = =1, [3rd quadrant]  θ =
-1 4
5π 5π
Therefore, z2 = 2 (cos +isin )
4 4
1 π
 For z3 = 1-i: tanθ = =-1, [4th quadrant]  θ = 
-1 4
π π
Therefore, z3 = 2 (cos  +isin  )
4 4
-1 3π
 For z4 =-1+i: tanθ = =-1, [2nd quadrant]  θ =
1 4
3π 3π
Therefore, z4 = 2 (cos +isin )
4 4

 TRANSFORMATION FROM z = r (cosθ+isinθ) TO z = x+yi

This is much easier! We just perform the operations! In fact, the

polar form is also Cartesian:
z = r(cosθ+isinθ) = rcosθ + irsinθ x=rcosθ, y=rsinθ

For example,
π π 1 3
the Cartesian form of z = 2(cos +isin ) is 2( +i ) = 1+ 3 i
3 3 2 2

NOTICE for the GDC

GDC transforms one form to another. For Casio use:

Run-Matrix – OPTN – COMPLEX - ►r<θ or ►a+bi

13
TOPIC 1: NUMBER AND ALGEBRA Christos Nikolaidis

 CIS FORM: z=rcisθ

There is an abbreviation for the polar form

z = r (cosθ+isinθ).
It is sometimes written as
z = rcisθ.
π π π
For example, z =2(cos +isin ) = 2cis .
3 3 3

 EULER’S FORM: z=reiθ

Another abbreviation is due to Euler.

We define†
eiθ=cosθ+isinθ

Consequently, the trigonometric form z=r(cosθ+isinθ) obtains the

form
z=reiθ

π
π π i
For example, z =2(cos +isin ) = 2 e 3 .
3 3

EXAMPLE 5
Write down all the possible forms of z1=1+i , z2 =3+4i, z3 =3+4i

Cartesian Polar form cis form Euler form

π π π π
1+i 2(cos +isin ) 2 cis 2e
i
4
4 4 4

3+4i 5[cos(0.927)+isin(0.927)] 5cis(0.927) 5 e0.927i

3-4i 5[cos(-0.927)+isin(-0.927)] 5cis(-0.927) 5 e-0.927i

† This is not accidental! It can be shown that eiθ follows all known exponential properties

14
TOPIC 1: NUMBER AND ALGEBRA Christos Nikolaidis

NOTICE
 Any complex number with modulus 1 has polar form
z = cisθ = cosθ+isinθ
(Indeed, |z|= cos 2 θ  sin 2 θ =1)

For any z given in polar form we know |z|. For example,

if z = 23cis3π = 23(cos3π+isin3π) then |z|=23

 For real numbers  a (on the horizontal, real axis)

the polar form is immediate, since
- the argument of a positive real number is 0
- the argument of a negative real number is π.

1 cis0 ei0 -1 cisπ eiπ

2 2cis0 2 ei0 -2 2cisπ 2 eiπ

3 3cis0 3 ei0 -3 3cisπ 3 eiπ

... ...

 For imaginary numbers of the form ai : (on the imaginary axis)
the argument is either π/2 or -π/2.

π  π
i cis eiπ/2 -i cis -  e-iπ/2
2  2
π  π
2i 2cis 2eiπ/2 -2i 2cis -  2e-iπ/2
2  2
π  π
3i 3cis 3eiπ/2 -3i 3cis -  3e-iπ/2
2  2
... ...

 The conjugate of z = r (cosθ+isinθ) is z = r (cosθ-isinθ)

Attention!! The latter is not in polar form. We must have a (+)

sign in front of i. However, we know that
cos(-θ) = cosθ and sin(-θ) = -sinθ

Hence z = r [cos(-θ)+isin(-θ)]

Indeed, if arg(z)=θ then arg( z )=-θ

15
TOPIC 1: NUMBER AND ALGEBRA Christos Nikolaidis

1.10 PRODUCTS, QUOTIENTS AND POWERS IN POLAR FORM

The polar form is convenient for multiplication, division and powers

of complex numbers:

For z1 = r1 cisθ1 and z2 = r2 cisθ2

z1 z2 = r1r2 cis(θ1+θ2)

z1 r
= 1 cis(θ1-θ2)
z2 r2

Notice that for z = r cisθ,

z2 = z z = r r cis(θ+θ) = r2 cis(2θ)
z3 = z2z = r2r cis(2θ+θ) = r3 cis(3θ)

We can generalize this result as follows:

For z = r cisθ
zn = rn cis(nθ) [De Moivre law]

NOTICE:
The modulus of z1z2 is r1r2: |z1z2| = |z1||z2|
z1 r z1 z1
The modulus of is 1 : =
z2 r2 z2 z2
The modulus of zn is rn: |zn| = |z|n
Also
The argument of z1z2 is θ1+θ2: arg(z1z2)=arg(z1)+arg(z2)
z1 z1
The argument of is θ1-θ2: arg( )=arg(z1)-arg(z2)
z2 z2
The argument of zn is nθ: arg(zn)=narg(z)

In other words,
the modulus |z| preserves the operations
the argument arg(z) behaves like log(z)

16
TOPIC 1: NUMBER AND ALGEBRA Christos Nikolaidis

For example, if z = r cisθ and w = ρ cisφ

z3 r3
the modulus of is
w4 ρ4
z3
the argument of is 3θ–4φ
w4
In other words
z3 r3
= cis(3θ–4φ)
w4 ρ4

EXAMPLE 1
π π 1 3
Let z =2cis and w =cis (In fact, z = 3 +i and w = + i)
6 3 2 2

Then
π π 3π π
 zw =2.1 cis( + ) =2cis( ) = 2cis( )
6 3 6 2
π π
=2[cos( )+isin( )]=2(0+i) = 2i
2 2
z 2 π π π
 = cis( - ) = 2cis(- )
w 1 6 3 6
π π 3 1
= 2[cos(- )+isin(- )] =2( - i)= 3 -i
6 6 2 2
π π
 z2 =22 cis(2 ) = 4cis
6 3
π π 1 3
= 4(cos +isin ) =4( + i)=2+2 3 i
3 3 2 2
π π
 z3=23 cis(3 ) = 8cis =8i
6 2
π
 z6=26 cis(6 ) = 64cisπ =-64
6
π
 w6=16 cis(6 ) = cis2π = cis0 = 1
3
z 6 26 π π
 3
= 3 cis(6 -3 ) =64(cis0)= 64
w 1 6 3
60π
 z60 = 260cis =260 cis10π =260 cis0 =260
6

17
TOPIC 1: NUMBER AND ALGEBRA Christos Nikolaidis

EXAMPLE 2

Find (1+i)10 (without using a GDC!)

Solution
It is more convenient to use the polar form of 1+i:

r = 1+1  2
π
tanθ = 1  θ =
4
Thus
π
1+i= 2 cis
4
Then
10π 5π
(1+i)10=( 2 )10 cis =25 cis
4 2
5π π
But =2π+ , thus
2 2
π
(1+i)10 = 25 cis =32i
2

EXAMPLE 3
Let z = cisθ = cosθ+isinθ.

We calculate z2 in two different ways:

 De Moivre’s theorem gives

z2 = cis2θ = cos2θ+isin2θ

 The identity (a+b)2 gives

z2 =(cosθ+isinθ)2

=cos2θ+2cosθ(isinθ)+(isinθ)2

=(cos2θ-sin2θ) +i(2sinθcosθ)

By comparing the two results we obtain the identities

cos2θ = cos2θ-sin2θ
sin2θ = 2sinθcosθ

18
TOPIC 1: NUMBER AND ALGEBRA Christos Nikolaidis

 GEOMETRICAL INTERPRETATION OF MULTIPLICATION

Let P be the point on the Complex plane which represents

w=2cisθ
- P is on a circle of radius 2 (centre at the origin)
- The line OP forms an angle θ with x-axis.

 If we multiply w by z=3, the result is zw=6cisθ.

The image point Q is on a circle of radius 6:

(enlargement by scale factor 3)

19
TOPIC 1: NUMBER AND ALGEBRA Christos Nikolaidis

 If we multiply w by z= cisφ, the result is zw=2cis(θ+φ).

The image line OQ forms an angle θ+φ with x-axis

(rotation by angle φ)

 If we multiply w by z=3cisφ, the result is zw=6cis(θ+φ).

Q is on a circle of radius 6; OQ forms an angle θ+φ with x-axis

(rotation by angle φ and enlargement by scale factor 3)

20
TOPIC 1: NUMBER AND ALGEBRA Christos Nikolaidis

1.11 ADDING SINUSOIDAL FUNCTIONS

In this section we use the polar form (in fact the Euler form) in
order to express

A1cos(x+θ1) + Α2cos(x+θ2) as a single function Αcos(x+θ)

A1sin(x+θ1) + Α2sin(x+θ2) as a single function Αsin(x+θ)

We use the fact that

Acos(x+θ) and Asin(x+θ)
are the real and the imaginary parts, respectively,
of the complex number
A[cos(x+θ)+isin(x+θ)], that is of Aei(x+θ)

For example, let

π
f(x)=4cosx and g(x)=3cos(x+ )
4
We want to express f(x)+g(x) in the form Acos(x+θ)

But f(x)=4cosx is the real part of 4eix

π
π i(x  )
g(x)=3cos(x+ ) is the real part of 3 e 4
4

Then f(x)+g(x) is the real part of the complex number

π π
i(x  ) i
4eix+3 e 4 = 4eix+3eix e 4

π
i
= eix(4+3 e 4 )
= eix(6.48e0.334i) [by GDC]
= 6.48ei(x+0.334)
Therefore,
f(x)+g(x) = 6.48cos(x+0.334)

Remark
Use your GDC to check that the graphs of the functions
π
y = 4cosx + 3cos(x+ )
4
y = 6.48cos(x+0.334)
coincide.

21
TOPIC 1: NUMBER AND ALGEBRA Christos Nikolaidis

In many applications in physics a sine wave (or sinusoid) is a curve

that describes a periodic oscillation (e.g. wind wave, sound wave,
light wave). It is a function of time (usually in seconds) of the form

Asin(ωt+φ) [or Acos(ωt+φ)]

where
A = amplitude
ω = angular frequency (ω=2πf, f=the number of cycles per second)
φ = phase

When we add sine waves of the same angular frequency we obtain

a sine curve of the same angular frequency.

Our previous example can be viewed as follows: for 2 waves

π
V1(t)=4cosωt, V2(t)=3cos(ωt+ )
4
the sum has the form
V(t)=6.48cos(ωt+0.334)

which is a wave of the same frequency.

EXAMPLE 1
Consider the waves of angular frequency 4
V1(t)=7sin(4t-1) and V2(t)=2sin(4t+3)
Express V1(t)+V2(t) in the form V(t)=Asin(4t+φ)
Solution
V1(t) is the imaginary part of z1=7ei(4t-1)
V2(t) is the imaginary part of z2=2ei(4t+3)
Then
z1+z2=7e4ite-i +2e4ite3i
= e4it (7e-i +2e3i)
= e4it(5.89e-1.26i) [by GDC]
= 5.89ei(4t-1.26)

But V(t) is the imaginary part of z1+z2, that is

V(t) = 5.89sin(4t-1.26)

22
TOPIC 1: NUMBER AND ALGEBRA Christos Nikolaidis

23
TOPIC 1: NUMBER AND ALGEBRA Christos Nikolaidis

24
TOPIC 1: NUMBER AND ALGEBRA Christos Nikolaidis

1.12 DEFINITION OF A MATRIX – MATRIX OPERATIONS

 A matrix is simply a rectangular array of numbers (called

elements). Some typical examples:
 4 8 
 
 1 2 7 4 6   2.5 4 
     0
 3 0 5   2 11  1/2 
 
 1 3.1 


A matrix is usually denoted by a capital letter, say

 4 8 
 
 1 2 7 4 6   2.5 4 
A=   B=   C= 
 3 0 5   2 11  0 1/2 
 
 1 3.1 

The first matrix above has 2 rows and 3 columns.

We say that A is a 2×3 (“two by three”) matrix, or otherwise that

the order of A is 2×3.

Likewise, B is a 2×2 matrix while C is a 4×2 matrix.

The general form of a 2×3 matrix is

 a11 a12 a13 

A=  
 a 21 a 22 a 23 

(notice that a23 for example is the element in row 2 and column 3)

 SQUARE MATRICES

No of rows = No of columns

The order of a square matrix is nxn, eg. 2×2, 3×3, 4×4 etc.

(in this case we may also say: “a square matrix of order n”.

For example,

25
TOPIC 1: NUMBER AND ALGEBRA Christos Nikolaidis

 a 11 a 12 a 13 
 4 6  
   a 21 a 22 a 23 
5 2 a a 32 a 33 
 31

are square matrices of order 2 and 3 respectively.

We also say that the elements a11, a22, a33 … (indicated above)
form the main diagonal of the square matrix.

 ROW MATRICES

Matrices of order 1×n

For example

A= 2 7 3 is a 1×3 row matrix

B=(4 2 6 -1 0 6) is a 1×6 row matrix

 COLUMN MATRICES (or VECTORS)

Matrices of order m×1

For example,

1
  - 1 
A=  4  is a 3×1 vector B=   is a 2×1 vector
2  3 
 

Notice: Matrices of order 1×1 are also defined, for example C=(5).

 THE ZERO MATRIX 0

All elements are 0

0 0 0
The 2×3 zero matrix is 0 =  
0 0 0
0 0
The 2×2 zero matrix is O =   and so on!
0 0

26
TOPIC 1: NUMBER AND ALGEBRA Christos Nikolaidis

 EQUAL MATRICES: A=B

A=B if  A and B have the same order

 the corresponding elements are equal

EXAMPLE 1

2 x  1 2 3  y 7  s t u
A=   B=   C=   D=  
 a 3 2 3 0  0 3 v w 7

It cannot be A=B since A is 2×2 while B is 2×3

A=C implies y=2, x=7, a=0

(3=3 holds anyway!)

B D since 0  7 (although both are 2×3)

Let A and B have the same order. We define some new matrices:

 THE SUM A+B

we simply add the corresponding elements

 THE DIFFERENCE A-B

we simply subtract the corresponding elements

 THE SCALAR PRODUCT nA (n is a scalar, i.e. a number)

we simply multiply each element of A by n

27
TOPIC 1: NUMBER AND ALGEBRA Christos Nikolaidis

EXAMPLE 2
1 2 3 3 5 0
A=   B=  
4 1 5  2 3 7 
Then

4 7 3  - 2 - 3 3   2 3 -3
A+B=   A-B=   B-A=  
 6 4 12   2 - 2 - 2 - 2 2 2 

 3 6 9   -3 -6 -9  -1 -2 -3
3A=   -3A=   -A=-1A=  
 12 3 15  -12 -3 -15 -4 -1 -5

1  1/2 1 3/2   0.5 1 1.5 

A=   or 0.5A=  
2  2 1/2 5/2   1 0.5 2.5 

Finally,

1 2 3 3 5 0
2A+3B = 2   +3  
4 1 5  2 3 7 
 2 4 6   9 15 0 
=   +  
 8 2 10   6 9 21 

 11 19 6 
=  
 14 11 31 

EXAMPLE 3

1 2  2 0 
Let A=   and B=   . Find the matrix X if
3 4 5 - 1

2A+X=3B

Method 1:

a b 
X must be of order 2×2. Suppose that X=  
 c d

Then

1 2  a b  2 0 
2A+X=3B  2   +   =3  
3 4   c d 5 - 1

28
TOPIC 1: NUMBER AND ALGEBRA Christos Nikolaidis

 2 4 a b   6 0 
   +   =  
 6 8   c d   15 - 3 
 2  a 4  b  6 0 
   =  
 6  c 8  d   15 - 3 

Hence,

2+a=6  a=4 4+b=0  b=-4

6+c=15  c=9 8+d=-3  d=-11

Therefore,

4 - 4 
X=  
 9 - 11 

Method 2:

We solve for X (as a usual equation)

2A+X=3B  X=3B-2A

2 0  1 2 
 X=3   -2  
5 - 1 3 4
 6 0  2 4
 X=   -  
 15 - 3   6 8 
4 - 4 
 X=  
 9 - 11 

The following operation is the most important one for matrices

 THE PRODUCT OF TWO MATRICES: AB

1 2   2 3
    = ?
3 4  5 1
 2 6
It is NOT   as someone would expect!! Here, we do not
 15 4 
multiply the corresponding elements.

29
TOPIC 1: NUMBER AND ALGEBRA Christos Nikolaidis

Let us start by a SPECIAL CASE. We multiply a row-matrix by a

column-matrix:

x
 
a b c  y  = (ax+by+cz)
 z
 

Notice that we multiply the corresponding elements and add the

results. For example,

3
 
2 3 5   4  = (2.3+3.4+5.1)=(23)
1
 

Now we are ready to multiply two matrices A and B in general.

First of all, the orders of A and B must be as follows:
A B
m × k k × n
That is,
number of columns of A = number of rows of B.

The order of the new matrix AB will be m × n

For example,

order of A order of B order of AB

3× 5 5×8 3×8

3× 2 2×1 3×1

2× 2 2×2 2×2

2× 3 2×3 not defined

The multiplication takes place as follows:

 we multiply rows of A by columns of B

 (row i) x (column j) will give the element aij

Ok, it seems complicated!!! Relax, take it easy!!!! Look at the

following description:

30
TOPIC 1: NUMBER AND ALGEBRA Christos Nikolaidis

Let
a x 
 1 2 3  
A=   B=  b y 
 4 5 6 c z
 
Notice that the order of AB is expected to be 2×2

 For the first row of AB:

multiply row 1 of A by each column of B separately:

 a x 
 1 2 3   1a  2b  3c 1x  2y 3z 
AB =    b y  =  
 4 5 6  c z     
 

 For the second row of AB:

multiply row 2 of A by each column of B separately:

 a x 
 1 2 3   1a 2b3c 1x 2y3z
AB =    b y  =  

 4 5 6  c z   4a 5b5c 4x 5y6z
 

EXAMPLE 4
2 5 
 1 2 3  
A=   B=  3 3 
 4 5 6 1 2
 
Then
2   5
1 2 3   =  11  1 2 3    17 
  3       3  =  
  1        
   2  
2   5
  =       
3  = 
  3   29    
 4 5 6  1     4 5 6  2  
 47 
  

In fact, we calculate all the resulting elements in one step:

2 5 
 1 2 3    11 17 
AB=    3 3  =  
 4 5 6   1 2   29 47 
 

31
TOPIC 1: NUMBER AND ALGEBRA Christos Nikolaidis

EXAMPLE 5

Consider again the matrices A and B above. Let us find the product
BA. Notice that the order is expected to be 3x3.

We obtain
2 5   2  20 4  25 6  30 
   1 2 3  
BA=  3 3    =  3  12 6  15 9  18 
 1 2   4 5 6   1  8 2  10 3  12 
   
 22 29 36 
 
=  15 21 27 
 9 12 15 
 

NOTICE: In general

AB  BA

They may be of different order, or perhaps only one the products

could be defined (eg. if A is a 2x3 and B is 3x5)

Even if both matrices A and B are square matrices, say 2x2, the
resulting 2x2 matrices AB and BA are not equal in general.

EXAMPLE 6

1 2  1 1
Let A=   and B=   . Then
3 4 0 2
1 2   1 1  1 5   1 1  1 2   4 6
AB=     =   while BA=     =  
 3 4   0 2   3 11   0 2  3 4   6 8

Sometimes though, it happens AB=BA. Then we say that matrices

A and B commute.

32
TOPIC 1: NUMBER AND ALGEBRA Christos Nikolaidis

EXAMPLE 7

1 2  5 4 
Let A=   and B=   . Then
3 4  6 11 
 1 2   5 4  17 26  5 4   1 2   17 26
AB=     =   and BA=     =  
 3 4   6 11   39 56  6 11   3 4   39 56

The following square matrix plays a key role in our theory.

 THE IDENTITY MATRIX I

1 in the main diagonal

0 elsewhere

Namely,
1 0 0 0
1 0 0  
1 0   0 1 0 0
I=   or I=  0 1 0  or I=  etc
0 1 0 0 1 0 0 1 0
   
0 0 0 1 


EXAMPLE 8
a b  1 0
Let A=   and I=   .
 c d 0 1
Then
a b   1 0 a b   1 0 a b  a b 
AI=     =   =A and IA=     =   =A
 c d 0 1  c d 0 1  c d  c d

In general, for any matrix A

AI=A and IA=A

(provided that the orders of A and I are appropriate!)

In other words, the identity matrix I plays the role of 1 (unit)

when we multiply matrices!

33
TOPIC 1: NUMBER AND ALGEBRA Christos Nikolaidis

 PROPERTIES OF MATRIX OPERATIONS

The following properties hold

For addition of matrices

A+B = B +A Commutative law

(A+B)+C = A+(B+C) Associative law

A+O = A = O+A O is the identity for addition

For multiplication of matrices

(AB)C = A(BC) Associative law

AI = A = IA I is the identity for multiplication

Remember that the commutative law does not hold for

multiplication, i.e. AB ≠ BA in general.

For addition and multiplication of matrices together

(A+B)C = AC+BC
Distributive laws
C(A+B) = CA + CB

For multiplication by scalars m,nR

m(A+B) = mA+mB
Distributive laws
(m+n)A = mA+nA

m(nA) = (mn)A

34
TOPIC 1: NUMBER AND ALGEBRA Christos Nikolaidis

1.13 THE DETERMINANT detA – THE INVERSE A-1

 2x2 DETERMINANT

a b 
Let A=   . The determinant of A is the number
 c d

detA= ad-bc
It is also denoted by

a b  a b
det   =ad-bc or =ad-bc
 c d c d

EXAMPLE 1

5 2  5 2
Let A=   . Then detA= = 5  4  2  3 =14
 3 4 3 4

2 - 4 2 -4
Let B=   . Then detB= =  12  12 =0
3 - 6 3 -6

EXAMPLE 2

Solve the equation

x -1
=0
2 x- 3
It is
x(x-3)+2=0  x2-3x+2=0  x=1 or x=2

 3x3 DETERMINANT
Let
 a1 a2 a3 
 
A=  b1 b2 b3  .
c c2 c3 
 1

In exams the 33 determinant, detA, will be asked only by GDC.

35
TOPIC 1: NUMBER AND ALGEBRA Christos Nikolaidis

Although it is not in the syllabus, just for information we mention

the definition of the determinant of A. It is the number

detA = a1b2c3 + a2b3c1 + a3b1c2 – a1b3c2 – a2b1c3 – a3b2c1

Ok, I know, it looks horrible!!!!

A more elegant way to estimate detA is

b2 b3 b b3 b b2
a1 - a2 1 + a3 1
c2 c3 c1 c3 c1 c2

 We multiply the elements of the first row, a1, a2, a3, by three
smaller determinants respectively.

 For a1, the corresponding 2x2 determinant can be obtained if

we eliminate the row and the column of a1

 a1 a2 a3 
 
 b1 b2 b3 
c c2 c3 
 1

Similarly for a2 and a3.

 We alternate the signs (+ - +)

EXAMPLE 3

 2 3 4
 
Let A=  5 6 7  . Then
1 2 8
 
6 7 5 7 5 6
detA = 2 -3 +4 = 2.34-3.33+4.4 = -15
2 8 1 8 1 2

Notice: The following terminology is also used

If detA=0 we say that the matrix A is singular.

If detA  0 we say that the matrix A is non-singular.

36
TOPIC 1: NUMBER AND ALGEBRA Christos Nikolaidis

Consider the following special cases:

 a x y a 0 0  a 0 0
     
A=  0 b z  B=  x b 0  C=  0 b 0 
0 0 c y z c  0 0 c 
     

Matrices like A are called upper-triangular. [0’s below main diagonal]

Matrices like B are called lower-triangular. [0’s above main diagonal]

Matrices like C are called diagonal. [0’s below & above main diagonal]

For the determinant of such a matrix we just multiply the

elements of the main diagonal.

detA=a.b.c detB= a.b.c detC= a.b.c

For example,
2 0 0 2 2 5
0 3 0  2  3  4  24 0 3 - 8  2  3  4  24
0 0 4 0 0 4

 THE INVERSE A-1 OF A 2X2 MATRIX

a b 
Let A=   . Given that detA  0 (and only then),
 c d

we define the inverse of A as follows

1  d - b
A-1 =  
detA - c a 

EXAMPLE 5

2 6
Let A=   . Then detA=2 and the inverse matrix is
1 4
1  4 - 6  2 - 3
A-1 =   =  
2 - 1 2  - 1/2 1 

2 8
Let B=   . Then detB=0 and hence, B-1 is not defined.
 1 4 

37
TOPIC 1: NUMBER AND ALGEBRA Christos Nikolaidis

To justify the term “inverse”, let’s multiply the matrices A and A-1
of the example above.

2 6  2 - 3  1 0 
AA-1=     =   = I
 1 4  - 1/2 1   0 1 
 2 - 3  2 6   1 0 
A-1A=     =   = I
- 1/2 1   1 4   0 1 

This is not accidental! In general

AA-1 = I and A-1A = I

NOTICE: Compare with numbers

NUMBERS MATRICES

The inverse of number a is a-1 The inverse of a matrix A is A-1

aa-1=1 and a-1a=1 AA-1=I and A-1A=I

Is any number invertible? Is any Matrix invertible?

NO. Only if a  0 NO. Only if detA  0

(if a=0, a-1 is not defined) (if detA=0, A-1 is not defined)

If a is invertible then If A is invertible then

1 1  d - b
a-1= A-1=  
a detA - c a 

NOTICE

If AB=I or BA=I we know that A is invertible and B is the inverse of

A, that is
A-1 = B

(B is also invertible and B-1 = A).

38
TOPIC 1: NUMBER AND ALGEBRA Christos Nikolaidis

 THE INVERSE A-1 OF A 3X3 MATRIX

The explicit formula for A-1 is out of our scope! It is enough to

know that

 A-1 exists only if detA  0

 AA-1=I and A-1A=I

 If AB=I or BA=I then B is the inverse of A

 A-1 can be found by GDC.

EXAMPLE 6

2 5 1 0 3 - 2 
   
Let A=  3 2 0  , B=  0 - 4 3 
4 3 0  1 14 - 11 
   

a) Find AB and BA

b) Find the inverse of A

c) Find the inverse of B

Solution

a) We can easily verify that AB=I and BA=I

b) Clearly A-1=B (the GDC also gives the same result)

c) Similarly, B-1=A (the GDC also gives the same result)

Notice that we cannot divide matrices, for example

B
is not defined.
A
However, we can multiply B by A-1, either as BA-1 or as A-1B,
according to the situation.

The following example will be characteristic:

39
TOPIC 1: NUMBER AND ALGEBRA Christos Nikolaidis

EXAMPLE 6

2 6  20 28
Let A=   and C=   . Find B given that AB=C
1 4  13 18 
Solution

Method 1: (analytical)

a b   2 6   a b   2a  6c 2b  6d 
Let B=   . Then AB=     =  
 c d  1 4   c d   a  4c b  4d 

Then AB=C implies

 2a  6c 2b  6d   20 28
 = 
 a  4c b  4d   13 18 
That is,
2a+6c=20 2b+6d=28
a+4c=13 b+4d=18

The first two equations give a=1, c=3

The second two equations give b=2, d=4.
Therefore,
1 2 
B=   .
3 4

A much more practical method is the following

Method 2: (solve for B)

 2 - 3
The matrix A is invertible with A-1 =   .
- 1/2 1 

Thus, we may multiply both parts of AB=C by A-1 on the left

AB=C  A-1AB=A-1C  IB= A-1C  B=A-1C

Hence
 2 - 3   20 28  1 2 
B=     =  
- 1/2 1   13 18   3 4 

40
TOPIC 1: NUMBER AND ALGEBRA Christos Nikolaidis

1.14 MATRIX EQUATIONS – THE LINEAR SYSTEM AX=B

 MATRIX EQUATIONS

An equation of matrices may involve an unknown matrix X.

Suppose that the matrices A, B and C are given.

Find the unknown matrix X in each of the following equations:

Matrix equation Solution

A+X=B X=B-A

2A+X=3B X=3B-2A
1
-3A+2X=5B 2X=3A+5B  X= (3A+5B)
2

AX=B X=A-1B (mind the order!)

XA=B X=BA-1 (mind the order!)

AXB=C X=A-1CB-1

AX+B=A AX=A-B  X=A-1(A-B)  X=I-A-1B

XA-B=XC XA-XC=B  X(A-C)=B  X=B(A-C)-1

AX+X=B (A+I)X=B  X=(A+I)-1B

(provided that the sizes of the matrices involved are appropriate)

Let us solve in detail some of the above equations:

EXAMPLE 7

2 5  3 4
Let A=   and B=   .
1 3 1 2 

Solve the equations

(a) -3A+2X=5B (b) AX =B (c) XA=B

(d) AX+B=A (e) AX+X=B

41
TOPIC 1: NUMBER AND ALGEBRA Christos Nikolaidis

Solution

(a) -3A+2X=5B  X = 1/2(3A+5B)

 21 35   21/2 35/2 
 X = 1/2   =  
 8 19   4 19/2 

(b) AX =B  X = A-1B

 3 - 5  3 4
 X =    
- 1 2   1 2 
 4 2
 X =  
- 1 0 

(c) XA =B  X = BA-1

3 4  3 - 5 
 X =    
 1 2  - 1 2 
5 -7 
 X =  
 1 - 1

(d) AX+B=A  X = I-A-1B

1 0  4 2
 X =   -  
0 1 - 1 0 
- 3 - 2 
 X =  
 1 1 

(e) AX+X=B  X = (A+I)-1B

1
3 5  3 4 
 X =    
1 4  1 2 
1 4 -5 3 4 
 X=    
7 - 1 3  1 2 
1  7 6   1 6/7 
 X=   = 
7  0 2   0 2/7 

42
TOPIC 1: NUMBER AND ALGEBRA Christos Nikolaidis

 SYSTEMS OF LINEAR EQUATIONS – THE EQUATION AX=B

Consider the 2×2 system (2 equations, 2 unknowns)

2x+3y=9
4x+7y=19

If we solve it in the traditional way (or by GDC) we will find:

x=3 and y=1 (check!)

We can describe this system by using matrices.

Let

 2 3
A=   the matrix of coefficients
4 7
x
X=   the vector of unknowns
 y
 9 
B=   the vector of constants
 19 

The equation of matrices AX=B is equivalent to

 2 3  x   9   2x  3y   9 
    =   , that is   =  
 4 7   y   19   4x  7y   19 

which gives in fact the system of linear equations above!

Consider now a 3×3 system:

5x+11y-21z = -22
x +2y -4z = -4
3x -2y +3z = 11
Again, if

 5 11 - 21  x - 22 
     
A=  1 2 4  X=  y  B=  - 4 
3  2 3   z  11 
    

the system can be written in the form AX=B.

43
TOPIC 1: NUMBER AND ALGEBRA Christos Nikolaidis

In general, any system of n linear equations and n unknowns (that

is n×n) can be expressed in the form

AX=B

Hence, if A is invertible (that is if detA  0) the solution is given by

X=A-1B

EXAMPLE 1

Consider the 2x2 system given above

2x+3y=9
4x+7y=19

which can be written in the form AX=B.

2 3
Since detA= =2  0,
4 7

1  7 - 3  7/2 - 3/2 
A-1=   =  
2 - 4 2   - 2 1 

The solution of the system is given by

7/2 - 3/2   9   3
X = A-1B =     =  
-2 1   19  1

In other words, x=3 and y=1.

EXAMPLE 2

Consider the 3x3 system given above

5x+11y-21z = -22
x +2y -4z = -4
3x -2y +3z = 11

which can be written in the form AX=B.

44
TOPIC 1: NUMBER AND ALGEBRA Christos Nikolaidis

The GDC gives

 2/7  9/7 2/7 

 
A-1 =  15/7  78/7 1/7 
 8/7  43/7 1/7 
 

The solution of the system is given by

 2/7  9/7 2/7  - 22   2

    
X = A-1B =  15/7  78/7 1/7   - 4  = - 1 
 8/7  43/7 1/7   11   1
    

EXAMPLE 3
0 3 - 2  2 5 1 1 1 1
       
Let A=  0 - 4 3 , A΄=  3 2 0  , B=  1  , C=  1 1 
 1 14 - 11  4 3 0 1 1 1
       

a) Find AA΄. What do you deduce?

b) Write down the system of three linear equations which

corresponds to the matrix equation AX=B

c) Solve the system of linear equations AX=B

d) Solve the matrix equation AX=C (this is not a system)

Solution
a) We easily obtain AA΄=I. So A and A΄ are inverse to each other.
A΄= A-1
b) 3y -2z = 1
-4y +3z = 1
x+14y-11z = 1
 2 5 1  1  8
     
c) AX=B  X=A-1B  X=  3 2 0  1   X=  5 
 4 3 0  1  7 
    
That is x=8, y=5, z=7
 2 5 1  1 1 8 8
    
d) AX=C  X=A-1C  X=  3 2 0  1 1  X= 5 5
 4 3 0  1 1 7 7
    

45
TOPIC 1: NUMBER AND ALGEBRA Christos Nikolaidis

Notice though that if A is not invertible (A-1 does not exist) the
system AX=B cannot be solved in this way. In general,

for the system AX=B

detA  0
 UNIQUE SOLUTION X=A-1B
(A-1 exists)

detA=0  NO SOLUTION, or
(A-1 doesn’t exist)
 INFINITELY MANY () SOLUTIONS

REMARK. Compare with numbers. For the equation ax=b

a 0
b
1  Unique solution x= .
(a-1= exists) a
a
 No solution
a=0 (e.g. 0x=5 has no solution)
(a-1 doesn’t exist)  Infinitely many solutions
(e.g. 0x=0, true for any xєR)

In IB exams, A will always be invertible (detA≠0, unique solution).

However, it is worth to see two examples of systems of linear

equations, where detA=0

EXAMPLE 4
Consider the systems
(a) x+2y=1 (b) x+2y=1
2x+4y=5 2x+4y=2

1 2
For both detA= =0,
2 4
so the systems have either no solution or an  number of solutions.

46
TOPIC 1: NUMBER AND ALGEBRA Christos Nikolaidis

We multiply the first solution by 2 and obtain 2x+4y=2. Hence the

two systems take the equivalent form
(a) 2x+4y=2 (b) 2x+4y=2
2x+4y=5 2x+4y=2

 System (a) has no solution (impossible) [check also by GDC]

 System (b) reduces to just one equation: x+2y=1.

There are infinitely many solutions (all points of the line x+2y=1)

If we solve for x: x=1-2y,

y  R (free variable). [check also by GDC]

x
In fact, for each value of y we obtain a different solution   :
 y
1 -1  -3 
for y=0:   , for y=1:   , for y=2:   etc
0 1 2

For 3x3 systems with detA=0, the GDC gives the answer.

EXAMPLE 5
Consider the systems
(a) 2x +3y +3z = 3 (b) 2x +3y +3z = 3
x + y -2z = 4 x + y -2z = 4
5x +7y +4z = 5 5x +7y +4z = 10
2 3 3
For both systems detA  1 2 - 2 =0
5 7 4
The GDC gives
(a) the system has no solution.
(b) the system has infinitely many solutions
x = 14 + 16z
y = -5 + 7z
z  R (free variable)

47
TOPIC 1: NUMBER AND ALGEBRA Christos Nikolaidis

1.15 TRANSFORMATION MATRICES

Any 2×2 matrix M maps a point P(x,y) to another point P΄(x΄,y΄).

x
For our purpose we express the point (x,y) as a vector   .
 y
a b 
Let M=   . Then
 c d
 a b   x   ax + by 
    = 
 c d   y   cx + dy 

We call
x  = ax + by

y = cx + dy

x  x 
In this way, M maps   to  
 y  y 
1 2 1
For example, if M=   , for the vector   , we obtain
3 4 1
 1 2   1  3
    = 
 3 4   1  7 

Thus, the matrix M maps the point A(1,1) to the point A΄(3,7).

In general,

 1 2   x   x + 2y   x    x + 2y 
    =  i.e.   =  
 3 4   y   3x + 4y   y   3x + 4y 

In this way we can find the image of several points. M maps

0 0 1 3 -1  3

 ֏  ,  ֏  ,  ֏  , etc.
0 0 1 7  2 5 

or, if we express the points in the traditional coordinate-form,

O(0,0) to O΄(0,0),

A(1,1) to A΄(3,7),

B(1,2) to B΄(3,5), etc.

48
TOPIC 1: NUMBER AND ALGEBRA Christos Nikolaidis

This kind of mapping is known as linear transformation. A linear

transformation maps, not only points to points but a whole shape
to a new shape, e.g. a line to a line, a triangle to a triangle, etc.

For example, the transformation above maps

the straight line passing through O(0,0) and A(1,1)

to the straight line passing through O΄(0,0) and A΄(3,7).

It also maps

the triangle OAB with vertices O(0,0), A(1,1) and B(-1,2)

to the triangle O΄A΄B΄ with vertices O΄(0,0), A΄(3,7) and B΄(3,5).

49
TOPIC 1: NUMBER AND ALGEBRA Christos Nikolaidis

 BASIC TRANSFORMATION MATRICES

 k 0 horizontal stretch
 
0 1 with a scale factor of k

1 0 vertical stretch
 
0 k  with a scale factor of k

 k 0 enlargement,
 
0 k  with a scale factor of k, centre (0,0)

 cosθ -sinθ  anticlockwise rotation

 
 sinθ cosθ  of angle θ about the origin

 cosθ sinθ  clockwise rotation

 
-sinθ cosθ  of angle θ about the origin

 cos2θ sin2θ  reflection

 
 sin2θ -cos2θ  in the line y = mx where m=tanθ

EXAMPLE 1
3 0 x  3x 
The matrix A=   maps  y  to  y  . Indeed,
0 1    

 3 0   x   3x 
    = 
 0 1   y  y 

Thus, it stretches each point horizontally with a scale factor of 3.

It also stretches any shape horizontally with a scale factor of 3.
For example, it maps
the triangle OAB with vertices O(0,0), A(2,0), B(1,1)
to the triangle O΄A΄B΄ with vertices O΄(0,0), A΄(6,0), B΄(3,1)

50
TOPIC 1: NUMBER AND ALGEBRA Christos Nikolaidis

EXAMPLE 2
Consider again the triangle OAB with O(0,0), A(2,0), B(1,1).

The image of the triangle under the following transformations is

shown in the 3rd column.

1 0 vertical stretch
 
0 2 with a scale factor of 2

enlargement,
2 0
  with a scale factor of 2,
0 2
centre (0,0)

clockwise rotation
 0 1
  of angle θ=9Ο°
-1 0 
about the origin

reflection
0 1
  in the line y = x
1 0
(m=tan45°=1)

51
TOPIC 1: NUMBER AND ALGEBRA Christos Nikolaidis

 COMPOSITION OF TRANSFORMATIONS
x
If vector   is transformed
 y
by matrix A,
then by matrix B,
then by matrix C,
we obtain
x x x x
  ֏ A   ֏ BA   ֏ CBA  
 y  y  y  y

Hence, the composition the 3 transformations can be achieved by

the product of the 3 matrices (in the reverse order).

EXAMPLE 3
x
Let   be transformed by
 y
3 0
a horizontal stretch with a scale factor of 3: A=  
0 1
0 1
and a clockwise rotation of angle θ=9Ο°: B=  
-1 0

x  3 0   x   3x 
Then   ֏    = 
 y  0 1   y  y 
 3x   0 1   3x   y 
and then  ֏  = 
 y  -1 0   y  -3x 

We obtain the same result by applying the transformation matrix

 0 1 3 0  0 1
BA=    = 
-1 0   0 1  -3 0 
Indeed,
 0 1 x  y 
    = .
-3 0   y  -3x 

In general, the composition of n transformations with matrices

A1, A2, …, An
can be achieved by the product of matrices An ⋯ A2A1

52
TOPIC 1: NUMBER AND ALGEBRA Christos Nikolaidis

 AN INTERESTING PROPERTY OF THE DETERMINANT.

Suppose that a transformation matrix M maps a closed shape S to
a new shape S΄. Then
(Area of S΄) = |detM| × (Area of S)

EXAMPLE 4
In Example 2, the original triangle OAB has area 1.

vertical stretch

detM=2
1 0
M=  
0 2 Area of image = 2

enlargement

detM=4
2 0
M=  
0 2 Area of image = 4

clockwise rotation 9Ο°

detM=1
 0 1
M=  
-1 0  Area of image = 1

Reflection in y = x

detM=-1
0 1
M=  
1 0 Area of image = 1

53
TOPIC 1: NUMBER AND ALGEBRA Christos Nikolaidis

 AFFINE TRANSFORMATIONS
We can extend the linear transformations described in this
e
paragraph by also adding a translation vector   :
f 
 a b   x   e   ax + by   e   ax + by + e 
    +  = + =  
 c d   y   f   cx + dy   f   cx + dy + f 

a b 
If   is the identity matrix, we simply have a translation
 c d
 x  e  x + e
  + = 
 y f   y+f 
These transformations are known as affine transformations.

EXAMPLE 5
The transformation
1 0  x   3  x   3  x +3
    +  =  +  = 
 0 1   y 5   y 5   y+5 
translates each point (x,y)
horizontally by 3 units to the right
vertically by 5 units up

For example, the image of the point (1,2) is the point (4,7).

EXAMPLE 6
The transformation
 1 2   x   5   x + 2y + 5 
    +  = 
 3 4   y   6   3x + 4y +6 
maps
0 5  1 6 1  8 
  ֏  ,   ֏  ,  ֏ , etc.
0 6 0 9  1  13 

54
TOPIC 1: NUMBER AND ALGEBRA Christos Nikolaidis

1.16 EIGENVALUES AND EIGENVECTORS

Consider
1 3
A   
4 2
3
Observe what happens when we multiply Α by vector u    .
4
 1 3   3   15  3
Au          5    5u
 4 2   4   20  4

We obtain a multiple of u. In such a case we say that 5 is an

eigenvalue of the matrix A and u is a corresponding eigenvector.
Our task here is to find all those characteristic vectors.

Let A be a square nn matrix. The value λ  R is said to be an

eigenvalue of Α if there exists a non-zero vector u such that

Au  λu

Such a vector is called eigenvector (corresponding to eigenvalue λ).

 FINDING EIGENVALUES AND EIGENVECTORS

Notice that

λ is an eigenvalue of Α  Au  λu for some u  0

 Au  λu  0 for some u  0
 (A  λI)u  0 for some u  0
 the system (A  λI)X = 0 has a nonzero solution
 A  λI has no inverse
 det(A  λI)  0

a b 
For a 22 matrix A    , the expression det(A  λI) is
 c d
aλ b
= (a-λ)(d-λ)-bc=0 = λ2-(a+d)λ+ ad-bc
c dλ

This is known as the characteristic polynomial of matrix A.

Notice that the constant term is detA.

55
TOPIC 1: NUMBER AND ALGEBRA Christos Nikolaidis

In practice:
 We find the eigenvalues by solving det(A  λI)  0
 For each eigenvalue λ, we solve the corresponding system
(A  λI)X =0
to find the eigenvectors.

The following example will clarify the process:

EXAMPLE 1
1 3
Consider again A   
4 2
1λ 3
det(A  λI) = = λ 2  3λ  10
4 2λ

There are two eigenvalues: λ  5 , λ  2 .

-4 3
 For λ  5 , the determinant becomes . Solve the system
4 -3
- 4x  3y  0
4x  3y  0

In fact, it is enough to use the 1st equation to find the ratio x/y:
x 3
- 4x  3y  0  4x  3y  
y 4
The corresponding eigenvectors are

 x  3 3
     t [i.e.   as well as all its multiples)]
 y  4   4

3 3
 For λ  2 , the determinant becomes . Solve the system
4 4
3x  3y  0
4x  4y  0

Again we use the 1st equation to find the ratio x/y:

x
3x  3y  0  x  y  0  x  -y   -1
y
The corresponding eigenvectors are
 x  - 1  -1 
     t [i.e.   as well as all its multiples)]
 y  1  1

56
TOPIC 1: NUMBER AND ALGEBRA Christos Nikolaidis

NOTICE

The characteristic polynomial for 22 matrices is a quadratic so it

has either two or one or no real eigenvalues. For example,

3  1
 The matrix A    has only one eigenvalue, λ  4 (check!)
1 5 
So there is only one eigenvector (together with all its multiples).
 x  - 1 
     t
 y  1 
- 1  1 
 The matrix A    has no real eigenvalues (check!).
 2 1 
So there are no eigenvectors with real coefficients.

These two cases are out of the scope of this course. We only
consider matrices with two distinct eigenvalues.

 DIAGONALISATION
For two matrices A and B we say that B is similar to A if

B  P 1 AP

for some invertible matrix P. We write B ∼ A

NOTICE
 A ∼ B: any matrix is similar to itself, since A  I 1 AI

 B ∼ A  A ∼ B: if B is similar to A, then A is similar to B, since

B  P 1 AP  A  PBP 1 [notice that P  (P 1 )1 ]

Thus we can simply say that A and B are similar.

 C ∼ B and B ∼ A  C ∼ A: if C is similar to B and B is similar to A,

then C is similar to A. Indeed,

C  Q1BQ and B  P 1 AP implies

C  Q 1 (P 1 AP)Q  (PQ) 1 A(PQ) [notice that (PQ)1  Q1P 1 ]

57
TOPIC 1: NUMBER AND ALGEBRA Christos Nikolaidis

Any 22 matrix with two distinct eigenvalues λ1 and λ2 is similar

to the diagonal matrix
λ 0
D   1 
0 λ 2 
We are going to explain the process of diagonalization by using the
example 1.

The matrix
1 3
A   
4 2

has two eigenvalues λ  5 , λ  2 with corresponding eigenvectors

3 - 1 
  t and   t
4  1 
Let
5 0   3  1
D    and P   
0  2 4 1 
Then

P 1 AP = D

Indeed, by using a GDC we can verify that

 1 1   1 3   3  1 5 0 
P 1 AP   7 7     
 4 3   4 2   4 1   0  2 
 7 7

NOTICE.
We could also consider the eigenvalues in the opposite order.
Then
- 2 0  - 1 3 
D    and P   
 0 5  1 4

and finally

- 2 0 
P 1 AP   
 0 5

58
TOPIC 1: NUMBER AND ALGEBRA Christos Nikolaidis

A nice application of the diagonalization is derived by the following

observation:
P 1 AP = D implies A  PDP 1
Then
A n  (PDP 1 ) (PDP 1 ) ⋯ (PDP 1 )  PD n P 1

But for a diagonal matrix D, it is easy to find the power D n :

λ n 0 
Dn   1
 0 λ 2 
n


EXAMPLE 2
1 3
For A    , find an expression of A n in terms of n.
 4 2 

Diagonalization gave us

A  PDP 1
where
5 0   3  1 1  1 1
D    , P    , P 1   
0  2 4 1  7   4 3 
Then

A n  PD n P 1

 3  1 5 n 0  1  1 1
      
4 1   0 ( 2) n  7   4 3 

1 35 n  ( 2) n   1 1
    
7  4  5 n ( 2) n    4 3

1  3  5 n  4  (2) n 3  5 n  3  ( 2) n 
  
7  4  5 n  4  ( 2) n 4  5 n  3  ( 2) n 
For example

 277 261   836761 837273 

A 4    and A 9   
 348 364   1116364 1115852 

(check by your GDC!).

59
TOPIC 1: NUMBER AND ALGEBRA Christos Nikolaidis

 A NICE PROPERTY OF THE CHARACTERISTIC POLYNOMIAL

Let us close this section by a surprising property of the

characteristic polynomial. We have seen that the characteristic
polynomial of the matrix
1 3
A   
4 2

is
1λ 3
det(A  λI) = = λ 2  3λ  10
4 2λ
A well-known result (known as Cayley-Hamilton theorem) says
that the matrix A satisfies the characteristic equation
A 2  3A  10 I  O .
(where I is the identity matrix).
Indeed,
2
1 3 1 3 1 0
A 2  3A  10 I     3   10 
 4 2   4 2  0 1

 13 9   3 9   10 0  0 0
         =  
 12 16   12 6   0 10  0 0

This relation helps us to express, the inverse A-1 as well as the

powers An in terms of A:
1
A 2  3A  10 I  O  A 2  3A  10 I  I  (A 2  3A)
10

If we multiply by A-1 we obtain

1
A 1  (A  3I)
10

Also, we can express A2, A3, A3 in terms of A and I as follows:

A 2  3A  10 I  O  A 2  3A  10I

A 3  A 2 A  (3A  10I)A  3A 2  10A  3(3A  10I)  10A  19A  30I

A 4  A 3 A  (19A  30I)A  19A 2  30A  19(3A  10I)  30A  87A  190I

etc.

60