Introduction to Mathematical Modelling

2204NSC

Owen Jepps

[email protected]

Trimester 2, 2021
Course Information

Convenors:
GC: Erik Streed ([email protected])
NA: Owen Jepps ([email protected])
Video Lectures:
Owen Jepps ([email protected])
Workshops:
NA: Owen Jepps ([email protected])
GC: TBA
Delivery
Watch online lectures, attempt problems by Friday workshops
NA Workshop: Friday 9-11 am (N79 2.03A)
GC Workshop: Friday 4pm-6pm (G30 1.19)
Course Information: Assessment

Assessment:
Final examination: 40%,
Assignments: 20% (Assg 1) + 25% (Assg 2)
Fortnightly activities 15% (available from week 3)

To pass, you must:

get 50% overall in the course; and
get 40% in the final exam

NOTE: To get a High Distinction in this subject, you will need to

go beyond the application of the Mathematical Problem Solving
methods to the problems we look at in the course, and applying
them to previously unseen problems
Your Exam Question Will Be . . .

Consider the following scenario

..
.
[previously unseen problem]
..
.

Develop a mathematical model of this scenario, to answer the

question posed
General Note on Assessment

In this course assessment will be based both on the correctness of

your answers and how clearly you define your problem. Good
modelling results are not just “correct”, they must also be clear to
an “end user”.

This means that to get good marks you must spell out
assumptions, write your mathematical steps clearly and correctly,
explain things clearly and, very importantly, write out your
conclusions clearly in an accessible way. We will spend time
developing template(s) to help with this.
Course Outline

Basic modelling concepts

Dimensional Analysis
Applications of derivatives
Related Rates
Optimisation problems
Applications of integrals
Di↵erential equations
Transport phenomena
Applications of partial di↵erential equations
Reference Sources

Any first year calculus book, e.g. Larson et al, Calculus.

Boyce, W. E. and Di Prima, R. C., Elementary Di↵erential
Equations and Boundary Value Problems, (8th Edition), Wiley
(2005). (QA 371.B773 2005).
Kreyszig, E., Advanced Engineering Mathematics, (9th
Edition), Wiley (2006). (QA 401.K7 2006).
Bender, E. A., An Introduction to Mathematical Modelling,
Wiley (1978). (QA 402.B4 1978).
O’Neil, P. V., Elements of Advanced Engineering
Mathematics, Cengage Learning (2010).
Mathematical Modelling/Problem Solving
What is a Mathematical Model?

A very formal definition:

A mathematical model is an abstract, simplified, mathematical

construct related to a part of reality and created for a particular
purpose.
(E.A. Bender, An introduction to mathematical modelling )

All models are wrong, but some are useful

(statistician George E. P. Box)
Why use Mathematical Models?

They help us make predictions

It may cost too much to run the experiment
It may be too dangerous to run the experiment
It may be too time consuming to simply ‘see what happens’
They can help determine ‘where’ we should be concentrating
our experiments
They can help to make predictions about what will happen
when we ‘change something’ within our system
We can run them multiple times to fine tune the process
What makes mathematical models useful?

We can use the language of mathematics.

1 We must formulate our ideas precisely and so are less likely to
let implicit assumptions slip by.
2 We have a concise language which encourages manipulation.
3 We have a large body of theory available to us.
4 We have high speed computers available for carrying out
calculations.
There is a trade-o↵ between points 3 and 4 because theory is good
for drawing general conclusions from simple models whereas
computers are useful for drawing specific conclusions from
complicated models.
Properties of Models

When we begin to formulate a model, we need to divide the

universe into three distinct parts:

1 Things whose e↵ects are neglected.

These are the modelling assumptions.

2 Things that a↵ect the model but whose behaviour the model
is not designed to study.
These are the inputs to the model.

3 Things the model is designed to study or predict.

These are the outputs from the model.
Goldilocks

We have to balance these three aspects to the model: the

neglected, the input and the output.

Neglect too much and you risk that

the model is misleading
the model is completely wrong

Include too much and you risk that:

the model is too hard to solve
the model takes too long to solve (even using computers)
the inputs are too ‘hard’ to collect

We need to be wary of either bad predictions, predictions that are

not believable and predictions that are contradictory.
How to build a mathematical model (Bender)

Model building involves imagination and skill. Here are some

guidelines as to how to build a model:
1 Formulate the Problem.
write down what you want to know, and what you do know
2 Outline the Model.
develop a way of relating what you do know and want to know
3 Is it useful?
are you in the Goldilocks zone?
4 Test the Model.
are your results sensible?
How to build a mathematical model (Polyà)

1 Understand the Problem.

write down what you want to know, and what you do know
2 Devise the Plan.
develop a way of relating what you do know and want to know
3 Carry out the Plan
did your plan work, or does it need revision?
4 Reflect on your Answer.
are your results sensible?
limiting cases?
is there a better way to solve?
Testing the Model

It is important not to treat the model as a perfect replica of

the actual system. It is risky to apply a model to problems
with assumptions that di↵er greatly from modelled.

Sometimes it is difficult to know which factors to ignore or

how accurately the input variables need to be determined.

We can start with a crude model and rough data estimates to

see which factors are important in the model and to determine
the accuracy required for the input variables.

Check that reasonable inputs don’t lead to obvious problems

with the model (e.g. complex numbers, wrong sign in values
or rates of change, wrong limiting cases).
Why do mathematical modelling?

Why develop and study mathematical models?

1 What is the most efficient way to divide the fuel between the
stages of a multistage rocket?
2 What would be the e↵ect of a very bad nuclear reactor
accident?
3 How large a meteor was needed to produce Meteor Crater in
Arizona?
Meteor Crater
Example problem: hair growth

Estimate how fast human hair grows on average, in km/h

1 Formulate the Problem.

write down what you want to know, and what you do know
2 Outline the Model.
how can we relate what we know to what we want
3 Is it useful?
are you in the Goldilocks zone?
4 Test the Model.
are your results sensible?
Example problems: for start of Workshop next week

Estimate the number of piano tuners in Brisbane.

Half of a snowball melts in an hour. How long does it take for
the remainder to melt?
A common technique when no models are available is to
collect data, try to fit curves and then treat the curves as if
they were a model or even an explanation. Would you have
faith in predictions made from such models? Give reasons for
your answer.
Australia’s population 2000-2050

What influences/factors might we try to incorporate in a model of

Australia’s population 2000-2050?
Example Problem: population growth

How will the population of Australia grow over the next few
generations?

1 What historical data can we find?

1 Google ‘Population of Australia’

2 INED Developed Countries Demography

3 ABS Population clock

2 What mathematical model can we construct?

Modelling our Example Problem

How will the population of Australia grow over the next few
generations?
1 Let the input variables be the net reproduction rate per
individual, r , the time, t, and N0 the size of the population at
t=0
2 The net reproduction rate is the birth rate minus the death
rate or the fractional rate of change of the population size
1 dN
r= (1)
N dt
3 Here the output variable is the size of the population at time
t, N(t).
Modelling our Example Problem

We have ignored time lag e↵ects (an assumption).

We have also ignored the possibility that system parameters

(e.g. the fraction of the population that is of reproductive
age) vary with t.

However, the model can be useful if it fits the real world.

The solution is
N(t) = N0 e rt .

Unless r = 0, the population will eventually either die out

(r < 0) or fill the universe (r > 0).
Limitations of our Example Problem Model

How good a model is this?

Study the long term behaviour of this model, i.e. the limiting
cases t ! 0, t ! 1).

These tests indicate we should reject the model for long term
growth, but it may be valid for short term growth.
Long Time Behaviour for our Example Problem

How can we fix the long-time behaviour of our model, to stop it

growing indefinitely?

This is necessary in a number of ‘population’ models, where

factors such as overcrowding or shortage of essential resources
e↵ectively cap the population growth

As the population increases we expect the death rate to

exceed the birth rate. Hence we can replace r by r (N) which
is a strictly decreasing function of N for large N, and negative
when N is very large:
1 dN
= r (N) (2)
N dt
Long Time Behaviour for our Example Problem

1 dN
= r (N)
N dt
Obtaining r (N) may be difficult to estimate accurately,
depending on the problem.

A common estimate is r (N) = N0 N, for parameter N0 .

If r (N0 ) = 0 for some N0 it can be shown that N(t) ! N0 as

t ! 1. This is a robust prediction since very few
assumptions have been made about r (N).
An Example Problem

The cycle of steps to develop the model can be repeated, since the
model still has drawbacks.
For example, N(t) can only approach N0 as t increases. In
reality, it might overshoot or fluctuate around the one value
due to random events.
We can introduce time lags. For example if the death rate, m,
is not age dependent and the birth rate b changes from zero
to a constant at p, then we get
dN
= mN(t) + bN(t p)
dt
We could make both m and b functions of N(t), N(t p) or
some weighted average of N on the interval [t p, t].
Another Example

The manager of a large commercial printing company asks your

advice on how many sales people to employ.

Qualitatively:
1 more sales peoples will mean increased sales overhead,
2 fewer sales people may mean losing potential customers.

The problem as stated is unanswerable.

Another Example

We need to understand more, we need to ask more questions:

What are the production limitations of the company?
What are the goals of the management?
Maximum profit?
Maximum “empire” with satisfactory profit?
Something else?
We need clear answers to these questions otherwise our
recommendations might be inaccurate.
A better way to proceed is to provide a description of the
consequences of sales forces of various sizes, thus leaving the final
decision to management.
Reminder: Steps for Mathematical Modelling

Model building requires imagination as well as mathematical skill.

Recall this overview for mathematical problem solving:.

1 Formulate the Problem.

write down what you want to know, and what you do know
2 Outline the Model.
how can we relate what we know to what we want
3 Is it useful?
are you in the Goldilocks zone?
4 Test the Model.
are your results sensible?