Chapter

11
Transition matrices and
Leslie matrices

Chapter objectives
 How do you construct a transition matrix from a transition diagram and
vice versa?
 How do you construct a transition matrix to model the transitions in a
population?
 How do you use a matrix recurrence relation, S0 = initial state matrix,
Sn+1 = TSn , to generate a sequence of state matrices?
 How do you informally identify the equilibrium state or steady-state matrix
in the case of regular state matrices?
 How do you use a matrix recurrence relation S0 = initial state matrix,
Sn+1 = TSn + B to model systems that include external additions or
reductions at each step of the process?
 How do you use and interpret Leslie matrices to analyse population growth?

In this chapter we use matrices to model proportional change of the numbers in a

particular state to itself and other states from one time to the next.
For example, in the first example of this chapter, there are two states which in this case
are the towns in which rental cars finish up each day.
In sections A-E of this chapter we look at transition matrices which satisfy certain
conditions and describe the proportional change.
In section F of the chapter Leslie matrices are used to model population change in a
particular way.
 11A Transition matrices - setting up a transition matrix 551

11A Transition matrices - setting up a transition

matrix
Learning intentions
 To be able to set up a transition matrix from a diagram or written information.

A car rental firm has two branches: one in Bendigo and one in Colac. Cars are usually rented
and returned in the same town. However, a small percentage of cars rented in Bendigo each
week are returned in Colac, and vice versa. The diagram below describes what happens on a
weekly basis.
B – Bendigo
20%
C – Colac

80% B C 90%

10%

What does this diagram tell us?

From week to week:
 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
Bendigo Colac
⎡ ⎤
Bendigo ⎢⎢⎢ 0.8 0.1 ⎥⎥⎥
Returned to ⎢⎢⎣ ⎥⎥⎦
Colac 0.2 0.9

This matrix is an example of a transition matrix (T). It describes the way in which
transitions are made between two states:
 state 1: the rental car is based in Bendigo.
 state 2: the rental car is based in Colac.
Note: In this situation, where the total number of cars remains constant, the columns in a transitional matrix
will always add to one (100%). For example, if 80% of cars are returned to Bendigo, then 20% must be
returned to Colac.
552 Chapter 11  Transition matrices and Leslie matrices

Example 1 Setting up a transition matrix

The diagram gives the weekly 80%

return rates of rental cars at three
A – Albury
locations: Albury, Wodonga and B – Benalla
Benalla. Construct a transition
W W – Wodonga
15%
matrix that describes the week- 70% 10%
by-week return rates at each of
11%
the three locations. Convert the A 5% B
percentages to proportions. 20%

12% 77%

Explanation Solution
1 There are three locations from which
Rented in
the cars can be rented and returned:
A W B
Albury (A), Wodonga (W) and Benalla ⎡ ⎤
A ⎢⎢⎢ ⎥⎥⎥
(B). To account for all the possibilities, ⎢⎢ ⎥⎥⎥
Returned to w ⎢⎢⎢⎢⎢ ⎥⎥⎥
⎥⎥⎦
a 3 × 3 matrix is needed. Construct a ⎢⎣
blank matrix labelling the rows and B
columns A, W and B, respectively.
Column labels indicate where the car
was rented. The row labels indicate
where the cars were returned to.
2 Complete the matrix by writing
A W B
each of the percentages (converted ⎡ ⎤
A ⎢⎢⎢ 0.7 ⎥⎥⎥
⎢⎢ ⎥⎥⎥
w ⎢⎢⎢⎢⎢ 0.1
to proportions) into the appropriate ⎥⎥⎥
locations. Start with column A and ⎢⎣ ⎥⎥⎦
write in values for each row: 0.7 (70%), B 0.2
0.1 (10%) and 0.2 (20%).
3 Mentally check your answer by
A W B
summing columns; they should ⎡ ⎤
A ⎢⎢⎢ 0.7 0.05 0.12⎥⎥⎥
sum to 1. ⎢⎢ ⎥⎥
w ⎢⎢⎢⎢⎢ 0.1 0.8 0.11⎥⎥⎥⎥⎥
⎢⎣ ⎥⎦
B 0.2 0.15 0.77

Example 2 Setting up a transition matrix

A factory has a large number of machines. Machines can be in one of two states:
operating or 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
 11A 11A Transition matrices - setting up a transition matrix 553

 5% of machines that are broken are repaired and start operating again
 95% of machines that are broken stay broken.

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.
Explanation Solution
1 There are two machine states: operating (O) or broken
Start
(B). To account for all the possibilities, a 2 × 2 transition
⎡ O B ⎤
matrix is needed. Construct a blank matrix, labelling the O ⎢⎢⎢ ⎥⎥⎥
End ⎢⎢⎣ ⎥⎥⎦
rows and columns O and B, respectively. B

2 Complete the matrix by writing each of the percentages

(converted to proportions) into the appropriate locations. O B
⎡ ⎤
Start with column O and write in the values for each O ⎢⎢⎢0.85 ⎥⎥⎥
⎢⎢⎣ ⎥⎥⎦
row: 0.85 (85%) and 0.15 (15%). B 0.15

3 Mentally check your answer by summing the columns;

they should sum to 1. O B
⎡ ⎤
O ⎢⎢⎢0.85 0.05⎥⎥⎥
⎢⎢⎣ ⎥⎥⎦
B 0.15 0.95

Exercise 11A

Setting up a transition matrix from a transition diagram

Example 1 1 The diagrams below describe a series of transitions between the states indicated.
Construct a 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% b 30%

40% A B 45% 70% X Y 75%

55% 25%

c 70% d 45%

10% Y 15% 25% B 20%

X 15% 23% Z A 35% 20% C

30% 30%
60% 55% 45% 65%
22% 15%
 554 Chapter 11  Transition matrices and Leslie matrices 11A

Example 2 2 A factory has a large number of machines which can be in one of two states, operating
(O) or broken down (B). It is known that that an operating machine breaks down by the
end of the day on 4% of the days, and that 98% of machines which have broken down
are repaired by the end of the day.
Complete the 2 × 2 transition matrix T to describe this.
Today
O B
⎡ ⎤
O ⎢⎢⎢⎢ ⎥⎥⎥
⎥⎥⎥ Tomorrow
T = ⎢⎢⎣ ⎦
B

3 A large company has 1640 employees, 60% of whom currently work full-time (F) and
40% of whom currently work part-time (P). Every year 20% of full-time workers move
to part-time work, and 14% of part-time workers move to full-time work.
Complete the 2 × 2 transition matrix T to describe this.
This year
F P
⎡ ⎤
F ⎢⎢⎢⎢ ⎥⎥⎥
⎥⎥⎥
T = ⎢⎢⎣ ⎦ Next year
P

Exam 1 style questions

4 In a particular newsagent, the two top-selling newspapers are the Argus and the
Bastion. The transition diagram below shows the way shoppers at this newsagent
change their newspaper choice from today to tomorrow.
70%

30% A B 35%

65%

A transition matrix that provides the same information as the transition diagram is
A Today B Today C Today
A B A B A B
⎡ ⎤ ⎡ ⎤ ⎡ ⎤
A ⎢⎢⎢⎢65% 30%⎥⎥⎥⎥ A ⎢⎢⎢⎢30% 65%⎥⎥⎥⎥ A ⎢⎢⎢⎢30% 65%⎥⎥⎥⎥
Tomorrow ⎢⎢⎣ ⎥⎥⎦ Tomorrow ⎢⎢⎣ ⎥⎥⎦ Tomorrow ⎢⎢⎣ ⎥⎥⎦
B 70% 35% B 70% 35% B 35% 70%
D Today E Today
A B A B
⎡ ⎤ ⎡ ⎤
A ⎢⎢⎢⎢30% 35%⎥⎥⎥⎥ A ⎢⎢⎢⎢65% 70%⎥⎥⎥⎥
Tomorrow ⎢⎢⎣ ⎥⎥⎦ Tomorrow ⎢⎢⎣ ⎥⎥⎦
B 65% 70% B 35% 30%
 11B Interpreting transition matrices 555

11B Interpreting transition matrices

Learning intentions
 To be able to interpret a transition matrix and a transition diagram.

Let us return to the car rental problem at the start of this section. As we saw then, the
following transition matrix, T , and its transition diagram can be used to describe the weekly
pattern of rental car returns in Bendigo and Colac.

Rented in B – Bendigo
⎡ B C ⎤ C – Colac
⎢⎢⎢0.80 0.10⎥⎥⎥ B 20%
⎢ ⎥⎥⎦
T = ⎢⎣ Returned to 80% B C 90%
0.20 0.90 C 10%

Using this information alone, a number of predictions can be made.

For example, if 50 cars are rented in Bendigo this week, the transition matrix predicts that:
 80% or 40 of these cars will be returned to Bendigo next week (0.80 × 50 = 40)
 20% or 10 of these cars will be returned to Colac next week (0.20 × 50 = 10).

Further, if 40 cars are rented in Colac this week, the transition matrix predicts that:
 10% or 4 of these cars will be returned to Bendigo next week (0.10 × 40 = 4)
 90% or 36 of these cars will be returned to Colac next week (0.90 × 40 = 36).

Example 3 Interpreting a transition matrix

The following transition matrix, T , and its transition diagram can be used to describe the
weekly pattern of rental car returns in three locations: Albury, Wodonga and Benalla.

A W B
⎡ ⎤ A = Albury
⎢⎢⎢ 0.7 0.05 0.12⎥⎥⎥ A 80% B = Benalla
⎢⎢⎢ ⎥⎥
T = ⎢⎢⎢⎢ 0.1 0.8 0.11⎥⎥⎥⎥⎥ W Returned to W = Wodonga
⎢⎣ ⎥⎦
0.2 0.15 0.77 B 10% W 15%
70%
A 5% 11% B
20%
77%
12%

Use the transition matrix T and its transition diagram to answer the following questions.
a What percentage of cars rented in Wodonga each week are predicted to be returned to:
i Albury? ii Benalla? iii Wodonga?
 556 Chapter 11  Transition matrices and Leslie matrices 11B

b Two hundred cars were rented in Albury this week. How many of these cars do we
expect to be returned to:
i Albury? ii Benalla? iii Wodonga?
c What percentage of cars rented in Benalla each week are not expected to be returned to
Benalla?
d One hundred and sixty cars were rented in Albury this week. How many of these cars
are expected to be returned to either Benalla or Wodonga?

Solution
a i 0.5 or 5% ii 0.15 or 15% iii 0.80 or 80%
b i 0.70 × 200 = 140 cars ii 0.20 × 200 = 40 cars iii 0.10 × 200 = 20 cars
c 11 + 12 = 23% or 100 − 77 = 23%
d 20% of 160 + 10% of 160 = 48 cars

Exercise 11B

Interpreting transition matrices

Example 3 1 When people go to the movies they buy either a bag of popcorn (P) or an ice cream (I).
Experience has shown that:
 85% of people who buy popcorn this time will buy popcorn next time
 15% of people who buy popcorn this time will buy an ice cream next time
 75% of people who buy an ice cream this time will buy an ice cream next time
 25% of people who buy ice cream this time will buy popcorn next time.
a Construct a transition matrix and transition diagram that can be used to describe this
situation. Use the models below.
This time
P I
⎡ ⎤
⎢⎢ ⎥⎥⎥ P
T = ⎢⎢⎢⎣ ⎥⎥⎦ Next time P I
I

b Eighty people are seen buying popcorn at the movies. How many of these are
expected to buy popcorn next time they go to the movies?
c Sixty people are seen buying an ice cream at the movies. How many of these are
expected to buy popcorn next time they go to the movies?
d On another occasion, 120 people are seen buying popcorn and 40 are seen buying
an ice cream. How many of these are expected to buy an ice cream next time they
attend the movies?
11B 11B Interpreting transition matrices 557

2 On Windy Island, sea birds are observed nesting at three sites: A, B and C. The
following transition matrix and accompanying transition diagram can be used to predict
the movement of these sea birds between these sites from year to year.
This year 10%

A B C 100% 80%
⎡ ⎤
⎢⎢⎢ 1.0 0.10 0.05⎥⎥⎥ A A B
⎢⎢ ⎥⎥ 5%
T = ⎢⎢⎢⎢⎢ 0 0.80 0.05⎥⎥⎥⎥⎥ B Next year
⎢⎣ ⎥⎦ 5% 10%
0 0.10 0.90 C C 90%

a What percentage of sea birds nesting at site B this year were expected to nest at:
i site A next year? ii site B next year? iii site C next year?
b This year, 850 sea birds were observed nesting at site B. How many of these are
expected to:
i still nest at site B next year? ii move to site A to nest next year?
c This year, 1150 sea birds were observed nesting at site A. How many of these birds
are expected to nest at:
i site A next year? ii site B next year? iii site C next year?
d What does the ‘1’ in column A, row A of the transition matrix indicate?

3 A car insurance company finds that:

 22% of car drivers involved in an accident this year (A) are also expected to be
involved in an accident next year
 9% of drivers who are not involved in an accident this year (N) are expected to be
involved in an accident next year.
The transition diagram that can be used to describe this situation is shown below.
78%
This year
22% 91%
A N
⎡ ⎤
⎢⎢0.22 0.09⎥⎥⎥ A A N
T = ⎢⎢⎢⎣ ⎥⎥⎦ Next year
0.78 0.91 N
9%
a In 2015, 84 000 drivers insured with the company were not involved in an accident.
i How many of these drivers were not expected to be involved in an accident in
2016?
ii How many of these drivers were expected to be involved in an accident in 2016?
b In 2015, 25 000 drivers insured with the company were involved in an accident.
i How many of these drivers were expected to be involved in an accident in 2016?
ii How many of these drivers were expected to be involved in an accident in 2017?
iii How many of these drivers were expected to be involved in an accident in 2018?
558 Chapter 11  Transition matrices and Leslie matrices 11B

4 Fleas can move beteen Now

three locations A, B and C.
A B C
The way a flea moves after ⎡ ⎤
⎢⎢⎢0.60 0.10 0.70⎥⎥⎥ A
5 seconds in a location can ⎢⎢⎢ ⎥⎥
T = ⎢⎢⎢⎢0.20 0.80 0.10⎥⎥⎥⎥⎥ B After 5 seconds
be exactly described by the ⎢⎣ ⎥⎦
transition matrix. 0.20 0.10 0.20 C
The move is not dependent on any previous move.
a If there are 30 fleas at location A at the beginning of the 5-second period, how many
fleas would you expect to
i stay at A ii go to B iii go to C
at the end of the 5-second period?
b If there were 60 fleas at each of the locations how many fleas would you expect to
have at
i A ii B iii C
after one 5-second period?
⎡ ⎤
⎢⎢⎢60⎥⎥⎥
⎢⎢⎢ ⎥⎥⎥
c Find the product T ⎢⎢⎢⎢60⎥⎥⎥⎥ and comment.
⎢⎢⎢ ⎥⎥⎥
⎣ ⎦
60
d At the conclusion of the first 5-second period there are 30 fleas at C.
i How many of these go to A in the next 5-second period?
ii How many of these go to B in the next 5-second period?
iii How many of these go to C in the next 5-second period?
⎡ ⎤
⎢⎢⎢60⎥⎥⎥
 ⎢⎢⎢ ⎥⎥⎥
e Evaluate the product 1 1 1 T ⎢⎢⎢⎢60⎥⎥⎥⎥ and comment.
⎢⎢⎢ ⎥⎥⎥
⎣ ⎦
60

Exam 1 style questions

This week
5 There are 120 students in a Year
S O F
12 class. Each week every student ⎡ ⎤
⎢⎢⎢0.6 0.2 0.4⎥⎥⎥ S
participates in one of three activities: ⎢⎢ ⎥⎥
T = ⎢⎢⎢⎢⎢0.1 0.7 0.1⎥⎥⎥⎥⎥ O Next week
Sport (S ), Outdoor Activities (O) ⎢⎣ ⎥⎦
or First Aid (F). 0.3 0.1 0.5 F

The activities that the children select

each week change according to the
transition matrix opposite.
From the transition matrix it can be concluded that:
A in the first week of the program, eighty students do Sport, twenty students children
do Outdoor activities and twenty Students do First Aid.
B at least 50% of the students do not change their activities from the first week to the
second week.
 11B 11C Transition matrices – using recursion 559

C in the long term, all of the children will choose the same activity.
D Sport is the most popular activity in the first week
E 40% of the students will do First Aid each week.

6 Warren text messages a friend each Today

week day of this week. His friends A B C D E
are Arthur (A), Belinda (B), Connie ⎡ ⎤
⎢⎢⎢0 1 0 0 0⎥⎥⎥ A
(C), Danielle (D) and Eleanor (E). ⎢⎢⎢ ⎥⎥
⎢⎢⎢0
⎢⎢⎢ 0 0 1 0⎥⎥⎥⎥⎥ B
On Monday, Warren will send a text ⎥⎥
message to Connie. T = ⎢⎢⎢⎢⎢1 0 0 0 0⎥⎥⎥⎥⎥ C Tomorrow
⎢⎢⎢ ⎥⎥⎥
⎢⎢⎢0 0 0 0 1⎥⎥⎥⎥ D
Based on the transition matrix, the ⎢⎢⎣ ⎥⎦
order in which Warren will text 0 0 1 0 0 E
message each of his friends for the
next four days is:

A Arthur, Belinda, Connie, Danielle B Danielle, Belinda , Arthur, Connie

C Danielle, Belinda, Arthur, Eleanor D Eleanor, Arthur, Danielle, Belinda
E Eleanor, Danielle, Belinda, Arthur

11C Transition matrices – using recursion

Learning intentions
 To be able to use a matrix recurrence relation: S0 = initial state matrix, Sn+1 = T Sn , to
generate a sequence of state matrices.
 To be able to informally identify the equilibrium state or steady-state matrix in the
case of regular state matrices.

We return again to the car rental problem. The car rental firm now plans to buy 90 new cars.
Fifty will be based in Bendigo and 40 in Colac.
Given this pattern of rental car returns, the first question the manager would like answered is:
‘If we start with 50 cars in Bendigo, and 40 cars in Colac, how many cars will be
available for rent at both towns after 1 week, 2 weeks, etc?’
You have met this type of problem earlier when doing financial modelling (Chapter 8). For
example, if we invest $1000 at an interest rate of 5% per annum, how much will we have
after 1 year, 2 years, 3 years, etc?
We solved this type of problem by using a recurrence relation to model the growth in our
investment year-by-year. We do the same with the car rental problem, the only difference
being that we are now working with matrices.
560 Chapter 11  Transition matrices and Leslie matrices

Constructing a matrix recurrence relation

A recurrence relation must have a starting point.
⎡ ⎤
⎢⎢⎢50⎥⎥⎥
In this case it is the initial state matrix: S 0 = ⎢⎢⎢⎣ ⎥⎥⎥⎦
40
Generating S 1

To find out the⎡ number ⎤ of cars in Bendigo and Colac after 1 week, we use the transition
⎢⎢⎢0.8 0.1⎥⎥⎥
matrix T = ⎢⎢⎢⎣ ⎥⎥⎥ to generate the next state matrix in the sequence, S 1 , as follows:
0.2 0.9⎦
S1 = T S0
⎡ ⎤⎡ ⎤ ⎡ ⎤
⎢⎢⎢0.8 0.1⎥⎥⎥ ⎢⎢⎢50⎥⎥⎥ ⎢⎢⎢0.8 × 50 + 0.1 × 40⎥⎥⎥
= ⎢⎢⎣ ⎢ ⎥ ⎢ ⎥
⎥⎥ ⎢⎢ ⎥⎥ = ⎢⎢⎢ ⎥⎥⎥
0.2 0.9⎦ ⎣40⎦ ⎣0.2 × 50 + 0.9 × 40⎦
⎡ ⎤
⎢⎢⎢44⎥⎥⎥
or S 1 = ⎢⎢⎢⎣ ⎥⎥⎥⎦
46
Thus, after 1 week we predict that there will be 44 cars in Bendigo and 46 in Colac.
Generating S 2

Following the same pattern, after 2 weeks;

⎡ ⎤⎡ ⎤ ⎡ ⎤
⎢⎢⎢0.8 0.1⎥⎥⎥ ⎢⎢⎢44⎥⎥⎥ ⎢⎢⎢39.8⎥⎥⎥
S 2 = T S 1 = ⎢⎢⎢⎣ ⎥⎥⎥ ⎢⎢⎢ ⎥⎥⎥ = ⎢⎢⎢
⎦⎣ ⎦ ⎣
⎥⎥⎥
⎦
0.2 0.9 46 50.2
Thus, after 2 weeks we predict that there will be 39.8 cars in Bendigo and 50.2 in Colac.
Generating S 3

After 3 weeks:
⎡ ⎤⎡ ⎤ ⎡ ⎤
⎢⎢⎢0.8 0.1⎥⎥⎥ ⎢⎢⎢39.8⎥⎥⎥ ⎢⎢⎢36.9⎥⎥⎥
S 3 = T S 2 = ⎢⎢⎢⎣ ⎥⎥⎥ ⎢⎢⎢
⎦⎣
⎥⎥⎥ = ⎢⎢⎢
⎦ ⎣
⎥⎥⎥
⎦
0.2 0.9 50.2 53.1
Thus, after 3 weeks we predict that there will be 36.9 cars in Bendigo and 53.1 in Colac.
A pattern is now emerging. So far we have seen that:
S1 = T S0
S2 = T S1
S3 = T S2
If we continue this pattern we have:
S4 = T S3
S5 = T S4
or, more generally, S n+1 = T S n .
With this rule as a starting point, we now have a recurrence relation that will enable us to
model and analyse the car rental problem on a step-by-step basis.
 11C Transition matrices – using recursion 561

Recurrence relation
S 0 = intial value, S n+1 = T S n

Let us return to the factory problem in Example 2.

Example 4 Using a recursion relation to calculate state matrices step-by-step

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.
At the start, 80 machines are operating and 20 are broken.
Use the recursion relation
S 0 = intial value, S n+1 = T S n
where
⎡ ⎤ ⎡ ⎤
⎢⎢⎢80⎥⎥⎥ ⎢⎢⎢0.85 0.05⎥⎥⎥
⎢ ⎥
S 0 = ⎢⎣⎢ ⎥⎦⎥ and T = ⎢⎣⎢ ⎢ ⎥⎥⎥
⎦
20 0.15 0.95

to determine the number of operational and broken machines after 1 day and after 3 days.
Explanation Solution
⎡ ⎤ ⎡ ⎤
⎢⎢⎢80⎥⎥⎥ ⎢⎢⎢0.85 0.05⎥⎥⎥
1 Write down a column matrix S 0 = ⎢⎢⎢⎣ ⎥⎥⎥⎦ T = ⎢⎢⎢⎣ ⎥⎥⎥
⎦
20 0.15 0.95
with S 0 representing the initial
operational state of the machines,
and the transition matrix.
⎡ ⎤⎡ ⎤ ⎡ ⎤
⎢⎢⎢0.85 0.05⎥⎥⎥ ⎢⎢⎢80⎥⎥⎥ ⎢⎢⎢69⎥⎥⎥
2 Use the rule S n+1 = T S n to S 1 = T S 0 = ⎢⎢⎣ ⎢ ⎥⎥⎥ ⎢⎢⎢ ⎥⎥⎥ = ⎢⎢⎢ ⎥⎥⎥
⎦⎣ ⎦ ⎣ ⎦
0.15 0.95 20 31
determine the operational state
After 1 day, 69 machines are operational and
of the machines after one day by
31 are broken.
forming the product S 1 = T S 0 and
evaluate.
⎡ ⎤⎡ ⎤ ⎡ ⎤
⎢⎢⎢0.85 0.05⎥⎥⎥ ⎢⎢⎢69⎥⎥⎥ ⎢⎢⎢60.2⎥⎥⎥
3 To find the operational state of the S 2 = T S 1 = ⎢⎢⎢⎣ ⎥⎥⎥ ⎢⎢⎢ ⎥⎥⎥ = ⎢⎢⎢
⎦⎣ ⎦ ⎣
⎥⎥⎥
⎦
0.15 0.95 31 39.8
machines after 3 days, we must ⎡ ⎤⎡ ⎤ ⎡ ⎤
⎢⎢⎢0.85 0.05⎥⎥⎥ ⎢⎢⎢60.2⎥⎥⎥ ⎢⎢⎢53.16⎥⎥⎥
S 3 = T S 2 = ⎢⎢⎢⎣ ⎥⎥⎥ ⎢⎢⎢ ⎥⎥⎥ = ⎢⎢⎢ ⎥⎥⎥
first find the operating state of the
⎦⎣ ⎦ ⎣ ⎦
machines after 2 days (S 2 ) and 0.15 0.95 39.8 46.84
use this matrix to find S 3 using After 3 days, 53 machines are operating and
S3 = T S2 . 47 are broken.
562 Chapter 11  Transition matrices and Leslie matrices

Calculator hint: In practice, generating ⎡ ⎤ ⎡ ⎤

⎢⎢⎢80⎥⎥⎥ ⎢⎢⎢80.⎥⎥⎥
matrices recursively is performed on your CAS ⎢⎢⎢ ⎥⎥⎥ → s0 ⎢⎢⎢ ⎥⎥⎥
calculator as shown opposite for the calculations ⎣ ⎦ ⎣ ⎦
20 20.
performed in Example 11. ⎡ ⎤ ⎡ ⎤
⎢⎢⎢0.85 0.05⎥⎥⎥ ⎢⎢⎢0.85 0.05⎥⎥⎥
⎢⎢⎢ ⎥⎥⎥ → t ⎢⎢⎢ ⎥⎥⎥
⎣ ⎦ ⎣ ⎦
0.15 0.95 0.15 0.95
⎡ ⎤
⎢⎢⎢69.⎥⎥⎥
t.s0 ⎢⎢⎣⎢ ⎥⎥⎦⎥
31.
⎡ ⎤ ⎡ ⎤
⎢⎢⎢69⎥⎥⎥ ⎢⎢⎢60 2⎥⎥⎥
⎢⎢⎢ ⎥⎥⎥ ⎢⎢⎢ ⎥⎥⎥
⎣ ⎦ ⎣ ⎦
31 39 8

and so on.

A rule for determining the state matrix of a system after

n steps
While we can use the recurrence relation:
S 0 = intial value, S n+1 = T S n
to generate state matrices step-by-step, there is a more efficient method when we need to de-
termine the state matrix after a large number of steps.
If we follow through the process step-by-step we have:
S1 = T S0
S 2 = T S 1 = T (T S 0 ) = T 2 S 0
S 3 = T S 2 = T (T S 1 ) = T 2 S 1 = T 2 (T S 0 ) = T 3 S 0
Continuing the process
S4 = T 4 S0
S5 = T 5 S0
or more generally, Sn = Tn S0 .
We now have a simple rule for finding the value, S n , of the state matrix after n steps.

A rule for finding the state matrix after n steps

If the recurrence rule for determining state matrices is
S 0 = initial state matrix, S n+1 = T S n ,
the state matrix after n steps (or transitions) is given by S n = T n S 0 .

Let us return to the factory problem we analysed in Example 2.

 11C Transition matrices – using recursion 563

Example 5 Determining the nth state of a system using the rule Sn = Tn S0

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.
Initially, 80 machines are operating and 20 are broken, so:
⎡ ⎤ ⎡ ⎤
⎢⎢⎢80⎥⎥⎥ ⎢⎢⎢0.85 0.05⎥⎥⎥
S 0 = ⎢⎢⎢⎣ ⎥⎥⎥⎦ and T = ⎢⎢⎢⎣ ⎥⎥⎥
⎦
20 0.15 0.95

Determine the number of operational and broken machines after 10 days.

Explanation Solution
1 Write down the transition matrix, T , and ⎡ ⎤ ⎡ ⎤
⎢⎢⎢0.85 0.05⎥⎥⎥ ⎢ ⎥
T = ⎢⎢⎢⎣ ⎥⎥⎥ S 0 = ⎢⎢⎢⎢⎢80⎥⎥⎥⎥⎥
initial state matrix, S 0 . Enter the matrices ⎦ ⎣ ⎦
0.15 0.95 20
into your 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 S n = T n S 0 and ∴ S 10 = T 10 S 0
substitute n = 10 to give S 10 = T 10 S 0 .
3 Enter the expression T 10 S into your ⎡ ⎤ ⎡ ⎤
⎢⎢⎢0.85 0.05⎥⎥⎥ ⎢⎢⎢0.85 0.05.⎥⎥⎥
calculator and evaluate. ⎢⎢⎢⎣ ⎥⎥⎥ → t
⎦
⎢⎢⎢
⎣
⎥⎥⎥
⎦
0.15 0.95 0.15 0.95.
⎡ ⎤ ⎡ ⎤
⎢⎢⎢⎢80⎥⎥⎥⎥ ⎢⎢⎢⎢80.⎥⎥⎥⎥
⎢⎢⎣ ⎥⎥⎦ → s ⎢⎢⎣ ⎥⎥⎦
20 20.
⎡ ⎤
⎢⎢⎢30.9056⎥⎥⎥
t10 · s ⎢⎢⎢ ⎥⎥⎥
⎣ ⎦
69.0944
⎡ ⎤
4 Write down your answer in matrix form and ⎢⎢⎢30.9⎥⎥⎥
⎢⎢⎢ ⎥⎥⎥
= ⎢⎢
⎣69.1⎥⎥⎦
then in words. S 10

After 10 days, 31 machines will be

operational and 69 broken.

Using the inverse matrix of a transition matrix

In the above we have seen how to move from left to right in the sequence of state matrices
by multiplying by the transition matrix.
S 0 , S 1 , S 2 , . . . S n , S n+1 . . .
We can move from right to left through the transition states by using the inverse of the
transition matrix. In general, the inverse is not a transition matrix.
S n+1 = T S n and S n = T −1 S n+1
564 Chapter 11  Transition matrices and Leslie matrices

Example 6
⎡ ⎤
⎢⎢⎢0.6 0.3⎥⎥⎥
We have a transition matrix T = ⎢⎢⎣ ⎢ ⎥⎥⎥
⎦
0.4 0.7
⎡ ⎤
⎢⎢⎢25 587⎥⎥⎥
and we know that the state matrix S 4 = ⎢⎢⎣⎢ ⎥⎥⎥.
⎦
34 413
Determine S 3 and S 2 .
Solution
⎡ ⎤
⎢⎢⎢ 7 −1⎥⎥⎥⎥
We know that S 4 = T S 3 . Hence S 3 = T −1 S 4 . First T −1 = ⎢⎢⎢⎣ 34 ⎥⎥⎦.
−3 2
You should hold this in your calculator and then
S 3 = T −1 S 4 and S 2 = T −1 S 3
⎡ ⎤⎡ ⎤ ⎡ ⎤⎡ ⎤
⎢⎢⎢ 7 −1⎥⎥⎥ ⎢⎢⎢25 587⎥⎥⎥ ⎢⎢⎢ 7 −1⎥⎥⎥ ⎢⎢⎢25 290⎥⎥⎥
= ⎢⎢⎢⎣ 34 ⎥⎥⎥ ⎢⎢⎢
⎦⎣
⎥⎥⎥
⎦ = ⎢⎢⎢⎣ 34 ⎥⎥⎥ ⎢⎢⎢
⎦⎣
⎥⎥⎥
⎦
− 3 2 34 413 − 3 2 34 710
⎡ ⎤ ⎡ ⎤
⎢⎢⎢25 290⎥⎥⎥ ⎢⎢⎢24 300⎥⎥⎥
= ⎢⎢⎢⎣ ⎥⎥⎥⎦ = ⎢⎢⎢⎣ ⎥⎥⎥⎦
34 710 35 700
Note: To calculate S 2 given S 3 we could have used:
2
S 2 = T −1 S 4

The steady-state solution

A second question a manager might like answered about the car rental is as follows.
‘Will the number of rental cars available from each location vary from week to week or
will they settle down to some fixed value?’
To investigate this question, we start by listing the state matrices from week 0 to week 15.

Week 0 1 2 3 4–11 12 13 14 15
⎡ ⎤ ⎡ ⎤ ⎡ ⎤ ⎡ ⎤ ⎡ ⎤ ⎡ ⎤ ⎡ ⎤ ⎡ ⎤
⎢⎢⎢50⎥⎥⎥ ⎢⎢⎢44⎥⎥⎥ ⎢⎢⎢39.8⎥⎥⎥ ⎢⎢⎢36.9⎥⎥⎥ ⎢⎢⎢30.3⎥⎥⎥ ⎢⎢⎢30.2⎥⎥⎥ ⎢⎢⎢30.1⎥⎥⎥ ⎢⎢⎢30.1⎥⎥⎥
State matrix ⎢⎢⎣⎢ ⎥⎥⎦⎥ ⎢⎢⎣⎢ ⎥⎥⎦⎥ ⎢⎢⎣⎢ ⎥⎥⎥
⎦ ⎢⎢⎣⎢ ⎥⎥⎥
⎦ ... ⎢⎢⎣⎢ ⎥⎥⎥
⎦ ⎢⎢⎣⎢ ⎥⎥⎥
⎦ ⎢⎢⎣⎢ ⎥⎥⎥
⎦ ⎢⎢⎣⎢ ⎥⎥⎥
⎦
40 46 50.2 53.1 59.7 59.8 59.9 59.9
What you should notice is that, as the weeks go by, the number of cars at each of the locations
starts to settle down. We call this the steady- or equilibrium- state solution.
For the rental car problem, the 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, which means the
numbers of cars at each location will not change from then on.
 11C Transition matrices – using recursion 565

This can be seen more clearly in the graph below (the points have been joined to guide the eye).
initial value: steady-state value:
Bendigo 50 Colac 60
60

Number of cars
50 Colac
40 Bendigo
30
20 initial value: steady-state value:
10 Colac 40 Bendigo 30
0
1 2 3 4 5 6 7 8 9 10 11 12 13 14 15
Weeks

In summary, even though the number of cars returned to each location varied from day to day,
the numbers at each location eventually settled down to an equilibrium or steady-state solution.
In the steady state, the number of cars at each location remained the same.

Important
1 In the steady state, cars are still moving between Bendigo and Colac, but the number
of cars rented in Bendigo and returned to Colac is balanced by the number of cars
rented in Colac and returned to Bendigo. Because of this balance, the steady state is
also called the equilibrium state.
2 For a system to have a steady state, the transition matrix must be regular and the
columns must add up to 1. A regular matrix is one whose powers never contain
any zero elements. In practical terms, this means that every state represented in the
transition matrix is accessible, either directly or indirectly from every other state.

A strategy for estimating the steady-state solution

In the car rental problem we found that, even though the number of cars returned to each location
initially varied from day to day, it eventually settled down so the number of cars at each location
remained the same.
Although we arrived at this conclusion by repeated calculations, we can arrive at the solution
much faster by using the rule S n = T n S 0 to find the nth state.

Estimating the steady state solution

If S 0 is the initial state matrix, then the steady-state matrix, S , is given by
S = T n S0
as n tends to infinity (∞).
Note: While in practice we cannot evaluate T n for n = ∞, we find that, depending on the circumstances,
large values of n can often give a very close approximation to the steady-state solution.
566 Chapter 11  Transition matrices and Leslie matrices

Example 7 Estimating the steady-state solution

For the car rental problem:

⎡ ⎤ ⎡ ⎤
⎢⎢⎢50⎥⎥⎥ ⎢⎢⎢0.8 0.1⎥⎥⎥
⎢ ⎥
S 0 = ⎢⎣⎢ ⎥⎦⎥ and T = ⎢⎣⎢ ⎢ ⎥⎥⎥
⎦
40 0.2 0.9

Estimate the steady-state solution by calculating S n for n = 10, 15, 17 and 18.
Explanation Solution
1 Write down the transition matrix T
⎡ ⎤ ⎡ ⎤
and initial state matrix S 0 . Enter the ⎢⎢⎢0.8 0.1⎥⎥⎥ ⎢⎢50⎥⎥
matrices into your calculator. Use T T = ⎢⎢⎣ ⎢ ⎥⎥⎦ , S 0 = ⎢⎢⎢⎣⎢ ⎥⎥⎥⎥⎦
⎥
0.2 0.9 40
and S .
2 Use the rule S n = T n S 0 to write Sn = T n S0
down the expression for the nth state ⎡ ⎤
⎢⎢⎢30.6⎥⎥⎥
for n = 10. ∴ S 10 = T S 0 = ⎢⎣⎢
10 ⎢ ⎥⎥⎥
⎦
59.4
3 Enter the expression T 10 S into your ⎡ ⎤
⎢⎢⎢30.565⎥⎥⎥
calculator and evaluate. t10 · s ⎢⎢⎢ ⎥⎥⎥
⎣ ⎦
4 Repeat the process for n = 15, 17 59.435
⎡ ⎤
and 18. ⎢⎢⎢⎢30.095⎥⎥⎥⎥
t15 · s ⎢⎢⎣ ⎥⎥⎦
59.905
⎡ ⎤
⎢⎢⎢30.047⎥⎥⎥
t17 · s ⎢⎢⎢ ⎥⎥⎥
⎣ ⎦
59.953
⎡ ⎤
⎢⎢⎢30.033⎥⎥⎥
t18 · s ⎢⎢⎢ ⎥⎥⎥
⎣ ⎦
59.967

5 Write down your answer in matrix ⎡ ⎤ ⎡ ⎤ ⎡ ⎤

⎢⎢⎢30.1⎥⎥⎥ ⎢⎢⎢30.0⎥⎥⎥ ⎢⎢⎢30.0⎥⎥⎥
form and then in words. This result S 15 = ⎢⎢⎢⎣ ⎥⎥⎥ ,
⎦ S 17 = ⎢⎢⎢⎣ ⎥⎥⎥ ,
⎦ S 18 = ⎢⎢⎢⎣ ⎥⎥⎥
⎦
agrees with the graphical result 59.9 60.0 60.0
arrived at earlier. The estimated steady-state solution is 30 cars
based in Bendigo and 60 cars based in Colac.
Note: To establish a steady state to a given degree of accuracy, in this case one decimal place, at least
two successive state matrices must agree to this degree of accuracy.
 11C 11C Transition matrices – using recursion 567

Skill-
sheet Exercise 11C

Calculating state matrices step-by-step and by rule

⎡ ⎤ ⎡ ⎤
⎢⎢⎢200⎥⎥⎥ ⎢⎢⎢0.7 0.6⎥⎥⎥
Example 4 1 For the initial state matrix S 0 = ⎢⎢⎢⎣ ⎥⎥⎥ and the transition matrix T = ⎢⎢⎢
⎦ ⎣
⎥⎥⎥: use the
⎦
400 0.3 0.4
recursion relation: S 0 = initial state matrix, S n+1 = T S n , to determine:
a S1 b S2 c S3
⎡ ⎤ ⎡ ⎤
⎢⎢⎢200⎥⎥⎥ ⎢⎢⎢0.7 0.6⎥⎥⎥
Example 5 2 For the initial state matrix S 0 = ⎢⎢⎢⎣ ⎥⎥⎥ and the transition matrix T = ⎢⎢⎢
⎦ ⎣
⎥⎥⎥: use the
⎦
400 0.3 0.4
recursion relation: S 0 = initial state matrix, S n+1 = T S 0 , to determine:
n

a S5 b S7 c S 12
⎡ ⎤
⎢⎢⎢0.65 0.4⎥⎥⎥
Example 6 3 We have a transition matrix T = ⎢⎢⎢⎣ ⎥⎥⎥
⎦
0.35 0.6
⎡ ⎤
⎢⎢⎢⎢5461⎥⎥⎥⎥
and we know that the state matrix S 5 = ⎢⎢⎣ ⎥⎥⎦.
4779
Determine S 4 and S 3 .
⎡ ⎤ ⎡ ⎤
⎢⎢⎢100⎥⎥⎥ ⎢⎢⎢0.9 0.2⎥⎥⎥
4 For the initial state matrix S 0 = ⎢⎣⎢ ⎢ ⎥
⎥⎦⎥ and the transition matrix T = ⎢⎣⎢ ⎢ ⎥⎥⎥:
Example 7
⎦
200 0.1 0.8
a use the recursion relation: S 0 = initial state matrix, S n+1 = T S n , to determine:
i S1 ii S 2 iii S 3
b determine the value of T 5
c use the rule S n = T n S 0 to determine:
i S2 ii S 3 iii S 7
d by calculating S n = T n S 0 for n = 10, 15, 21 and 22, show that the steady-state matrix
⎡ ⎤
⎢⎢⎢200⎥⎥⎥
is close to ⎢⎢⎢⎣ ⎥⎥⎥⎦.
100
⎡ ⎤ ⎡ ⎤
⎢⎢⎢100⎥⎥⎥ ⎢⎢⎢0.7 0.4 0.1⎥⎥⎥
⎢⎢⎢ ⎥⎥⎥ ⎢⎢⎢ ⎥⎥⎥
5 For the initial state matrix S 0 = ⎢⎢⎢⎢⎢200⎥⎥⎥⎥⎥ and the transition matrix T = ⎢⎢⎢⎢⎢0.2 0.1 0.3⎥⎥⎥⎥⎥:
⎢⎢⎣ ⎥⎥⎦ ⎢⎢⎣ ⎥⎥⎦
300 0.1 0.5 0.6
a use a recurrence relation to determine:
i S1 ii S 2 iii S 3
b use the relationship S n = T n S 0 to determine:
i S2 ii S 3 iii S 7
c by calculating S n = T n S 0 for n = 10, 15, 17 and 18, show that the steady-state matrix
⎡ ⎤
⎢⎢⎢247.1⎥⎥⎥
⎢⎢⎢ ⎥⎥⎥
is close to ⎢⎢⎢⎢⎢129.4⎥⎥⎥⎥⎥.
⎢⎢⎣ ⎥⎥⎦
223.5
568 Chapter 11  Transition matrices and Leslie matrices 11C

Practical applications of transition matrices

6 Two fast-food outlets, Jill’s and Pete’s, are located in a small town.
In a given week:
 80% of people who go to Jill’s return the next 20%
week
 20% of people who go to Jill’s go to Pete’s the 80% J P 75%
next week
 25% of people who go to Pete’s go to Jill’s the 25%
next week
 75% of people who go to Pete’s return the next week.
a Construct a transition matrix to describe this situation. Call the matrix T .
b Initially, 400 people eat at Jill’s and 400 eat at Pete’s. Write down a column matrix
S 0 that describes this situation.
c How many of these people do we expect to go to Jill’s the next week? How many to
Pete’s?
d How many do we expect to go to Jill’s after 5 weeks? How many to Pete’s?
e In the long term, how many do we expect to eat at Jill’s each week? How many of
these people do we expect to eat at Pete’s?

7 Imagine that we live in a world in which people are either ‘happy’ or ‘unhappy’, but
the way people feel can change from day to day.
In this world:
 90% of people who are happy today 10%
will be happy tomorrow
 10% of people who are happy today 90% H U 40%
will be unhappy tomorrow
 40% of people who are unhappy 60%
today will be unhappy tomorrow
 60% of people who are unhappy today will be happy tomorrow.

a Construct a transition matrix to describe this situation. Call the matrix T .

b On a given day, out of 2000 people, 1500 are happy and 500 are unhappy. Write
down a column matrix, S 0 , that describes this situation.
c The next day, how many of these people do we expect to be ‘happy’ and how many
‘unhappy’?
d After 4 days, how many of these 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’?
11C
C 11C Transition matrices – using recursion 569

8 In another model of this world, people can be ‘happy’, ‘neither happy nor sad’, or
‘sad’, but the way people feel can change from day to day.
The transition matrix opposite shows how H N S
people’s feelings may vary from day to day ⎡ ⎤
H ⎢⎢⎢0.80 0.40 0.35⎥⎥⎥
in this world, and the proportions of people ⎢⎢ ⎥⎥
T = N ⎢⎢⎢⎢⎢0.15 0.30 0.40⎥⎥⎥⎥⎥
involved. ⎢⎣ ⎥⎦
S 0.05 0.30 0.25

In the transition matrix, the columns define the situation today and the rows define the
situation tomorrow.
a On a given day, out of 2000 people, 1200 are ‘happy’, 600 are ‘neither happy nor
sad’ and 200 are ‘sad’. Write down a matrix, S 0 , that describes this situation.
b The next day, how many people do we expect to be happy?
c After 5 days, how many people do we expect to be happy?
d In the long term, how many of the 2000 people do we expect to be happy?

Exam 1 style questions

9 Students at a boarding school have a choices of two breakfast cereals, Crispies (C)
and Krunchies (K). The change in the percentage of students who have each cereal on
consecutive days is descibed by the transition matrix T shown below.
Today
C K
⎡ ⎤
C ⎢⎢⎢⎢0.42 0.56⎥⎥⎥⎥
T = ⎢⎢⎣ ⎥⎥⎦ Tomorrow
K 0.58 0.44

On Monday 25% of the students ate Crispies. What percentage of the students ate
Krunchies on Tuesday?
A 47.5% B 48% C 50.25% D 52.5% E 62%

10 A factory employs the same number of workers each day. The workers are allocated to
work with either machine A or machine B. The workers may be allocated to work on a
different machine from day to day, as shown in the transition matrix below.
A B
⎡ ⎤
A ⎢⎢⎢⎢0.32 0.16⎥⎥⎥⎥
T = ⎢⎢⎣ ⎥⎥⎦
B 0.68 0.84
Machine A has 72 workers each day on it. Each day, the number of workers machine B
will be
A 24 B 36 C 72 D 288 E 306
 570 Chapter 11  Transition matrices and Leslie matrices 11C

11 Consider the matrix recurrence relation below.

⎡ ⎤ ⎡ ⎤
⎢⎢⎢10⎥⎥⎥ ⎢⎢⎢ x 0.4 y ⎥⎥⎥
⎢⎢⎢ ⎥⎥⎥ ⎢⎢⎢ ⎥⎥⎥
S 0 = ⎢⎢⎢⎢⎢20⎥⎥⎥⎥⎥ , S n+1 = T S n where T = ⎢⎢⎢⎢⎢0.6 z 0.4⎥⎥⎥⎥⎥
⎢⎢⎣ ⎥⎥⎦ ⎢⎢⎣ ⎥⎥⎦
30 0.1 0.2 w
⎡ ⎤
⎢⎢⎢14⎥⎥⎥
⎢⎢⎢ ⎥⎥⎥
Matrix T is a regular transition matrix. Given that S 1 = ⎢⎢⎢⎢⎢26⎥⎥⎥⎥⎥ which of the following is
⎢⎢⎣ ⎥⎥⎦
20
true?
A x = 0.3, y = 0.1, z = 0.4, w = 0.5 B x = 0.3, y = 0.3, z = 0.4, w = 0.7
C x = 0.2, y = 0.7, z = 0.3, w = 0.3 D x = 0.2, y = 0.8, z = 0.3, w = 0.2
E x = 0.3, y = 0.6, z = 0.4, w = 0.4

11D Transition matrices – using the rule Sn+1 = TSn + B

Learning intentions
 To be able to use a matrix recurrence relation S0 = initial state matrix, Sn+1 = T Sn + B
to model systems that include external additions or reductions at each step of the
process.

To date, we have only considered matrix recurrence models of the form

S 0 = initial state matrix, S n+1 = T S n
This recurrence model can be used to model situations where the total number of objects in the
system, like cars, machines, people or birds, remains unchanged. For example, in the car rental
problem 90 cars are available for rental. But what happens if management wants to increase the
total number of cars available for rent by adding, say, an extra car at each location each week?
To allow for this situation we need to use the matrix recurrence relation:
S 0 = initial state matrix, S n+1 = T S n + B
where B is a column matrix.
The next example applies this model to the rental car problem.

Example 8 Determining the nth state of a system using the rule Sn = T n S0 + B

A rental starts with 90 cars, 50 located at Bendigo and 40 located at Colac.

Cars are usually rented and returned in the B – Bendigo
same town. However, a small percentage
C – Colac
of cars rented in Bendigo are returned 20%
in Colac and vice versa. The transition 80% B C 90%
10%
diagram opposite gives these percentages.
 11D Transition matrices – using the rule Sn+1 = TSn + B 571

To increase the number of cars, two extra cars are added to the rental fleet at each location
each week. The recurrence relation that can be used to model this situation is:
⎡ ⎤ ⎡ ⎤ ⎡ ⎤
⎢⎢⎢50⎥⎥⎥ ⎢⎢⎢0.8 0.1⎥⎥⎥ ⎢ ⎥
S 0 = ⎢⎢⎢⎣ ⎥⎥⎥⎦ , S n+1 = T S n + B where T = ⎢⎢⎢⎣ ⎥⎥⎥ and B = ⎢⎢⎢⎢⎢2⎥⎥⎥⎥⎥
⎦ ⎣ ⎦
40 0.2 0.9 2
Determine the number of cars at Bendigo and Colac after:
a 1 week b 2 weeks.

Explanation Solution
a Use the rule S 1 = T S o + B to determine S 1 = T S0 + B
the state matrix after 1 week and write ⎡ ⎤⎡ ⎤ ⎡ ⎤
⎢⎢⎢0.8 0.1⎥⎥⎥ ⎢⎢⎢50⎥⎥⎥ ⎢⎢⎢2⎥⎥⎥
your conclusion. = ⎢⎢⎢⎣ ⎥⎥⎥ ⎢⎢⎢ ⎥⎥⎥ + ⎢⎢⎢ ⎥⎥⎥
⎦⎣ ⎦ ⎣ ⎦
0.2 0.9 40 2
⎡ ⎤ ⎡ ⎤ ⎡ ⎤
⎢⎢⎢44⎥⎥⎥ ⎢⎢⎢2⎥⎥⎥ ⎢⎢⎢46⎥⎥⎥
= ⎢⎢⎢⎣ ⎥⎥⎥⎦ + ⎢⎢⎢⎣ ⎥⎥⎥⎦ = ⎢⎢⎢⎣ ⎥⎥⎥⎦
46 2 48
Thus, we predict that there will be 46 cars
in Bendigo and 48 cars in Colac.
b Use the rule S 2 = T S 1 + B to determine S 2 = T S1 + B
the state matrix after 2 weeks and write ⎡ ⎤⎡ ⎤ ⎡ ⎤
⎢⎢⎢0.8 0.1⎥⎥⎥ ⎢⎢⎢46⎥⎥⎥ ⎢⎢⎢2⎥⎥⎥
your conclusion. = ⎢⎢⎢⎣ ⎥⎥⎥ ⎢⎢⎢ ⎥⎥⎥ + ⎢⎢⎢ ⎥⎥⎥
⎦⎣ ⎦ ⎣ ⎦
0.2 0.9 48 2
⎡ ⎤ ⎡ ⎤ ⎡ ⎤
⎢⎢⎢41.6⎥⎥⎥ ⎢⎢⎢2⎥⎥⎥ ⎢⎢⎢43.6⎥⎥⎥
= ⎢⎢⎢⎣ ⎥⎥⎥⎦ + ⎢⎢⎢⎣ ⎥⎥⎥⎦ = ⎢⎢⎢⎣ ⎥⎥⎥⎦
52.4 2 54.4
Thus, we predict that there will be 43.6
cars in Bendigo and 54.4 cars in Colac.

Unfortunately, the recurrence rule S n+1 = T S n + B does not lead to a simple rule for the state
matrix after n steps, so we need to work our way through this sort of problem step-by-step.

Using the inverse matrix of a transition matrix

In the above we have seen how to move from left to right in the sequence of state matrices by
applying S n+1 = T S n + B
S 0 , S 1 , S 2 , . . . S n , S n+1 . . .
We can move from right to left through the transition states by the observation
S n+1 = T S n + B and S n = T −1 (S n+1 − B)
Question 4 in the Exercise can be completed in this way.
 572 Chapter 11  Transition matrices and Leslie matrices 11D

Skill-
sheet Exercise 11D

Using a recurrence rule to calculate state matrices

⎡ ⎤ ⎡ ⎤
⎢⎢⎢0.6 0.2⎥⎥⎥ ⎢ ⎥
1 For the transition matrix T = ⎢⎢⎢⎣ ⎥⎥⎥ and the state matrix S 0 = ⎢⎢⎢⎢⎢100⎥⎥⎥⎥⎥:
Example 7
⎦ ⎣ ⎦
0.4 0.8 100
a use the recurrence rule S n+1 = T S n to determine:
i S1 ii S 3 ⎡ ⎤
⎢⎢⎢10⎥⎥⎥
b use the recurrence rule S n+1 = T S n + R, where R = ⎢⎢⎢⎣ ⎥⎥⎥⎦, to determine:
5
i S1 ii S 2
⎡ ⎤
⎢⎢⎢−20⎥⎥⎥
c use the recurrence rule S n+1 = T S n − B, where B = ⎢⎣⎢ ⎢ ⎥⎥⎥, to determine:
⎦
20
i S1 ii S 2

Practical application
2 On Windy Island, sea birds are observed nesting at three sites: A, B and C. The
following transition matrix and accompanying transition diagram can be used to predict
the movement of sea birds between these sites from year to year.
This year 10%
A B C 100% 80%
⎡ ⎤
⎢⎢⎢ 1.0 0.10 0.05⎥⎥⎥ A A B
⎢⎢ ⎥⎥
T = ⎢⎢⎢⎢⎢ 0 0.80 0.05⎥⎥⎥⎥⎥ B Next year 5%
⎢⎣ ⎥⎦
5% 10%
0 0.10 0.90 C
C 90%
⎡ ⎤
⎢⎢⎢10 000⎥⎥⎥
⎢⎢⎢ ⎥⎥⎥
Initially, 10 000 sea birds were observed nesting at each site, so S 0 = ⎢⎢⎢⎢⎢10 000⎥⎥⎥⎥⎥.
⎢⎢⎣ ⎥⎥⎦
10 000
a Use the recurrence rule S n+1 = T S n to:
i determine S 1 , the state matrix after 1 year
ii predict the number of sea birds nesting at site B after 2 years.
b Without calculation, write down the number of sea birds predicted to nest at each of
the three sites in the long term. Explain why this can be done without calculation.
c To help solve the problem of having all the birds eventually nesting at site A, the
ranger suggests that 2000 sea birds could be removed from site A each year and
relocated in equal numbers to sites B and C.
The state matrix, S 2 , is now given by
S2 = T S1 + N
11D 11D Transition matrices – using the rule Sn+1 = TSn + B 573

⎡ ⎤ ⎡ ⎤ ⎡ ⎤
⎢⎢⎢10000⎥⎥⎥ ⎢⎢⎢1.0 0.10 0.05⎥⎥⎥ ⎢⎢⎢−2000⎥⎥⎥
⎢⎢⎢ ⎥⎥⎥ ⎢⎢⎢ ⎥⎥⎥ ⎢⎢⎢ ⎥⎥⎥
where S 1 = ⎢⎢⎢⎢⎢10000⎥⎥⎥⎥⎥ , T = ⎢⎢⎢⎢⎢ 0 0.80 0.05⎥⎥⎥⎥⎥ and N = ⎢⎢⎢⎢⎢ 1000 ⎥⎥⎥⎥⎥ .
⎢⎢⎣ ⎥⎥⎦ ⎢⎢⎣ ⎥⎥⎦ ⎢⎢⎣ ⎥⎥⎦
10000 0 0.10 0.90 1000
Evaluate:
i S2 ii S 3 (assuming that S 3 = T S 2 + N) iii S 4 (assuming that S 4 = T S 3 + N).

Exam 1 style questions

3 The matrix S n+1 is determined from the matrix S n using the recurrence relation
S n+1 = T × S n − C, where
⎡ ⎤ ⎡ ⎤ ⎡ ⎤
⎢⎢⎢0.6 0.75 0.1⎥⎥⎥ ⎢⎢⎢2000⎥⎥⎥ ⎢⎢⎢1975⎥⎥⎥
⎢⎢⎢ ⎥⎥⎥ ⎢⎢⎢ ⎥⎥⎥ ⎢⎢⎢ ⎥⎥⎥
T = ⎢⎢⎢⎢⎢0.2 0.2 0.1⎥⎥⎥⎥⎥ , S 0 = ⎢⎢⎢⎢⎢1000⎥⎥⎥⎥⎥ , S 1 = ⎢⎢⎢⎢⎢ 650 ⎥⎥⎥⎥⎥ .
⎢⎢⎣ ⎥⎥⎦ ⎢⎢⎣ ⎥⎥⎦ ⎢⎢⎣ ⎥⎥⎦
0.2 0.05 0.8 1000 1125
and C is a column matrix. Matrix S 2 =
⎡ ⎤ ⎡ ⎤ ⎡ ⎤ ⎡ ⎤ ⎡ ⎤
⎢⎢⎢⎢ 1860 ⎥⎥⎥⎥ ⎢⎢⎢ 1700 ⎥⎥⎥
⎢⎢⎢ ⎥⎥⎥
⎢⎢⎢ 1710 ⎥⎥⎥
⎢⎢⎢ ⎥⎥⎥
⎢⎢⎢ 1600 ⎥⎥⎥
⎢⎢⎢ ⎥⎥⎥
⎢⎢⎢ 1650 ⎥⎥⎥
⎢⎢⎢ ⎥⎥⎥
⎢⎢ ⎥⎥
A ⎢⎢⎢⎢ 687.5 ⎥⎥⎥⎥ B ⎢⎢⎢⎢ 560.5 ⎥⎥⎥⎥ C ⎢⎢⎢⎢ 587.5 ⎥⎥⎥⎥ D ⎢⎢⎢⎢1725.5⎥⎥⎥⎥ E ⎢⎢⎢⎢ 550.5 ⎥⎥⎥⎥
⎢⎢⎢ ⎥⎥⎥ ⎢⎢⎢ ⎥⎥⎥ ⎢⎢⎢ ⎥⎥⎥ ⎢⎢⎢ ⎥⎥⎥ ⎢⎢⎢ ⎥⎥⎥
⎣ ⎦ ⎣ ⎦ ⎣ ⎦ ⎣ ⎦ ⎣ ⎦
1452.5 1022.5 1202.5 1200.5 1032.5

4 Supporters of a football team attend home games. There are 3 areas, bays A, B and C,
where they sit. There is considerable moving of position from game to game and the
numbers attending the home games gradually decline as the year progresses. Let Xn
be the state matrix that shows the number of supporters in each bay n weeks into the
the season. The number of supporters in each location can be determined by the matrix
recurrence relation
Xn+1 = T Xn − D
where
This game
⎡ ⎤
⎡
A B C
⎤ ⎢⎢⎢70⎥⎥⎥
⎢⎢⎢ 0.1 0.2 0.5 ⎥⎥⎥ A ⎢⎢⎢ ⎥⎥⎥
⎢⎢ ⎥⎥ and D = ⎢⎢⎢⎢⎢70⎥⎥⎥⎥⎥
T = ⎢⎢⎢⎢⎢ 0.3 0.7 0.2 ⎥⎥⎥⎥⎥ B Next game ⎢⎢⎣ ⎥⎥⎦
⎢⎣ ⎥⎦ 70
0.6 0.1 0.3 C

⎡ ⎤
⎢⎢⎢ 9830 ⎥⎥⎥
⎢⎢⎢ ⎥⎥⎥
If X3 = ⎢⎢⎢⎢⎢11 130⎥⎥⎥⎥⎥ then X2 =
⎢⎢⎣ ⎥⎥⎦
7830
⎡ ⎤ ⎡ ⎤ ⎡ ⎤ ⎡ ⎤ ⎡ ⎤
⎢⎢⎢12370⎥⎥⎥ ⎢⎢⎢ 4000 ⎥⎥⎥ ⎢⎢⎢12 000⎥⎥⎥ ⎢⎢⎢ 7500 ⎥⎥⎥ ⎢⎢⎢ 79300 ⎥⎥⎥
⎢⎢⎢ ⎥⎥⎥ ⎢⎢⎢ ⎥⎥⎥ ⎢⎢⎢ ⎥⎥⎥ ⎢⎢⎢ ⎥⎥⎥ ⎢⎢⎢ ⎥⎥⎥
A ⎢⎢⎢⎢ 9510 ⎥⎥⎥⎥ B ⎢⎢⎢⎢10 000⎥⎥⎥⎥ C ⎢⎢⎢⎢11 000⎥⎥⎥⎥ D ⎢⎢⎢⎢ 8600 ⎥⎥⎥⎥ E ⎢⎢⎢⎢245232⎥⎥⎥⎥
⎢⎢⎢⎣ ⎥⎥⎥⎦ ⎢⎢⎢⎣ ⎥⎥⎥⎦ ⎢⎢⎢
⎣
⎥⎥⎥
⎦
⎢⎢⎢
⎣
⎥⎥⎥
⎦
⎢⎢⎢
⎣
⎥⎥⎥
⎦
4650 15 000 5000 12 000 231011
 574 Chapter 11  Transition matrices and Leslie matrices

11E Leslie matrices

Learning intentions
 To use and interpret Leslie matrices to analyse changes in population over time.

Leslie matrices are used to construct discrete models of population growth. In particular, they
are used to model changes in the sizes of different age groups within a population.

The general setting

Leslie matrices were developed by Patrick Holt Leslie (1900–1972) while he was working in
the Bureau of Animal Population at the University of Oxford. They are used by biologists and
ecologists to model changes over time in various animal populations.
Age groups First the population is divided into age groups. The time period for each age
group is the same length. Together they cover the life span of the population. For example,
in a study of a human population, we could use a time period of 10 years and consider eleven
age groups as follows:

Age group(i) 1 2 3 4 ··· 10 11

Age range (years) 0–10 10–20 20–30 30–40 ··· 90–100 100–110

Note: Only the females of the species are counted in the population, as they are the ones who give birth to
the new members of the population.

A Leslie matrix is a transition matrix that can be used to describe the way population changes
over time. It takes into account two factors for the females in each age group: the birth rate, bi ,
and survival rate, si , where i is the number of the age group.
Birth rates We ignore migration, and so the population growth is entirely due to new female
births. The birth rate, bi , for age group i is the average number of female offspring from a mother
in age group i during one time period. For example, average birth rate of women in age group 4
(20 − 30 years) might be 1.7 female children for the 10 year period.
Survival rates The survival rate, s i , for age group i is the proportion of the population in
age group i that progress to age group i + 1. Note that 0 ≤ s i ≤ 1.
For example, the survival rate for age group 2 might be 0.95, that is 95% of females in this
10 − 20 year age group would survive to progress to age group 3, 20 − 30 years.
Note: The survival rate of the last age group (100 − 110) is taken to be 0.

A simple example
We start with a simple example where the life span of the species is 9 years. We will divide the
population into three age groups. This means we use a time period of 3 years.
 11E Leslie matrices 575

Age group(i) 1 2 3
Age range (years) 0–3 3–6 6–9

A Leslie matrix for three age groups is a 3 × 3 matrix of the form

⎡ ⎤
⎢⎢⎢b1 b2 b3 ⎥⎥⎥
⎢⎢⎢ ⎥⎥⎥
L = ⎢⎢⎢⎢⎢ s1 0 0 ⎥⎥⎥⎥⎥
⎢⎢⎣ ⎥⎥⎦
0 s2 0
Suppose that the survival rates are s1 = 0.6 and s2 = 0.3, and that the birth rates are b1 = 0,
b2 = 2.3 and b3 = 0.4. Then the Leslie matrix is
From age group i
1 2 3
⎡ ⎤
⎢⎢⎢0 2.3 0.4 ⎥⎥⎥ 1 birth rate
⎢⎢⎢ ⎥⎥⎥
L = ⎢⎢⎢⎢⎢0.6 0 0 ⎥⎥⎥⎥⎥ 2 survival rate To age group i + 1
⎢⎢⎣ ⎥⎥⎦
0 0.3 0 3 survival rate

Life cycle transition diagram

The above Leslie matrix can be represented by a diagram which we will refer to as a Life cycle
transition diagram.

0.4
2.3

1 2 3
0.6 0.3

The population state matrix

The population state matrix is a column matrix that lists the number in each age group at a
given time.
The initial population state matrix is denoted by S 0 . Suppose that for our example, initially
the population has 400 females in each age group. We represent the initial population state
matrix S 0 , as a 3 × 1 column matrix as shown below.
Age group
⎡ ⎤
⎢⎢⎢400⎥⎥⎥ 1
⎢⎢⎢ ⎥⎥⎥
S 0 = ⎢⎢⎢⎢⎢400⎥⎥⎥⎥⎥ 2
⎢⎢⎣ ⎥⎥⎦
400 3

We can now use the Leslie matrix, L, in combination with the initial state matrix S 0 to generate
the state matrix after one time period, S 1 , to find the size of each age group after one time period
(3 years) as follows:
576 Chapter 11  Transition matrices and Leslie matrices

⎡ ⎤⎡ ⎤ ⎡ ⎤
⎢⎢⎢0 2.3 0.4 ⎥⎥⎥ ⎢⎢⎢400⎥⎥⎥ ⎢⎢⎢1080⎥⎥⎥
⎢⎢⎢ ⎥⎥⎥ ⎢⎢⎢ ⎥⎥⎥ ⎢⎢⎢ ⎥⎥⎥
S 1 = LS 0 = ⎢⎢⎢⎢⎢0.6 0 0 ⎥⎥⎥⎥⎥ ⎢⎢⎢⎢⎢400⎥⎥⎥⎥⎥ = ⎢⎢⎢⎢⎢ 240⎥⎥⎥⎥⎥
⎢⎢⎣ ⎥⎥⎦ ⎢⎢⎣ ⎥⎥⎦ ⎢⎢⎣ ⎥⎥⎦
0 0.3 0 400 120

Thus after one time period, there are 1080 females in age group 1, 240 in age group 2 and 120
in age group 3 and the total population size has increased from 1200 (= 400 + 400 + 400) to
1440 (= 1080 + 240 + 120). Similarly, to find the number in each age group after two time
periods we calculate S 2 from S 1 as follows:
⎡ ⎤⎡ ⎤ ⎡ ⎤
⎢⎢⎢0 2.3 0.4 ⎥⎥⎥ ⎢⎢⎢1080⎥⎥⎥ ⎢⎢⎢600⎥⎥⎥
⎢⎢⎢ ⎥⎥⎥ ⎢⎢⎢ ⎥⎥⎥ ⎢⎢⎢ ⎥⎥⎥
S 2 = LS 1 = ⎢⎢⎢⎢⎢0.6 0 0 ⎥⎥⎥⎥⎥ ⎢⎢⎢⎢⎢ 240⎥⎥⎥⎥⎥ = ⎢⎢⎢⎢⎢648⎥⎥⎥⎥⎥
⎢⎢⎣ ⎥⎥⎦ ⎢⎢⎣ ⎥⎥⎦ ⎢⎢⎣ ⎥⎥⎦
0 0.3 0 120 72
Thus, after two time periods, there are 600 females in age group 1648 in age group 2 and 72 in
age group 3 and the over-all population size has decreased to 1320.
Finding the population matrix Sn after n time periods.
To speed up the process we can make use of the explicit formula for the state matrix S n after n
time periods. Notice that there is a pattern when calculating the population state matrices:
S 1 = LS 0
S 2 = LS 1 = L2 S 0
S 3 = LS 2 = L3 S 0
..
.
S n+1 = LS n = Ln S 0
In general, we can find the population matrix S n using the rule
S n = Ln S 0
Using this rule, to find S 3 , we have
⎡ ⎤3 ⎡ ⎤ ⎡ ⎤
⎢⎢⎢0 2.3 0.4⎥⎥⎥⎥ ⎢⎢⎢⎢400⎥⎥⎥⎥ ⎢⎢⎢⎢1519.2⎥⎥⎥⎥
⎢⎢⎢ ⎥⎥ ⎢⎢ ⎥⎥ ⎢⎢ ⎥⎥
S 3 = L3 S 0 = ⎢⎢⎢⎢⎢0.6 0 0 ⎥⎥⎥⎥⎥ ⎢⎢⎢⎢⎢400⎥⎥⎥⎥⎥ = ⎢⎢⎢⎢⎢ 360⎥⎥⎥⎥⎥
⎢⎢⎣ ⎥⎥⎦ ⎢⎢⎣ ⎥⎥⎦ ⎢⎢⎣ ⎥⎥⎦
0 0.3 0 400 194.4
Continuing in this way, we can see the change over time in the total population and in the
distribution of the age groups.

Change in the population over time

Time period 0 1 2 3 4 5
Age 0–3 years 400 1080 600 1519.2 905.76 2139.70
Age 3–6 years 400 240 648 360.0 911.52 543.46
Age 6–9 years 400 120 72 194.4 108.00 273.46
Total 1200 1440 1320 2073.6 1925.28 2956.61
 11E Leslie matrices 577

Leslie matrices
An m × m Leslie matrix has the form
⎡ ⎤
⎢⎢⎢b1 b2 b3 · · · bm−1 bm ⎥⎥⎥
⎢⎢⎢ ⎥⎥⎥
⎢⎢⎢ s 0 0 · · · 0 0 ⎥⎥⎥
⎢⎢⎢⎢ 1 ⎥⎥⎥⎥
⎢⎢⎢ 0 s · · · ⎥⎥⎥
⎢⎢⎢ 2 0 0 0 ⎥⎥⎥
L = ⎢⎢⎢ ⎥
⎢⎢⎢ 0 0 s3 · · · 0 0 ⎥⎥⎥⎥⎥
⎢⎢⎢ . .. ⎥⎥⎥⎥⎥
⎢⎢⎢ . .. .. . . ..
⎢⎢⎢ . . . . . . ⎥⎥⎥
⎢⎣ ⎥⎥⎦
0 0 0 · · · sm−1 0
where:
 m is the number of age groups being considered
 s i , the survival rate, is the proportion of the population in age group i that progress to
age group i + 1
 bi , the birth rate, is the average number of female offspring from a mother in
age group i during one time period.
Leslie matrix and its interpretation
From age group
1 2 3 4
⎡ ⎤
⎢⎢⎢ 0 1.4 1.2 0.3⎥⎥⎥ 1
⎢⎢⎢ ⎥⎥⎥
⎢⎢⎢0.6 0 ⎥⎥⎥ 2
⎢ 0 0 ⎥⎥⎥
L = ⎢⎢⎢⎢ ⎥⎥⎥ To age group
⎢⎢⎢ 0 0.5 0 0 ⎥
⎥ 3
⎢⎢⎢ ⎥⎥⎥
⎣ ⎦
0 0 0.1 0 4

This is a Leslie matrix with 4 age groups. The corresponding life-cycle transition
diagram is shown here.
0.3
1.2
1.4
1 2 3 4
0.6 0.5 0.1

Recursive rules

The population matrix S n is an m × 1 matrix representing the size of each age group after
n time periods. This is calculated using a recursive formula
S 0 is the initial state matrix, S n+1 = LS n
or the explicit rule
S n = Ln S 0
578 Chapter 11  Transition matrices and Leslie matrices

Example 9 Determining state matrices and life cycle diagrams

Use the Leslie matrix and initial state matrix below to answer the following questions.
From age group
1 2 3 4
⎡ ⎤ ⎡ ⎤
⎢⎢⎢ 0 1.8 2.6 0.1⎥⎥⎥ 1 ⎢⎢⎢1000⎥⎥⎥
⎢⎢⎢ ⎥⎥⎥ ⎢⎢⎢ ⎥
⎢⎢⎢0.2 0 ⎥⎥⎥ 2 ⎢⎢⎢ 0 ⎥⎥⎥⎥⎥
⎢ 0 0 ⎥⎥⎥ ⎢ ⎥⎥⎥
L = ⎢⎢⎢⎢ ⎥⎥⎥ S 0 = ⎢⎢⎢⎢
⎢⎢⎢ 0 ⎥⎥⎥⎥⎥
To age group
⎢⎢⎢ 0 0.4 0 0 ⎥
⎥ 3
⎢⎢⎢ ⎥⎥⎥ ⎢⎢⎢⎣ ⎥⎥⎥⎦
⎣ ⎦
0 0 0.3 0 4 0
a Write down
i the birth rate for age group 2 ii the survival rate for age group 3
b Complete a life cycle diagram for this Leslie matrix.
c Evaluate the following population state matrices.
S 1 , S 5 and S 20
⎡ ⎤
⎢⎢⎢9.53⎥⎥⎥
⎢⎢⎢⎢ ⎥⎥
⎢⎢⎢2.42⎥⎥⎥⎥⎥
= ⎢⎢⎢⎢ ⎥⎥, determine S 17
⎢⎢⎢1.22⎥⎥⎥⎥⎥
d Given that S 16
⎢⎢⎢ ⎥⎥⎥
⎣ ⎦
0.16

Explanation Solution
a i The birth rate for age group 2 is Birth rate for age group 2 = 1.8
given in the matrix position, row 1,
column 2.
ii The survival rate for age group 3 is Survival rate for age group 3 = 0.3
given in the matrix position, row 4,
column 3.
b Survival rates 0.1
s1 = 0.2, s2 = 0.4, s3 = 0.3 2.6
Birth rates 1.8
1 2 3 4
b2 = 1.8, b3 = 2.6, b4 = 0.1 0.2 0.4 0.3

c S 1 = LS 0 Using a calculator.
⎡ ⎤ ⎡ ⎤ ⎡ ⎤
S5 = L S0
5 ⎢⎢⎢ 0 ⎥⎥⎥ ⎢⎢⎢149.76⎥⎥⎥ ⎢⎢⎢3.84⎥⎥⎥
⎢⎢⎢⎢ ⎥⎥ ⎢⎢⎢⎢ ⎥⎥ ⎢⎢⎢ ⎥
⎢⎢⎢0.97⎥⎥⎥⎥⎥
S 20 = L20 S 0 ⎢⎢⎢200⎥⎥⎥⎥⎥ ⎢⎢⎢ 26.4 ⎥⎥⎥⎥⎥
S 1 = ⎢⎢⎢⎢ ⎥⎥, S = ⎢⎢ ⎥⎥, S = ⎢⎢⎢ ⎥⎥⎥
⎢⎢⎢ 0 ⎥⎥⎥⎥⎥ 5 ⎢⎢⎢⎢⎢ 16.64 ⎥⎥⎥⎥⎥ 20 ⎢⎢⎢⎢⎢0.49⎥⎥⎥⎥⎥
⎢⎢⎢ ⎥⎥⎥ ⎢⎢⎢ ⎥⎥⎥ ⎢⎢⎢ ⎥⎥⎥
⎣ ⎦ ⎣ ⎦ ⎣ ⎦
0 8.64 0.19
 11E Leslie matrices 579

d S 17 = LS 16 ⎡ ⎤
⎢⎢⎢7.54⎥⎥⎥
⎢⎢⎢⎢ ⎥⎥
(Further investigation would reveal that ⎢⎢⎢1.91⎥⎥⎥⎥⎥
= ⎢⎢⎢⎢ ⎥⎥
⎢⎢⎢0.97⎥⎥⎥⎥⎥
the population continues to decrease S 17
over time.) ⎢⎢⎢ ⎥⎥⎥
⎣ ⎦
0.37

Example 10 Entering information in a Leslie matrix and state matrix

Information about a population of female goats is given in the following table.

Age group (years) 0−1 1−2 2−3 3−4 4−5

Initial population 10 25 40 20 15
Birth rates 0 0.2 0.9 0 0
Survival rates 0.6 0.7 0.5 0.2 0

a Write down the initial population state matrix, S 0 .

b Using the information in the table above, write down a Leslie matrix to describe the
change in population of female goats.
c Construct a life cycle transition diagram for this Leslie matrix.
d Determine the number of 3 − 4 year old female goats in the population after 3 years.
Round your answer to the nearest whole number.

Explanation Solution
a Enter the initial population ⎡ ⎤
⎢⎢⎢10⎥⎥⎥
numbers into a 5 × 1 matrix. ⎢⎢⎢⎢ ⎥⎥⎥⎥
⎢⎢⎢25⎥⎥⎥
⎢⎢⎢ ⎥⎥⎥
S 0 = ⎢⎢⎢⎢⎢40⎥⎥⎥⎥⎥
⎢⎢⎢ ⎥⎥⎥
⎢⎢⎢⎢20⎥⎥⎥⎥
⎢⎢⎣ ⎥⎥⎦
15

b Enter the birthrates and survival ⎡ ⎤

⎢⎢⎢ 0 0.2 0.9 0 0⎥⎥⎥⎥
rates into a 5 × 5 matrix. ⎢⎢⎢ ⎥⎥
⎢⎢⎢0.6 0 0⎥⎥⎥⎥⎥
⎢⎢⎢⎢ 0 0
⎥⎥
L = ⎢⎢⎢⎢ 0 0.7 0⎢ 0 0⎥⎥⎥⎥⎥
⎢⎢⎢ ⎥⎥⎥
⎢⎢⎢ 0 0 0.5 0 0⎥⎥⎥⎥
⎢⎢⎢ ⎥⎥⎦
⎣
0 0 0 0.2 0
580 Chapter 11  Transition matrices and Leslie matrices

c Survival rates
0 0
s1 = 0.6, s2 = 0.7, s3 = 0.5, 0.9
s4 = 0.2 0.2
1 2 3 4 5
Birth rates 0.6 0.7 0.5 0.2
b2 = 0, 2, b3 = 0.9, b4 = 0,
b5 = 0

⎡ ⎤3 ⎡ ⎤ ⎡ ⎤
⎢⎢⎢ 0 0.2 0.9 0 0⎥⎥⎥⎥ ⎢⎢⎢⎢10⎥⎥⎥⎥ ⎢⎢⎢⎢ 8.7 ⎥⎥⎥⎥
⎢⎢⎢ ⎥⎥ ⎢⎢ ⎥⎥ ⎢⎢ ⎥⎥
⎢⎢⎢0.6 0 0⎥⎥⎥⎥⎥ ⎢⎢⎢⎢⎢25⎥⎥⎥⎥⎥ ⎢⎢⎢⎢⎢10.17⎥⎥⎥⎥⎥
⎢⎢⎢ 0 0
⎥⎥ ⎢⎢ ⎥⎥ ⎢⎢ ⎥⎥
⎢⎢⎢
d S 3 = L3 S 0 ⎢⎢⎢⎢ 0 0.7 0 0 0⎥⎥⎥⎥⎥ ⎢⎢⎢⎢⎢40⎥⎥⎥⎥⎥ = ⎢⎢⎢⎢⎢17.22⎥⎥⎥⎥⎥
⎢⎢⎢ ⎥⎥⎥ ⎢⎢⎢ ⎥⎥⎥ ⎢⎢⎢ ⎥⎥⎥
⎢⎢⎢ 0 0 0.5 0 0⎥⎥⎥⎥ ⎢⎢⎢⎢20⎥⎥⎥⎥ ⎢⎢⎢⎢ 2.1 ⎥⎥⎥⎥
⎢⎢⎣ ⎥⎥⎦ ⎢⎢⎣ ⎥⎥⎦ ⎢⎢⎣ ⎥⎥⎦
0 0 0 0.2 0 15 1.75

There are two goats in the 3-4 year old age group
in this population.

Long term (limiting) behaviour of population numbers

The following examples demonstrate numerical techniques for modelling the use of Leslie
matrices.

Example 11 Limiting behaviour for Leslie matrices

Consider the following Leslie matrix L and initial population matrix S 0 :

⎡ ⎤ ⎡ ⎤
⎢⎢⎢0 4 4 ⎥⎥⎥ ⎢⎢⎢1000⎥⎥⎥
⎢⎢⎢ ⎥⎥⎥ ⎢⎢⎢ ⎥⎥⎥
L = ⎢⎢⎢⎢⎢0.25 0 0⎥⎥⎥⎥⎥ and S 0 = ⎢⎢⎢⎢⎢ 0⎥⎥⎥⎥⎥
⎢⎢⎣ ⎥⎥⎦ ⎢⎢⎣ ⎥⎥⎦
0 0.5 0 0
a Find
i S5 ii S 10 iii S 50

Premultiply each of these state matrices by 1 1 1 to calculate the total populations
at each of these stages and comment.
b Determine S 25 and S 26 . Divide each age group population for S 26 by the correspond-
ing age group population for S 25 and show that S 26 ≈ 1.1915S 25 and comment.

Solution
a
⎡ ⎤ ⎡ ⎤ ⎡ ⎤
⎢⎢⎢1000⎥⎥⎥ ⎢⎢⎢ 2500⎥⎥⎥ ⎢⎢⎢ 2 777 063⎥⎥⎥
⎢⎢⎢ ⎥⎥⎥ ⎢⎢⎢ ⎥⎥⎥ ⎢⎢⎢ ⎥⎥⎥
i S 5 = ⎢⎢⎢⎢ 250⎥⎥⎥⎥ ii S 10 = ⎢⎢⎢⎢⎢531.25⎥⎥⎥⎥⎥ iii S 50 = ⎢⎢⎢⎢⎢582 688.05⎥⎥⎥⎥⎥
⎢⎢⎢ ⎥⎥⎥ ⎢⎢⎣ ⎥⎥⎦ ⎢⎢⎣ ⎥⎥⎦
⎣ ⎦
62.5 218.75 244521.18
 11E Leslie matrices 581

Finding the total populations according to this model


i 1 1 1 S 5 = [1312.5]. The population is approximately 1312 after 5 years.

ii 1 1 1 S 10 = [3250]. The population is approximately 3250 after 10 years.

iii 1 1 1 S 50 = [3 604 272.2]. The population is approximately 3 604 272
after 50 years.
A numerical investigation reveals that the population increases without bound.
b We calculate S 25 and S 26 :
⎡ ⎤ ⎡ ⎤
⎢⎢⎢34 781.25⎥⎥⎥ ⎢⎢⎢41 441.41⎥⎥⎥
⎢⎢⎢ ⎥⎥⎥ ⎢⎢⎢ ⎥⎥⎥
S 25 = L25 S 0 = ⎢⎢⎢⎢⎢ 7297.85⎥⎥⎥⎥⎥ , S 26 = L26 S 0 = ⎢⎢⎢⎢⎢ 8695.31⎥⎥⎥⎥⎥
⎢⎢⎣ ⎥⎥⎦ ⎢⎢⎣ ⎥⎥⎦
3062.50 3648.93
Then we can find the rate of increase in each age group during the 26th time period:
41 441.41 8695.31 3648.93
≈ ≈ ≈ 1.1915
34 781.25 7297.85 3062.50
This suggests that the age-group proportions have stabilised after the 25 time periods.
The long-term growth rate is approximately 1.19. (That is, after a certain stage, the
population is increasing by 19% each time period.)
Note: Try different entries in S 0 to see if you get the same behaviour. The long-term growth rate is
largely dependent on the Leslie matrix L.

Limiting behaviour of Leslie matrices

Often we will find that, after a long enough time, the proportion of the population in
each age group does not change from one time period to the next. This happens if we can
find a real number k such that LS n+1 = kS n for some sufficiently large n. This does not
happen with every Leslie matrix as we see in Example 13.

Example 12 A Leslie matrix and state matrix with constant rate of increase

Consider the following Leslie matrix L and initial population matrix S 0 :

⎡ ⎤ ⎡ ⎤
⎢⎢⎢0 2.3 0.4 ⎥⎥⎥ ⎢⎢⎢1600⎥⎥⎥
⎢⎢⎢ ⎥⎥⎥ ⎢⎢⎢ ⎥⎥⎥
L = ⎢⎢⎢⎢⎢0.6 0 0 ⎥⎥⎥⎥⎥ and S 0 = ⎢⎢⎢⎢⎢ 800⎥⎥⎥⎥⎥ .
⎢⎢⎣ ⎥⎥⎦ ⎢⎢⎣ ⎥⎥⎦
0 0.3 0 200
a Determine
i S 10 ii S 14 iii S 15
b The rate of increase of the population is a constant and each of the age group
populations increase in the same way. Find this rate by comparing S 14 and S 15 .
c Confirm the ratio of the age group populations stays constant for S 0 , S 1 and S 10 at
8:4:1
582 Chapter 11  Transition matrices and Leslie matrices

Solution
⎡ ⎤ ⎡ ⎤ ⎡ ⎤
a ⎢⎢⎢9906.78⎥⎥⎥ ⎢⎢⎢20542.70⎥⎥⎥ ⎢⎢⎢ 24651.24⎥⎥⎥
⎢⎢⎢ ⎥⎥⎥ ⎢⎢⎢ ⎥⎥⎥ ⎢⎢⎢ ⎥⎥⎥
i S 10 = ⎢⎢⎢⎢⎢4953.39⎥⎥⎥⎥⎥ ii S 14 = ⎢⎢⎢⎢⎢10271.35⎥⎥⎥⎥⎥ iii S 15 = ⎢⎢⎢⎢⎢ 12325.62⎥⎥⎥⎥⎥
⎢⎢⎣ ⎥⎥⎦ ⎢⎢⎣ ⎥⎥⎦ ⎢⎢⎣ ⎥⎥⎦
1238.35 2567.84 3081.4043
By comparing S 14 and S 15 ,
24651.24 12325.62 3081.4043
= = ≈ 1.2
20542.70 10271.35 2567.84
we find that the growth rate is 1.2.
c 8 : 4 : 1 = 1600 : 800 : 200 ≈ 9906.78 : 4953.39 : 1238.35
b

Example 13 Periodic, increasing and decreasing rates of change

Consider the following Leslie matrix L and initial population matrix S 0 :

⎡ ⎤ ⎡ ⎤
⎢⎢⎢0 0 b ⎥⎥⎥ ⎢⎢⎢1000⎥⎥⎥
⎢⎢⎢ ⎥⎥⎥ ⎢⎢⎢ ⎥⎥⎥
3

L = ⎢⎢⎢⎢⎢0.25 0 0 ⎥⎥⎥⎥⎥ and S 0 = ⎢⎢⎢⎢⎢ 0⎥⎥⎥⎥⎥

⎢⎢⎣ ⎥⎥⎦ ⎢⎢⎣ ⎥⎥⎦
0 0.5 0 0
Investigate the long-term behaviour of the population if:
a b3 = 8 b b3 = 4 c b3 = 10

Solution
a Let b3 = 8. Use your calculator to store the matrices L and S 0 . Then compute:
⎡ ⎤ ⎡ ⎤ ⎡ ⎤
⎢⎢⎢ 0⎥⎥⎥ ⎢⎢⎢ 0⎥⎥⎥⎥ ⎢⎢⎢1000⎥⎥⎥
⎢⎢⎢ ⎥⎥⎥ ⎢⎢⎢ ⎥⎥ ⎢⎢⎢ ⎥⎥⎥
S 1 = LS 0 = ⎢⎢⎢⎢⎢250⎥⎥⎥⎥⎥ , S 2 = L2 S 0 = ⎢⎢⎢⎢⎢ 0⎥⎥⎥⎥⎥ , S 3 = L3 S 0 = ⎢⎢⎢⎢⎢ 0⎥⎥⎥⎥⎥
⎢⎢⎣ ⎥⎥⎦ ⎢⎢⎣ ⎥⎥⎦ ⎢⎢⎣ ⎥⎥⎦
0 125 0
The population will continue to cycle through these three states; this is because L3 = I.
b Let b3 = 4. Then a numerical investigation suggests that the population decreases over
the long term:
⎡ ⎤ ⎡ ⎤ ⎡ ⎤
⎢⎢⎢ 0⎥⎥⎥ ⎢⎢⎢ 0 ⎥⎥⎥ ⎢⎢⎢0 ⎥⎥⎥
⎢⎢⎢ ⎥⎥⎥ ⎢⎢⎢ ⎥⎥⎥ ⎢⎢⎢ ⎥⎥⎥
S 1 = LS 0 = ⎢⎢⎢⎢⎢250⎥⎥⎥⎥⎥ , S 5 = L5 S 0 = ⎢⎢⎢⎢⎢ 0 ⎥⎥⎥⎥⎥ , S 50 = L50 S 0 = ⎢⎢⎢⎢⎢0 ⎥⎥⎥
⎥⎥⎥
⎢⎢⎣ ⎥⎥⎦ ⎢⎢⎣ ⎥⎥⎦ ⎢⎢⎣ ⎥⎦
0 62.5 0.0019
c Let b3 = 10. Then a numerical investigation suggests that the population increases over
the long term:
⎡ ⎤ ⎡ ⎤ ⎡ ⎤
⎢⎢⎢ 0⎥⎥⎥ ⎢⎢⎢ 0 ⎥⎥⎥ ⎢⎢⎢ 0 ⎥⎥⎥
⎢⎢⎢ ⎥⎥⎥ ⎢⎢⎢ ⎥⎥⎥ ⎢⎢⎢ ⎥⎥⎥
S 1 = LS 0 = ⎢⎢⎢⎢⎢250⎥⎥⎥⎥⎥ , S 5 = L5 S 0 = ⎢⎢⎢⎢⎢ 0 ⎥⎥⎥ ,
⎥⎥⎥⎥ S 50 = L50 S 0 = ⎢⎢⎢⎢⎢ 0 ⎥⎥⎥⎥⎥
⎢⎢⎣ ⎥⎥⎦ ⎢⎢⎣ ⎦ ⎢⎢⎣ ⎥⎥⎦
0 156.25 4440.89

Note: A population can increase, decrease, become constant or oscillate.

 11E 11E Leslie matrices 583

Skill-
sheet Exercise 11E

Determining state matrices and life cycle diagrams

Example 8 1 Use the Leslie matrix and initial state matrix below to answer the following questions.
From age group
1 2 3 4
⎡ ⎤ ⎡ ⎤
⎢⎢⎢ 0 1.9 2.1 1.1⎥⎥⎥ 1 ⎢⎢⎢100⎥⎥⎥
⎢⎢⎢ ⎥⎥⎥ ⎢⎢⎢ ⎥
⎢⎢⎢0.7 0 ⎥⎥⎥ 2 ⎢⎢⎢100⎥⎥⎥⎥⎥
⎢ 0 0 ⎥⎥⎥ ⎢ ⎥⎥⎥
L = ⎢⎢⎢⎢ ⎥⎥⎥ S 0 = ⎢⎢⎢⎢
⎢⎢⎢100⎥⎥⎥⎥⎥
To age group
⎢⎢⎢ 0 0.5 0 0 ⎥
⎥ 3
⎢⎢⎢ ⎥⎥⎥ ⎢⎢⎢⎣ ⎥⎥⎥⎦
⎣ ⎦
0 0 0.6 0 4 100
a Write down
i The birth rate for age group 2 ii The survival rate for age group 3
b Complete the life cycle diagram for this Leslie matrix.
c Evaluate the following population state matrices.
i S1 ii S 3 iii S 20
⎡ ⎤
⎢⎢⎢2613⎥⎥⎥
⎢⎢⎢ ⎥
⎢⎢⎢1200⎥⎥⎥⎥⎥
⎢ ⎥⎥⎥
d Given that S 7 = ⎢⎢⎢⎢
⎢⎢⎢ 485 ⎥⎥⎥⎥⎥
determine S 8 . Give your values correct to the nearest whole
⎢⎢⎢ ⎥⎥⎥
⎣ ⎦
168
number.

2 Complete the life cycle diagram corresponding to each of the following Leslie
matrices:
⎡ ⎤ ⎡ ⎤
⎢⎢⎢ 0 2.9 3.1 2.1⎥⎥⎥ ⎡ ⎤ ⎢⎢⎢ 0 0 3 8⎥⎥⎥⎥
⎢⎢⎢ ⎥⎥⎥ ⎢⎢⎢ 0 0 0.42 ⎥⎥⎥ ⎢⎢⎢ ⎥⎥
⎢⎢⎢0.8 0 ⎥
⎥ ⎢ ⎥ ⎢
⎢ 0⎥⎥⎥⎥⎥
⎢ 0 0 ⎥⎥⎥ ⎢⎢ ⎥⎥ ⎢⎢0.4 0 0
a ⎢⎢⎢⎢ ⎥⎥⎥ b ⎢⎢⎢⎢0.6 0 0 ⎥⎥⎥⎥⎥ c ⎢⎢⎢⎢ ⎥⎥⎥
⎢⎢⎢⎢ 0 0.7 0 0 ⎥⎥⎥⎥ ⎢
⎢⎣⎢ ⎥⎦⎥ ⎢⎢⎢⎢ 0 0.5 0 0⎥⎥⎥⎥
⎢⎢⎣ ⎥⎥⎦ 0 0.75 0 ⎢⎢⎣ ⎥⎥⎦
0 0 0.5 0 0 0 0.25 0
3 Construct the Leslie matrix corresponding to each life cycle diagram.
a 2.4 b 3
1.3 2.3
1 2 3 1 2 3
0.7 0.6 0.6 0.3

c 0.6
2.6
1.4
1 2 3 4
0.5 0.4 0.05
 584 Chapter 11  Transition matrices and Leslie matrices 11E

Entering information in a Leslie matrix and state matrix

Example 9 4 Information about a population of female kangaroos in a particular area is given in the
following table.

Age group (years) 0−4 4−8 8 − 12 12 − 16 4−5

Initial population 15 20 30 15 10
Birth rates 0 0.2 0.9 1.1 0
Survival rates 0.8 0.9 0.7 0.8 0

a Write down the initial population state matrix, S 0 .

b Write down the Leslie matrix.
c Complete the life cycle diagram for this Leslie matrix.
d Determine the population state matrix after
i one year, (S 1 ) ii after 5 years, (S 5 ).
e Determine the number of 4 − 8 year old female kangaroos in the population after 5
years.
f State the initial total population.

g Use multiplication of state matrices by the matrix 1 1 1 1 1 to find the total
population after
i one year, ii after 5 years iii after 10 years.
h It is conjectured that the population is increasing by about 10% per annum.
Calculate each of the following and comment
i 1.1 × 90 ii 1.15 × 90 iii 1.110 × 90.

5 Information about a population of female locusts is given in the following table.

Stage Eggs Nymphs Adults

Initial population 0 0 50
Birth rates 0 0 1000
Survival rates 0.02 0.05 0

a Write down the initial population state matrix, S 0 .

b Write down the Leslie matrix.
c Construct a time life-cycle transition diagram for this Leslie matrix.
d Determine the population state matrix after
i one year, (S 1 ) ii after 3 years, (S 3 ) iii after 4 years, (S 4 ).
 11E 11E Leslie matrices 585

e If the Initial population is now:

Stage Eggs Nymphs Adults

Initial population 50 100 50

find the populations of each after

i one year (S 1 ) ii after 3 years (S 3 ). iii after 4 years (S 4 ).

Limiting behaviour for Leslie matrices

Example 10 6 Consider the following Leslie matrix L and initial population state matrix S 0 :
⎡ ⎤ ⎡ ⎤
⎢⎢⎢0 2 1⎥⎥⎥⎥ ⎢⎢⎢204⎥⎥⎥
⎢⎢⎢ ⎥⎥ ⎢
⎢⎢ ⎥⎥⎥
L = ⎢⎢⎢⎢⎢0.5 0 0⎥⎥⎥⎥⎥ and S 0 = ⎢⎢⎢⎢⎢ 96⎥⎥⎥⎥⎥
⎢⎢⎣ ⎥⎥⎦ ⎢⎢⎣ ⎥⎥⎦
0 0.25 0 23
a Find
i S5 ii S 10 iii S 20

Premultiply each of these state matrices by 1 1 1 to calculate the total
populations at each of these stages and comment.
b Determine S 20 and S 21 . Divide each age group population for S 21 by the
corresponding age group population for S 20 and show that S 21 ≈ 1.057S 20 and
comment.

Example 11 7 A Leslie matrix that models a certain population of female animals is

⎡ ⎤
⎢⎢⎢0 2.5 1 ⎥⎥⎥
⎢⎢⎢ ⎥⎥⎥
L = ⎢⎢⎢⎢⎢0.6 0 0⎥⎥⎥⎥⎥
⎢⎢⎣ ⎥⎥⎦
0 0.25 0
where the animals have a maximum life span of 9 years, and the population has been
divided into three age groups of 3 years each.
a Assume that each age group initially consists of 400 females. What is the number of
females in each age group after:
i 3 years ii 6 years iii 9 years?
⎡ ⎤
⎢⎢⎢767⎥⎥⎥
⎢⎢⎢ ⎥⎥⎥
b Now assume that the initial population is 1200 and S 0 = ⎢⎢⎢⎢362⎥⎥⎥⎥ . Find S n after
⎢⎢⎢ ⎥⎥⎥
⎣ ⎦
71
i 3 years ii 6 years iii 9 years.
 586 Chapter 11  Transition matrices and Leslie matrices 11E

c Calculate
i 1.27S 0 ii 1.272 S 0 iii 1.273 S 0
Compare these answers to the answers of part b

Example 12 8 Consider the following Leslie matrix L and initial population matrix S 0 :
⎡ ⎤ ⎡ ⎤
⎢⎢⎢ 0 0 12⎥⎥⎥ ⎢⎢⎢1200⎥⎥⎥
⎢⎢⎢ ⎥⎥⎥ ⎢⎢⎢ ⎥⎥⎥
L = ⎢⎢⎢⎢⎢ 14 0 0 ⎥⎥⎥⎥⎥ and S 0 = ⎢⎢⎢⎢⎢ 0⎥⎥⎥⎥⎥
⎢⎢⎣ ⎥⎥⎦ ⎢⎢⎣ ⎥⎥⎦
0 13 0 0
a Find:
i LS 0 ii L2 S 0 iii L3 S 0
b Comment on these results in terms of the population behaviour. Try using a different
initial population matrix S 0 .
c Now investigate for each of the following Leslie matrices. Comment on population
increase or decrease.
⎡ ⎤ ⎡ ⎤
⎢⎢⎢ 0 0 6⎥⎥⎥ ⎢⎢⎢ 0 0 15⎥⎥⎥⎥
⎢⎢⎢ ⎥⎥⎥ ⎢⎢⎢ ⎥⎥
i L = ⎢⎢⎢⎢ 14 0 0⎥⎥⎥⎥ ii L = ⎢⎢⎢⎢ 14 0 0 ⎥⎥⎥⎥⎥
⎢⎢⎢ ⎥⎥⎥ ⎢⎢⎢ ⎥⎥⎦
⎣ ⎦ ⎣
0 13 0 0 1
3 0

9 A Leslie matrix that models a certain population of female insects is

⎡ ⎤
⎢⎢⎢0 3 2 2⎥⎥⎥
⎢⎢⎢⎢ ⎥⎥⎥⎥
⎢⎢⎢0.5 0 0 0 ⎥⎥⎥
L = ⎢⎢⎢⎢ ⎥⎥⎥
⎥⎥⎥
⎢⎢⎢0 0.5 0 0 ⎥⎥⎥
⎢⎢⎢ ⎥⎦
⎣
0 0 0.1 0
where the insects have a maximum life span of 4 months, and the population has been
divided into four age groups of 1 month each.
Assume that each age group initially consists of 400 females. What is the number of
females in each age group after:
a 1 month b 2 months c 3 months?

10 For a certain species of fish, we consider three age groups each of one year in length.
These fish reproduce only during their third year and then die. Assume that 20% of fish
survive their first year and that 50% of these survivors make it to reproduction age. The
initial population consists of 1000 newborns.
a Investigate what happens for each of the following values of b3 :
i b3 = 10 ii b3 = 15 iii b3 = 6
b For b3 = 20, determine the long-term growth rate and the proportion of fish in each
age group.
11E 11E Leslie matrices 587

Exam 1 style questions

11 The Leslie matrix for a certain endangered species is:

⎡ ⎤
⎢⎢⎢0.9 2.5 0.4⎥⎥⎥
⎢⎢⎢ ⎥⎥⎥
L = ⎢⎢⎢⎢⎢0.3 0 0 ⎥⎥⎥⎥⎥
⎢⎢⎣ ⎥⎥⎦
0 0.45 0
Some of the species were moved into a sanctuary. The initial female population in the
sanctuary is given by
⎡ ⎤
⎢⎢⎢130⎥⎥⎥
⎢⎢⎢ ⎥⎥⎥
S 0 = ⎢⎢⎢⎢⎢ 40 ⎥⎥⎥⎥⎥
⎢⎢⎣ ⎥⎥⎦
20
The best estimate of the total female population after 7 years is
A 1000 B 1500 C 2000 D 2500 E 3000
⎡ ⎤
⎢⎢⎢0 2 b⎥⎥⎥
⎢⎢⎢ ⎥⎥⎥
12 The Leslie matrix L = ⎢⎢⎢⎢c 0 0⎥⎥⎥⎥ satisfies the matrix equation
⎢⎢⎢ ⎥⎥⎥⎦
⎣
0 d 0
⎡ ⎤ ⎡ ⎤
⎢⎢⎢16⎥⎥⎥ ⎢⎢⎢16⎥⎥⎥
⎢⎢⎢ ⎥⎥⎥ ⎢⎢⎢ ⎥⎥⎥
L ⎢⎢⎢⎢⎢ 4 ⎥⎥⎥⎥⎥ = ⎢⎢⎢⎢⎢ 4 ⎥⎥⎥⎥⎥
⎢⎢⎣ ⎥⎥⎦ ⎢⎢⎣ ⎥⎥⎦
2 2
The values of b, c and d are
1 1 1 1 1 1
A b = 2, d = , c = B b = 2, d = , c = C b = 4, d = ,c =
4 2 2 4 4 2
1 1 1 1
D b = 4, d = , c = E b = 2, d = , c =
2 4 4 2
13 A population of birds is modelled by using the Leslie matrix
⎡ ⎤
⎢⎢⎢ 0 2 1.5⎥⎥⎥⎥
⎢⎢⎢ ⎥⎥
L = ⎢⎢⎢⎢⎢0.44 0 0 ⎥⎥⎥⎥⎥
⎢⎢⎣ ⎥⎥⎦
0 0.55 0
The growth has reached the point where the rates of growth of the different age groups
⎡ ⎤
⎢⎢⎢1000⎥⎥⎥
⎢⎢⎢ ⎥⎥⎥
of the population are constant and the state matrix at this point is S k = ⎢⎢⎢⎢⎢ 400 ⎥⎥⎥⎥⎥. The rate
⎢⎢⎣ ⎥⎥⎦
200
of growth per time period is
A 10% B 11% C 12%
D 13% E 14%
Review 588 Chapter 11  Transition matrices and Leslie matrices

Key ideas and chapter summary

State matrix A state matrix S n is a column matrix whose elements represent the
Transition nth state of a dynamic system defined by a recurrence relation of the
Assign-
ment matrixes
form: S 0 = initial state, S n+1 = T S n . Here T is a square matrix called a
transition matrix.

Steady-state The steady-state matrix, S, represents the equilibrium state of a

matrix system. For regular matrices, this equilibrium state of a system can be
estimated by calculating T n S 0 for a large value of n.

Leslie matrices  An m × m Leslie matrix has the form

⎡ ⎤
⎢⎢⎢b1 b2 b3 · · · bm−1 bm ⎥⎥⎥
⎢⎢⎢ ⎥⎥⎥
⎢⎢⎢ s ··· 0 ⎥⎥⎥⎥⎥
⎢⎢⎢⎢ 1 0 0 0
⎥⎥
⎢⎢⎢ 0 s ··· 0 ⎥⎥⎥⎥⎥
⎢ 0 0
L = ⎢⎢⎢⎢⎢
2
⎥⎥
⎢⎢⎢ 0 0 s3 ··· 0 0 ⎥⎥⎥⎥⎥
⎢⎢⎢ . .. ⎥⎥⎥⎥⎥
⎢⎢⎢ . .. .. .. ..
⎢⎢⎢ . . . . . . ⎥⎥⎥
⎢⎣ ⎥⎥⎦
0 0 0 · · · sm−1 0
where:
• m is the number of age groups being considered
• si , the survival rate, is the proportion of the population in
age group i that progress to age group i + 1
• bi , the birth rate, is the average number of female offspring from a
mother in age group i during one time period.
 The population matrix S n is an m × 1 matrix representing the size
of each age group after n time periods. This is calculated using a
recursive formula
S 0 is the initial state matrix, S n+1 = LS n
or the explicit rule
S n = Ln S 0

Skills checklist

Download this checklist from the Interactive Textbook, then print it and fill it out to check
Check- your skills. 
list

11A 1 I can set up a transition matrix from a diagram.

See Example 1, and Exercise 11A Question 1

 Chapter 11 Review 589

Review
11A 2 I can set up a transition matrix from a written information.

See Example 2, and Exercise 11A Question 2

11B 3 I can interpret a transition matrix and a transition diagram.

See Example 3, and Exercise 11B Question 1

11C 4 I can use a recurrence relation to calculate state matrices step by step

See Example 4, and Exercise 11C Question 1

11C 5 I can use a recurrence relation Sn+1 = Tn S0 to determine the nth state.

See Example 5, and Exercise 11C Question 2

11C 6 I can use the inverse of a transition matrix.

See Example 6, and Exercise 11C Question 4

11C 7 I can estimate steady state solution for suitable transition matrices.

See Example 7, and Exercise 11C Question 4

11D 8 I can use the matrix recurrence relation S0 = initial state matrix, Sn+1 = TSn + B.

See Example 8, and Exercise 11D Question 1

11E 9 I can determine state matrices and construct life cycle diagrams in situations
modelled by Leslie matrices.

See Example 9, and Exercise 11E Question 1

11E 10 I can enter information into a Leslie matrix from written information.

See Example 10, and Exercise 11E Question 2

11E 11 I can use numerical techniques to consider the limiting behaviour of Leslie
matrices.

See Example 12, and Exercise 11E Question 3

11E 12 I can identify the properties of a Leslie matrix and the state matrices when
there is a constant rate.

See Example 13, and Exercise 11E Question 4

Multiple choice questions

1 The transition matrix that can be used to represent 25%
the information in the diagram shown is:
75% A B 95%

5%
Review 590 Chapter 11  Transition matrices and Leslie matrices

From From
⎡ A B ⎤ ⎡ A B ⎤
A To A ⎢⎢⎢0.75 0.25⎥⎥⎥ B To A ⎢⎢⎢0.75 0.05⎥⎥⎥
⎢⎢⎣ ⎥⎥⎦ ⎢⎢⎣ ⎥⎥⎦
B 0.05 0.95 B 0.25 0.95

From From
A B
⎡ A B ⎤
⎡ ⎤
C To A ⎢⎢⎢0.75 0.25⎥⎥⎥ D To A ⎢⎢⎢0.75 0.95⎥⎥⎥
⎢⎢⎣ ⎥⎥⎦ ⎢⎢⎣ ⎥⎥⎦
B 0.95 0.05 B 0.25 0.05

From
A B
⎡ ⎤
E To A ⎢⎢⎢0.25 0.05⎥⎥⎥
⎢⎢⎣ ⎥⎥⎦
B 0.75 0.95

2 The transition matrix that can be used to represent 60%

the information in the diagram shown is:
X Y Z 10% Y 35%
⎡ ⎤ 75%
A To X ⎢⎢⎢0.75 0.05 0.30⎥⎥⎥
⎢⎢⎢ ⎥⎥
X 5% 20%
Y ⎢⎢⎢⎢0.10 0.60 0.20⎥⎥⎥⎥⎥ Z
⎢⎣ ⎥⎦ 15%
Z 0.15 0.35 0.50 50%
30%

X Y Z X Y Z
⎡ ⎤ ⎡ ⎤
B To X ⎢⎢⎢0.75 0.10 0.15⎥⎥⎥ C To X ⎢⎢⎢0.75 0.10 0.15⎥⎥⎥
⎢⎢⎢ ⎥⎥ ⎢⎢⎢ ⎥⎥
Y ⎢⎢⎢⎢0.60 0.05 0.35⎥⎥⎥⎥⎥ Y ⎢⎢⎢⎢0.10 0.05 0.35⎥⎥⎥⎥⎥
⎢⎣ ⎥⎦ ⎢⎣ ⎥⎦
Z 0.50 0.30 0.20 Z 0.50 0.30 0.20

X Y Z X Y Z
⎡ ⎤ ⎡ ⎤
D To X ⎢⎢⎢0.75 0.05 0.15⎥⎥⎥ E To X ⎢⎢⎢0.75 0.05 0.15⎥⎥⎥
⎢⎢⎢ ⎥⎥ ⎢⎢⎢ ⎥⎥
Y ⎢⎢⎢⎢0.10 0.60 0.20⎥⎥⎥⎥⎥ Y ⎢⎢⎢⎢0.15 0.35 0.50⎥⎥⎥⎥⎥
⎢⎣ ⎥⎦ ⎢⎣ ⎥⎦
Z 0.15 0.35 0.50 Z 0.10 0.60 0.20

The following information is needed for Questions 3 to 8.

⎡ ⎤ ⎡ ⎤
⎢⎢⎢0.6 0.5⎥⎥⎥ ⎢⎢⎢100⎥⎥⎥
3 A system is defined by a transition matrix T = ⎢⎢⎢⎣ ⎥⎥⎥⎦ with S 0 = ⎢⎢⎢⎣ ⎥⎥⎥.
⎦
0.4 0.5 200
For this system, S 1 =
⎡ ⎤ ⎡ ⎤ ⎡ ⎤ ⎡ ⎤ ⎡ ⎤
⎢⎢⎢ 60 ⎥⎥⎥ ⎢⎢⎢140⎥⎥⎥ ⎢⎢⎢160⎥⎥⎥ ⎢⎢⎢166⎥⎥⎥ ⎢⎢⎢200⎥⎥⎥
A ⎢⎢⎢⎣ ⎥⎥⎥
⎦ B ⎢⎢⎢ ⎥⎥⎥
⎣ ⎦ C ⎢⎢⎢
⎣
⎥⎥⎥
⎦ D ⎢⎢⎢
⎣
⎥⎥⎥
⎦ E ⎢⎢⎢ ⎥⎥⎥
⎣ ⎦
200 160 140 144 100
 Chapter 11 Review 591

Review
4 For this system, T 2 is:
⎡ ⎤ ⎡ ⎤ ⎡ ⎤ ⎡ ⎤
⎢⎢⎢⎢0.36 0.25⎥⎥⎥⎥ ⎢⎢⎢0.56 0.55⎥⎥⎥
⎢⎢⎢ ⎥⎥⎥
⎢⎢⎢0.6 0.5⎥⎥⎥
⎢ ⎥⎥⎥
⎢⎢⎢1.2 1.0⎥⎥⎥
⎢ ⎥⎥⎥
A ⎢⎢⎣ ⎥⎥⎦ B ⎣ ⎦ C ⎢⎢⎣ ⎦ D ⎢⎢⎣ ⎦
0.16 0.25 0.44 0.45 0.4 0.5 0.8 1.0
E not defined

5 For this system, S 3 is closest to:

⎡ ⎤ ⎡ ⎤ ⎡ ⎤ ⎡ ⎤ ⎡ ⎤
⎢⎢⎢⎢160⎥⎥⎥⎥ ⎢⎢⎢166.6⎥⎥⎥ ⎢⎢⎢166.7⎥⎥⎥ ⎢⎢⎢640⎥⎥⎥ ⎢⎢⎢400⎥⎥⎥
A ⎢⎢⎣ ⎥⎥⎦ B ⎢⎢⎢⎣ ⎥⎥⎥
⎦ C ⎢⎢⎢⎣ ⎥⎥⎥
⎦ D ⎢⎢⎢⎣ ⎥⎥⎥
⎦ E ⎢⎢⎢⎣ ⎥⎥⎥
⎦
140 133.4 133.3 560 800

6 For this system, the steady-state matrix is closest to:

⎡ ⎤ ⎡ ⎤ ⎡ ⎤ ⎡ ⎤ ⎡ ⎤
⎢⎢⎢166.5⎥⎥⎥ ⎢⎢⎢166.6⎥⎥⎥ ⎢⎢⎢166.7⎥⎥⎥ ⎢⎢⎢166.8⎥⎥⎥ ⎢⎢⎢166.9⎥⎥⎥
A ⎢⎢⎣ ⎢ ⎥
⎥⎥⎦ B ⎢⎢⎣ ⎢ ⎥
⎥⎥⎦ C ⎢⎢⎣ ⎢ ⎥
⎥⎥⎦ D ⎢⎢⎣ ⎢ ⎥⎥⎥ E ⎢⎢⎣ ⎢ ⎥⎥⎥
⎦ ⎦
133.5 133.4 133.3 133.2 133.1
⎡ ⎤
⎢⎢⎢10⎥⎥⎥
7 If L1 = T S 0 + B, where B = ⎢⎢⎢⎣ ⎥⎥⎥⎦, then L1 equals:
20
⎡ ⎤ ⎡ ⎤ ⎡ ⎤ ⎡ ⎤ ⎡ ⎤
⎢⎢⎢ 70 ⎥⎥⎥ ⎢⎢⎢150⎥⎥⎥ ⎢⎢⎢170⎥⎥⎥ ⎢⎢⎢176⎥⎥⎥ ⎢⎢⎢210⎥⎥⎥
A ⎢⎢⎢⎣ ⎥⎥⎥
⎦ B ⎢⎢⎢⎣ ⎥⎥⎥
⎦ C ⎢⎢⎢⎣ ⎥⎥⎥
⎦ D ⎢⎢⎢⎣ ⎥⎥⎥
⎦ E ⎢⎢⎢⎣ ⎥⎥⎥
⎦
220 180 160 164 120
⎡ ⎤
⎢⎢⎢10⎥⎥⎥
8 If P1 = T S 0 − 2B, where B = ⎢⎢⎢⎣ ⎥⎥⎥⎦, then P1 equals:
20
⎡ ⎤ ⎡ ⎤ ⎡ ⎤ ⎡ ⎤ ⎡ ⎤
⎢⎢⎢140⎥⎥⎥ ⎢⎢⎢170⎥⎥⎥ ⎢⎢⎢180⎥⎥⎥ ⎢⎢⎢170⎥⎥⎥ ⎢⎢⎢180⎥⎥⎥
A ⎢⎢⎢⎣ ⎥⎥⎥
⎦ B ⎢⎢⎢⎣ ⎥⎥⎥
⎦ C ⎢⎢⎢⎣ ⎥⎥⎥
⎦ D ⎢⎢⎢⎣ ⎥⎥⎥
⎦ E ⎢⎢⎢⎣ ⎥⎥⎥
⎦
100 100 100 160 180

9 A system of state matrices S n is defined by the matrix equation S n+1 = GS n where

⎡ ⎤
⎢⎢⎢ 0 −0.5⎥⎥⎥
G = ⎢⎢⎢⎣ ⎥⎥⎥⎦.
1.5 0.5
⎡ ⎤
⎢⎢⎢10⎥⎥⎥
If S 1 = ⎢⎢⎣⎢ ⎥⎥⎦⎥ , then S 2 equals:
20
⎡ ⎤ ⎡ ⎤ ⎡ ⎤ ⎡ ⎤ ⎡ ⎤
⎢⎢⎢−12.5⎥⎥⎥ ⎢⎢⎢−10⎥⎥⎥ ⎢⎢⎢10⎥⎥⎥ ⎢⎢⎢10⎥⎥⎥ ⎢⎢⎢15⎥⎥⎥
A ⎢⎢⎢⎣ ⎥⎥⎥
⎦ B ⎢⎢⎢
⎣
⎥⎥⎥
⎦ C ⎢⎢⎢ ⎥⎥⎥
⎣ ⎦ D ⎢⎢⎢ ⎥⎥⎥
⎣ ⎦ E ⎢⎢⎢ ⎥⎥⎥
⎣ ⎦
−2.5 25 20 25 30
⎡ ⎤ ⎡ ⎤
⎢⎢⎢0.5 0.6⎥⎥⎥ ⎢ ⎥
10 T = ⎢⎢⎢⎣ ⎥⎥⎥ is a transition matrix. S 5 = ⎢⎢⎢⎢⎢22⎥⎥⎥⎥⎥ is a state matrix.
⎦ ⎣ ⎦
0.5 0.4 18
If S 5 = T S 4 , then S 4 equals:
⎡ ⎤ ⎡ ⎤ ⎡ ⎤ ⎡ ⎤ ⎡ ⎤
⎢⎢⎢18⎥⎥⎥ ⎢⎢⎢20⎥⎥⎥ ⎢⎢⎢21.8⎥⎥⎥ ⎢⎢⎢22⎥⎥⎥ ⎢⎢⎢18.2⎥⎥⎥
⎢
A ⎢⎢⎣ ⎥⎥⎦ ⎥ ⎢
B ⎢⎢⎣ ⎥⎥⎦ ⎥ C ⎢⎢⎣ ⎢ ⎥
⎥⎥⎦ D ⎢⎢⎢⎣ ⎥⎥⎥⎦ E ⎢⎣⎢ ⎢ ⎥⎥⎥
⎦
22 20 18.2 18 21.2
Review 592 Chapter 11  Transition matrices and Leslie matrices

11 A large population of birds lives on a remote island. Every night each bird settles at either
location A or location B.
On the first night the number of birds at each location was the same. On each subsequent
night, a percentage of birds changed the location at which they settled.
The movement of birds between the two locations is described A B
by the transition matrix T shown opposite. Assume this pattern ⎡ ⎤
A ⎢⎢⎢⎢0.8 0⎥⎥⎥⎥
of movement continues. T= ⎢⎢⎣ ⎥⎥⎦
B 0.2 1
In the long term, the number of birds that settle at location A will:

A not change B gradually decrease to zero C gradually increase

D eventually settle at around 20% of the island’s bird population
E eventually settle at around 80% of the island’s bird population

Use the following information to answer Questions 12 and 13.

Two politicians, Rob and Anna, are the only
Number of people
candidates for a forthcoming election. At the
planing to vote
beginning of the election campaign, people
Candidate for candidate
were asked for whom they planned to vote.
Rob 5692
The numbers were as per the table.
Anna 3450

During the election campaign, it is expected 24%

that people may change the candidate that
they plan to vote for each week according to
75% Rob Anna 76%
the transition diagram shown.

25%

12 The total number of people who are expected to change the candidate that they plan to vote
for 1 week after the election campaign begins is:
A 828 B 1423 C 2251 D 4269 E 6891

13 The election campaign will run for 10 weeks. If people continue to follow this pattern of
changing the candidate they plan to vote for, the expected winner after 10 weeks will be:
A Rob by about 50 votes B Rob by about 100 votes
C Rob by fewer than 10 votes D Anna by about 100 votes
E Anna by about 200 votes
 Chapter 11 Review 593

Review
Written response questions
1 The Diisco (D) and the Spin (S) are two large music venues in the same city. They both
open on the same Saturday night and will open on every Saturday night.
⎡ ⎤
The matrix A1 opposite is the attendance matrix for the first ⎢⎢⎢500⎥⎥⎥ D
Saturday. This matrix shows the number of people who attended A1 = ⎢⎢⎢⎣ ⎥⎥⎥
⎦
240 S
the first Diisco and the number of people who attended the Spin.
The number of people expected to attend the second Saturday for each venue can be
determined using the matrix equation
A2 = GA1
This Saturday
D S
⎡ ⎤
⎢⎢⎢1.2 −0.4⎥⎥⎥ D Next Saturday
where G is the matrix G = ⎢⎣ ⎢ ⎥⎥⎦
0.2 0.6 S

a i Determine A2 , the attendance matrix for the second Saturday.

ii What was the total attendance on the second Saturday.
Assume that the attendance matrices for successive Saturdays can be determined as
follows:
A3 = GA2 , A4 = GA3 , and so on such that An+1 = GAn
b Determine the attendance matrix (with the elements written correct to the nearest whole
number) for the eighth Saturday.
c Describe the way in which the number of people attending the Diiscoo is expected to
change over the next 80 or so Saturdays.
Suppose instead that 500 people attend the first Diiscoo, and 490 people attend the Spin.
d Describe the way in which the attendance at both venues changes if attendance follows
this prediction.

2 Suppose that the trees in a forest are classified into three age groups: young trees
(0–15 years), middle-aged trees (16–30 years) and old trees (more than 30 years). A time
period is 15 years, and it is assumed that in each time period:
 10% of young trees, 20% of middle-aged trees and 40% of old trees die
 surviving trees enter into the next age group; old trees remain old
 dead trees are replaced by young trees.
Review 594 Chapter 11  Transition matrices and Leslie matrices

Complete the 3 × 3 transition matrix T to describe this.

Y M O
⎡ ⎤
Y ⎢⎢⎢⎢ ⎥⎥⎥
⎥⎥⎥
⎢⎢⎢ ⎥⎥⎥
M ⎢⎢⎢⎢ ⎥⎥⎥
⎢⎢⎣ ⎥⎦
O

3 The following table represents a study of a particular population of marsupials, which has
been divided into eight age groups. The table gives the initial population, birth rate and
survival rate for each age group.

Age group 1 2 3 4 5 6 7 8
Initial population 0 100 100 50 0 0 0 0
Birth rate 0 0.1 0.9 0.2 0 0 0 0
Survival rate 0.98 0.95 0.95 0.9 0.7 0.5 0.1 0

a Write down the Leslie matrix for this population.

b Calculate S2 and S3 .
c Estimate the long-term growth rate of the population.

4 The growth of algae in a particular lake is being studied to protect the ecology from a
disastrous algal bloom. The algae can live for up to four days. So the population is divided
into four age groups of one day each. The fertility rates and survival rates are being
monitored so that the population can be modelled using a Leslie matrix.
At the beginning of the study in late winter (day 0), it was observed that the algae
concentration in the lake was 3200 cells per millilitre of water, with equal numbers in each
age group. The fertility rates on the four days of life were 0.2, 0.5, 0.6 and 0.4 respectively.
The survival rate for each of the first three days of life was 0.7.
a Write down a Leslie matrix to represent this particular model.
b Find the population matrix for cells per millilitre of water on day 20, correct to three
significant figures.
c Find the population matrix for cells per millilitre of water on day 21. Hence find the
rate of change in the algae concentration per day at this stage.
d With the coming of spring on day 21, the fertility rates increased to 0.3, 0.6, 0.7 and 0.5;
the survival rate remained unchanged. Find the population matrix after a further
three weeks (i.e. on day 42).
e With the arrival of warmer weather on day 42, the fertility rates increased to 0.3, 0.7,
0.8 and 0.5; the survival rate increased to 0.85. Suppose that an algal bloom is declared
if the concentration of algae reaches 100 000 cells per millilitre of water. Using trial
and error, find the day of the study on which an algal bloom was declared.