Академический Документы
Профессиональный Документы
Культура Документы
C H A P T E R
27
MODULE 6
Matrices and
E
applications II
PL
What is an identity matrix?
What is the inverse of a matrix and how is it determined?
How do we use inverse matrices to solve sets of simultaneous equations?
How do we define the power of a matrix?
What is a transition matrix and how is it used?
Just as the product of a number and one, is the same number (1 × 3 = 3 × 1 = 3), the product
of a square matrix and the identity matrix is the same matrix, IA = AI = A.
Cambridge University Press • Uncorrected Sample pages • 978-0-521-61328-6 • 2008 © Jones, Evans, Lipson
722 and McMenamin
TI-Nspire & Casio ClassPad material in collaboration with Brown
P1: FXS/ABE P2: FXS
9780521740517c27.xml CUAU031-EVANS September 7, 2008 7:15
Solution
1 For a 2× 2 matrix, the identity matrix
1 2 1 0
1 0 A = I=
is I = . Write down A and I. 3 4 0 1
0 1
E
2 Form the product AI and evaluate. 1 2 1 0
Use a calculator if you wish. AI =
3 4 0 1
1×1+2×0 1×0+2×1 1 2
= =
3×1+4×0 3×0+4×1
PL
3 Form the product IA and evaluate.
Use a calculator if you wish.
∴ AI = A
IA =
=
1 0
0 1
1 2
3 4
1×1+0×3 1×2+0×4
0×1+1×3 0×2+1×4
∴ IA = A
=
1
3
3 4
2
4
M
The inverse matrix A−1
Having defined the properties of the identity matrix, we are now in the position to define a new
matrix concept, the inverse matrix.
The inverse of a square matrix A, is called A−1 and has the property that
A A−1 = A−1 A = I .
Before we investigate how to calculate an inverse matrix, we need to show you that they
actually exist.
Solution
2 3 −1 5 −3
1 Write down A and A−1 . A= A =
3 5 −3 2
Cambridge University Press • Uncorrected Sample pages • 978-0-521-61328-6 • 2008 © Jones, Evans, Lipson
TI-Nspire & Casio ClassPad material in collaboration with Brown and McMenamin
P1: FXS/ABE P2: FXS
9780521740517c27.xml CUAU031-EVANS September 7, 2008 7:15
2 Form the product A A−1 2 3 5 −3
and evaluate. AA−1 =
3 5 −3 2
Use a calculator if you
2 × 5 + 3 × (−3) 2 × (−3) + 3 × 2 1 0
wish. = =
3 × 5 + (−3) × 5 3 × (−3) + 5 × 2 0 1
AA−1 = I
3 Form the product A−1 A 5 −3 2 3
and evaluate. A −1 A =
−3 2 3 5
E
Use a calculator if you
5 × 2 + (−3) × 3 5 × 3 + (−3) × 5 1 0
wish. = =
(−3) × 2 + 2 × 3 (−3) × 3 + 2 × 5 0 1
∴ A −1 A = I
Like matrix multiplication, finding the inverse of a matrix generally requires long and tedious
If A =
a b
PL
computations. Fortunately, a graphics calculator will do it for us. However, not all square
matrices have inverses. To see why, we need to introduce another new matrix concept, the
determinant of a square matrix and see how it is related to finding the inverse of a square
matrix. To keep things manageable, we will restrict ourselves initially to 2 × 2 matrices.
of a 2 × 2 matrix
The determinant
, then the determinant of matrix A, written as det (A) or |A|, is given by
M
c d
a b
det (A) = =a×d −b×c
c d
Find thedeterminant
of the matrices:
2 3 2 3 2 4
a A= b B= c C=
3 5 2 3 2 3
Solution
1 Write down the matrix 2 3
2 3
a A= ∴ det ( A ) = =2×5−3×3=1
and use the rule 3 5 3 5
a b
det (A) = 2 3
c d 2 3
b B = ∴ det ( B ) = =2×3−2×3=0
= a × d − b × c. 2 3 2 3
2 Evaluate. 2 4
2 4
c C = ∴ det (C ) = = 2 × 3 − 2 × 4 = −2
2 3 2 3
Cambridge University Press • Uncorrected Sample pages • 978-0-521-61328-6 • 2008 © Jones, Evans, Lipson
TI-Nspire & Casio ClassPad material in collaboration with Brown and McMenamin
P1: FXS/ABE P2: FXS
9780521740517c27.xml CUAU031-EVANS September 7, 2008 7:15
From Example 3, we can see that the determinant of a matrix is a number that can take on both
positive and negative values as well as being zero. As we will now see, for a matrix to have an
inverse, its determinant must be non-zero.
of a2 × 2 matrix
The inverse
a b
If A = , then its inverse A−1 , is given by
c d
E
1 d −b 1 d −b
A−1 = =
ad − bc −c a det (A) −c a
1
provided = 0, that is, provided det (A) = 0.
ad − bc
Example 4
a A=
Solution
2 2
3 4
PL
Finding the inverse of a 2 × 2 matrix
M
1 Write down the matrix a 2 2
A =
and use the rule 3 4
a b
2 2
det (A) =
c d ∴ det ( A ) = =2×4−3×2=2
3 4
= a × d − b × c.
Evaluate.
SA
2 Write down the matrix and use the b 2 3
a b B =
2 3
rule det (B) = = a × d − b × c.
c d 2 3
Evaluate. ∴ det ( B ) = =2×3−2×3=0
2 3
As the determinant of the matrix is zero,
Inverse not defined
the matrix has no inverse.
Cambridge University Press • Uncorrected Sample pages • 978-0-521-61328-6 • 2008 © Jones, Evans, Lipson
TI-Nspire & Casio ClassPad material in collaboration with Brown and McMenamin
P1: FXS/ABE P2: FXS
9780521740517c27.xml CUAU031-EVANS September 7, 2008 7:15
3 Write down the matrix and c 2 4
use the rule C =
2 3
a b
2 4
det (C) =
c d ∴ det (C ) = = 2 × 3 − 2 × 4 = −2
2 3
= a × d − b × c.
Evaluate. −1
1 d −b
∴C =
Use the formula det (C ) −c a
1 d −b 1 3 −4 −1.5
C −1 = =
=
2
det (C) −c a −2 −2 −1
E
2 1
to evaluate.
PL
There are rules for finding the inverse of a square matrix of any size, but they are extremely
complicated and take huge amounts of time to compute by hand. So, in practice, we use a
calculator to find the inverse of all but a (2 × 2) matrix. In fact, you can use a calculator to
compute the inverse of a (2 × 2) matrix if you wish, but it is often just as quick to do it by
hand. The same goes for calculating determinants, although the determinant of a (2 × 2)
matrix is much more quickly computed by hand.
How to find the determinant and inverse of a matrix using the TI-Nspire CAS
M
1 2 3
If A = 4 1 0 , find det (A) and A−1 .
2 0 2
Steps
SA
Cambridge University Press • Uncorrected Sample pages • 978-0-521-61328-6 • 2008 © Jones, Evans, Lipson
TI-Nspire & Casio ClassPad material in collaboration with Brown and McMenamin
P1: FXS/ABE P2: FXS
9780521740517c27.xml CUAU031-EVANS September 7, 2008 7:15
Notes:
1 When the elements in the matrix to be
E
inverted are whole numbers, the elements
of the inverse will always be whole
numbers or fractions. If this is the case,
and the inverse you obtain with your
calculator contains decimals, it is worth
PL
converting it to fractional form.
Although the mode is set to APPRX (or
decimal), we can make the display show
fractions using exact( . . ..) in front of the
command. exact( . . . ) can also be pasted
from the Catalog ( ).
2 If the matrix has no inverse, the
calculator will respond with the error
message Singular matrix.
M
How to find the determinant and inverse of a matrix using the ClassPad
1 2 3
SA
If A = 4 1 0 , find det (A) and A−1 .
2 0 2
Steps
1 Enter the matrix A into your
calculator.
Note: Change the status of the
calculator to Standard in order for
fractions to be displayed. Tapping on
Decimal will change the calculator to
Standard.
Cambridge University Press • Uncorrected Sample pages • 978-0-521-61328-6 • 2008 © Jones, Evans, Lipson
TI-Nspire & Casio ClassPad material in collaboration with Brown and McMenamin
P1: FXS/ABE P2: FXS
9780521740517c27.xml CUAU031-EVANS September 7, 2008 7:15
E
3 To calculate the inverse
matrix A−1 , type A ∧ − 1.
PL
Press E to evaluate.
Note: If the matrix has no inverse,
the calculator will respond with the
message Undefined.
M
Exercise 27A
1 a Write down the:
SA
2 Show that each of the following pairs of matrices are inverses by multiplying one by the
other. Use a calculator if you wish.
1 1 2 −1 2 2 2 −1
a and b and
1 2 −1 1 3 4 −1.5 1
9 7 −3 7 0 2 0 0.5
c and d and
4 3 4 −9 2 0 0.5 0
Cambridge University Press • Uncorrected Sample pages • 978-0-521-61328-6 • 2008 © Jones, Evans, Lipson
TI-Nspire & Casio ClassPad material in collaboration with Brown and McMenamin
P1: FXS/ABE P2: FXS
9780521740517c27.xml CUAU031-EVANS September 7, 2008 7:15
E
−2 4 0 3 1 4 4 2
1 0 0 −2
i I = j J=
0 1 −2 4
a A=
1.1 2.2
0 3.0
1 1 0
d D = 1 1 0
0 1 1
PL
b B=
0.2 −0.1
10
2 1
e E = 0 1 −1
0 0
4
0
1
c C= 2
1
2
1
2
4
4x + 2y = 5
SA
3x + 2y = 2
A−1 AX = A−1 C
or I X = A−1 C
or X = A−1 C
Cambridge University
Note: The order ofPress • Uncorrected
multiplication Sample here.
is important A−1 C not C A−1 . • 2008 © Jones, Evans, Lipson
pagesIt•is978-0-521-61328-6
TI-Nspire & Casio ClassPad material in collaboration with Brown and McMenamin
P1: FXS/ABE P2: FXS
9780521740517c27.xml CUAU031-EVANS September 7, 2008 7:15
−1 1 −1 x 5
Noting A = and substituting X = and C = ,
−1.5 2 y 2
x 1 −1 5
=
y −1.5 2 2
1 × 5 + (−1) × 2 3
= = (or use a calculator)
(−1.5) × 5 + 2 × 2 −3.5
x 3
∴ = or x = 3 and y = −3.5
E
y −3.5
Thus, using the matrices, the process of solving a pair of simultaneous linear equations can be
reduced to finding the inverse of a matrix and performing a matrix multiplication. While there
is not a lot to gain over algebraic methods for a pair of simultaneous equations, the same
PL
matrix method can be used to solve three, four or more simultaneous linear equations. This is
an extremely difficult task using algebraic methods.
Before we go on to practise the method, it is important for you to know when it does not
work and why. First, we need to recognise that not all pairs of simultaneous linear equations
have solutions. This can happen for two reasons:
the equations are inconsistent, their graphs are parallel and thus do not cross
the equations are dependent, the graphs coincide and thus do not cross at a single point
Inconsistent equations y
M
For example, the pair of simultaneous linear equations
2
x + 2y = 4
2x + 4y = 6 1.5
x + 2y = 4
have no solution. We can see why, by plotting their graphs
2x + 4y = 6
(opposite). When plotted, the two equations are parallel
x
SA
Dependent equations
For example, the pair of simultaneous linear equations
y
x + 2y = 4
2x + 4y = 8
2
x + 2y = 4
have no solution. We can see why, by plotting a graph
(opposite). When plotted, the graphs of the two equations
coincide, with no single point of intersection. 2x + 4y = 8
E
There is no unique solution. Rather, there are an x
O 4
infinite number of possible solutions, all the pairs of
points that lie on the line. When this happens, we
say that the equations are dependent.
If we try to solve these equations using matrix Calculator hint
1 2
2 4
PL
methods by solving the matrix equation:
x
y
=
4
8
.
det(A) =
1 2
2 4
= 1 × 4 − 2 × 2 = 0.
To construct this plot on a graphics
calculator, rewrite the equations in
the form y = mx + c.
x + 2y = 4
or 2y = 4 − x or y = 2 − 0.5x
2x + 4y = 8
or 4y = 8 − 2x or y = 2 − 0.5x
M
As a general rule, if a pair of linear equations is
dependent, det(A) = 0 and the equations have
no unique solution.
SA
X = A−1 C
Note: The order of multiplication is important here: X = A−1 C not CA−1
For the set of equations:
ax + by = e a b x e
A= , X= and C = .
cx + dy = f c d y f
For the set of equations:
ax + by + cz = j a b c x j
d x + ey + f z = k A = d e f , X = y and C = k ,
gx + hy + i z = l g h i z l
and so on.
Cambridge University Press • Uncorrected Sample pages • 978-0-521-61328-6 • 2008 © Jones, Evans, Lipson
TI-Nspire & Casio ClassPad material in collaboration with Brown and McMenamin
P1: FXS/ABE P2: FXS
9780521740517c27.xml CUAU031-EVANS September 7, 2008 7:15
Solution
3 4 x 6
1 Rewrite the equations in matrix form. =
E
2 3 y 4
2 By identifying the matrices, A, X, and
3 4 x 6
C, rewrite the matrix equation in the Let A = , X = , C =
2 3 y 4
form AX = C.
Then AX = C or X = A −1 C (det ( A ) = 0)
PL
3 Provided det (A) = 0, the solution
X = A−1 C.
However, not all pairs of simultaneous linear equations have unique solutions, as we will see in
SA
3x + 4.5y = 9
2x + 3y = 4
Cambridge University Press • Uncorrected Sample pages • 978-0-521-61328-6 • 2008 © Jones, Evans, Lipson
TI-Nspire & Casio ClassPad material in collaboration with Brown and McMenamin
P1: FXS/ABE P2: FXS
9780521740517c27.xml CUAU031-EVANS September 7, 2008 7:15
Solution
1 Rewrite the equations in matrix form.
3 4.5 x 9
=
2 3 y 4
2 By identifying the matrices, A, X,
3 4.5 x 9
and C, rewrite the matrix equation Let A = , X = , C = .
2 3 y 4
in the form AX = C.
3 Provided det (A) = 0, the solution Then AX = C or X = A −1 C (det ( A ) = 0)
in matrix form is X = A−1 C.
E
In this case:
3 4.5
det (A) = = 0,
2 3
but let us see what happens.
your calculator.
PL
4 Enter the matrices A and C into
3x + 4y − 2z = −5
2x + 3y = −1
x + 2y + 3z = 3
Solution
1 Rewrite the equations in matrix form. 3 4 −2 x −5
2 3 0 y = −1
1 2 3 z 3
2 By identifying the matrices, 3 4 −2 x −5
A, X, and C, rewrite the matrix Let A = 2 3 0 , X = y , C = −1
equation in the form AX = C. 1 2 3 z 3
Note: There is no z term in the second equation, so its coefficient is zero.
Cambridge University Press • Uncorrected Sample pages • 978-0-521-61328-6 • 2008 © Jones, Evans, Lipson
TI-Nspire & Casio ClassPad material in collaboration with Brown and McMenamin
P1: FXS/ABE P2: FXS
9780521740517c27.xml CUAU031-EVANS September 7, 2008 7:15
E
matrix product A−1 C.
Example 8
PL
6 Write down your answer.
Application
−11
Therefore, X = 7
0
or x = −11, y = 7 and + z = 0.
M
A manufacturer makes two sorts of orange flavoured chocolates, House Brand and Orange
Delights. The number of kilograms of House Brand (x) and the number of kilograms of
Orange Delights (y) that can be made from 80 kg of chocolate and 120 kg of orange fill can be
found by solving the following system of equations:
0.3x + 0.5y = 80
SA
Solution
1 Rewrite equations in matrix form.
0.3 0.5 x 80
=
0.7 0.5 y 120
2 By identifying the matrices, A, X, 0.3 0.5 x 80
and C, rewrite the matrix equation
Let A = ,X = ,C = .
0.7 0.5 y 120
in the form AX = C.
3 Provided det(A) = 0, the Then AX = C
solution X = A−1 C. or X = A −1 C (det(A ) = 0)
Cambridge University Press • Uncorrected Sample pages • 978-0-521-61328-6 • 2008 © Jones, Evans, Lipson
TI-Nspire & Casio ClassPad material in collaboration with Brown and McMenamin
P1: FXS/ABE P2: FXS
9780521740517c27.xml CUAU031-EVANS September 7, 2008 7:15
5 Write down your answer. 100
Therefore, X = or x = 100 and y = 100.
100
E
Exercise 27B
a 3x + 2y = 2
g
2x + 5y = 4
x − 3y = 7
−2x + y = 4
5x − 2z = 3
x −y+ z=2
x +y+ z=1
PL
1 Write each of the following systems of linear equations in matrix form.
b 3x + 5y = 6
h
2x + 4y = 3
e −3x − 2y = 2
x + 2y = −1
x
2x
+ y
− y
− 2z
+ z
+ w
− w
=3
=2
c x + 2y = 1
2x − 3y = 2
f 3x + 4y − 2z = 5
2x + 3y + 5z = 2
x + 2y + 3z = 3
x + 2y + z + w =1
M
2x − 3y + 2z − 2w =0
2 Give two explanations of how a system of two linear equations can have no unique solution.
3 What is the condition for the matrix equation AX = C not to have a unique solution?
3 1 x 2 4 2 x 2 −1 1 x 0
a = b = c =
2 1 y 1 2 1 y 1 2 −1 y 1
0 −1 x 2 2.5 1 x 2 1 1 x 0
d = e = f =
2 1 y 4 5 2 y 1 2 2 y 2
1 2 3 x 1
−1 1 x 0 5 2 x 10
g = h = i 2 4 6 y = 0
2 −1 y 1 2 1 y 4
1 1 1 z 1
1 0 3 x 1 −1 0 2 x 2
j 0 1 0 y = 0 k 0 1 1 y = 1
0 1 1 z 1 0 1 1 z 3
Cambridge University Press • Uncorrected Sample pages • 978-0-521-61328-6 • 2008 © Jones, Evans, Lipson
TI-Nspire & Casio ClassPad material in collaboration with Brown and McMenamin
P1: FXS/ABE P2: FXS
9780521740517c27.xml CUAU031-EVANS September 7, 2008 7:15
E
1 1 1 z 2
1 0 3 x 1 −1 0 2 x 3
i 0 1 0 y = 5 j 0 1 1 y = 2
2 0 6 z 1 0 1 −1 z −1
6 Solve the following systems of linear equations using matrix methods. (Not all have
solutions.)
a 3x − 5y = 9
e
2x − 3y = 12
PL b x + 3y = 4
x + 5y = 10
c x + 3y = 4
2x + 6y = 5
x + 4y − 2z = 8 f x + y − z = 0 g x + y − z = 2 h −2y + z + w =−5
2x + 3y + 2z = 4
x + 2y − z = 3
x + y+z=2
x + 2y − z = −3 x−
y+z=1
z=2
x + y + z − w =2
x+ y+ 2w =0
x + y + z + w =0
d 3x − 2y = 4
10x − 6y = 9
7 An outdoor clothing manufacturer makes two sorts of jackets, Polarbear and Polarfox. The
M
number of Polarbear jackets (x) and the number of Polarfox jackets (y) that can be made
from 150 metres of fabric and 150 hours of worker time can be found by solving the
following system of equations:
2x + 2y = 150
2.4x + 1.8y = 150
SA
Cambridge University Press • Uncorrected Sample pages • 978-0-521-61328-6 • 2008 © Jones, Evans, Lipson
TI-Nspire & Casio ClassPad material in collaboration with Brown and McMenamin
P1: FXS/ABE P2: FXS
9780521740517c27.xml CUAU031-EVANS September 7, 2008 7:15
E
we define
A2 as A × A,
A3 as A × A × A,
A4 as A × A × A × A and so on.
PL
Only square matrices can be raised to a power.
Finally, just as we define 20 to be 1, A0 is defined to be I, the identity matrix.
A2 = A × A
A3 = A × A × A and so on.
M
A0 is defined as I, the appropriate identity matrix.
Solution
2 3
1 Write down the matrix. A =
4 3
2 Either:
Evaluate A3 by multiplying out. A3 =
A × A × A
Write down your answer. 92 93
∴ A3 =
124 123
or:
Enter the matrix into your calculator. A ∧3
A3 =
Evaluate A3 by raising to a 92 93
power. Write down your answer. ∴A =
3
124 123
Cambridge University Press • Uncorrected Sample pages • 978-0-521-61328-6 • 2008 © Jones, Evans, Lipson
TI-Nspire & Casio ClassPad material in collaboration with Brown and McMenamin
P1: FXS/ABE P2: FXS
9780521740517c27.xml CUAU031-EVANS September 7, 2008 7:15
Solution
1 0 −1 1 0 1
1 Write down the matrices A = , B = C =
2 −1 2 1 1 1
E
2 Enter the matrices A, B and C
into your calculator.
c AB2 − 3C 2 =
3 −9
Exercise 27C
2 1
1 If A = determine A2 , A3 , A4 and A7 .
1 3
−1 1
2 If A = determine A4 , A5 , A6 and A7 .
1 2
2 1 −1 1 0 1
3 If A = , B= and C = , evaluate
1 3 1 2 1 −2
a A + 2B − C 2 b AB − 2C 2 c (A + B + 2C)2
d 4A + 3B 2 − C 3 e (A − B)3 − C 3
Cambridge University Press • Uncorrected Sample pages • 978-0-521-61328-6 • 2008 © Jones, Evans, Lipson
TI-Nspire & Casio ClassPad material in collaboration with Brown and McMenamin
P1: FXS/ABE P2: FXS
9780521740517c27.xml CUAU031-EVANS September 7, 2008 7:15
B = Bendigo
20%
E
C = Colac
80% B 90%
C
10%
0.8 (or 80%) of cars rented each week in Bendigo are returned to Bendigo
0.2 (or 20%) of cars rented each week in Bendigo are returned to Colac
0.1 (or 10%) of cars rented each week in Colac are returned to Bendigo
0.9 (or 90%) of cars rented each week in Colac are returned to Colac
The percentages (written as proportions) are summarised in the form of the matrix below.
rented in
M
Bendigo
Colac
Bendigo 0.8 0.1
returned to
Colac 0.2 0.9
This matrix is an example of a transition matrix. It describes the way in which transitions are
made between two states:
SA
Cambridge University Press • Uncorrected Sample pages • 978-0-521-61328-6 • 2008 © Jones, Evans, Lipson
TI-Nspire & Casio ClassPad material in collaboration with Brown and McMenamin
P1: FXS/ABE P2: FXS
9780521740517c27.xml CUAU031-EVANS September 7, 2008 7:15
Solution rented in
1 There are three locations from which the cars can A W B
be rented and returned, Albury (A), Wodonga (W) A
and Benalla (B). To take into account all the returned to W
possibilities, a 3 × 3 transition matrix is needed. B
Constuct a blank matrix labelling the rows and
columns A, B and C respectively. Column labels
indicate where the car was rented. The row
E
labels indicate where the cars were returned to.
2 Complete the matrix by writing each of the
A W B
percentages (converted to proportions) into the A 0.7
appropriate locations in the matrix. Start with W 0.1
Column A and write in values for each row: B 0.2
PL
0.7 (70%), 0.1 (10%) and 0.2 (20%).
3 Mentally check your answer by summing columns;
they should sum to one.
A factory has a large number of machines. Machines can be in one of two states, operating or
M
broken. Broken machines are repaired and come back into operation and vice versa. On a
given day:
85% of machines that are operational stay operating
15% of machines that are operating break down
5% of machines that are broken are repaired and start operating
95% of machines that are broken stay broken.
SA
Construct a transition matrix to describe this situation. Use the columns to define the situation
at the ‘Start’ of the day and the rows to describe the situation at the ‘End’ of the day.
start
Solution
O B
1 There are two machine states, operating (O) or broken (B). O
To take into account all the possibilities, a 2 × 2 transition end
B
matrix is needed. Constuct a blank matrix, labelling the rows
and columns O and B respectively.
2 Complete the matrix by writing each of the percentages O B
O 0.85
(converted to proportions) into the appropriate locations
B 0.15
in the matrix. Start with Column O and write in values for
each row: 0.85 (85%) and 0.15 (15%).
3 Mentally check your answer by summing columns; they O B
O 0.85 0.05
should sum to one.
B 0.15 0.95
Cambridge University Press • Uncorrected Sample pages • 978-0-521-61328-6 • 2008 © Jones, Evans, Lipson
TI-Nspire & Casio ClassPad material in collaboration with Brown and McMenamin
P1: FXS/ABE P2: FXS
9780521740517c27.xml CUAU031-EVANS September 7, 2008 7:15
B = Bendigo
20%
C = Colac
E
80% B 90%
C
10%
The car rental firm now plans to buy 90 new cars and base 50 in Bendigo and 40 in Colac.
PL
Given this pattern of rental car returns, the questions the manager would like answered are:
‘If we start off with 50 cars in Bendigo, and 40 cars in Colac, how many cars will be
available for rent at Bendigo and Colac after one week, two weeks, etc?’
‘What will happen in the long term? Will the numbers of cars available for rent each week
from each location vary from week to week or will it settle down to some fixed value?’
Using our knowledge of matrices, we can answer both these questions.
The transition matrix that describes how the cars move between Bendigo and Colac is
0.8 0.1
T =
M
0.2 0.9
After one week, the number of cars at each branch will change.
Construct a new state matrix S1 to show the number of cars at each branch after one week.
Then,
S1 = T S0
0.8 0.1 50
=
0.2 0.9 40
0.8 × 50 + 0.1 × 40
=
0.2 × 50 + 0.9 × 40
44
or S1 =
46
Thus, after one week there will be 44 cars in Bendigo and 46 in Colac.
Cambridge University Press • Uncorrected Sample pages • 978-0-521-61328-6 • 2008 © Jones, Evans, Lipson
TI-Nspire & Casio ClassPad material in collaboration with Brown and McMenamin
P1: FXS/ABE P2: FXS
9780521740517c27.xml CUAU031-EVANS September 7, 2008 7:15
Thus, after two weeks (theoretically) there will be 39.8 cars in Bendigo and 50.2 in Colac.
After three weeks
0.8 0.1 39.8 36.9
S3 = T S2 = =
E
0.2 0.9 50.2 53.1
Thus, after three weeks (theoretically) there will be 36.9 cars in Bendigo and 53.1 in Colac.
If we keep on with the process we find the following:
PL
...
30.4
59.6
30.3
59.7
30.2
59.8
30.1
59.9
30.1
59.9
What you should notice is that as the weeks go by, the number of cars at each of the
starts to settle down to what we call the steady state solution given by the matrix
locations
30.1
59.9
The theoretical steady state solution is 30.1 (in practice 30) cars at the Bendigo branch and
59.9 (in practice 60) cars at the Colac branch and, it will not change from then on. This can
be seen in the graph (the points have been joined to guide the eye).
.
M
Initial value Bendigo: 50 Steady state value: Colac 60
60
Colac
50
Number of cars
40
SA
Bendigo
30
Initial value Colac: 40 Steady state value: Bendigo 30
20
10
0
1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16
Weeks
Cambridge University Press • Uncorrected Sample pages • 978-0-521-61328-6 • 2008 © Jones, Evans, Lipson
TI-Nspire & Casio ClassPad material in collaboration with Brown and McMenamin
P1: FXS/ABE P2: FXS
9780521740517c27.xml CUAU031-EVANS September 7, 2008 7:15
Let us return to the factory problem in Example 12. The factory has a large number of
machines. The machines can be in one of two states, operating (O), or broken (B). Broken
machines are repaired and come back into operation and vice versa. On a given day, the
situation is described by the transition matrix. Columns define the machine states at the start of
the day.
O B
E
O 0.85 0.05
T =
B 0.15 0.95
At the start of a particular day, 80 machines are operating and 20 are broken. How many
machines are in operation and how many are broken after:
a one day?
Solution
PL
b three days?
S0 =
0.85 0.05
0.15 0.95
80
20
M
3 To determine the operational state 0.85 0.05 80 69
of the machines after one day, form S1 = T S0 = =
0.15 0.95 20 31
the product S1 = T S0 and evaluate.
4 Write down your conclusion. After one day, 69 machines are operational
and 31 are broken.
SA
Cambridge University Press • Uncorrected Sample pages • 978-0-521-61328-6 • 2008 © Jones, Evans, Lipson
TI-Nspire & Casio ClassPad material in collaboration with Brown and McMenamin
P1: FXS/ABE P2: FXS
9780521740517c27.xml CUAU031-EVANS September 7, 2008 7:15
While we can use repeated matrix multiplication to work out successive states of an evolving
situation such as the car rental problem, it is clearly extremely tedious. Fortunately there is a
more efficient way of doing this and working out what happens in the long term.
If we follow through the process step by step we have:
S1 = T S0
S2 = T S1 = T (T S0 ) = T 2 S0
S3 = T S2 = T (T S1 ) = T 2 S1 = T 2 (T S0 ) = T 3 S0 and so on.
E
Sn = T n S0
We now have a simple rule for determining the state matrix after n steps.
Example 14
Solution
B
0.05
0.95
M
1 Write down the transition matrix T and initial
0.85 0.05 80
state matrix S0 . Enter the matrices into your T = S0 =
0.15 0.95 20
calculator. Use T and S.
2 To find out how many machines are in Sn = T n S0
operation and how many are broken after
10 days, write down the rule Sn = T n S0 and ∴ S10 = T 10 S0
SA
Cambridge University Press • Uncorrected Sample pages • 978-0-521-61328-6 • 2008 © Jones, Evans, Lipson
TI-Nspire & Casio ClassPad material in collaboration with Brown and McMenamin
P1: FXS/ABE P2: FXS
9780521740517c27.xml CUAU031-EVANS September 7, 2008 7:15
Number of cars
40
location varied from day to day, eventually Bendigo
30
it settled down to a steady state solution in 20
Initial value Colac: 40 Steady state value: Bendigo 30
Although we arrived at this conclusion by repeated calculations, we can arrive at the solution
E
much faster.
S = T n S0
PL
as n tends to infinity (∞).
While in practice we cannot evaluate T n for n = ∞, we find that in many situations, values
of n around 15 to 30 will give a very close approximation to the steady state solution.
Solution
0.8 0.1 50
1 Write down the transition matrix T and T = S0 =
0.2 0.9 40
initial state matrix S0 . Enter the matrices
SA
Cambridge University Press • Uncorrected Sample pages • 978-0-521-61328-6 • 2008 © Jones, Evans, Lipson
TI-Nspire & Casio ClassPad material in collaboration with Brown and McMenamin
P1: FXS/ABE P2: FXS
9780521740517c27.xml CUAU031-EVANS September 7, 2008 7:15
5 Write down your answer in matrix 30.6 30.1 30.0
form, then in words. This result S10 = , S15 = , S20 =
59.4 59.9 60.0
agrees with the graphical result
There appears to be a steady state solution
arrived at earlier.
with 30 cars at Bendigo and 60 at Colac.
Exercise 27D
E
1 Complete the following transition matrices:
0.75 0.05 0.90 0.15 0.80 0.50 0.33
a b c d
0.25 0.85 0.20 0.65
2 The diagrams below describe a series of transitions between the states indicated. Construct a
c
40% A
PL
transition matrix that can be used to represent each of these diagrams. Use columns to define
the starting points. Convert the percentages to proportions.
a 60%
55%
70%
B 45%
b
d
70% X
30%
25%
45%
Y 75%
M
10% Y 25% B
15% 45% 20%
60%
15% 23% 35% 20%
X Z A C
30% 30%
55% 65%
22% 15%
3 Two fast-food outlets, Jill’s and Pete’s, are located in a small town. In a given week:
SA
J P
J
Next week:
P
Cambridge University Press • Uncorrected Sample pages • 978-0-521-61328-6 • 2008 © Jones, Evans, Lipson
TI-Nspire & Casio ClassPad material in collaboration with Brown and McMenamin
P1: FXS/ABE P2: FXS
9780521740517c27.xml CUAU031-EVANS September 7, 2008 7:15
c How many of the 800 people do we expect to go to Jill’s in the next week? How many go
to Pete’s?
d How many of the 800 people do we expect to go to Jill’s after five weeks’ time? How
many go to Pete’s?
e In the longer term, how many of the 800 people do we expect to go to Jill’s? How many
go to Pete’s?
4 Imagine we live in a world in which people are either ‘happy’ or ‘unhappy’. However, the
way people feel changes from day to day. In the imagined world:
E
r 90% of people who are happy today will be happy tomorrow
r 10% of people who are happy today will be unhappy tomorrow
r 40% of people who are unhappy today will be happy tomorrow
r 60% of people who are unhappy today will be unhappy tomorrow
b
Tomorrow:
PL
a Construct a transition matrix T of the form
H
U
Today
H
On a given day, out of 2000 people, 1500 are happy and 500 are unhappy. Write down a
column matrix S0 that describes this situation.
M
c The next day, how many people do we expect to be ‘happy’ and how many ‘unhappy’?
d After four days, how many people do we expect to be ‘happy’ and how many ‘unhappy’?
e In the long term, how many people do we expect to be ‘happy’ and how many ‘unhappy’?
5 In another model of this world, people can be ‘happy’, ‘neither happy nor sad’, or ‘sad’.
However, the way people feel changes from day to day. In the imagined world:
r 80% of people who are happy today will be happy tomorrow
SA
r 15% of people who are happy today will be neither happy nor sad tomorrow
r 5% of people who are happy today will be sad tomorrow
r 40% of people who are neither happy nor sad today will be happy tomorrow
r 30% of people who are neither happy nor sad today will be the same tomorrow
r 30% of people who are neither happy nor sad today will be sad tomorrow
r 35% of people who are sad today will be happy tomorrow
r 40% of people who are sad today will be neither happy nor sad tomorrow
r 25% of people who are sad today will be sad tomorrow
a Construct a transition matrix T to describe this situation. Use the columns to define the
situation today and the rows to describe the situation tomorrow.
Cambridge University Press • Uncorrected Sample pages • 978-0-521-61328-6 • 2008 © Jones, Evans, Lipson
TI-Nspire & Casio ClassPad material in collaboration with Brown and McMenamin
P1: FXS/ABE P2: FXS
9780521740517c27.xml CUAU031-EVANS September 7, 2008 7:15
b On a given day, out of 2000 people, 1200 are ‘happy’, 600 are ‘neither happy nor sad’ and
200 are ‘sad’. Write down a column matrix S0 that describes this situation.
c The next day, how many people do we expect to be happy, neither happy nor sad, or sad?
d After five days, how many people do we expect to be happy, neither happy nor sad, or sad?
e In the long term, how many of the 2000 people do we expect to be happy, neither happy
nor sad, or sad?
0.9 0.2 100
6 For the transition matrix T = and an initial state matrix S0 = :
0.1 0.8 200
E
a Use the relationship Sn = T Sn−1 to determine: i S1 ii S2 iii S3
b Determine the value of T 5 .
c Use the relationship Sn = T n Sn−1 to determine i S2 ii S3 iii S7
d Calculate Sn = T S0 for n = 10, 15, 20 and 25 to show that the steady state solution is
n
200
close to
100
PL
.
0.7 0.4 0.1
300
7 For the transition matrix T = 0.2 0.1 0.3 and an initial state matrix S0 = 200 :
0.1 0.5 0.6
c Calculate Sn = T S0 for n = 10, 15, 20 and 25 to show that the steady state solution is
n
M
247.1
close to 129.4 .
223.5
SA
Cambridge University Press • Uncorrected Sample pages • 978-0-521-61328-6 • 2008 © Jones, Evans, Lipson
TI-Nspire & Casio ClassPad material in collaboration with Brown and McMenamin
P1: FXS/ABE P2: FXS
9780521740517c27.xml CUAU031-EVANS September 7, 2008 7:15
Review
Key ideas and chapter summary
The identity matrix An identity (unit) matrix I is a square matrix with ones
down the leading diagonal and zeros elsewhere.
1 0 0 0
1 0 0
1 0
0 1 0 0
0 1 0 are all examples of
0 1 0 0 1 0
0 0 1 identity matrices
0 0 0 1
E
The determinant of a matrix The determinant of a matrix A is written det(A) or |A|.
a b a b
If A = , then det(A) = =a×d −b×c
c d c d
Only square matrices have determinants.
the determinant.
1 3
3 4
then
= 1 × 4 − 3 × 3 = −5
For higher order matrices, a calculator is used to calculate
1
provided = 0, that is, provided det (A) = 0.
ad − bc
1 3
For example, if A = then
3 4
1 d −b 1 4 −3
A−1 = =
det(A) −c a −5 −3 1
−0.8 0.6
=
0.6 −0.2
For higher order matrices, a calculator is used to determine
the inverse.
Cambridge University Press • Uncorrected Sample pages • 978-0-521-61328-6 • 2008 © Jones, Evans, Lipson
TI-Nspire & Casio ClassPad material in collaboration with Brown and McMenamin
P1: FXS/ABE P2: FXS
9780521740517c27.xml CUAU031-EVANS September 7, 2008 7:15
ax + by = e
Representing systems of linear The set of linear equations can be written in
cx + dy = f
equations with matrices
a b x e
matrix form as = .
c d y f
x + 3y = 1
For example, the set of equations
3x + 4y = 2
1 3 x 1
can written in matrix form as =
E
3 4 y 2
A similar pattern follows for sets of three, four, five etc.,
equations.
Matrix solution of a system Provided det(A) = 0, the set of linear equations defined
of linear equations by AX = C has the solution X = A−1 C. The order of
PL multiplication is important.
is given by
x
y
=
=
1 3
3 4
−0.8
1 3
3 4
−1
0.6 −0.2
0.6
1
2
x
y
1
2
=
=
1
2
0.4
0.2
A4 = A × A × A × A and so on.
Only square matrices can be raised to a power.
A0 is defined to be I, the identity matrix.
Transition matrix Transition matrices are square and have the property that
the sum
of the columns
equals one. For example,
0.7 0.1
T = could be a transition matrix.
0.3 0.9
Initial state matrix The initial state matrix S0 defines the starting state of a
100
system. For a two-state system, S0 = , could be an
200
initial state matrix.
Cambridge University Press • Uncorrected Sample pages • 978-0-521-61328-6 • 2008 © Jones, Evans, Lipson
TI-Nspire & Casio ClassPad material in collaboration with Brown and McMenamin
P1: FXS/ABE P2: FXS
9780521740517c27.xml CUAU031-EVANS September 7, 2008 7:15
Review
Steady-state solution matrix The steady-state matrix S represents the final state of a
system. The final state of a system can be estimated by
calculating T n S0 , for a large value of n.
Skills check
E
recognise an identity matrix
calculate the determinant of a matrix
know the properties of an inverse matrix
find the inverse of a square matrix using a calculator
use determinants to test a system of linear equations for solutions
PL
use inverse matrices to solve systems of linear equations
determine the power of a square matrix using a calculator
find the steady-state solution of a system given the initial state S0 , and the transition
matrix T
Multiple-choice questions
A 1 B 0 C 1 D 2 E 4
−1
3 Y =
0.5 0 4 −2 1 0 1 −4 2
A B C D
−0.5 1 −2 1 0 1 8 2 −1
E not defined
4 U −1=
0.5 0 4 −2 1 0 1 −4 2
A B C D
−0.5 1 −2 1 0 1 8 2 −1
E not defined
2
5 3U =
8 0 4 0 12 0 24 0
A B C D
7 1 3 1 9 3 21 3
E not defined
Cambridge University Press • Uncorrected Sample pages • 978-0-521-61328-6 • 2008 © Jones, Evans, Lipson
TI-Nspire & Casio ClassPad material in collaboration with Brown and McMenamin
P1: FXS/ABE P2: FXS
9780521740517c27.xml CUAU031-EVANS September 7, 2008 7:15
75% A B 95%
E
5%
The transition matrix that can be used to represent the information in the diagram
above is:
From From
A
A To: A 0.75
B 0.05
A
C To: A 0.75
B 0.25
PL From
From
B
0.25
0.95
B
0.95
0.05
A
B To: A 0.75
B 0.25
A
D To: A 0.75
B 0.95
B
0.05
0.95
From
B
0.25
0.05
M
A B
E To: A 0.25 0.05
B 0.75 0.95
9 60%
SA
10% Y 35%
75%
5% 20%
X Z
15%
50%
30%
The transition matrix that can be used to represent the information in the diagram
above is:
From From
X Y Z X Y Z
A To: X 0.75 0.05 0.30 B To: X 0.75 0.10 0.15
Y 0.10 0.60 0.20 Y 0.60 0.05 0.35
Z 0.15 0.35 0.50 Z 0.50 0.30 0.20
Cambridge University Press • Uncorrected Sample pages • 978-0-521-61328-6 • 2008 © Jones, Evans, Lipson
TI-Nspire & Casio ClassPad material in collaboration with Brown and McMenamin
P1: FXS/ABE P2: FXS
9780521740517c27.xml CUAU031-EVANS September 7, 2008 7:15
Review
From From
X Y Z X Y Z
C To: X 0.75 0.10 0.15 D To: X 0.75 0.05 0.15
Y 0.60 0.05 0.35 Y 0.10 0.60 0.20
Z 0.50 0.30 0.20 Z 0.15 0.35 0.50
From
X Y Z
E
E To: X 0.75 0.05 0.15
Y 0.15 0.35 0.50
Z 0.10 0.60 0.20
10 Which of the following systems of linear equations has a unique solution?
x − 3y = 6 2x + 2y = 6 4x − 3y = 6
C
I
2x + y = 3
A I only
−3 1
2 3
2 0
x
y
y
PL
11 The linear equations
2 2
x
=
=
II
B I and II only
3
6
3
4x + 4y = 3
2x − 3y = 6
2x + y = 3
6
C II only
D
−3 1
2 −3
III
8x − 12y = 3
D I and III only
=
3
6
6
E all
2 1 y 3
M
2 3 x 6
E =
2 1 y 3
Cambridge University Press • Uncorrected Sample pages • 978-0-521-61328-6 • 2008 © Jones, Evans, Lipson
TI-Nspire & Casio ClassPad material in collaboration with Brown and McMenamin
P1: FXS/ABE P2: FXS
9780521740517c27.xml CUAU031-EVANS September 7, 2008 7:15
E
Extended-response questions
3x + 2y = 7
i
2 3
PL
4x − 2y = 0
using matrix methods.
a Write the equations in matrix form.
b The solution is given by X = A−1 C. Write down the matrices A, A−1 , X , and C.
c Solve the equations.
d Use the determinant test to show that the following systems of linear equations do
not have a unique solution.
x
=
6
ii
2 2
x
=
6
M
6 9 y 3 8 8 y 24
1 −3 x 6
iii =
2 −6 y 3
2 We wish to solve the following system of linear equations
x − 2y + z = 0
SA
3x + 2y − z = 4
2x − y + z = 3
using matrix methods.
a Write the equations in matrix form.
b The solution is given by X = A−1 C. Write down the matrices A, A−1 , X , and C.
c Solve the equations.
0.15 0.75 400
3 For the transition matrix T = and initial state matrix S0 = :
0.85 0.25 800
a Use the relationship Sn = T Sn−1 to determine: i S1 ii S2 iii S3
4
b Determine the value of T .
c Use the relationship Sn = T n Sn−1 to determine: i S2 ii S3 iii S6
d Calculate Sn = T n S0 for n = 10, 15, 30 and 40 to show that the steady state
562.5
solution is close to .
637.5
Cambridge University Press • Uncorrected Sample pages • 978-0-521-61328-6 • 2008 © Jones, Evans, Lipson
TI-Nspire & Casio ClassPad material in collaboration with Brown and McMenamin
P1: FXS/ABE P2: FXS
9780521740517c27.xml CUAU031-EVANS September 7, 2008 7:15
Review
4 Experience at a fitness centre shows that:
r 70% of members who use the centre in a given week will also use the
centre the next week
r 30% of members who use the centre in a given week will not use the
centre the next week
r 40% of members who do not use the centre in a given week will use the
centre in the next week
E
r 60% of members who do not use the centre in a given week will not use
the centre in the next week
a Construct a transition matrix T of the form:
This week
c
d
PL
Next week: Use
Not use
This week, 400 members used the fitness centre and 100 did not use the fitness
centre. Write down a column matrix S0 that describes this situation.
How many members do you expect to use the centre next week?
How many members do you expect to use the centre after five weeks?
e In the long term, how many members do you expect to use the fitness centre each
M
week?
SA
Cambridge University Press • Uncorrected Sample pages • 978-0-521-61328-6 • 2008 © Jones, Evans, Lipson
TI-Nspire & Casio ClassPad material in collaboration with Brown and McMenamin