MM ToXe

lS Taye m eam aat: Lea (e-s-e

storing and analysing data Is it possible to create a snowflake using mathematical formulae?
The Koch snowflake is an attempt to describe the patterns like those .
found in an ice crystal. p 4 us
When an ice crystal is magnified we see that smaller versions of the
Concepts S ice crystal can be found at each level of magnification. This property
m Systems is called self-similarity and structures that display it are called fractals.
m Modelling By identifying and describing the smallest piece in this pattern with
a mathematical process we can create the entire structure simply by
Microconcepts carrying out this process repeatedly. . g™
Matrix Each stage of the Koch snowflake is described by the iterative process
Array shown.
Order of matrix Stage 0
We begin by dividing the line segment in stage 0 into thirds and
Element creating an equilateral triangle in the middle third of the segment.
Dimensions of matrix e Write down step-by-step instructions describing geometrically how you _/A\E Stage 1
Eigenvalue could use the shape in stage 1 to create the shapes in stages 2, 3 and 4.
Eigenvector e Construct graphs of stages 1 and 2 of the Koch snowflake on
Fractal
Stage 2
square millimetre graph paper using the scale 1 unit = 30mm
Determinant like the one shown. Determine the exact coordinates of each y
Inverse vertex and endpoint for each figure.
Identity e Is there a pattern in the values of the coordinates of each of these 1
Linear system points that you could use to determine the coordinates of the o4
How can urban planners determine the points on stage 3?
Transformations
population of different cities if people te
Iterative function systems Save your work as you will need the results to check your answers 1 2 3
are constantly moving between them? el to the investigation at the end of the unit. y
1
How can athletes ensure Developing inquiry skills
their diet provides all the
Where else do fractals appear in the real world? 0
nutrients they need?

edd
Can we use fractal formulae to model real-life situations? How might x
these be useful? 7 2 3
Think about the questions in this opening problem and answer any

iu
you can. As you work through the chapter, you will gain mathematical
knowledge and skills that will help you to answer them all.

Veal
Before you start
Click here for help
You should know how to: Skills check with this skills check

1 Identify and write the general term of a 1 Find a formula for the nth term in the
geometric sequence.
sequence: {18 -12, 8,- 16...)
eg The general term of 4, -8, 16, —32, ...
isu, =4x (-2)"1 2 Determine the sum of the first 15 terms of
2 Find the sum of geometric series. each sequence in question 2.
eg The sum of 8 terms of the series 3 Write an expression for S, for the series
12+6+341.5
+... is
1 1y (ay +++ then use it to
12(1-0.5*) o+o(t}+6{2) (3)
S, =————_ = 23.9. 2 2 2
1-0.5
find S,,-

372 373
hi
 ea eer Cea. Er ING DATA 9.1
9.1 Introduction to matrices and Matrix addition
Three models of a particular brand of television undergo two
matrix operations

BEL Ty
manufacturing processes each carried out in different factories. The
elements of matrices A and B represent the shipping and manufacturing
Computers store large amounts of data using a programming structure costs (in USD) for each of the three models at each of two factories.
called an array. Understanding how data is stored and manipulated

Coty
leads you to a study of matrices and matrix operations. Manufacturing Shipping Manufacturing Shipping
125 35 Model 1 300 50 Model 1
AmatrixA with dimensions m X nis a rectangular array of real (or complex) International- A= 275 40 IModel2 and B= 210 35 \Model 2
numbers containing 7 rows and 1 columns where 4; refers to the mindedness 180 55 Model 3 325 65 Model 3
element located in row 7 and column j. Such a matrix is said to have English mathematician
orderm Xn. and lawyer, James Adding a,, + b, , = 425 indicates that the total manufacturing cost of
Column 1 Column 2 Column n Sylvester, introduced Model lis $425. Similarly, adding a, , + b, , = 120 and indicates that
L | od the term “matrix” in the
19th century and his
the total shipping cost of Model 3 is $120.
441 44,2 + An \— owt friend, Arthur Cayley,
a0),
4214 42,2 + 82,9 +f— row2 advanced the algebraic
A=-| - A ; “There is no branch of
aspect of matrices. Thus, 125 30 300 50 425
&8Qo

LNT E NN
A+B=|275 40/+/210 35/=|485 mathematics, however
180 55 325 65 505 3
m1 Am,2 == &m,n me towm abstract, which may not aEa
someday be applied I
to phenomena of the »J

LPPEL
a
To add or subtract two or more matrices, they must be of the same order. You real world.”— Nikolai
-1 3 4 2) add or subtract corresponding elements. Lobatchevsky
For example, matrix P=} 1 5 0 7 has dimensions 3 x 4 where
Where does the power
-3 4 -2 6) of mathematics come
The zero matrix 0 is the matrix whose entries are all zero.
a, ,is the element located in the 2nd row and 3rd column and has a from? Is it from its
0 0°
value of 0. Note that matrix P contains 3 x 4 = 12 elements. For 2 x 2 matrices is the zero matrix, for 2 x 3 matrices it is ability to communicate
, as a language, from the
If m=n (ie the number of columns is equal to the number of rows)
0 0 0) axiomatic proofs or from
then we say that P is a square matrix of order n. its abstract nature?
0 0 0;
Consider the 3 x 5 matrix C showing the cost (in euros) of manufac-
turing 3 different products for 5 different manufacturers. For any matrix A, A+0=AandA-0O=A.

Manufactuer 1 Manufactuer2 Manufactuer 3 Manufactuer4 Manufactuer 5

Scalar multiplication for matrices
Product 1 3.50 4.25 3.82 4.10 3.95
Consider the situation where manufacturer A increases both its
c= Product 2 8.25 10.50 75 9.45 7.80 shipping and manufacturing costs by 10%. That is, each new cost is
Product 3 10.50 9.75 8.95 11.00 10.80 110% of its old cost. Therefore, multiplying each value by 1.10 gives a
e matrix representing the new costs.
For example, the cost of producing product 3 with manufacturer 125(1.10) 35(1.10)) (137.50 38.50>
4 is given by element a, , and is €11.00.
1.10A =| 275(1.10) 40(1.10) |=| 302.50 44 HINT
180(1.10) 55(1.10) 198 60.5,
Ascalarisa
In this way, we define scalar multiplication of matrices. quantity that is one-
Matrices Aand Bare equal if and only if the matrices have the same
dimensional. Some
dimensions and their corresponding elements are equal.
examples of scalars
Given a matrix A and a real number k then KA is obtained by multiplying are temperature and
every element of A by k where kis referred to as a scalar. weight.

374 375
 Dee Ue eh ee ee ING DATA 9 ° 2

32450» a Determine the points:

cent
Find aand bit2P-50=0, P=(_) 1 2b)
: Jana a= (4a -1>
Oa 4 Thie matrixix PP=|= 1872?
34175 describes
lescribesthethe prices
pri i(xly efg

‘ 19250, 3 Hf
1 2b a -l 0 0) The zero matrix is a matrix all of whose (in USD) of four types of cars. The sales tax ili, —X. 2
2 3c 1) -1 l-1 04)}- lo 0 elements are 0. rate is 7.5% of the purchase price. %
0. i S
/ a Write down T= kP where T represents b show each of these the points on a ry
2-5a 4b+5) (0 0 the amount of tax paid for each car. te . . | b i ;
6c+5 o | lo o b The matrix C represents the cost of . a rtedoec oh by a ne See
each car including tax. Write a matrix P y KA.
2 equation that describes C in terms of P. 6 Find real values of y and A if
2-5a=O0>a=— If two matrices are equal then their
-8 12 -9 12
5 corresponding elements are equal. 5A column vector can be regarded asa 2 * 1 uMA+AB= where A=
-5 matrix. The point P = (-3,6) is written in - 4 -6 3
4b+5=0>b=— 3) _»y
4 matrix form as X-( k and B= (; 2 :

6c+5=0¢=—2 ° 3
6
rate

‘ er es
9.2 Matrix multiplication and 23
Er
et eae @) properties PE
e oO

1 Given that 3 Consider the matrices R, S, T, U, V, Consider a matrix F whose elements represent the number of each of
3 W, and X below. three types of medium size drinks (frappé, cappuccino, and iced coffee)
a (? 3 | B=|-4 0 -1 3 2 1 sold at a café on each day of the first week in June.
4 2 6} l = =
> R=) 3 ; S= ‘ : > Frappé Cappuccino Iced coffee
8 5 2) > - 28 32 16
and C=|-4 0 3, write down the values 3-9 8 -3 -2 Sun 34 51 8
1) -l 7y r=() 3 U=|6 7 2 Mon 40 31
o 1 -l Tues 15
of a,» 4,5, b, 1, c,,andc, |.
—- : W= Wed 37 24 20
2 The matrix Q describes the number of units v= -2 0 _ 1-2 3
of each of three products that are produced -4 -8 3 2 -1 Thurs} 75 47 29
by each of four manufacturers where 4: > Fri 38 29 19
represents the amount of product i produced xX-( 7 = } Sat
by manufacturer j. -6 4 47 34 12
420 250 145 0 ) Find each of the following if possible. If not The price for a frappe is €4.00, cappuccino is €3.50 and iced coffee is
Q=|340 575 420 100 possible, explain why. Check your answers €2.50.
using technology.
200 375 425 235, a ow . R-wW © 4U+s The following calculation can be used to determine the revenue R for
a Write down the dimensions of Q. each day:
b Write down the value of In4 and describe d Lp ly da revenue = # frappes < price + # cappuccinos x price + #iced ; oan
its meaning. 3 2 9 ~ sold P sold P coffee P
c Sake 4 on maths T representing tne total Sunday $264 = 28 x 400 + 32 x 3.50 + 16 x 2.50
number of products produced by ea
manufacturer.

376 Sig
 MODELLING WITH MATRICES: STORING AND ANA' ING DATA 9.2
Similarly, the revenue for each day is obtained by summing
the products of the number of each type of drink sold and the Investigation 1
| ; 2 4) 2 -1 3° i

BEL Ty
corresponding prices for each drink on each day of the week. That is,
Consider the matricesA = 1a } B= c= ,
28(4.00) + 32(3.50) + 16(2.50) 264 » 0 3 4) 2,
34(4.00) + 51(3.50) + 8(2.50) 334.50 1 2 -2 -2
40(4.00) + 31(3.50) + 15(2.50) 306 D=|0 3 1 |andE=| 1

Coty
R =| 37(4.00)+ 24(3.50)+ 20(2.50) |=} 282 -1 -3 1 -1
75(4.00) + 47(3.50) + 29(2.50) 537 1 (FETED Find the products of AB, AC, BD, DE, and BE.
38(4.00) + 29(3.50) +19(2.50) 301
47(4.00) + 34(3.50) + 12(2.50) 337 , 2 Complete the table below to determine the dimensions (order) of the
products AB, AC, BD, DE, and BE.
A more compact way of expressing R would be to represent the price
Dimensions of first Dimensionsof Dimensions of
4.00» Product matrix second matrix — product
of each drink as a 3 x 1 matrix P =| 3.50 . AB 2x2 2x3
2.50, AC |
BD
28 32 16 64 Sunday’s revenue = 28(4.00)
+ 32(3.50)+16(2.50)
DE &8Qo

LNT E NN
34 51 8 334.50 4.00 3
BE aEa
40 31 15 |(4.00 306 =(28 32 16)| 3.50 |= 264 I
so that R=WP=/37 24 20] 3.50 |=| 282 2.50 »J

LPPEL
3 (EETEI Based upon the results in question 1 explain how you could a
75 47 29|(2.50 537 determine the dimensions of the product of two matrices using the
38 29 19 G01) Friday's revenue = 38(4.00) + 29(3.50)+19(2.50) dimensions of the two matrices being multiplied?

47 34 12 337, 4.00 4 Explain why the products BA, CB, and DB cannot be determined.

=(38 29 19)| 3.50 |=301 5 Cir When does the product of two matrices exist? TOK
2.50 How are mathematical
definitions different
If A has dimensions m x n and B has dimensionsp x q then from definitions in other
If P= AB then each element ofP (named asP,) is found by summing the
e the product AB is defined only if the number of columns in A is equal to areas of knowledge?
products of the elements in row of A with the elements in columnj of B. the number of rows in B How are mathematical
definitions different
(thatis, 2 =p)
from properties,
-1 2° e and when the product does exist, the dimensions of the product is m x q. axioms, or theorems?
A 1 5 -3
For example, ta-(i) 4 » Jane B= -3 4 then
21,
> —3
-1 2 Example 2
= 364
C. 4 2 A diet research project consists of adults and children of both sexes. The number of
2 1 participants in the survey is given by the matrix:
_ (Me) +5(-3)-3(2) (2) +5(4)-3(1))_ (-22. 19 Adults — Children
© (=1(-1)
+ 4(-3) + 2(2) -1(2)
+ 4(4) + 20) ) -7 ret _ ( 75 180 } Male
“(110-250 JFemale
Continued on next page
oO

378 379
 ea eer Cea. Er ING DATA 9.2
Orne number of daily grams of protein, fat, and carbohydrates consumed by each child and Properties of multiplication for matrices i) Fa
adult is given by the matrix: For matrices A, B, and C: 3
Protein Fat Carbohydrate ¢ Non-commutative AB # BA s
(15 20 25 Adult © Associative property A(BC) = (AB)C 2
“(8 16 20 Jchild © Distributive property A(B + C)=AB + ACand (B+ C)A=BA+ CA
©
a Determine AB. b Explain the meaning of AB,, c¢ Explain why BA does not exist. ~
a
i)
75 180\(15 20 25 | Multiply the matrices. These properties only hold when the products are defined.
a AB=
110 250). 8 16 20 e InA(B + C), B + Cis pre-multiplied by A.
e¢ In(B+C)A, B + Cis post-multiplied by A.
_ (7515) +180(8) —75(20)+180(16) _75(25) +180(20) »
(oe eee 110(20) +250(16) 110(25)+250(20),
2565 4380 5475 Multiplicative identity for matrices
The multiplicative identity for real numbers is 1 since ax 1=1xa=a
ea 6200 ne
for any real number a.
b AB,, = 110(25) + 250(20) = 7750 and represents the total Total carbohydrates If Ais any square matrix and Tis the identity matrix, then A XI=Ix A=A.
number of carbohydrates consumed by females in the project. consumed by the In other words, if any square matrix is pre- or post-multiplied by the
110 adult females identity matrix, then the answer is the original matrix. cay
FE
c BA is not possible since the number of columns in B is not
is 110(25) and the
equal to the number of rows in A.
total carbohydrates rs}
The multiplicative identity ofa square 1 X 1 matrix
A is given by 3 &
consumed by 250 child
y arn
females is 250(20). 10-0 @2
[0 1 0

00 1,
Note that both A xI, =A and, x A=A hold only in the case where A is a
Investigation 2 square matrix.
; ee -1 5% —2 1% . Unless needed for clarity, 7 is not normally written, and Tis used alone to
Use the matrices A = ,B= and C= to answer each question. denote the identity matrix, whatever the size.
4 1) -2, 3-4,
1 Multiplication of real numbers a and b is commutative since ab = ba, for example 5 x 4=4 x 5=20.
Determine AB and BA. Is matrix multiplication commutative? rcise 9B
Multiplication of real numbers is distributive since a(b + c) = ab + acand (b +c)a=ab +ac.
1 Using the matrices 2 Given that Pis a 4 x m matrix, Qisa
2 Determine each of the following then describe your observations.
) 2 2 41 2 xn matrix, determine the values of m
i A(B+C) ii (B+Q)A iii AB+AC iv BA+CA a-() ?} “(2 0 9) and n if PQ is a 4 x 3 matrix.
3 Based upon your observations is matrix multiplication distributive? Is A(B + C) = AB + BC? Isit
1 0 1-2 -1 3 Find a matrix that has the effect of
correct to write (B + C)A=AB + CA?
c-|-1 2| p-|3 0 -4 summing the entries in every row ofa
4 Multiplication of real numbers is associative which means that a(bc) = (ab)c. Is matrix multiplication
associative? You can answer this by showing whether or not A(BC) = (AB)C. -2 3 3201 aS
3x3 matrix. Thatis,if A=|d e f | find
\ 2 3 -l ghi
5 Consider the marix I= [ : E=| 1), and F=|2 44,
d 3 5-2 (a+b+c) \
a Find AIandIA b Write down what you notice.
find each of the following matrices if possible. a matrix B such that AB = (a+ Ge f)
6 Gin How do matrix and matrix multiplication work in terms of the commutative and If it is not possible, state the reason. (g +h+ i)
associative properties?
a (C+F)A b DE c ABC
d CF e 2A-3I,
380 381
 ea eer Cea. ANALYSING DATA 9.3
4 . 3 2 «1% New federal laws require that the manu-
Exercise 9C
4 Consider the matrix A= facturer reduce its daily emissions of carbon
-1 5 -6,

BEL Ty
monoxide, sulfur dioxide, and nitric oxide
a Find a matrix B such that AB= A. Complete each exercise using the definitions In algebra, (a + b)(a — b) = a? — b? and
by 60%, 20%, and 40% of its current
and properties you learned in this section then (a + b)? = a? + 2ab + b? where a, be R.
b Find a matrix C such that CA = A. levels. The manufacturer takes corrective
check your answers using technology. Expand (T + X)(T— X) and (T+ X)? using
¢ Explain why A does not have a measures to reduce emissions to meet the
minimum daily standards. properties of matrices. Explain why

Coty
multiplicative identity. 1 wa-() jana a -(? Find
5 A manufacturer makes three types of
a Let matrix N be a matrix whose entries 3°10, -l1 1, (T+ X)(T— X) # T? —- X* and
are the total daily number of kilograms of each if possible. If not possible, explain why. (T+ X)? #T? + 2TX + X?
products P,, Py and P, at each of its four
pollutants released for each product after
plant locations L,, Ly, Ly and L. 4" Matrix A the corrective measures. Write down the a AB b BA Show that(SU)? # S?U2. Explain why using
matrix R such that N= AR then find N. c A d Be matrix multiplication.
gives the daily amount (in kilograms) of
carbon monoxide, sulfur dioxide, and nitric The daily cost (in USD) of removing each Use the following matrices to answer Show that 2(RT) = (2R)T= R(2T).
oxide produced during the manufacturing kilogram of carbon monoxide, sulfur questions 2-6. What does this suggest about scalar
process of each product. dioxide, and nitric oxide at each of the four 0 -1 3 21 multiplication?
carbon sulfur nitric plants is given by the matrix C: R=-|3 7/)s=|4 1 3 3 5a
monoxide dioxide oxide L Ll, lL, ly 7 yo l ? Consider the matrices A = 12 and
5 4 -7 5 2 2,
B 20 35 15} 10 8 12 8 carbon monoxide 2 -5>
A= P,| 18 28 12 C=|3 24 3 sulfur dioxide 12 4-3 2 a-(3 30 Earo)
= = .
Pl 45 72 32, 7 9 6 11 | nitric oxide -(2, 3 U=|5 -4 2) . ae
01-1 a Show that AB=BA= I. = BS
b Find AC and describe the meaning of its b Determine (2AB)!°. con
entries. _{1 2 3 X= 7-2 Crs
W= 3 2 -1) “ l6 4 8 Michael makes the following conjectures
for all square matrices A and B.
2 Calculate each of the following if possible. a (AB)"=A"B" b (kA)"= kA"
+ nti If not ibl lain why.
Investigation 3 ROE Possile, EXPlain WAY. If you agree with Michael's claims, prove
a SR b TR ¢ WR+X d W(S-U) them. If you disagree, write down counter-
ioe! 2 0 2 2 0 10
Consider the matrix A= z | B= o | C= D= and I, = e (S—U)W examples.
4 7 Oo -3 30 0 -l ol
3 Use the properties of matrices to show that
ays } UR+R=(U+J)R. Verify your answer.
1 Using the fact that A2= A x A showthaA?
t # [;: 7 Pp
2. Find B, C?, and (1).

3 IfA=
a b*
under what conditions isA* =
ayaa
2 p2
z p Does this same condition apply to all 9.3 Solving systems of equations
J c
nXn matrices? Explain. using matrices
4 Determine D2, D3, and D+. Write a formula for DX where ke Z*. One of the many useful applications of matrix multiplication is that
it gives you an efficient way to express and solve systems of linear
5 Without using technology write your own definition for the zeroth power of a square matrix A°. Explain equations. Any system containing n variables and n equations can be
your reasoning. How does the GDC interpret A° and is this result consistent with your definition? expressed as a matrix equation.

For example,

ra-(% a 1be hen Al k = Ax Axx 4. {ee

\-ae47y=23
hse uivalent to (
a -3
7S}
7 hy) (23)
, i factors
of A
PO eto [ 48-3t-2z=0] 4 -3 -2Vs 0)
Inte casewhereD=c=0then a! =(* ats at { 2s + 2t + 3z = 6} is equivalent to 2 2 3 4tl=|-6.
J
I 6s+t-z=2 \ 6 1 -l)hz 2,

382 383
 ea eer Cea. Er ING DATA 9 . 3

In general, an nxn system of equations can be expressed in the form © .

AX = B where A is an v x n matrix containing the coefficients, B is b Consider that you are given a matrix A = ab where a, b, cand dare

BEL Ty
an x 1 matrix containing the constants, and X is an n x 1 matrix cd, ue
containing the variables. real numbers and you wish to determine the multiplicative inverse of A.
Since you are interested in the values of the variables that solve all Let B=(" x po
equations in the system, our goal is to determine X. At first glance you yz ‘be Cae tal cate NYS c a

Coty
notice that AX = B looks similar to the more familiar linear equation such (ee eee
. we A
as 2x = 10, where you would simply divide both sides by 2. However, i Use the factthat AB=I to show that aw y = ana { ax+bz= |
solving the matrix equation by dividing both sides by the matrix A would cw+dy=0 ax+dz=1
ii Sob h system in i to find the val if w, x, y, and zint ifa,
not lead you to X = 3. Unfortunately, matrix division is not defined in Lapa eS GMT TASER NAIA Te AAU UES
A b, c,andd.
the same way as other matrix operations. The key to unlocking the Verify that AB= and BA =I.
answer to this problem utilizes the multiplicative inverse property.
1
Recall that if ais a real number then — is referred to as the
a
The multiplicative inverse of a 2 x 2 matrix jl
multiplicative inverse of a since ax Le land 1 a=1. For example, P a by NUE CTD
E 4 a 3Y 2 Investigation 4 reveals that if A = ( q then the inverse of A is mindedness
a and Fj are multiplicative inverses since(-2\3 }- land cureh ele given by: c / Carl Gauss first used cay
2 snares L (db theterm “determinant” R=
B23 = 1. Note that th t diti fi hich b A'= where ad — bc #0 in 1801. $2
5 |\ales ote that there are two conditions for which a number Matitmennods ad—be\-c a, Eta
as
has a multiplicative inverse. First, both the number and its inverse were ueea teson . . ab . . allo
: . simultaneous linear You can thus determine the inverse of using the following
must have a product equal to 1 and second, this product must satisfy Seca steps cd
the commutative property of multiplication. second century BC
Similarly, n x n matrices AandB are called multiplicative inverses in China in the book 4 Subtract the product of the values on the minor diagonal from the
if AB=1I, and BA =I, where I, is the identity matrix. For example “Nine Chapters on product of the numbers on the main diagonal.
the Mathematical Art”
5 il 1 -1) ; F
A= ( Jana B= ( are multiplicative inverses since written during the Han
) 2 Interchange the elements on the main diagonal and change the
sign of the numbers on the minor diagonal.
3.1/1 -1 1 0° 1 -1/3 1 1 0)
AB= = and BA= = E
2 1-2 3 ol, -2 3 f2 1) \o 1)
Thus, the central question is, given a square matrix A how do you find \
its multiplicative inverse? The following investigation will guide you 3 Divide each element of (“ 0 by (ad — bc).
-c a,
through this process for a 2 x 2 matrix.

You will see in later sections of this chapter as well as in future EXAM HINT
Investigation 4 chapters that the value (ad — bc) has geometric significance. For this
Finding the
Determine the multiplicative inverse ofA = ( ‘ ma reason, this value is given a name and is defined as the determinant
multiplicative
of matrix A and is written as det A or |A|. Restating the formula for
A7! you now have: inverse for square
w x matrices beyond a
ai LetB= ( pe the multiplicative inverse of A. Use the fact that 1(d -b)
y Zz Al=— where
ad — bc #0 2X2 requires using
|Al\-c a,
AB=Ito show that technology, and you
{ 8w-6y=1 { 8x-6z=0 In the case where |A| = 0, A“! does not exist and A is said to be a will need to know
and , singular matrix or non-invertible. how to do this for
\-sw+4y=0 \-sr+4z=1
ii Solve each system in ito find the values of w, x, y, and z. Check your
your exams.
answer by verifying
that AB=I and BA=I. ©

384 385
 ea eer Cea. Er ING DATA 9.3
AX=B
Example 3 EXAM HINT TOK
») A'AX=A"'B (pre-multiplying both sidesby A”)

BEL Ty
In examinations
Find the inverse of ( 1 -2 Do you think that
3-4 j you should be able 1,X = A'B (multiplicative inverse property) one form of symbolic
to demonstrate the X=A'B (Multiplicative identityproperty) representation is
use the formula preferable to another?
1 -4 2)

Coty
yo 1 ( d —-b)
1x (-4)-(-2)x3 | -3 1)
\ [Al -c a)
_1f-4 2|_| -2 1 to compute the Example 5
Either of these two forms can be
inverse of a2 x2
—-2(-3 1) (-15 05, used, choose whichever is most Solve the systems of equations by first forming a matrix equation.
matrix but are
appropriate for the context.
expected to use the 4s—3t-2z7=0
op b { . z= 0 |
matrix utility using 425+ 2t+3z=-6}
\-3x47y =23
technology to find I 6s+t-z=2 |
the inverse of a
Example 4 43 matrix that is larger
thana2x2. 1Ox—5y=35| (10 -5Yx)_ (35) Rewriting the system in the form.
Use technology to determine the inverse of P=|2 2 AX=B
-3x+7y = 23 -3 7 )\y) 23)
6 1 &8Qo

LNT E NN
10 -5)'(10 -5\/x) (10 -5)"(35) 3
Verify that PP! = P-'P=I,. Pre-multiplying both sides by A aEa
3 7) (3 7hy) 3 7) a3) I
»J
x) 1 (7 5\35

LPPEL
ei i a
12 12 12 y) 70-15(3 1023 AA+=1,
Pet, 2 A x) 1 360
B) 15 15
Hil 7 y) 55\335 Lx=X

6 30 30, 72
ly wy
x)_}ul
a3 2 12
1 1
12
1
12 100
y) | 671l The calculation | 19-5 | | 3>
3 7 23)
/
could

have been done directly on a GDC.

pp'=|22 3 )-2 -2 4 jelo 1 ol=1 Use GDC to obtain the result.
co 1 1 3 1515 001
7k 7 The solution to the system is x = 2 and y= =
6 30 30
b 4s—3t-2z7=0
i 1 1 2s+2t+3z=-6 =>
2 | 12 12 4-3 -2 100
6s+t-z=2
HE? mt) 7, =. 2 2 3 |=|0 1 O|=I, 4 -3 -2\s 0»
2 15 15 6 1 1 001
1 ou 2 - 2
6
2
1
3
-lhz)
(2
')\¢t |=|-6

6 30 30

Continued on next page

°
You are now ready to unlock the solution to the problem proposed at
the beginning of section 9.3, which was to solve the matrix equation
AX =B forX.

386 387
 ea eer Cea. Er ING DATA 9 . 3

© 29) (18) (5 )(75) (13 29 18 5 75 13>

Ss} 4 -3 -2)'(0> 8 |,| 38 |,| 72 || 45 || 60 | M=| 8 38 72 45 60 a
tl]=(2 2 3 -6 1 9 )\74 1 53 1 9 74 1 53, s

e cpt 2, To encode the message, you can choose any 3 x 3 matrix for our Ff
pel 721 e
3 cipher. For example, if you choose C=| 0 3 -—1 |then our =
a 34 -2
encrypted message E is
B) 7 2 129 18 5 75 13 220 211 253 616 264
The solution to the system is E=CM=|0 3 -1] 8 38 72 45 60|=| 23 105 142 134 127
s=-},t=4,and2=-$ 3 4 2) 1 9 74 1.53) |-57 80 125 -47 95
That is, the bank password is stored in the document in matrix form as

220 211 253 616 264°

Applications
i i to cryptography 23 105 142 134 127. a
Protecting sensitive electronic information such as a password relies -57 80 125 -47 95, rs 8
upon a process called data encryption. Data encryption protects data by To decrypt the password requires you to solve E = CM for M: Fy FA
translating it into an unrecognizable form using an algorithm called a er
cipher. E=C(M@C'E=C'CM ..M=C'E 29
pays
To encrypt your bank account password “CharLie578Sam*!” you assign 7 2 1)'(220 211 253 616 264
a number to each lowercase and uppercase letter along with the special M=C'E=|0 3 -l 23 105 142 134 127
characters and the digits 0 through 9 as shown below. -3 4 -2} |-57. 80 125 -47. 95

ajbjc|d/je|/fjigih/i j [k I}/m{/n|o/;pi/q ri[s[tlulv[wlx [giz 2 8g -5)\(220 211 253 616 264»

1)};2);3/4/5]6)2/8]9 |10/)11 | 12 113 14/15 16 17/18/19 20/21 | 22 (23 24 |25 | 26 =|3 -ll 7 23 105 142 134 127

alei[clolelrl[el[u{[i[s[k([t[m[nl[olelo[r|s|tlulv|w|x|via@ 9 34 21)\-57 80 125 -47 95,

27/28/29 30/31 32 33 34/35 3637/38/39 40 41 42 43 44 45 46 a7 48 4g |50 51 52 29 18 5 75 13
=| 8 38 72 45 60
tlel#[s[a[>[elei*[e{:[s[7[.fol[1[2]3]4]s [67/8 |9 | space 1974 «1°53
5354/55/56 /57 58/59/60 61 62 63/64 65/66 67 68 69 70 71 72/73 74 75/76| 77
So the decrypted password is
Using table above converts the password “CharLie578Sam*!”: 29 8 1 18 38 9 5 72 74 75 45 1 13 60 53

mar T c har Lie 5 7 8 Sam * !

BSS c h a r L i e 5 ? | 8 Soa m |* !
‘word! | it | til
Converted |
=, 29 8 |1 (18 |38 9 |S 72 |74 75 45 1 (13 (60 53
il | | 1 i
Next, you decide upon the length of the block you will use to divide HINT 1 Determine the multiplicative inverse of each 31 ‘
premessword. Ble aan you wil use a lock of three which will If the word length matrix using A? = 1 ( d 7} Verify c ‘ ;
divide the word into five blocks giving the following 3 x 5 matrix M. was not a factor of5 detAl—-c a 1 3
then the remaining your answer by showing that AA~! =I 2 4,
spaces could simply 2 -4% b -2 -3)
be filled with zeros. 2 13 5s, -1 0,
388 389
 ea eer Cea. ANALYSING DATA 9.4
2 Determine the multiplicative inverse of 5 A manufacturer wishes to produce four 224 288 84 192 279 75 38 158 103 96 83 220
each matrix using technology expressing different products A, B, C, and D. The
Encrypted message: | 144 202 50 109 182 49 28 80 71 60 50 150

BEL Ty
the elements as exact values. Verify your table below shows the number of minutes
answer by showing that AA“! =I. required on each of four machines to 83 163 49 89 174 44 15 79 51 51 47 131
produce each product.
2 -1 4° 1.2 -2.5 3 12

NN Ww
a/l 2 0 b |08 -1 4 Machine | Product Product Product Product Encryption matrix: | 1 1

Coty
-l1 1 -2, 1.5 -0.75 1 A B iS D ol
I 2 | 41 ]a 3
2 8 Develop your own cipher using matrices to encrypt a sensitive piece of information. Design
ll 1 | 3 | 2 4 your own conversion table and encryption matrix by assigning different numbers to each of
3 4 2 Mm 2 | 4 2 2 the characters and numbers. Check your work by having a friend decrypt your message by
in es -1 Vv 3 | 4 1 2 giving them your encrypted message, encryption key, and conversion table.
3
The maximum amount of time available
ao 1 for each machine I, II, II, and IV is
2 J
240 minutes, 380 minutes, 280 minutes,
Write each system in the form of and 400 minutes respectively. Calculate
AX =B then solve the system using how many of each product can be manu- 9.4 Transformations of the plane
X=A™"'B. factured if each machine uses all of its For this section you need to be familiar with the ideas of rotation,
x+y+3z=30 | ( available time. reflection and enlargement, as illustrated in the example below. 9 BH
Qo
5x-7y =12 }
a {3x+2y-z=20} b | 6 a Jonathan makes the conjecture that
—2x+5y = 20
laxty+z=10 | (AB)~! = A“'B-!. Show that Jonathan’s Example 6 oe
conjecture does not hold using the as
2a+b-c+4d=-2 | -1 2 4 1 Draw the image of the triangle shown after the following y ts
5b-3c=4 matrices A = and B= é transformations: 3
ar 4a-3b+d=1.75 |
2 -3 31
b Michelle claims that (AB)! = B-'A™! a rotation of 90° counter clockwise (or anticlockwise) about (0, 0) 2
| a+2c-8d =-0.5]) does hold using the matrices A and B b reflection in the line x=0 1 Lil
For questions 4-6 create a system of above. c enlargement scale factor 2 centre (0, 0)
equations that represents each situation i Verify that Michelle’s claim is correct d a stretch parallel to the x-axis, scale factor 2. 123 4
then solve the system using matrices. Be
ii Choosing your own 2 x 2 matrices for
sure to define all variables. The rotation can be done using tracing paper or
A and B show that (AB)! =B1A™.
drawing an L shape from the origin to the point
In 2016, Sonya invested a total of $175,000 in iii. Prove that (AB)! = B-1A“ using the and rotating the L shape the required angle.
three different index funds FF, andF,. After fact that (AB)(AB)"! =I,
one year, the combined value of all of her
investments was $181,615. Data collected
on each of these investments showed that
each investment made average annual gains
of 2.5%, 4.8%, and 3.5% respectively during
the year. If Sonya invested twice as much
money in F, than in Fi, calculate the amount
Reflection in the line x = 0 means all points are
she invested in each fund. transformed to a position an equal distance the
other side of the mirror line.
Using the conversion table on page 386 decrypt the famous quote by James Joseph Sylvester
[James Joseph Sylvester (1814-1897) — an English mathematician who made fundamental
contributions to matrix theory, invariant theory, number theory, partition theory, and
combinatorics.]

Continued on next page

390 391
 MODELLING WITH MATRICES: STORING AND ANA' ING DATA 9.4
Oo; By considering the images of (1, 0) and (0, 1) suggest a matrix that represents an enlargement scale
cy An enlargement scale factor 2 centre (0, 0) means

BEL ea |
factor 2, centre (0, 0).
4 the image of each point moves twice as far from
the origin as before. Their new positions can easily Verify your answer by multiplying the points from the triangle above by this matrix and seeing if all the
3 be calculated by multiplying all their coordinates coordinates are multiplied by two.
2 by 2. How can you use the points (1, 0) and (0, 1) to find a transformation matrix?
1 All standard transformations of the xy-plane, including rotations, reflections, stretches and enlargements

CST
jp x can be represented by 2 x 2 matrices provided the point (0, 0) is invariant. These are often referred to as
O 1234 5 6 linear transformations.
ac
dy A stretch scale factor 2 parallel to the x-axis (or A matrix representing a linear transformation can be written (; d here (a, b) is the image of (1, 0)
4 horizontally) just multiplies all the x-coordinates
Py by 2 and leaves the y-coordinates unchanged. and (c, d) the image of (0, 1).
Similarly for a stretch parallel to the y-axis. 6 Find the matrix that represents a rotation of 90° clockwise about (0, 0).
2 1 |
? Find the matrix that represents a stretch parallel to the x-axis with a scale factor of2 and the y-axis
1 invariant.
$aa x Using the same method some general formulae can be found.
oO 1234 5 6
8 a Inthe diagram below A and B are the images of (1, 0)

LPPEL LNT E NN
pue fiuyawioag
and (0, 1) undera counter clockwise rotation of @
about (0, 0). Use the diagram to show this rotation is
_ (cos@ —sin@ »
Investigation 5 represented by the matrix
sind cosO ,
F

Points in the plane can be represented by their position vectors. In the b Whatwill be the matrix for a clockwise rotation of
x
1 magnitude 0?
diagram below, for example, the position vector ofA is
1 c Hence write down the matrix that represents a rotation
,
of 60° clockwise about (0, 0).
The position vectors of the vertices of the triangle ABC can be putin a single
123 9 a The line y=mx can be written as y= (tana)x, where &
matrix | is the angle made with the x-axis.
i 2 "}
In the diagram below A and B represent the images of (1, 0) and
. 1 0Y1 2 3 (0, 1) respectively under a reflection in the line y= (tana)x.
1 Find the product
of A
1M 21 a Explain why the image of (1, 0) has coordinates (cos20,
Let the columns of the new matrix be the position vectors of the image of triangle ABC under the sin2q@).
1 0% b By finding OBC in terms of o find the image of (0, 1) under
transformation represented by the matrix ( 1 the transformation.
TY
2 Onacopy of the diagram above draw the triangle ABC and its image after the transformation. c Hence verify that the matrix that represents a reflection in
cos2a sin2a \
Whatis the transformation? the line y= (tana)xis P
sin2a -cos2a ,
1 0 \
3. Inthe same way use the matrix ( to find the image of (1, 0) and (0, 1) under this transformation. d By first finding o, determine the matrix that represents a
/ reflection in the line y = v3x.
What do you notice about the image matrix?
10 By considering the images of (1, 0) and (0, 1) find the general
4 Test your conjecture by considering the image of (1, 0) and (0, 1) under this transformation represented matrices for:

by the matrix (: } a aone-way stretch, parallel to the x-axis, scale factor k

cd HINT
b anenlargement scale factor k, centre (0, 0).
i>) These formulae are all
in the formula book

392 393
 ea eer Cea. ANALYSING DATA 9.4

Example 7 Example 8

BEL Ty
a Write down the matrix that represents a rotation of 45° anticlockwise about the Find the 2 x 2 matrix that will transform the point (2, 1) to (1, 4) and the point (1, -3)
1 to (4, 9).
origin given that cos 45° = sin45° = Di
2 a b°
Let T= 3

Coty
b Write down the matrix that represents a horizontal stretch of scale factor V2. cd,
c¢ Hence find the image of (2, 3) after a rotation of 45° anticlockwise about (0, 0) followed

e
(0 82) easeass
by a horizontal stretch of scale factor 2. Therefore, Write down the matrix equation
d Write down the single matrix that represents a rotation of 45° anticlockwise about (0, 0) that describes the transformation.
followed by a horizontal stretch of scale factor N2e

an te -s ta o
\ Develop two 2 x 2 systems of

a
aL
v2 v2 This is using the formula for an (“)e(¢ B Y ) (f )e sa -a ve e equations and solve each system
using elimination or technology.
a anticlockwise rotation.
RR, 2
ee
=
1 << 7b=-7 -landa=1
a-3b=4).(2)

b | v2 0| 2c+d=4 &8Qo

LNT E NN
This is using the formula for a one-way ana 7" © 7d=-14 d=-2andc=3 3
Oo stretch. Both these formulae are in the c-3d =9) .(-2) aEa
formula book. I
»J

LPPEL
Therefore, T = 1-1 a
1 1 3 -2

- (v201 of] 1
a
1 3
i
The point is first rotated and then
.
its image

a +=a),
= is stretched.

& Investigation 6
v2 of 1 | | -1 Consider the triangle ABC with vertices A(0, 0), B(2, 2), C(— 1,5).
=lo af v2 ls
1 Show that AABC isa right-angled triangle and find its area.
S| 2,
v2 2 Triangle ABC is enlarged by a factor of 3 to obtain triangle A’B’C..

a Use matrix multiplication to determine the coordinates of A’B’C.

[ae] Bit 3
‘OnF! |
ae
1
BB
1 =
v2
=
V2,
) Determine the area of triangle A’B’C’. By what factor has the area
of triangle ABC changed? By what factor will the area change if it is
enlarged by a factor of k?
1 2%
3 Triangle ABC is transformed by T = (
2-1)
a_ Find the coordinates of the image points of triangle ABC.
Show that the new triangle is a right-angled triangle and determine its
area. By what factor has the area of triangle ABC changed under T?
Ifa transformation represented by a matrix
A is followed by a transformation c¢ Calculate det(T] and show that the area of triangle ABC enlarged by a
factor of |det(T)].
5°
represented by matrixB then the single matrix that represents both
transformations is BA.
Continued on next page

394 So)
 ea eer Cea. ANALYSING DATA 9.4
°o 4 Consider triangle ABC under the transformation T = (
21
4 a Write down the transformation matrix
for a reflection in the line y = (tana)x.
9 Let T be a linear transformation such that
-1 2
T= |. Find:

BEL Ty
-1 3 b Verify that a reflection followed by a 1 -2
a Use graphing software to show that area of A’B’C’ = 42. ls ABC a second reflection in the same line is a the coordinates of the image of the point
right triangle? How do you know? equivalent to the identity matrix.
P(-3, 4) under T
b Show that this area is equal to |det(7]| x Area of triangle ABC. 5 a Write down the exact matrix R fora b the coordinates of the point having an

Coty
\ rotation of 45° clockwise about the
2 image of (—12, 12).
5 Consider the transformation of triangle ABC under T =(? 3° Does the 1
origin, given that cos45° = sin 45° = —. 10 Find the 2 x 2 matrix that will transform.
, v2 the point (2, 1) to (1, 4) and the point
relationship |det(T)| x Area of triangle ABC hold for T? Explain why.
b Find the value of R® and interpret your (1, -3) to (4, 9).
result. ii A triangle with vertices (1, 1), (3, 1) and
6 a Find the area of the triangle with (3, 3) is rotated 5 clockwise about (0, 0)
In the investigation above you discovered a very useful relationship between vertices at (0,5), (4,5) and (2,7).
the area of a figure’s pre-image and the area of its image under a linear trans- and then reflected in the line y =—x. Let R
b Use the result, area of image = |det (T)| x
formation T. represent the rotation and T represent the
area of object to find the area of the
Area of the image =|det(T]| x Area of the pre-image image of the triangle: reflection.

i after a vertical stretch with scale a Write down the matrices R and T.
factor 2, followed by an enlargement, b Find the single matrix that represents R
Exercise 9E followed by T. &8Qo

LNT E NN
centre (0, 0), scale factor 3
3
Ht Find the transformation matrix for: ii after undergoing the transformation c Hence find the image of the triangle aEa
* after both these transformations. I
a A reflection in the x-axis B=(6,7) represented by the matrix 2 2) »J

LPPEL
6 g > d From a sketch or otherwise describe the a
b An anticlockwise rotation about the —3 5 )
C= (10,6) single transformation represented by RT.
origin of 45° 4
tT
? Prove that neither transformation will 12 Let E be a matrix representing an
c A clockwise rotation of 3 change the area of an object by considering enlargment scale factor 0.5, centre (0, 0).
the determinant for the general matrix for: a Write down the matrix E.
A reflection in the line y = —x
a areflection b a rotation. b Find the single matrix that represents a
~ea

A vertical stretch of scale factor 3

8 Let Tbe a linear transformation such that series of n enlargements scale factor 0.5,
An enlargement scale factor 2, centre a Use matrix multiplication to determine 305 centre (0, 0).
(0, 0) the image of triangle ABC when it is T= |. Find:
24
g A reflection in the line y= Bx. reflected across the x-axis to obtain
triangle A’B’C’. Sketch the triangle and a_ the coordinates of the image of the point
a Write down the transformation matrix P(14, 12) under T
its image on the same axis.
for an enlargement scale factor 4, centre
Triangle A’B’C’ is reflected across the b the coordinates of the point having an
(0, 0). image of (12, 12).
y-axis to obtain A”B”C”.
b Write down the transformation matrix
for a rotation of 180° about (0, 0). b Write a single matrix equation to find
the image of P under the composition of 13 The point P(—2, 5) undergoes two rotations R, followed by R,.
A rotation of 180° about the origin followed
by an enlargement of scale factor k where both reflections. Sketch triangle A”B”C” a Given that R, is a counterclockwise rotation of 20° and R, is a counterclockwise rotation of
k > 0 is often referred to as an enlargement on the same axis. 40°. Determine the coordinates of the image of P.
with scale factor —k. ¢ Show that if triangle ABC is reflected b Show that
c Find the product of the two matrices across the x-axis then again across the
cos60° —sin60° | x cos 20° cos40° —sin20°sin40° —cos40°
sin 20° — sin 40° cos 20° )/ x
found in parts a and b and hence x axis that the resulting image is triangle
ABC. sin60° cos60° y sin 40° cos 20° +.cos40° sin 20° —sin 40°sin 20° +cos40°cos20° y
explain why this definition is justified.
Consider the triangle ABC whose d What transformation is equivalent to c¢ Given that R, is a counterclockwise rotation of « and R, is a counterclockwise rotation of 0,
vertices are described by the matrix reflection in the x-axis followed by show that
reflection in the y-axis. cos(a@ + @) = cosacos@— sinasin@ and sin(a + 0) = sin@cosa + cosasind.
2 6 10
P= |
(i 7 :|

396 397
 MODELLING WITH MATRICES: STORING AND ANA' ING DATA 9.4
Affine transformations
Points in the plane can also be transformed using vectors

BEL Ty
(translations). For example, the point (3, 2) under the

mel? as 3) (1 4)
translation will have position vector + =
3, 2) (3) la,

Coty
or coordinates (4, —1).
An Affine transformation consists of a linear transformation
(represented by a matrix) and a translation and is of the
form Ax + b where x is the point being acted on. Affine
transformations are often used to describe or create objects
for which the whole object is the same (or approximately
the same) as part of itself (self-similar), many fractals are
self-similar. A famous example is the Barnsley Fern.

Investigation 7
Example 9
Consider the series of squares Sy Sy Sy ... where the y
The square PQRS with vertices P(—2, 4), Q(—6, —3), R(1, -7), (5, 0) undergoes a sides of each successive square is one-half of the previous 10 &8Qo

LNT E NN
transformation described by square, as shown. Each successive square is formed by a 8 3
aEa
1 transformation of the form AX +b where
A is a 2 x 2 matrix
-—— 0 \ €
, 3 and bis a 2X1 column vector. s e»a

LPPEL
(5 y ) o f LW[ }: [ ,
Let (x, ),¥,,;) beapointin S, , and (x, y,,) be the image 4 So |
2 of this pointin S_ n such that 2 Sy
, S, |
1 9 x,
a"T= Al")
ae
(=) Hb. 0 2 4 6 8 10 12 14
+>
16
x
2
a By multiplying two transformation matrices verify that 1 _ is equivalent to an
0 = 1 By considering the points(0, 0), (0, 8), and (8, 0) in S, and their corresponding images in S, write
2
/ down and solve a system of equations to determine A and b.
enlargement of scale factor 0.5 followed by a reflection in the y-axis. 2 Using the coordinates ofa point in the interior of S| along with the coordinates of its image in S_ verify
b Determine the coordinates of the vertices of the image of PQRS. your answer in question 1.
¢ Draw the square PQRS and its image on the same axes.
Determine the coordinates of the image of (6, 6) in S,.
ie 0 ie "] (*: 0) Take each point in turn and find 4 a Apply the transformation to find (x,, y,) in terms of (x), Yo).
0 05 ol = 0 05, its
its image.
i . The ful working
The full king for for thi the
point (—2, 4) is shown. b Apply the transformation again to find (x,, y,) in terms of (x,, y,).

CH? OOOH
\
= \ Fa My X)+8+44+2
¢ Hence showthat|“* |=| 8
Vs 1
3 Yo ,
(-6,-3) > (6,0.5) d__ Use the formula fora geometric series to find an expression for (x,,, y,,) in terms of (x), Yq).
(1-7) >(2.5,-1.5) 5 Use your conjecture to determine the coordinates of the image of (6, 2) in Se

(5,0) — (0.5,2)

398 399
 DELLING WITH MATRICES: STORING AI ANALYSING DATA 9.4

®
Let E be the enlargement matrix and R the d= Write down the matrix R.
rotation matrix. e Find ER.

BEL Ty
Ht The points A(-3, 4), B(2, -3), C(0, 4), ii matrix Q represents a clockwise ce i Find det(E). f Use your GDC to find the coordinates of
and D(5, 8) are transformed to A’, B’, C’, D’ rotation of 135° ii Find the area of T,. all three vertices in T,.
according to a linear transformation defined iii Hence find the total area covered by
iii. matrix R represents a reflection in the
-4 0 6) T, to Ty.

Coty
by X= + . Find: line y= 3x
1 -4 -2)
iv matrix S represents a stretch of factor
a_ the coordinates of the image points
2 parallel to the x-axis and a stretch
A’, BY, C’, D’
of factor 4 parallel to the y-axis.
b the coordinates of the point whose
image is P(-6, —3).
b Find a single transformation in the form
of AX + b that transforms a point (x, y) by
Developing inquiry skills
Under a transformation T the image (x’, y’) Q followed by P followed by R followed In the chapter opener, you found, using geometry, the exact values of the
of a point (x, y) is obtained by the matrix by S. coordinates of each vertex and endpoint of stages 1 and 2 of the Koch
equation Snowflake.
c Finda single transformation in the

GG Go
form of AX + b when a point (x, y) is Stage 1
\ y
translated by the vector [ 4 followed 14
the image of the point (—5, 8) J
rLVin Ee}
rate
aoc.

the point with an image of (3, 6) by R.

the image of the point (a, a) where ae R 5 A sequence of equilateral triangles (T,,) are *7 23
Firs
the point with an image of (a, a) where shown below. T, is formed from T,,_, by py ee cory
1 2 3 e2
aeR. rotating the triangle 60° clockwise about
A shape is transformed by first undergoing (0, 0) and enlarging by a factor 0.5 (centre You may have noticed that the basic building block of the Koch
a enlargement by a factor of 3 followed by (0, 0). Snowflake is given by stage 1—thatis, using a series of transformations
a rotation of 90° anti-clockwise then lastly on stage 4 you can create each colored piece of the snowflake in
stage 2.
2
translated by the vector (3} Stage 2

a Determine the 2 x 2 matrix T and the

2x 1 column vector b that maps the pre-
image point (x, y) to its image point (x, 7)

such that(; )- 7(* }. b

x ‘
2-15 1 -05°\Y 05 1
-0.5 4
y y
Ve
HES)
66
(4.8
JI6" .
The image point (x’, y’) is now reflected
across the origin to obtain the image point
T, is the triangle shown with two vertices at Na n/N
(-2, 0) and (0, 0). (0,0) (3.9) 9) (1,0) (2,0) 9] 8.9) (3,0)
(x",y") The perimeter of the shape formed from
b Determine the 2 x 2 matrix A and the these triangles is the distance around the 1 Describe the series of transformations that are necessary to map the
2 x 1 column vector c that maps the outside edge. red piece in stage 1 to the red piece in stage 2. Write your answer in
pre-image point (x, y) to its image point
a Find the perimeter of the shape shown the form of AX + bwhereA is a2 x 2 matrix, bisa 2 x 1 column
(x’,y”) such that(7 }. a?) c. made up of triangles T, to T,. vector, and X is a2 x 1 column vector of the vertices and endpoints
y y b Explain why the maximum perimeter is of stage 1. Name this transformation T,. By applying oe to each
4 a Find the 2 x 2 transformation matrices P, reached once T, is added to the diagram vertex and endpoint in stage 1, determine the image points of the red
Q, R, and S given that and find this value. piece in stage 2.
i matrix P represents a reflection in the 2 Repeat the steps in question 1 to determine the transformations of
x-axis the form AX + b that maps the vertices in stage 1 onto the blue,
orange, and purple pieces of stage 2. Name these transformations
T,, T,, and T, respectively. By applying these transformations
determine the coordinates of each vertex and endpoint in each stage.

400 401
 ea eer Cea. Er ING DATA 9.5
9.5 Representing systems You could represent the probabilities in the computer simulation given
in investigation 8 in two tree diagrams as below. These are called =
Some applications of probability consider sequences of events in which transition probabilities. 3
an outcome depends on the previous outcome. For example, the i
oe —? 0.5 —D Ay
probability of our inheriting a particular genetic trait is dependent on 25
the genes of our parents. Whether or not today sees heavy rainfall may D w
Di)
be affected by the type of weather from the previous day. 04 >W 05 ~w &
i)
Another convenient way to summarise and represent these 0.5
Investigation 8 probabilities is in this transition state diagram:
Sequences A and B are simulations of weather data over 21 days. You can also represent the same information in a 05 GW) (0) >06
W represents the event “It was wet today” and D represents the event “It was dry today”. transition matrix in which the column headings refer
to the current state of the system, shown here as the 0.4
A DDD DW DW WO DW DW ODWWW OD DD OD matrix T and the row headings the next (future) state.
B WoD DW W DW W WW WWW OD DOD W DW W W
The entries in the transition matrix show the probability of the
You can investigate this data to seeif there is any evidence of the weather tomorrow being dependent on the transition from the current state to the future state.
weather today as follows. Curent state
D Ww
In sequence A there are 8 wet days and 13 dry. You only consider the first 20 days however since you do not
D (06 05

Future state
have information about the weather on the 22nd day. From the first 20 days you have this data: {WD, WW, &8Qo

LNT E NN
"
WD, WD, WD, WW, WW, WD} to represent the days on which the weather was wet and then the weather on the w lo4 05 3
aEa
next day. From this set you can find these experimental conditional probabilities:
I
In the following investigation you will find out how to interpret »J

LPPEL
3 5
P(Wet tomorrow | Wet today) = 8 and P(Dry tomorrow | Wet today) =Fl repeated multiplication of a transition matrix. a

Show that for sequence B an estimate of the conditional probabilities are:

Investigation 9
4 3
P(Wet tomorrow | Dry today) = Fi and P(Dry tomorrow | Dry today) = FI
06 PD Coninuing with the situation described above You can use this tree diagram
Dl to find the probability that if is dry today, it is also dry two days from now by
One sequence was generated by a computer simulation programmed with these conditional probabilities: 0.6 on
oa w finding
P(Wet tomorrow | Dry today) =0.4 and P(Dry tomorrow | Dry today) =0.6 and 7 os bd 0.6X0.6+0.4
x 0.5 =0.56
P(Wet tomorrow | Wet today) =0.5 and P(Dry tomorrow | Wet today) =0.5 04 ~w < and the probability that if it is dry today it will be wet two days from now by
The other sequence was simulated by flippinga coin. Discuss in a small group which method may have finding
generated each sequence 08 *W 0.6 x 0.44 0.4 x 0.5 = 0.44
1 How did you decide?
1 Apply a tree diagram to find the probability that if is wet today, it is also wet two days from now.
2 Whyis it difficult to be sure?
2 Apply a tree diagram to find the probability that if is wet today, it is dry two days from now.
5 Groce How do you describe independent events?
0.6 0.5
3 Given T = |, Find T2.
0.4 0.5
The states of this system are “Dry” and “Wet” and the sequences A 4 How do you interpret the values in the matrix T7?
and B show the system changing state from one day to another. In this
system, you cannot predict the outcome of a given trial with absolute TOK Gan What is more efficient to find the probabilities in T7- matrix multiplication or drawing the two
certainty, but you can use probabilities to quantify the likely state of tree diagrams?
the system in any given trial. How do “believing Gay How do you interpret the values in the matrix T>?

oon o
that” and “believing in”
differ? Gana How can you interpret our results when multiplying by the transition matrix?
A Markov Chain is a system in which the probability of each event depends How does belief differ Gary What does 71 represent in T”?
only on the state of the previous event. from knowledge?
Gon How can the probabilities of Markov chains be represented?

402 403
 ODELLING WITH MATRICES: STORING AI Er ING DATA 9 . 5

iti matei IS asquare matrixix whirien Summarizes

pememaen i atransition
iti state International
a © 7 The most efficient method is to store the Fa
iagram. It describes the probabilitiesof moving from one state in a system mindedness matrix in a GDC as T and find T°, 3
to another state. . cf
r . Matrices play an a
The sum of each column of the transition matrix is 1. important role in the Fe
The transition matrix is extremely useful when there are more than two states projection of three- g [9-307 0.280 0.243 =
p , A . a 73 =|0.316 0.316 0.316 Cy
in the system and when you wish to predict states of the system after several dimensional images 0.377 0.405 0.441 S
time periods. into a two-dimensional Read off the number in the C column and =
screen, creating the The probability that a bicycle starting in the | the O row, which is 0.441 and write a
realistic seeming central business district at the end of three conclusion that interprets this number.
motions in computer- days is in the outer zone is 0.441.
based applications.

Example 10 From what y you have learned in this section, how can y you best
Dockless bicycle company Mathbike hires bicycles in a city through a mobile phone app. represent a system involving probabilities?
Users can unlock a bicycle with their smartphone, ride it to their destination then lock the bicycle. For visualising a system, would you choose a tree diagram, a transition state
Mathbike divides the city into three zones: Inner (I), Outer (O), and Central business district (C). diagram ora transition matrix?
By tracking their bicycles with GPS over several weeks, the company finds that at the end of each
&8Qo

LNT E NN
day: For calculating proba! s, would you choose a tree diagram, a transition
3
state diagram ora transition matrix? aEa
e 50% of the bicycles rented in zone C remained in zone C, 30% were left in Zone I, and 20% I
were left in Zone O »J

LPPEL
a
e 60% of the bicycles rented in zone O remained in zone O, 30% were left in zone I, and Exercise 9G
10% were left in zone C.
© 35% of bicycles rented in zone I remained in zone I, 35% were left in zone O, and 30% 1 For these transition matrices, construct the a_ Show this information in a transition
were left in zone C. corresponding transition diagrams: matrix T by writing down the blank
a Show this information in a transition state diagram. a Current state entries.
b Show this information in a transition matrix. x Y Current location
c¢ Determine the probability that after three days, a bicycle that started in C is now in O. eo x({ 1 c oP
sa 115
2
a 0.60 The diagram gives a quick way to check that g y 6 4 Te 3 c 02
the relationships given in the question are 2 11 5 ~ 8 P| 0.25
put in the right places. =
b Current location
A B Cc b Calculate T? and hence find the probability
A 0.2 0 0.47) that a person who buys Popsi now will
BT=|05 09 0 change to Ceko three weeks from now.
c 03 01 0.53, 3 > Upon further research, it is found that a
2. This transition state diagram shows the more realistic model is to create a third state
findings from market research of the buying N to represent a person buying neither Ceko
b ¢ Current ; location ° habits of consumers who shop weekly and nor Popsi.i The transition
iti i
state diagram isf
Bs Be careful to make the column headings buy either soft drink brand Ceko or Popsi. For
22 c 0.50 0.30 0.10 represent the current state of the system example. re Probability that a person buying
as
BS 1 _T=/ 030 035 030 and the row headings the future state. ko will buy Popsi the next week is 0.25,
8 005
& 0 0.20 035 0.60
&
0.75 G (c) (P) 08

02
404 405
 ODELLING WITH MATRICES: STORING AI Er ING DATA 9 . 6

a Complete the transition matrix: 5 Two candidates A and B are running for then a; is the probability that the system transitions from state j to
Current drink purchased an elected government office. Statistics state i after n stages. The sequence of transition matrices T”, n € Z* =
c P N gathered from a survey taken over several se . . . EI
aoa 0.65 weeks prior to the election showed that behaves in different ways according to T, just as the geometric me
Be . P « © sequence ur, n€ Z* behaves in different ways according to the 2
SE after each week approximately 4.1% of 1 =
ater e) T= 0.05 . value of r. cs
os the candidates who were planning to ry
3 0.04 vote for candidate A changed their mind om
and decided to vote for candidate B and ry
b Find the probability that a person buyin; i i =
uae PI ty that a p yng 3.2% of those who were planning to vote Investigation 10
neither Ceko nor Popsi now will buy
a for candidate B changed their mind and
Popsi five weeks from now. decided to vote for candidate A. 1 Use technology to find these powers of each transition matrix T.
4 A game begins in state 1 with two white
oe in one8 box and two blue in
. another. Five weeks before the election a final survey T T2 T? | T4 | T2% | T° | Term
coins .
ae showed that 45 520 people were planning to a \
A coin is chosen at random from each box . 5 1
, . vote for candidate A and 38745 people were — =
and put into the other then the process is . : los
« planning to vote for candidate B.
repeated many times. 6 4
a Write a matrix equation that describes To
°e oe 1 1® 9 e° °- owt the number of voters supporting each Z
.
candidate m weeks after the final survey. b 0.2 0 0.47
* State 2 State 3 b Assuming that the trends revealed in the 05 0.9 0 ae
. . Ea:
a Construct the transition matrix to show clecin coo up unt the day of the 0.3 0.1 0.53 os S|
the probabilities of transitioning from e ection etermine w! ich cane ‘ate wil ] 38
one state to another. . win. Calculate the margin of victory. c 0 05 0 ie
mw
b The game is made more sophisticated root <2
by starting with three white coins in 0 05 0
one box and three blue in another, d 70 008
represented as 3W3B. State the new .
number of states. the transition matrix 03 1 0 01
to show the probabilities of transitioning 03 0 1 01
from one state to another. 0400 0
c Investigate the transition matrices for A 025 0 0.65.
other situations such as 4W4B, 5W5B, 0 0 035
3W2B etc. 0751 05

2 What patterns are there in these sequences of matrices?

Each sequence above is classified by one of these terms: absorbing,
regular, periodic.
9g . 6 Re Pp rese nt i ng stea d y state An absorbing state is one that when entered, cannot transition to
another.
systems 3 What would a definition of the other terms be? Drawing
You have seen in section 9.5 that if Tis an m x m transition matrix TOK a transition state diagram may help you distinguish between the different
and if types.
\ Ifwe can find solutions P ,
My A. 3 Aim . . . 4 Identify the appropriate term for each.
a a a of higher dimensions,
. cic 23 2m can we reason that these 5 (EXCITE What happens to a regular Markov chain with transition
T"=| a, ds, as, Iam spaces exist beyond our matrix Tas n tends to infinity?
sense perception?

Ant An Ans, vee 4nm /

406 407
 ea eer Cea. Er ING DATA 9.6
Aregular transition matrix T has the property that there exists 7 € Z*such Investigation 11 Fe
that all entries in T” are greater than zero. For high powers of 1a regular 3
transition matrix converges to a matrix in which all the columns have the 1 Using the Ceko and Popsi matrix and vector above, use technology to complete the table. s
same values. In the rest of this chapter you only consider regular transition 2
matrices. n= Algebra Fs
no. of representation Vector representation oo
weeks ofS n Matrix representations of S n of S n =a
5
You can use this property to make predictions about the state of a (3 a
system after many time periods. In practical applications you often 0 S, 0
need to consider a population in which a certain number are in either
45, 45,
state at the beginning. For example, consider the transition matrix
0.75 0.2 \(81) 69.75
which represents the probabilities of transitioning between buying
Ceko or Popsi each week: 1 TS, 0.25 0.8)\45, 56.25,
Current location 0.75 0.2 \( 69.75 0.6125 0.31\(81) 63.5625»
2 TS,=TS, =
c P 1 0 0.25 0.8 )\56.25 0.3875 0.69 45, 62.4375,
2
g Cj} 075 0.2 5 0.75 0.2 63.5625) _ (0.536875 0.3705 \ 81 60.1594 )
T= 2
2 P/|025 08 3 TS,=TS, 0.25 0.8 62.4375) (0.463125 0.6295 \.45 65.8406 , a
eZ iy
a)
4 Er
A survey of 126 consumers showed that 81 bought Ceko and 45 Popsi.
10 th Etat
arn
20 | “2
49
The initial state vector S, shows the initial state of the system. |
2 Gury What are the converging steady state patterns in your results?

3 Granny Generalise these convergent patterns using the terms long-term probability matrix and
Number of
steady state vector.
consumers
4 Whatinterpretations can you make of these patterns regarding the long-term market share of each
c 81 brand?
So=
P | 45 5 What assumptions are made?

6 (IOSCe How can you use powers of the transition matrix to make predictions?
You can predict the market share after one week by finding
0.75 0.2\(81) (69.75%
S, =TS, = =
0.25 0.845) \56.25, If Pis a square regular transition matrix, there is a unique vector q such that
Pq=q. This q is called the steady-state vector for P.
Hence you can predict that after one week, between 69 and 70 of
these 126 consumers will purchase Ceko. Remember that some of
these customers will have initially bought Ceko and some will have
changed from Popsi. Also, some customers who originally bought Ceko Example 11
will have changed to Popsi. The multiplication S$, = TS, quantifies and Dockless bicycle company Mathbike open their business by distributing a number of
represents all the transitions.
110°
Similarly, you can multiply by T successively to predict the states in
bicycles in a city according to the initial state vector S$, =| 80 .
subsequent weeks, and look for patterns in our predictions.
S, = TS, = T’S,. Repeating this process establishes the equation 50
S,=T"S, : Continued on next page
©

408 409
 ee eh ATL eae 9.6
6. Calculate the likely number of bicycles in each zone of the city after five days using the
transition matrix given in example 10. Total population u= 4g (the steady state

BEL Ty
b Find the likely steady state of Mathbikes. (p) represented P probability
Sy by S, XS =v vector)
¢ Comment on how the owners of Mathbikes should reflect on their choice of the initial
state vector S,. State the limitations of this model. 63) 8 27

Coty
18) 54
a (0.50 0.30 0.10) (110 65.746
0.30 0.35 0.30}| 80 |=| 75.789 Find S, = 7°S, with a GDC
0.20 0.35 0.60}\ 50} \98.465}
There would be approximately 66
Make sure to interpret your findings.
Mathbikes in zone C, 76 in Zone I and
98 in Zone O.
0
b (0.50 0.30 0.10)°(110\ (65.263> ae ,
0.30 0.35 0.30 80 |=|75.789 Use your GDC to investigate higher powers Then repeat with your own 2 x 2 regular transition matrix T.
of the transition matrix and look for 3 Gury Does the steady state probability vector depend on the initial
0.20 0.35 0.60 50 98.947 , convergence. As with geometric sequences, state vector?
the rate of convergence depends on the
So, the system quickly reaches an equilibrium 4 What relationship exists between the steady state probability vector and &8Qo

LNT E NN
values you are given.
at 65 bikes in C, 76 in I and 99 in O. . . 3
A steady state matrix is reached earlier than the long-term probability matrix? aEa
this, after ten days. 11 I
0.6 02>.with “ »J
, 33

LPPEL
¢ This shows that Mathbike’s original
Think critically about the context. Let T = long-term probability matrix X = and a
distribution of bikes to zone I was a Y 0.4 08, 22
good guess, but they would have met Write a complete sentence that presents 33
the needs of the customers better by your findings.
placing 50 bikes in C and 110 in O at the u, \
letu= [ " |be the steady state probability vector. Hence,
start. Us)
This model does not account for broken 0.6u, + 0.2u, =u,
or stolen bikes. Nor does it account for
changing habits amongst the customers Tu =u >| 0.4u, + 0.8u, = u,.
due to weather patterns or competition u,+u,=1
from other companies.
5 Explain why each of the three equations hold and show that the solution of
1»

this system is U =

Investigation 12 3)
, ee 0.6 0.2> With your own regular 2 x 2 matrix T, find the long-term probability matrix
1 By investigating
the sequence T” show that if T = nao , then the by repeated multiplication and then form three equations to solve Tu =u
So) using technology.
wile

wie

long-term probability matrix is X = Repeat with your own regular 3 x 3 matrix.

Gury What do you notice?

WIN

WIN

Ginry How can you find the steady state probabilities? Give two
2 Fillin the table with examples of your own initial state vectors and find
different methods.
values for the steady state population vector and the steady state
probability vector. Oo

410 411
 ODELLING WITH MATRICES: STORING AI ANALYSING DATA 9.7
Litt E) a 9.7 Eigenvalues and eigenvectors

BEL Ty
st Find the long-term probability matrices for At the present time, 45% of the
the Markov processes with these transition customers are provided electricity by A, Investigation 13 TOK
matrices: 25% by B and 30% by C. \
The multi-billion dollar
0.24 0 0.31% a Find the percentage of the customers
Consider the transformation T = ( applied to the points on the line
; 0.13 0.71°

Coty
eigenvector.
b | 0.42 0.85 0.05 provided by each electricity provider
,
0.87 0.29, Ly:y=2x. Google's success derives
0.34 0.15 0.64, after two years.
in large part from its
b Calculate the long-term percentage i Write down the coordinates of any five points on L, then show that images
PageRank algorithm,
0.1 06 0.2 0.7* of customers who will be provided of these five points lie on the same line.
which ranks the
0.05 0.2 0.2 0.05 electricity by B. State your assumptions. Take a general point on the line with coordinates (x, 2x) and find its image. importance of webpages
0.4 0.05 0.2 0.2 ¢ By comparing the matrices T, T?, T° and ii Based upon your observations in part i write down the value of A for which according to an
0.45 0.15 0.4 0.05, T°, comment critically on the likelihood eigenvector of a weighted
that the transition probabilities remain 1 1Vx x) . link matrix
=A . What does this tell you about how T transforms
2 A regular transition matrix Tina constant in the long term. 24 2x How ethical is itto create
given Markov process is defined as A hire car owned by Mathcar is in one of mathematics for financial
the points on L, geometrically?
0.81 0.1 0.08 four possible states: A: functioning normally, gain?
‘
T=/0.09 0.15 0.5 B: functioning despite needing a minor also maps the points on line L.,: y =x to image points that
repair, C: functioning despite needing a , &8Qo

LNT E NN
0.1 0.75 0.42
major repair, D: broken down. Mathcar only are also on y=x. Determine the scale factor of the stretch that describes HINT 3
4, 1 \ aEa
takes a car out of service if it breaks down. . . - . 1 1Vx x) When a transformation I
Let u=|u, be the steady state probability this mapping by solving the equation sal. »J
matrix Tmaps each

LPPEL
The transition matrix A\x x) a
Usy point on one line toa
vector such that Tu =u. 0.9 0 0 0.96* point on a different
0.03 0.85 0 0 The investigation above reveals a special type of linear transformation line we say that the
a Find u by solving a system of equations. F= represents the
0.02 0.05 0.6 0 that maps points on a line back onto the same line and is described as a lines are invariant but
Demonstrate your answer for part a stretch by a factor A. In this section you will learn how this idea along the points themselves
is correct by finding the long-term 0.05 0.1 0.4 0.04, with our knowledge of matrix algebra will allow you to determine
probability matrix. are not.
probabilities of the car transitioning from one large powers of matrices. These results lay the foundation for our
3 Three electricity providers A, B and C state to another after a period of a month. study of modelling and analyzing dynamical systems and networks in
compete in a market. A survey of consumers chapters 12 and 14.
a Interpret the zeros in the matrix in the International-
has determined the probabilities that a person
context of the problem, and the entry in mindedness
using an electricity provider in one year will the fourth row fourth column. Eigenvalues and eigenvectors
either stay with that provider or will switch to Geometrically, you say that the lines y = 2x and y = x are invariant The prefix “eigen-” is
another next year. The probabilities are given b Verify C that Fis a regular Markov
under the transformation T. German, meaning “to
in this transition matrix: matrix.
own" or “is unique to”
c¢ Mathcar begins business with a fleet of In general, you are interested in determining the equations of indicating that that an
Current provider
500 cars in state A, 25 in B, 0 in C and the invariant lines and their associated scale factors, A, for a given eigenvalue is a unique
A B c 0 in D. Find how many cars are in each transformation matrix A. That is, you wish to solve the equation value associated
provider 0.8 0.05 0.06) state after 10 months. x x \ with a unique vector
after one al }-af or expressed more simply Ax = Ax. The values of A
a T=|0.12 0.92 0.1 d Comment on how Mathcar could change y y, named the eigenvector.
their practices so that changes are made are referred to as the eigenvalues of matrix A. Each eigenvalue Eigenvalues describe
Cc 0.08 0.03 0.84,
to F which improve the availability of of A is associated with a particular line that is invariant under the the characteristics
their cars. transformation A whose equation is described by the eigenvector x. of a transformation
and for this reason
Consider the problem of determining the eigenvalues and eigenvectors are also referred to as
14 characteristic values and
of the transformation matrix A =(: 5} eigenvectors referred to
as characteristic vectors.

412 413
 ea eer Cea. Er ING DATA 9.7
You would like to find the eigenvalues A and their associated t>

14 and therefore have the form x,=| ¢ where tis any real number.

BEL Ty
eigenvectors such that (: 5 s = Ax. Proceeding with the matrix
2)
algebra yields: That is, there are an infinite number of eigenvectors associated with
the eigenvalue of A, =—1. One particular eigenvector (when f= 2)

Coty
is x, (2:
= Similarly, there are an infinite number of eigenvectors
4
, s -Ax=0 (addition property of equality) associated with the eigenvalue of A, = 5 described by the points on the
14 line y = x and these vectors have the form x, [/
= ser tis any real
( : }s -A1,x=0 (identity property)
) r
14 1 0 number. One particular eigenvector (when t= 1) is x, (i
=
( }-a[ =0 (distributive property)

x= 0 (matrix addition) >

2 3-a a b* , . P
IfA= then the real number (is called an eigenvalue of A if there
1-a 4

pue fiuyawioag
¢ /
When the matrix ( > 3 a an inverse then both sides of the

LPPEL LNT E NN
exists a nonzero vectorx such that Ax = Ax where xis the associated
equation can be multiplied by the inverse matrix to give the solution 5 eigenvector with the eigenvalue A.
a NT
h ea
x= a | If looking for additional solutions they must occur when the mindedness
y ~
inverse does not exist which happens when the matrix (A - AJ) is Belgian/Dutch a-A Ob HINT
singular and therefore its determinant is equal to zero. mathematician Simon The eigenvalues of A are given by the solutions of cod 1 =0 The process for
Thus: Stevin use vectors in his
aa theoretical work on falling
referred to as the characteristic equation. finding the steady
state vector of a
W-a 64 bodies and his treatise
Markov chain, ie
De 3-A4 =0 “Principles of the art of
weighing” in the 16th Matrix algebra provides you with a tool for modelling, analyzing, and solving TH=u, is
(1-4)(3-A)-8=0 era predicting the long-term behaviour of systems that change over time equivalent to finding
WV? -42-5=0 (called dynamical systems). the eigenvector for
an eigenvalue of 1.
“A, ' = -landa,=5
2 4
1-4
Now, returning to the original equation ( > 3-a s = 0 to find the

corresponding eigenvector for each eigenvalue gives you: Example 12 .

Find the eigenvalues and corresponding eigenvectors of A = [ L
24 0 2x+4y=0
A=-10 *\. °o faxtay both equations give
y= ~*)
2 4hy 0 2x+4y=0

Solve PA ae 0
-4 4 0 -4x+4y=0 3 -2-A
Aa) <> *\- ° jax ay both equations give y = x
2 -2)\y 0 2x-2y=0
A (-2 -A) -3 =0 This can be expanded or solved directly
The eigenvectors associated with the eigenvalue of A,=—1 are
h=-30rl using technology.
described by the position vectors of the points on the line y = dy
2 When A= -3

Continued on next page

414 415
 ea eer Cea. Er ING DATA 9.7

3 1\x)_ o Using (A-AI) x=0 a4 0 i iw”

2 Let Pbe a matrix containing the eigenvectors x, and x, ofA and D = ( A h a is,P = (;

BEL Ty
2
3 ify) lo, and D=(? ° ‘
0 8,
=> 3x+y=0 or y=-3x An alternative equation to find the
a Find APand show that
AP= PD.

Coty
\
_ i torsis}O 1 |X |__3]x b_ Explain using the definition of eigenvalues and eigenvectors why this will always be the case.
Similarly when A= 1 Poly j.jo eigenvectors *s ° 2 | } {
3 -3)\y) lo y yy 3 a Use matrix algebra to show that if AP= PD then A= PDP!
x+y=Oor y=x b Use the result in part a to show that A* = PD>P-!
\ c Howdid you use the fact that matrix multiplication is associative?
Possible eigenvectors are therefore 1 In the first case any vector which has the d_ How did you use properties of inverses?
aL: 3) y-coordinate —3 x the x-coordinate will do. 4 a Forthe value of D given in question 2 find the value of D>.
and
1 , b Can you generalize your result for D”.

5 a Determine A’ by finding PD>P —'. Verify the result using the GDC to calculate A>.
b Can you generalize your result for A”?

PELL E NN
pue fiuyawioag
4 Determine the characteristic equation of _(1 0° d pH 0 -3> \ EXAM HINT
each matrix in the form A? + pA +q=0 then ch ol P= 3 40, a
Let A = ( containing real number entries and having distinct real
determine its eigenvalues. In an exam, matrices
c ,
\ will always have dis-
eit -2\ b p-(> 0° e r-(3 o7 eigenvalues A,and a, then you say that A is diagonalizable and can thus
ale | “ls, 0.6 0.3, be written in the form ofA= PDP! where P= (X, X,) is the matrix of tinct real eigenvalues.
a, 0»
G 1) 4 Consider the transition matrix eigenvectors and D = I
e¢ C=
=| 2 a b » § [ 0 dy rs Internationa
a r-(, \-b where 0<a<landO0<b<l. mindedness
2 Determine whether or not 4=-1 an 7a h8, . .
3 2 a Show that A= 1 is an eigenvalue of T Vectors developed quickly
eigenvalue of } and determine the other value of 4 in in the first two decades
Let A be a diagonalizable 2 x 2 matrix expressed in the form of
terms of a and b. of the 19th century with
3 Find the eigenvalues and eigenvectors of , 0)
b Determine eigenvectors for each Danish-Norwegian Caspar
each matrix. A= PDP*'then
A" = PD"P-'where P=(X, X,) and D =(* A
eigenvalue of T in terms of a and b. Wessel, Swiss Jean Robert
2)
5 -8 2 4) Argand and German Carl
a C= b Q= |
23 5 3) Friedrich Gauss.

Example 13
Diagonalization and powers of a matrix 12
a Find the diagonalization of A = (3 3}

b Hence find an expression for A‘ in the form PD*P™!.

Investigation 14
3 4 c Find an expression for A‘ as a product of 3 matrices with no exponents.
Consider the matrixA = 3 ,
-6 2 2 3 | Eigenvalues are the solutions of the
Ja-ar|-o-! =0@1°-1-6=0 characteristic equation |A — AI|=0
1 Show that the eigenvalues of A are A, =-land A, = 8 and that the corresponding eigenvectors are 3 -A
1 1) A, =3 and A,=-2
x,= (:) and x,-( i
Continued on next page

416 417
 ea eer Cea. ANALYSING DATA

There are an infinite number Investigation 15

BEL Ty
of eigenvectors of the form Consider the situation where people
t 2t ’ move between two neighbourhoods
x, = (") and x, = [ where ina particular city. Each year since
TIF
2015, 8% of people currently living in
te R you arbitrarily choose t= 1 (for

Coty
neighbourhood A move to neighbourhood
simplicity.)
Band 12% of people currently living in
neighbourhood B move to neighbourhood
A. You may assume any movement other
than between the two neighborhoods
exactly balances those arriving with those
leaving.

What will the population of each neighbourhood be in 2020 assuming the migration rates remain fixed?

1 Let Tbe the transition matrix where T,; represents the probability ofa person moving from
neighbourhood 7 to neighbourhoodj. Explain why the transition matrix is
Therefore: 1 A B
The 5 can be taken out as a factor
_ A(0.92 0.12% &8Qo

if aR

LNT E NN
or placed inside one of the matrices, &B (33 0.88 , 3
aEa
either is acceptable. €
e»a

waft 2p olesse
2 The populations of neighbourhoods A and B are 24,500 and 45,200 respectively at the beginning of

LPPEL
. . 0.92 0.12 \/ 24500
2015. After 1 year the population can be expressed as the matrix P, = TP, =
0.08 0.88 45200
Whatis the population of neighbourhood A after 1 year? Neighbourhood B?

rel ale slftscs Let S, = [ee

#
. The population after 7 years is given by the expression T'S...

sel ae oles 2,
Find the eigenvalues and eigenvectors for T.

4 Diagonalize T and use the relationship T”= PD”P"! to show that

pif? LY! 0 V0.2 0.2%
“(2 -1)l0 0.8" 0.4 -0.6

Return to transition matrices 5 Asn—c, whatwill D" = (

ro) .
approach? What will T” approach?
In section 9.5 you saw how to find the steady state of a dynamical system 0 0.8",
by considering high powers of the transition matrices or by solving the 6 Hence find the long-term population of the neighbourhoods.
equation Tu = u where u is the steady state probability vector. The administration of both neighborhoods would like a formula to tell them how many residents are
This section will show how the same results can be obtained using expected to be in each community 71 years after 2015.
diagonalization. Though this is a initially a longer process, once the ? Multiply out your expression found in question 3 to find a suitable formula for each of the school boards.
diagonalised form has been found it is much easier for you (or for
computers) to calculate high powers using this than by performing 8 Use your formula to write down the populations of neighborhood A in 2020.
successive multiplications.
An additional benefit is that it also gives you an easy way to find a
formula for the state of the system after n transitions.

418 419
 ea eer Cea. ANALYSING DATA

© ° If A has dimensions m x n and B has dimensionsp X q then

BEL Ty
2 1) © the product
AB is defined only if the number of columns in A is equal to the number of rows in B
1 Given that eigenvalues of R = [ > 3 “are 2 (thatis, 1 =p)
FA: company S keeps 3 of its customers while
© and when the product does exist, the dimensions of the product is m x q.
4, =-4 and A, =—1 write R in the form
; of them switch to company R. At the Properties of multiplication for matrices

Coty
R=PDP".
16 -35° For matrices A, B, and C:
2 Consider A= ‘ beginning of the 2005, company R had
6 -13, © Non-commutive AB # BA
6500 customers while company S had 5200
a Determine the eigenvalues and customers. © Associative property A(BC) = (AB)C
eigenvectors of A. State why A is © Distributive property A(B + C)=AB + ACand (B+ C) A=BA+CA
a Write down a transition matrix T
diagonalizable.
representing the proportion of the These properties only hold when the products are defined.
b Express A in the form of A= PDP". customers moving between the two
Find a general expression for A” in terms companies.
of n. The multiplicative identity ofa square 7 x 1 matrix
A is given by I, = . Note that
b Find the distribution of the market after
d_ Use your expression in part ¢ to find A‘. two years. Describe the change in this 00 1,
Verify your result using the matrix utility distribution from the 2005.
bothA x I, =A andI_ x A = A hold only in the case where
A is a square matrix.
on the GDC. ¢ Write Tin the form T= PAP!
0.4 0.75 If A is a square matrix then A‘ = Ax Ax-x A, &8Qo
3 Consider matrix T = (

LNT E NN
0.6 0.25)
Show that 7” = 1|8+9p" 8-9p" k factors of A 3
d aEa
a Express Tin the form T= PDP". 17(9-9p" 9+8p", a0 kg) I
In the case where b=c=0 then A‘ = =|? 7° »J

LPPEL
Find T* using your answer in part a. Verify 5 Od od,
where p= “Do a
0.4 0.75) The matrices which represent rotations, reflections, enlargements and one way stretches are all given
our result by findin; usin;
M4 Y «(oe oa] 8 e Hence, find an expression for the number in the IB formula book.
the matrix utility on the GDC. of customers buying from R after n years. A Markov Chain is a system in which the probability of each event depends only on the state of the
¢ Using the result from part b determine f Verify your formula by finding how previous event.
the long-term behavior of T. many customers are purchasing from Atransition matrix is a square matrix which summarizes a transition state diagram. It describes the
R after two years. probabilities of moving from one state to another in a stochastic system.
4 Suppose only two rival companies, R and
S, manufacture a certain product. Each year, g Find the long-term number of customers © The sum of each column of the transition matrix is 1.
1 buying from R.
company R keeps z of its customers while © The transition matrix is extremely useful when there are more than two states in the system and
when you wish to predict states of the system after several time periods.
2 of them switch to company S. Each year, Aregular transition matrix T has the property that there exists ne Z* such thatall entries in 7” are
greater than zero.
The initial state vector S, shows the initial state of the system.

@
If Pis a square regular transition matrix, there is a unique probability vector q such that Pq = q. This q

Chapter summary is called the steady-state vector for P.

a b>
© Toadd or subtract two or more matrices, they must be of the same order. You add or subtract LetA = ( containing real number entries and having distinct real eigenvalues uF and 4, then
c /
corresponding elements.
you say that A is diagonalizable and can thus be written in the form ofA= PDP! where P= (x, x,) is
© Given a matrixA and a real number k then KA is obtained by multiplying every element of A by k
where Kis referred to as a scalar. , .
the matrix of eigenvectors and D =
4, 0 \
L
e \fP=AB then each element ofP (named as P.) is found by summing the products of the elements in 0 4)
row i of A with the elements in columnj of B. Let A be a diagonalizable 2 x 2 matrix expressed in the form of A= PDP"! then
a 0
oO A” =PD"P“! where P= (x, x,) and D = rs
A zy

420 421
 DELLING WITH MATRICES: STORING AI Er ING DATA

Chapter review
5, Xn Xo ‘
Click here fora mixed y b Given that =C, a Find A”. (2 marks)
review exercise Vn Voy
b Hence find the matrix C such that

BEL Ty
1 -1 21 i write down C, AC=B. (3 marks)
5 -1 3 3 1 1 2 -3)
1 IfA= | B=|-1|,C= D= }E=/-2 1 O|F= find ii find C,, c,, and Cc,
-2 2 2 4 0 -5 -1 4 -2) Pa: Consider the following system of
-3 -1 4 iii find the sum of the areas of 33
equations:
Ty TT, —

Coty
each of the following without using the matrix utility on the GDC. If it does not exist state why it 5x + 3z= 23
does not.
a 2C+3D b A-B c DF d FE
Exam-style questions x—2y+5z7= 23
3y + 7z= 122
e FE? f (C+D)? g EB+B h ct \
8 Pi: Given
A= 5 and B= 4 0 a Write the system in the form AX = B,
i CDA j c-21, . 2 -3, 1 -2 where A, X and B are matrices.
find the following matrices: (2 marks)
a 2B marks) (2
b Find the matrix A7!. (2 marks)
2 Determine the values of a and b such that b Determine the coordinates of P’, Q’, R’, b A-B marks) (2
and Ss’. c AB marks) (2 c Hence solve the original system of
3a 2 wl 2b) 7g 6) d (A+B)? marks) (2 equations. (3 marks)
+ 4, a |= . ce Area of P’Q’R’S’ = k x Area of PQRS.
-1 2b 1 -£] (3 -9) Determine the value of k. -1) .
Pa: At a particular opening night screening
2
9 P4: Given that the matrix 4 =| * l is a4
3 Solve each equation for X. 6 The linear transformation 2 4 3-x,) - of Galaxy Wars 2, one cinema had a &8Qo

LNT E NN
singular, determine the possible value(s) full attendance of 750 people. It 3
2 1 3.05 x! -3 1\x 4 -1 th nt aEa
fae 3x— = = maps Pp: the potpoints of x. (5 marks) charged £8 for adults, £5 for children,
1 2 -7 0 y' -1 ahy 2
and £4 for OAPs. »J
I

1 3 3 1)

LPPEL
3 -5 2° (x,y) to (x,y). Find The number of OAPs attending was a
b |X =
-1 2 1 -10, a the coordinates of the image of (—3, 5) to pi:P=| 2 7? landg=|? 2. one-fifth of the total attendance.
b the coordinates of the point who image . Bou 1 3 The total amount received in entrance
@ ie
BF -l
+31, =
5 4 is (-7,0) 2 2 2 2) tickets was £4860.
at 2 2 6 ¢ the coordinates of the image of (a, 2a) a Find the matrix (PQ)>. (3 marks) Use a matrix method to determine the
4 Write each system as a matrix equation then where ae R. b Hence, find the smallest value of n number of children that attended the
solve the system using matrix algebra (if ? A series of triangles Ty T,, Ty... are formed n_(-l 0% screening. (9 marks)
possible). Write your answers in exact form. such that (PQ) -( 0-1:
by rotating each consecutive triangle anti- -1) ; 4 -2
clockwise by @ then enlarging the triangle Consider the matrix A-( i }
w=3+2y+2z (2 marks) 3 Pes
Ty =4x-20 by a factor of k. =
a b 3y-4z7=6-w 41% =
3x =2y+10 41 P2:a Show that the matrix 4 = a Find the eigenvalues and
A4w+5y-3=2z == 3 6, corresponding eigenvectors of A.
satisfies the equation (7 marks)
e 6x =5y+18
A? — 10A + 211= 0, where J is the b Hence find a matrix P and a matrix
27-9x =-7.5y 2 x 2 identity matrix. (4 marks) D such that D = P-!AP. (2 marks)
5 Parallelogram PQRS with vertices P(—4, —2), c Find a general expression for A” in
b Hence express A? in the form
Q(-1,7), R(8,4), T(5,—5) is translated by the terms of n. (5 marks)
pA+al. (3 marks)
v
¢ Hence express A‘ in the form
vector(5 then enlarged by a factor of i
J 2 a Given that the series of transformations TA + sI. (3 marks)
3 2 -1)
then reflected across the x-axis. can be described by [*: )- a 12 P41: The matrix A=|0 1 2 and
a Determine a single transformation of the Vn Vary =
form of AX + b that maps PQRS to its determine - 113,
image P’Q’R’S’. i the 2 x 2 matrix A 10 1%
ii the exact coordinates of the vertices B=|0 1 0
of T,. 101

422 423
 M a rC 0 m m p h 0 n e Ss oe Communication,

You will now find the stationary distribution using three different methods.
1 Exploration criteria: Mathematical
and Markov chains communication (B), Use of mathematics (E) Using technology
1B topic: Markov chains, Matrices Calculate X,, and X,, (that is P'’x, and P”°x,) using a calculator or computer.
2 This will give an indication of the convergence to a stationary distribution (to =
Po a reasonable degree of accuracy). a
S Example £
Why would this not be considered a sophisticated approach to the problem?
S
a=)
Aphone company, MarComm, has four package options, roo
+ A, B, Cand D. Solving a system of equations }
oo= Acustomer is initially allocated to package A for a year. =a
7) This method uses the fact that once the stationary distribution, 1, is reached,
g After this, every year, customers can either continue with multiplying by the transition matrix has no effect > Pr=n. Fs
‘3 their present option or change to one of the other options. 7, | n,| [x] 3
Ci]
a= The probabilities of each possible change are given in this . Tt, e Tt, 3=
transition state diagram: So, by letting 7 = |’ You can solve the system of equations ‘P} a| ole
FS= J =]
3 3 3 5
2
H
m, | m) |] 8Es 3
> Use this method with the information in the example to calculate m and to ra
verify that the result is equivalent to the answer obtained using the previous
method using technology.
Does this feel more sophisticated than the previous process using technology?
Justify your answer.

Using eigenvalues and eigenvectors

If the stationary distribution 7 is such that Px = 1, then as you have seen in
How would MarComm be able to determine the probabilities on the diagram? this chapter the matrix P must have an eigenvalue of 1 with corresponding
Represent this information in a transition matrix, P, of the system. eigenvector 7 scaled so that the sum of its elements is 1.
Calculate the eigenvalues of P from the example and the corresponding
If customers are all initially allocated to option A for the first year, then write down an initial state vector, Boy
eigenvectors (using technology).
wl How could you demonstrate that you understand fully what you are finding
in the form X, = 4 . here and that you understand the process that is taking place beyond just
y getting the correct answers?
z] How will you organize the mathematical calculations so that the
The probabilities for the next state can then be found by multiplying x, by the transition matrix P. process from posing the problem to answering the problem is logical?
Calculate the probabilities of being in each option after one year if the customer starts with option A. These methods give MarComm 7, the probabilities of a customer being in each
al of the four options in the long term.
b How could the company use this information?
Write down these probabilities in the form x, = al Clearly all three methods produce the same approximate vector 7 correct to
| four significant figures.
However, which method do you prefer?
How might MarComm find this information useful? Which method is the most sophisticated? Justify your answer.
Repeating this iteratively will give the probability of a customer being in any state in subsequent years.
Investigate the behaviour of the system over time. Extension
m1 Does a system always reach a
1.
Let 7 be the stationary distribution of the system, where z = \. etal
a How quickly will the process
{ , | y converge to the steady state?

424 425