See discussions, stats, and author profiles for this publication at: https://www.researchgate.

net/publication/328902712

Mathematical Modelling: Guidebook for teachers and teams

Book · November 2018

CITATIONS READS

0 173

1 author:

Peter Lawrence Galbraith

The University of Queensland
149 PUBLICATIONS   2,604 CITATIONS   

SEE PROFILE

Some of the authors of this publication are also working on these related projects:

Australian support material for the International Mathematical Modelling Challenge View project

Application of System Dynamics to problems involving Education View project

All content following this page was uploaded by Peter Lawrence Galbraith on 24 November 2018.

The user has requested enhancement of the downloaded file.

Mathematical
Modelling
A guidebook
for teachers
and teams

Peter Galbraith and Derek Holton

Contents
Preparing students for mathematical modelling  3
What is mathematical modelling?  4
Mathematical modelling framework  6
Introductory problem: Adapting a recipe  9
Junior modelling problems for upper primary, junior or middle secondary  12
Pancakes  13
Hyperthermia  14
Köchel numbers  16
Evacuating  18
Intermediate modelling problems for junior or middle secondary  20
Waste not, want not  21
Howzat!  24
How many trees make a newspaper?  27
Bushwalking  30
Farm dams I  32
Suggestions for student-generated modelling  36
Senior modelling problems for middle or senior secondary  37
Nuclear blast  38
Population growth  40
Fifteen-forty  44
Farm dams II  47
Super Size Me  51
Temporary traffic lights  55
Teamwork and report writing  57
Modelling report illustrating use of criteria  60
Relating modelling qualities to judging criteria  66
The International Mathematical Modeling Challenge (IM 2C)  67
Previous IM 2C questions  68
2015: Movie scheduling  68
2016: Record insurance  70
2017: Jet lag  78
2018: The best hospital  83
Additional example problems from the IM2C  84

Published 2018 for ACER and IM²C This work is licensed under the Creative Commons Attribution-NonCommercial-ShareAlike 4.0
International (CC-BY-NC_SA 4.0).
www.immchallenge.org.au
Visit http://creativecommons.org/licenses/by-nc-sa/4.0/

Australian Council for Educational Research You are free to: share, copy and redistribute the material in any medium or format; adapt, remix,
transform, and build upon the material.
19 Prospect Hill Road, Camberwell, VIC 3124 AUSTRALIA
Attribution — You must give appropriate credit, provide a link to the license, and indicate if changes
Phone: +61 3 9277 5555 were made. You may do so in any reasonable manner, but not in any way that suggests the licensor
www.acer.org endorses you or your use.

ABN 19 004 398 145 NonCommercial — You may not use the material for commercial purposes.
ShareAlike — If you remix, transform, or build upon the material, you must distribute your
Design by ACER Creative Services contributions under the same license as the original.

Australian Council for Educational Research 2

Preparing students
for mathematical
modelling
Ultimately, only life educates, and the deeper that life, the real world, burrows into
the school, the more dynamic and the more robust will be the educational process.
That the school has been locked away and walled in as if by a tall fence from life itself has
been its greatest failing. Education is just as meaningless outside the real world
as is a fire without oxygen, or as is breathing in a vacuum.
Vygotsky1

The material here has been developed These considerations have influenced the audience. At times suggestions along
with two purposes in mind. Firstly, it sets way in which the material is structured. these lines are included – but many others
out to provide a basis for the development The mainstream development works are possible and it is the hope that the
of workshops geared to those for whom systematically through, and illustrates contexts provide opportunity for individual
mathematical modelling is a new by example, those features deemed initiative and development wherever these
endeavour. Secondly, it provides support essential for the development of modelling are seen as relevant.
material for teachers already working in the proficiency. This procedure gives autonomy
This document aims to provide teachers
field. The material may be of interest to: to the user in deciding if, when, and
with material that will enable them to
how much of the additional material is
■■ individuals who want support prepare students to develop skills of
appropriate for purposes of the moment.
materials and ideas for teaching mathematical modelling. It is an espoused
In particular it means that those already
modelling up to senior secondary level goal of curricula that students are enabled
working in the field can skip content,
to apply the mathematics they learn to
■■ pre-service and in-service education or indeed sections, according to their
address problems in their personal lives, as
courses featuring components in priorities and background. Ultimately it
productive citizens, and in the workplace.
mathematical modelling means that users are not forced to engage
with everything that is included, while the In a specific international initiative to
■■ mentors who intend to enter teams in
mainstream development does aim to support this ideal, an International
the IM2C
provide a complete package, beginning Mathematical Modeling2 Challenge (IM2C)
■■ teacher organisations, such as the from no assumed knowledge and building was instigated in 2014. A specific purpose
Australian Association of Mathematics to sophisticated application of modelling of this material is to provide additional
Teachers. techniques. Finally, problems can often be support to assist teachers and schools in
adapted – simplified or extended according choosing and mentoring student teams to
to circumstance, intention, or target participate in the IM2C.

1
Vygotsky, L.S. (1926/1997). Educational psychology. Boca Raton, FL: CRC Press LLC.
2
Note the American spelling. We use this for the IM²C title as this is used internationally. We have also used American spelling
when we are using problems that have been sourced from America.

Australian Council for Educational Research 3

 What is mathematical
modelling?

It is the mark of an instructed mind to rest satisfied

with the precision which the nature of the subject
permits and not to seek an exactness where only
an approximation to the truth is possible.
 Aristotle

Mathematical models are used to help problem solving, the IM2C program begins In the field of education a variety of
make decisions in a variety of new from an assumption that mathematics meanings have been attached to the
developments. Governments consult is everywhere in the world around us – term mathematical modelling. At a very
mathematical modellers to predict the a challenge is to identify its presence, simplistic level the term ‘mathematical
consequences of new initiatives, such as access it, and apply it productively. The model’ has been used in the sense of
a change of the tax system; engineers OECD, the American Common Core a formula. Another common notion is
use mathematical models to build bridges Standards Initiative, the Australian associated with mathematical modelling
and multi-storey buildings; and social and Curriculum, and other national curricula used as a vehicle to achieve other
biological research relies on a great deal all avow that students should have a curricular goals. As such the purpose is
of modelling that uses statistics. For a mathematical preparation which equips not the construction of mathematical
good article on the importance and use of them to use their acquired knowledge models to solve problems per se, but
mathematical modelling, see ‘5 Reasons in their personal lives, as citizens, and in rather to contrive the embedding of
to Teach Mathematical Modeling’, by the workplace. some pre-determined mathematics in a
mathematics professor Rachel Levy.3 This contextual setting as a mechanism for the
Such a purpose implies two intersecting
is well worth the read for both teachers learning of certain mathematical concepts,
goals. Firstly, to develop a systematic
and students. procedures and so on.
and successful approach to addressing
Children are natural modellers. individual problems located in real-world Mathematical modelling approached
For example they know that often settings, and secondly, through this as content entails the construction of
‘1 = 2’ provides an equitable solution to the means, to enable students cumulatively mathematical models of natural and social
problem of sharing lollies of different sizes to become real-world problem solvers. phenomena that are problem driven, and
(one big lolly equals The latter means that they not only can where the choice of relevant mathematics
two smaller ones). Yet formal schooling address productively problems set by is itself part of the solving process. It
so often robs initiative in the name of others, but become able to identify and is this type of modelling that the IM2C
conformity. As with all those who value address problems that matter to them. initiative seeks to promote and reinforce.
mathematical modelling as real world

3
Levy, R. (2015, May 5). 5 reasons to teach mathematical modelling. American Scientist. https://www.americanscientist.org/blog/
macroscope/5-reasons-to-teach-mathematical-modeling

Australian Council for Educational Research 4

Mathematical modelling is: Here a mathematical model is represented in three parts:
■■ a process in which real-life situations and relations in these 1. an extra-mathematical domain (D)
situations are expressed by using mathematics (Haines & 2. some mathematical domain (M)
Crouch4 ), or 3. m
 apping from the extra-mathematical domain (D) to the
■■ a cyclical process in which real-life problems are translated mathematical domain (M), resulting in outcomes that are
into mathematical language, solved within a symbolic system, translated back to the extra-mathematical domain (D).
and the solutions tested back within the real-life system The extra-mathematical world is a useful way of representing
(Verschaffel, Greer, & De Corte5). that subset of the wider ‘real world’ that is relevant to a particular
In both instances, mathematical models are seen to move problem. By real world we mean everything that is to do with
beyond the physical characteristics of a real-life situation to nature, society, or culture, including everyday life, as well as school
examine its structural features through mathematics; it entails or university subjects, and scientific or other scholarly disciplines
the construction of mathematical models of natural and social different from mathematics.
phenomena that are problem-driven, and where the choice of Objects, relations, phenomena, assumptions, questions, etc.
relevant mathematics is itself part of the solving process. in D are identified and selected as relevant for the purpose
For an introductory example of the process for construction of and situation and are then mapped – translated – into objects,
a mathematical model, see ‘Adapting a recipe’ on p. 9. relations, phenomena, assumptions, questions, etc. pertaining
This includes a spatial representation mapping the modelling to M. Within M, mathematical deliberations, manipulations and
process for a simple problem. inferences are made, the outcomes of which are then translated
back to D and interpreted as conclusions concerning that domain.
To help ensure that we are all on the same theoretical page, Within the above activity, the use of technology is relevant
we refer to the approach to modelling summarised in a study whenever it can enhance the mathematical process – which
by the International Commission on Mathematical Instruction is often. This so-called modelling cycle, (to be elaborated and
(see Figure 1).6 illustrated shortly) may be iterated several times, on the basis of
validation and evaluation of the model in relation to the domain,
until the resulting conclusions concerning D, are satisfactory in
relation to the purpose of the model construction.

Figure 1: The iterative

mathematical modelling cycle

D Mapping M

Extra-mathematical world Mathematical world

4
Haines, C.R., & Crouch, R.M. (2007). Mathematical modelling and applications: Ability and competence frameworks. In W. Blum,
P. Galbraith, M. Niss, & H-W. Henn (Eds.), Modelling and applications in mathematics education: The 14th ICMI study (pp. 417–424).
New York: Springer.
5
Verschaffel, L., Greer, B. & De Corte, E. (2002). Everyday knowledge and mathematical modeling of school word problems. In K.
Gravemeijer, R., Lehrer, B., Oers, B., van and L. Verschaffel (Eds.), Symbolizing, modeling and tool use in mathematics education
(pp. 257–276). Dordrecht, Netherlands: Springer.
6
Niss, M., Blum, W., & Galbraith, P. (2007). Introduction. In W. Blum, P. Galbraith, M. Niss, & H-W. Henn (Eds.), Modelling and applications
in mathematics education: The 14th ICMI study (pp. 3–32). New York: Springer.

Australian Council for Educational Research 5

 Mathematical
modelling framework

The IM2C operates on the assumption that mathematics is The following list indicates stages in such a process that
everywhere in the world around us; the challenge is to identify its are sequential in terms of the progress of a solution.
presence, access it, and apply it productively. The IM2C exists to
1 
Describe the real-world problem. Identify and understand the
help students:
practical aspects of the situation.
■■ develop a systematic and successful approach to addressing 2  pecify the mathematical problem. Frame the real-world scenario
S
individual problems located in real-world settings, and as an appropriate, related mathematical question(s).
■■ through this development, enable students cumulatively to 3 Formulate the mathematical model. Make simplifying
become effective solvers of real-world problems. assumptions, choose variables, estimate magnitudes
of inputs, justify decisions made.
The desire is to produce students who can not only productively
address problems set by others, but become able to identify and 4 Solve the mathematics.
address problems themselves. 5 Interpret the solution. Consider mathematical results
in terms of their real-world meanings.
In order to be useful and applicable in practice (both in the
context of the IM2C, and more broadly), the cyclic process 6 Evaluate the model. Make a judgment as to the adequacy of
of modelling referenced above needs to be elaborated in a form the solution to the original question(s). Modify the model as
that can guide (scaffold) a systematic approach to individual necessary and repeat the cycle until an adequate solution has
problems. The approach needs to be authentic – that is to been found.
be consistent with the way professional modellers approach 7 Report on success or document how further research could
problems in their field. make adjustments and try for a better solution. Communicate
clearly and fully your suggestions to address the real-world
problem.

Figure 2: Mathematical modelling framework The interpretation and evaluation stages indicate the cyclic nature
of mathematical modelling.
Describe the real-world problem If the proposed first solution is not an adequate solution to
the original question, the problem needs to be readdressed by
repeating of earlier stages (stages 3 to 6) in sequence, and this
may need to be carried out several times before an adequate
Specify the mathematical problem solution is found.
Sometimes an extension or refinement of the original problem is
suggested by the outcome of a first modelling endeavour.
Formulate the mathematical model In this instance the question is re-specified, and further cycles of
activity are conducted with the new question.
Of particular importance is the realisation that it is the problem and
Solve the mathematics its solution that drive the modelling processes and the choices that
the modellers consequently make. For the modeller the activity is
usually anything but smoothly cyclic, because checking, testing
and evaluating mean that there
Interpret the solution is frequent movement within and between intermediate stages
of the total process – potentially making the development of
some models a very challenging exercise. Various versions of the
modelling cycle exist, but they all contain the same essentials and
Evaluate/validate the model
ordering of stages.

Report solution

Australian Council for Educational Research 6

Figure 2 shows a mathematical modelling framework. The central Prescriptive modelling
vertical arrows portray a logical imperative in the progression of
a solution through a sequence of stages (e.g., a model cannot be In prescriptive modelling, the purpose is not to explain or make
solved before it is formulated), while the horizontal arrows on the predictions about real-world phenomena but to organise or
right indicate that iterative backtracking may occur between any structure a situation, for example, the location of a new power
phases of the modelling cycle, as and when a need is identified. plant. Every prescriptive modelling project also has an inbuilt cyclic
expectation; any ‘solution’ needs to be subjected to sensitivity
There are two significant families of modelling: descriptive and testing, a procedure that tests how vulnerable the solution
prescriptive. is to changes in input variables. At a minimum, this involves
recalculations of the model using modified assumptions, slightly
amended parameter values, interpretation, and an evaluative
Descriptive modelling
judgement of the outcome. If small changes in parameter values
Descriptive modelling is the most common form of modelling. create big changes in outcome, the model needs to be totally
Usually, a real-world problem is specified, a model formulated revisited. (Within this Guidebook, prescriptive modelling problems
and the resulting mathematical results interpreted and evaluated include Howzat and the 2017 IM2C problem).
in terms of the needs of the original problem. Where necessary,
The methodology shown in the example problems is one that
further (cyclic) modelling activity is conducted until a satisfactory
can be applied to all problems, in that it follows the logic of the
solution is achieved. Descriptive modelling does not necessarily
structure of Figure 2. It does not mean that solutions using the
involve validation of the initial solution. For example, modelling
approach will be identical. Far from it, as different choices for
phenomena such as global warming or population projections
assumptions, parameter values, or mathematical development
are undertaken without access to future values of variables for
will lead to their own outcomes. What is common, is that these
purposes of evaluation. Other methods of evaluation are required,
outcomes must be held accountable in terms of mathematical and
for example, completeness of model structure in terms of known
modelling criteria.
influences, accuracy of data values used, and the reasonableness
of the resulting outcome. Refinements to and projections from Following this theme, the solutions to example problems are
such models will be ongoing, as fresh data become available developed in terms of the structure embedded in Figure 2. This
to improve the quality of inputs and to act as evaluation criteria means that specific lesson plans are not produced as there are
for earlier models. For example, estimating the impact of typically several ways in which assistance and scaffolding can
variations in fertility and immigration rates, from those assumed be provided within different phases of the modelling process.
in population projection models, are important in testing the In consequence, teachers have autonomy of choice, and the
robustness of predictions, and hence to prepare for possible opportunity to support students in ways deemed most suitable
future demands on educational, health and social services. Cyclic to the needs of the moment. Illustrations are provided through
activity associated with such modelling contexts is continuous the solution processes developed and recorded within the
and can extend over years. (Within this Guidebook, examples respective problems.
of descriptive modelling problems include Kochel Numbers,
Supersize Me, and Population Projections.)

Australian Council for Educational Research 7

To help learners conceptualise and develop a greater Two approaches to the introductory problem ‘Adapting a recipe’
understanding of mathematical modelling, example problems are are provided.
available for download and distribution.
Approach I Used as a quick introduction to the modelling process
Throughout these problems, example data sets and information for those with no previous experience of modelling. For this
sources are provided. This has been necessary to illustrate purpose, it is first presented as a pre-modelling exercise, and
workings and development of models. In general, in real-world the modelling process is introduced to formalise what has been
modelling scenarios, finding appropriate data sources are part done, but often overlooked. The identified process is then applied
of the task. When using these example problems, teachers or to other problems. In this approach, the modelling process is
mentors may choose to withhold the example data sets provided, highlighted and reporting is restricted to a simple summary of
and instead direct student activity to make obtaining data part of the new recipe. With this approach it is usually possible to move
the modelling task. quickly to other problems, often within half an hour.
The junior modelling problems involve the sorts of mathematical Approach II The second (fuller) approach is used when the
processes that we often undertake automatically. In doing so, we complete modelling process is followed by a detailed examination
can overlook that we have made assumptions that are so ‘obvious’ of what is involved in report writing. The example is again used
that we did not realise that we had made them. Students should be to introduce the modelling process, and illustrate the respective
encouraged to articulate the assumptions and processes required. stages. We first show the mathematics used with amplifying
Those already familiar with basic modelling approaches might words, give a summary, and then finally an idea of what a full
wish to skip to later material. report might look like. This approach is geared particularly to
introduce and illustrate reporting as a part of the totality.
Most students find it hard to write reports, probably because they
are given very few opportunities to write any mathematics at all.
As the result of many factors, answers seem to be valued more
highly than the mathematical arguments leading to them. To help
students develop report writing skills, we suggest that they be
encouraged to add words to answers. This can start with just the
odd word or two of explanation such as ‘since … then’, ‘because of
Pythagoras’ Theorem …’ or ‘so the number of apples is 11’.
As far as mathematical modelling is concerned, report writing
skills are essential and will be of value to students to learn. In many
jobs these days, when people often work in multi-skilled teams, it’s
often necessary for a member of the team to pass information to
the others both in oral and written form.
To ease students into report writing, we suggest that students new
to modelling are first given practice in writing summaries of no
more than a page to each of the modelling activities they do from
the following section. This can gradually be extended to full reports
as students go through the three stages of examples that follow.
To write a full report takes time. We suggest that you don’t insist
on every report being written in full; students may gain much from
working up a bullet summary that could be used to form the notes
for an oral report.

Australian Council for Educational Research 8

Introductory problem
Level: Upper primary

If this problem is used as an introduction to modelling (Approach I), give it to the group without further advice. Ask them to work on it for
a few minutes and come up with an amended recipe. They will have no knowledge of a modelling process or cycle; however, they will have
made assumptions and come up with mathematical answers, which need to be interpreted in practical terms if the recipe is to be realistic.
Discussing this will show the need for systematising the process (distribute the modelling diagram at this point). The boxes can be filled
out as the group translates and organises what individuals have done. Approach I is useful when the attention is to move quickly to other
problems – the main focus is the editing process. Approach II is useful when the construction of a report is a focus, as it illustrated aspects
of report writing using a simple problem.

Adapting a recipe
Describe the real-world problem

Chocolate mousse
185g cooking chocolate
1
4 cup hot water

1 teaspoon vanilla essence

5 large eggs
1
1 2 cups cream

Melt chocolate. Beat eggs.

Whip cream. Fold all ingredients
together. Chill in fridge for 1 hour.

Serves 4–6.

Specify the mathematical problem we note essential assumptions that underpin the model
development. Identifying them explicitly helps to emphasise their
Freda Nurke is planning a new cookery book where all her recipes
centrality to every modelling enterprise. Similarly, in the interpretation
are based around larger/extended families of 6 to 9 people.
box, real-world practicalities moderate precise mathematical results,
She would like advice on how she should change the chocolate
a step that we can easily subsume without recognising it.
mousse recipe from her previous book, as shown above.
The three arrows out of box 6 reflect that (in the general case)
Pre-modelling exercise (Approach I or Approach II) evaluation may be followed by a report, but may require instead
This simple problem requires a recipe to be adapted. Most people a revisiting of the problem context, the mathematical problem
would do this on the back of an envelope or even mentally with a identified – and indeed, other stages of the modelling process.
calculator to help. We would probably overlook the fact that we
had made assumptions and used a process that was so ‘obvious’
that we did not realise that we had done it.
Situational assumptions
Facilitate students’ discussion of what assumptions they will
Formulate the mathematical model need to make before creating a model. These might include:
Group members should be encouraged to recognise and
■ availability of all ingredients
articulate the assumptions they made in their earlier approach to
solving the real-world problem. ■ all the people being served have uniform appetites
■ serving size will be the same for the adapted recipe as it was
Introduce the mathematical modelling framework (see Figure 3) for the original recipe.
and encourage students to map out their response to the problem
using this template. Emphasise that the framework is cyclical, not strictly linear. Real-
world modelling will often require that students return to earlier
Using the template means that key elements can be represented. steps in the framework to consider new variables, source new
This can be useful for structuring a report. In the formulation box, data, reassess assumptions and test solutions.

Australian Council for Educational Research 9

Figure 3: Adapting a recipe using the mathematical modellling framework

Links between the real world

Real world and the mathematical world Mathematical world
(extra-mathematical domain)
(mapping)

3. Formulate the mathematical model

Assumptions:
• ingredients available
• uniform appetites
2. Specify the mathematical problem • same size servings as before
1. Describe the real-world problem
Calculate the quantities of ingredients Model equation:
Adapt recipe to cater for 6–9 people
required • new amount of ingredients = k × old
amount of ingredients
• k = new number of people ÷ old number
of people
3
• k=―
―
2

5. Interpret the solution 4. Solve the mathematics

• 275–280 grams of chocolate • 277.5 grams of cooking chocolate
3
• almost half a cup hot water • ―
― cup of hot water
6. Evaluate/validate the model 8
1
The proof of the pudding is in the eating! • 7 or 8 eggs • 1―
2
teaspoon of vanilla essence
1 • 7.5 large eggs
• 1―2
teaspoons vanilla
1
• 2 and a bit cups of cream • 2―
4
cups of cream

7. Report the solution

Communicate, use the model to explain,
predict, decide, recommend…
Present the amended recipe written out,
with assumptions defined, and explaining
how the problem was approached

Mathematical formulation Evaluate the model

k = new amount of people ÷ old amount of people The exercise can be extended to apply the model to
new amount of ingredients = k × old amount of ingredients related problems.

k = 6 ÷ 4 or k = 9 ÷ 6 What if there are two people coming to dinner? A dozen?

3 Or, working backwards: If I have half a carton of eggs with which
k= or k = 1.5
2 to make mousse, how many people can I invite for dessert?

Solve the mathematics Will the larger volume of mousse in the amended recipe take
longer to chill? If I have less time to prepare, how might I reduce
277.5 grams of cooking chocolate
chill time? (For example, use several smaller serving dishes
3
cup of hot water instead of one large dish.)
8
1
1 teaspoon of vanilla essence
2 Report the solution
7.5 large eggs Write out a report summary or full report. Examples are shown on
2
1
cups of cream the following page.
4

Interpret the solution Possible extensions

Consider the solution in real-world terms. How would you cook with Research real-world shopping limitations. Can I buy exactly 275
half an egg? Rewrite the ingredients list, for example: grams of chocolate, 7 eggs, 2 cups of cream? How many blocks
of chocolate, cartons of eggs, containers of cream would this
275 or 280 grams of cooking chocolate
recipe need? Visit supermarket websites to research product
almost half a cup of hot water sizes, and write out a shopping list.
1
1 2
teaspoons of vanilla essence Follow up this piece of modelling by a statistical exercise. Make
7 or 8 eggs two different lots of mousse and then have a blind testing to see
which version people prefer. This could be complicated by using a
2 and a bit cups of cream different recipe for mousse from a different cookbook.

www.immchallenge.org.au
Text by Peter Galbraith for IM²C and ACER. Available under Creative Commons non-commercial licence CC BY-NC 4.0.

Australian Council for Educational Research 10

 Dear Freda,
My team has been working on adapting your chocolate mousse recipe, and we
have enclosed our suggestion in our report below. Please feel free to contact
us if there is anything in our report that needs further explanation.
We wish you well with your new venture.
Yours sincerely
Fred Fixer Pty. Ltd.

Summary For example, there is no international standard for ‘cup’. Around

Using the recipe that you sent us that serves 4 to 6 people, the world the volume of a cup varies from 200ml up to nearly 290
we have produced a mathematical model in which our main ml (see https://en.wikipedia.org/wiki/Cup_(unit)). Further, it is
assumption is that the new amount of each ingredient will be subjective as to what a ‘large egg’ actually is. If the chef has only
equal to the old amount of that ingredient, multiplied by the small eggs, will 6 be sufficient in the original recipe given? The
ratio of the new number of people to the number of people in amount of cooking chocolate to use in the adapted recipe may
the original recipe. As a consequence of this, we have made a be unrealistically precise for most kitchen cooks; however, the
suggestion for the new recipe ingredients to create a dessert to resulting mousse may taste just as nice or even sweeter if more,
serve 6 to 9 people. This suggestion can be found at the end of rather than less, chocolate is used. And it may be that 1 teaspoon
our full report. of vanilla essence may do just as well for 4 to 6 people as 6 to 9.

It is not clear to us, though, how long the new mixture needs to be Importantly, it is not clear that the simple ratio method for
chilled before it will set satisfactorily. It would be worth your while calculating ingredients is suitable for cooking, and you may want
experimenting in your kitchen to determine an optimal chilling to look further into this. It might be better to make enough for
time for your recipe. 12 to 18 people and halve this when chilling is complete. The
remainder could then be put into the freezer for another day.
In the full report we detail our approach to the problem and
discuss other matters that might be considered. Finally, we recommend that you use the amounts proposed below.

Chocolate mousse recipe to serve 6–9 people:

Full report
275–280 grams of cooking chocolate
Situational assumptions: all ingredients are available; all the
people being served have uniform appetites; serving size will be almost half a cup of hot water
the same for the adapted recipe as it was for the original recipe.
1
1 teaspoons of vanilla essence
2
Mathematical assumption: new amount of each ingredient (now
denoted k) = old amount of ingredient × new number of people ÷ 7 or 8 eggs
recipe number of people.
2 and a bit cups of cream
Mathematical model: k = new amount of people ÷ old amount
of people

k = 6 ÷ 4 or k = 9 ÷ 6 therefore k = 3/2 or k = 1.5

There are elements in the recipe that may introduce inaccuracies

in the adaptation for 6 to 9 people.

www.immchallenge.org.au
Text by Peter Galbraith for IM²C and ACER. Available under Creative Commons non-commercial licence CC BY-NC 4.0.

Australian Council for Educational Research 11

Junior modelling problems
for upper primary, junior
or middle secondary

Australian Council for Educational Research 12

Example problem
Level: Upper primary
Junior modelling

Pancakes
Describe the real-world problem

Dear Fred,
I really appreciated the work you did for me in converting a chocolate mousse recipe for use on my new
cooking book Recipes for the Larger Family. I would be grateful if you could look at my recipe below for 4 to 6
pancakes and adjust it for 6 to 9 pancakes.
Naturally my company will recompense you for the time you spend on this work.
I look forward to an early reply.

Yours sincerely
M. Freda Nurke
International Chef and Author

Recipe for 4 to 6 pancakes

1 egg
1 cup self-raising flour
200 ml milk
Mix all ingredients thoroughly so that there are no lumps.
Heat pan until a drop of water dances on the pan.
Add oil to pan.
Add pancake mixture to pan.
Cook pancake on one side till brown, then turn over and
cook the other side.
Remove pancake from pan and put on a plate covered by
paper towel.
Put into warm oven and add subsequent pancakes to
the plate.

Model the mathematical problem

The modelling activity here is generated by a letter from a This example problem is presented without further comment, but
satisfied customer, Freda Nurke. can be solved in a similar way to the ‘Adapting a recipe’ example
problem, discussed on the previous pages.
The modelling goal is to produce a recipe for 6 to 9 pancakes
from the original recipe for 4 to 6 pancakes. It might be good if
the new menu could be tested against the old one for taste.

www.immchallenge.org.au
Text by Derek Holton and Peter Galbraith for IM²C and ACER. Available under Creative Commons non-commercial
licence CC BY-NC 4.0.

Australian Council for Educational Research 13

Example problem
Level: Upper primary or junior secondary

Hyperthermia

Describe the real-world problem

Danger for children left in hot cars Lane closure betwe

Monday, December 5, 2015. A King County man will face 1 February, 2015. Traffic restrict
police questioning over suspicion of negligence after to allow for the construction of
allegedly leaving his two-year-old grandson alone in a car one kilometre east of Highbur
in a shopping centre parking lot on Saturday. Police were traffic lane will be temporarily
forced to smash a window of the locked car to rescue the reasons, traffic will be restricted
toddler, who was on the verge of dehydration. Highway, through the work site
will be in place. Temporary traffi
delays can be expected.

Hyperthermia, or heat-related illness, occurs when a person’s

Child deaths from vehicular Circumstances leading to
body absorbs more heat from the environment than the body
heatstroke, United States, by year child vehicular death
can dissipate through cooling. The human body’s cooling
mechanisms include perspiration, which is a loss of fluid. Losing 2005 47 Child forgotten by adult
a lot of fluid results in dehydration. More than 400 children have 2006 29 54% of cases
died in cars from hyperthermia in the United States since 2005.
The children have ranged in age from 5 days to 14 years; more 2007 36
than half have been less than 24 months old. 2008 43
2009 33 Child playing unattended
in vehicle
2010 49 29% of cases
2011 33
2012 34
2013 44 Child intentionally left in
Source: Null, J. (2016). Heatstroke deaths of children in vehicles. vehicle by adult 17% of
Department of Meteorology and Climate Science, San Jose 2014 31 cases
State University. http://noheatstroke.org 2015 24

Australian Council for Educational Research 14

 Specify the mathematical problem Interpret the solution
Investigate why small children and animals are so much at risk in Applying this logic to the problem of children in cars suggests
locked cars in hot weather. that children (smaller people) will have a higher surface area/
volume ratio than adults (larger people) do. Therefore, they will
lose fluid more quickly. Therefore, they are at greater risk of
Formulate the mathematical model dehydration.
The rate of fluid loss from a body depends on (is proportional to)
its surface area, SA.
Evaluate the model
The amount of fluid in a body depends on (is proportional to) its The diagram below suggests how to extend the simple cube
volume, Vol. approach to construct more elaborate representations, using
cuboids, cylinders or spheres, to construct physical models of
So a critical factor is the surface area/volume ratio of the body.
children, adults, animals and so on. Associated mathematical
We begin by considering blocks of different dimensions.
development will call on knowledge of mensuration. There is
opportunity for students to construct representations of animals
of choice.

l l

SA small cube = 6 Vol small cube = 1 SA/Vol small cube = 6

SA large cube = 24 Vol large cube = 8 SA/Vol large cube = 3

Solve the mathematics

6

5
surface area: volume

3 Report the solution

2 Students should write out a report following the modelling
framework structure.
1

0
0 2 4 6 8 10

length of side of cube

Create a graph plotting the ratio for blocks with lengths from 1
to 10. This shows that smaller cubes have higher surface area/
volume ratios than larger cubes. That is, smaller cubes have a
greater surface area through which to lose fluid relative to the
volume of fluid they have to lose. Smaller cubes will lose fluid at a
greater rate than larger cubes.

www.immchallenge.org.au

Text by Peter Galbraith for IM²C and ACER. Available under Creative Commons non-commercial licence CC BY-NC 4.0.

Australian Council for Educational Research 15

Example problem
Level: Junior or middle secondary

Köchel numbers

Photo credit: Portrait of Mozart by Barbara Krafft

Describe the real-world problem

‘It is sobering to realise that when Mozart was my age he had already been
dead for three years.’
— Tom Lehrer, mathematician/comedian

Wolfgang Amadeus Mozart, one of the most influential composers of the Classical era, was born January 27, 1756, in Salzburg, and died
December 5, 1793, in Vienna. He composed more than 600 works over his career.
Ludwig von Köchel, Viennese botanist, mineralogist and educator, published an inclusive, chronological catalogue of Mozart’s work in 1862.
Köchel (K) numbers are assigned sequentially according to the date of composition. For example, Mozart’s opera The Magic Flute is given
the Köchel number 620, and is (approximately) the 620th piece of music Mozart composed.
Compositions completed at the same time are listed K69, K69a, and so on.

Specify the mathematical problem

A new composition by Mozart completed in April 1784 has come
to light. What Köchel number should it be given?

Australian Council for Educational Research 16

 Formulate the mathematical model Solve the mathematics
The date of the new composition (April 1784)
Data corresponds to x = 28.3
Given the information in the table, find a way of linking Köchel
numbers with the dates of completion. Substituting in y = 24.4x − 252.8 gives y = 436.5.

Date
K Composition Interpret the solution
completed
An appropriate Köchel number would be K = 437 or K437a or K437b.
65 Jan 1769 Dance music

176 Dec 1773 Dance music Evaluate the model

Referring to the table, a composition with K = 437 (April 1784)
271 Jan 1777 Piano concerto
should lie towards the upper end of the interval between K = 385
317 Mar 1779 Mass (July 1782) and K = 470 (April 1785). It does, so it is reasonable to
infer that the model is suited to its purpose.
385 July 1782 Symphony: Haffner
On the other hand, the model was derived using only nine points.
It would be reasonable to repeat the calculations using research
470 Apr 1785 Andante for strings
to locate say 20 points to use as a starting basis, and to note and
525 Aug 1787 Serenade: Eine Kleine Nachtmusik comment on any differences.

588 Jan 1790 Opera: Cosi fan tutte Refinement using technology
Students familiar with graphical calculator technology will likely
620 Sept 1791 Opera: Die Zauberflöte
identify the opportunity to use the regression facility to obtain the
line of best fit by technical means. This is of course a legitimate
approach, but one which should not be forced on those who
From the table above, we translate the completion dates to
are not familiar with the appropriate technology. (Diversion
Mozart’s age in years, knowing his birth date. (K65 corresponds
down unfamiliar technological paths has been shown to impede
to 13.0 and K620 corresponds to 35.7 etc.) These can be added
progress within modelling problems.)
as another column beside ‘date completed’ in the table. Plotting
the points on a grid with Mozart’s age along the horizontal axis In the present case, the application of the linear regression facility
(x), and Köchel numbers on the vertical axis (y), they are seen to is shown in the straight line drawn on the graph above. Its equation
lie approximately along a straight line. is y = 24.8x – 259.3.

Placing a ruler (by eye) through the points identifies a line which This leads to a value of K = 443 which is within 1.5% of the value
in our judgment best fits the set of points. The line can then be calculated using eye and hand. The final choice would involve
drawn – a so-called ‘line of best fit’. The equation to this line will searching out and examining the Köchel numbers already
give a link between age and Köchel number. assigned to compositions around the date in question.

There are several ways of finding this equation. For example, if

our ruled line passes through the endpoints (13.0, 65) and (35.7,
Report the solution
620) of the graph then its equation will be Y = 24.4x − 252.8
(working to one decimal place). The report should contain the foregoing, perhaps with a summary
statement that generalises the usefulness of the model beyond
MOZART the single case which is featured here. We can note also that a
formula such as this can be used generally to check for errors in
800
Köchel number

assigning Köchel numbers, by looking for wayward results when

600 the formula is applied to large numbers of data.
400 line fit
There should also be comment on reservations concerning
200 precision, on account of approximations used in the approach.
0
10 20 30 40
Age (years)

www.immchallenge.org.au
Text by Peter Galbraith and Alistair Carr for IM²C and ACER. Available under Creative Commons non-commercial
licence CC BY-NC 4.0.

Australian Council for Educational Research 17

Example problem
Level: Junior secondary
Intermediate modelling

Evacuating

Photo Credit: Jens Sohnrey

Describe the real-world problem

Australia’s tallest apartment block fails fire Lane closure b

safety compliance 1 February, 2015. Traffic
15 November, 2012. A recent evaluation has identified a potential fire to allow for the constru
safety hazard at Australia’s tallest residential building. The Q1 complex, an one kilometre east of
80-storey, $250-million skyscraper on Queensland’s Gold Coast, completed traffic lane will be tem
a decade ago, houses more than 500 apartments and more than 1000 reasons, traffic will be r
residents. The privately-conducted safety evaluation, reported by the ABC, Highway, through the w
indicated that one of Q1’s two emergency escape routes does not meet will be in place. Tempor
fire safety compliance standards set by the Australian building code. In the dasons, traffic will be re
event of a major fire, a design flaw could cause the northern stairwell to fill Highway, through the w
with smoke, putting hundreds of people at risk of asphyxiation. will encounter

Specify the mathematical problem ■ 527 residential apartments, including a mixture of 1-, 2- and
3-bedroom apartments
A bomb threat is received at 2 am and the building must be
evacuated. How long would this take if one stairwell cannot be used? ■ 1331 stairs per stairwell
■ 11 lifts
■ 2 stairwells
Formulate the mathematical model
Data Situational assumptions
This is information that defines important aspects of the ■ While building has two exit stairwells, only one is safe.
structure, and can be provided as background or left for students ■ Lifts are closed to avoid failure through overcrowding.
to source (however websites can give conflicting information).
■ All residents are mobile and hear the evacuation call.
This problem is based around estimates rather than exact
calculations and the information below is sufficiently accurate to ■ Building is highly populated and floors have equal numbers
supports this. The Q1 building contains: of residents.
■ 1-, 2- and 3-bedroom apartments are equally distributed.
■ More than 1000 residents (plus staff)
■ Number of stairs between levels is the same throughout.
■ 76 residential floors

Australian Council for Educational Research 18

Mathematical assumptions Total time (T) for evacuation = time for first occupant on floor 1 to
■ Number of floors = 76 leave building + delay until last occupant can move + time for last
occupant to leave building (this needs thinking about).
■ Number of apartments per floor = 527/76 ≈ 7
Show that:
■ Average number of evacuees per apartment (e): e = 2.5*
(*Include 0.5 to represent staff presence throughout building) T = (t + s/v) + (fN − 1)d + fs/v where N = 7e

■ Number of evacuees per floor using stairwell (N): N = 7e

Solve the mathematics
■ Number of steps per floor/one stairwell (s): s = 665/76 = 8.75 ≈ 9
e = 2.5; f = 76; s = 9; v = 0.5; d = 1; t = 10 gives T ≈ 45.4 min
■ Average speed on steps (moving carefully) (v): v = 0.5 steps/sec
For a lightly populated building assume e = 1.5; for faster or
■ Time delay between successive evacuees when moving slower steps vary ‘v’; for different separation distances between
(d): d = 1 sec moving evacuees vary ‘d’ and so on.

■ Time for first occupant on floor 1 to reach stairs (t): t = 10 sec

■ Values of v and d are chosen to reflect a balance between

Interpret the solution
speed and safety. Consider the time taken to evacuate residents from the building
under the various assumptions. Describe the influence of the
■ Evacuation rate is governed by access to stairs. What happens numerical changes in assumed values? What are the risks? What
in corridors is effectively irrelevant because of wait times. kind of robust estimate seems reasonable? Is it likely that the
residents could exit safely in the event of a fire?
The numbers assigned to the quantities like (e, v, d) need to be
estimated, and others (such as ‘s’) inferred. They need to be Estimate how much more quickly the building could be evacuated
‘reasonable’ in terms of the real context, and can be varied for if both stairwells were safe and operational. Is it likely that
different calculations. residents could exit safely if both stairwells were functional?

For example, the average number of evacuees per apartment has

been chosen with a fairly ‘full’ building in mind. Average speed on
Evaluate the model
steps, and time delay between people moving, need to take into
account that people crashing into one another must be avoided, Consider how outcomes are affected by changes in numbers
while otherwise moving as quickly as possible. Values here can of people, speed of descent and delay times. Are these what
be agreed by discussion, and/or acting out. (Students have used you would expect? What does the model suggest are the most
thought experiments effectively here.) Similarly, selection of other important influences on evacuation time? Is this consistent with
numerical values needs to be supported by discussion. intuition and experience?

Last
out Report the solution
A simple modelling report will address the time question as it is
written in the problem statement. It should contain all the above
X XXXXX...X Top floor components of the modelling problem and its solution, including
implications of further calculations using a range of different
parameter choices. A more comprehensive report, going further
than the question demands, would synthesise these data into a
X cohesive narrative, considering the implications for safety of the
X X X X X .............. XX Floor 2 residents of the apartment building, and considering if possible,
X
X fire safety compliance standards set by the Australian building
X code. The precise structure of a report will depend on how the
X X X X X .............. X X Floor 1
X problem was expressed.
X
X Ground floor (Exit level)
X

First
out

www.immchallenge.org.au

Text by Peter Galbraith for IM²C and ACER. Available under Creative Commons non-commercial licence CC BY-NC 4.0.

Australian Council for Educational Research 19

Intermediate modelling
problems for junior
or middle secondary

The problems in this section are intermediate in demand. They

are suitable for junior secondary to middle secondary students.
Application of the modelling process in these problems involves
more initiative and persistence. Mathematics is often initially
absent, and needs to be introduced by the modeller.
Classroom conditions for implementation will vary so it is
important that all students have the opportunity to become
modellers on an individual basis. For collaborative activity,
students should work in groups of about four and learn to divide
the work between them. They should start as a whole group and
decide what needs to be done; what assumptions need to be
made; what data are needed; which students will work on which
sections; and how they will construct a summary report. To make
such decisions they will need to be individually competent.
At this level more effort needs to be made on report writing. We
suggest that students should be encouraged and helped to make
summary reports for all of the models they tackle. Reports can be
based around the content recorded as they progress through the
different stages of the modelling process. It would also be useful
for different groups to make verbal presentations of their report
to the class. From time to time, a full report should be written and
preferably assessed.
To help this communication process, we recommend that more
writing be included in students’ work on curriculum-oriented
mathematics. Students should day-to-day be encouraged to add
comments to their work to show how the next line follows or
why they used a given approach. They should also present their
answers in word form. For instance, instead of finishing with
‘x = 5’, they should be encouraged to write ‘The largest possible
root of the original equation is 5’.
In addition, in class, students should be helped to communicate
their answers and methods to the whole class. So instead of
giving a verbal answer of ‘5’, they could say: ‘I used the quadratic
formula to solve the quadratic. The two solutions were -1 and
5. So the answer to the question “what is the bigger root of the
quadratic?” is 5.’
This more verbal approach to mathematics will stand students
in good stead at work and in social situations. If they are able
to explain and discuss the way they have tackled a problem,
mathematical or otherwise, they are more likely to be able to
convince people of their solution and be more able to appreciate
other people’s arguments. The end result should be a better
one for everyone.

Australian Council for Educational Research 20

Example problem
Level: Junior or middle secondary

Waste not, want not

Describe the real-world problem

On 16 February 2016, the Australian population

reached 24 million people. Waste generation Lane closure between Newto
rates are a function of population growth, the 1 February, 2015. Traffic restrictions will apply on the
level of urbanisation and per capita income and to allow for the construction of a new eastbound o
Australians now produce about 50 million tonnes one kilometre east of Highbury. To facilitate this w
of waste each year, averaging over 2 tonnes per traffic lane will be temporarily closed for approxim
person. There are more of us and we generate reasons, traffic will be restricted to one lane, on a 60
more waste per person, each year. Highway, through the work site where 55- and 25-kil
In the period 1996–2015 our population rose by will be in place. Temporary traffic signals will be in u
28% but waste generation increased by 170%. dasons, traffic
Source:will beRitchie
Mike restricted
– to one lane, on a 60
Waste is growing at a compound growth rate of Highway, through the work
Director, MRA site Group
Consulting where 55- and 25-kil
7.8% /Year. will encounter
https://blog.mraconsulting.com.
au/2016/04/20/state-of-waste-2016-
current-and-future-australian-trends/

Australia is one of the highest waste producers in the world, Specify the mathematical problem
recently ranked in the top five waste producing nations, on a per
Find estimates in kilograms (or tonnes) for the total amount of
person basis. In a year, we produce the equivalent of three million
waste produced in Australia over the next century.
garbage trucks full of compacted rubbish. Every year the average
Australian family produces enough rubbish to fill a three-bedroom
house from floor to ceiling.
Formulate the mathematical model
Some waste we are directly responsible for (e.g. household
waste); other waste is generated on our behalf (e.g. from Assumptions
manufacturing processes, the building industry, road construction ■ Total waste produced in a year = population × waste ÷ person.
etc.). It is convenient to describe total waste generation in terms
of amount per person. From the figures quoted, the total waste ■ Rates of increase that have applied over the past 20 years will
produced annually is currently 2.1 tonnes per head of population continue into the future.
(approximately). Of this, household waste accounts for about one-
■ Recycling or use of landfill do not enter at this point, as they
seventh of the total.
are about managing waste that has been already produced.
On the positive side, recycling is growing at a faster rate and for Here we are concerned with the production of total waste, to
the first time since 2005 we have seen a decline in tonnages of which re-cycling or landfill might later be applied.
waste sent to landfill (in the most progressive states). We now
recycle approximately 58% of all the waste we generate and Parameter values
landfill the rest. ■ Total waste is growing at a compound rate of 7.8% per year.
Several modelling problems are suggested by this background ■ Initial (2016) value for population is 24 000 000.
information, in terms of implications for the future. We consider
one such problem. ■ Initial (2016) amount of waste per person is 2.1 tonnes
per year.

■ Estimate of amount of total waste produced in 2016

= 24 000 000 × 2.1
= 50 400 000 tonnes (assumption 1)

Australian Council for Educational Research 21

Model development Interpret the solution
From the descriptions provided on the MRA website, rate of Column A contains the initial values of population, and of total
growth of waste follows a compound interest pattern. waste produced per person in 2016.

Column B contains the assumed annual rate of increase of waste.

Principle of compound growth
Amount next year (A1) = amount this year (A0) + added amount Column C contains the consecutive years over which the
(interest). calculation runs.

A1 = A0 + rA0 = A0 (1+r) where r = annual compounding Column D contains the total waste (in tonnes) produced on a
(interest) rate. yearly basis since 2016.

Then A2 = A1 + rA1 = A1 (1+r) gives the amount for the second Column E contains the accumulated waste produced since 2016.
year, and so on year by year.
The last entry in column E estimates the number of tonnes of
Applying this principle to total waste generation (W): waste produced over the century – it represents an answer to the
question originally posed.
W1 = W0 (1+w)

W2 = W1 (1+w) and so on
Evaluate the model
Where W0 =50 400 000; w =0.078. Writing out the estimate for total waste produced over the century
looks like this:
Applying assumption (2) we now extend predictions into
the future. From column E: Total waste produced over 100 years =
1 180 300 000 000 (over 1 trillion tonnes!)

Solve the mathematics This looks ‘rather high’ in practical terms, but we need something
closer to home to relate to.
This is readily achieved using a spreadsheet.
Column D tells us the amount of total waste generated on a yearly
A B C D E basis will be 85 440 000 000 tonnes per year in 100 years’ time.

Waste The background information says that over a 20-year period the
waste/year Total waste
1 Initial values growth year population grew by 28%.
(tonne/yr) (tonne)
rate
Population Assuming that population continues to grow at about the same
2 1 = A3×A6 =D2 rate it will not be too much in error to assume that in 100 years it
(persons)
will have grown to around 70 million.7
3 24000000 2 = D2×(1+$B$6) = (E2+D3)
Dividing 85 440 000 000 by 70 million gives 1220 tonnes produced
4 3 Copy Copy per person (approx) in the year 2115.

Waste/ Noting that household waste formed about 1/7 of total waste, and
5 Waste/yr 4 assuming the proportion remains stable, this gives a quantity of
person/yr
about 174 tonnes of household rubbish per person for the year 2115.
6 2.1 0.078 5
On average this amounts to the production of about 0.48 tonnes
″ ″ (480 kg) per day for each person. That sounds like a lot of garbage
(in more ways than one!)
″ ″
After checking the mathematics, we are sceptical about the model
predictions. What should be revisited?
101 100 8.5445E+10 1.1803E+12

7
It isn’t too much trouble to infer a compounding growth rate for population of 1.112 % p.a. from the given data, which when
projected over 100 years from 2016 gives a population of about 73 million

Australian Council for Educational Research 22

Model refinements Additional problem contexts to explore
Assuming that the initial values are reliable, and population growth
rates are fairly stable, we can identify growth rate for waste A. Recycling
production ‘w’ as speculative. On the one hand we have assumed How does Australia’s recycling compare to the rest of the world?
that the scale of growth observed over a 20-year period will
continue for the next century. Is this reasonable? But beyond this, Australians compete against the rest of the world in many ways
a closer inspection of the indicated growth rate (7.8% per year) and recycling is no exception. In some areas, such as newspaper
given on the website is not consistent with the associated figure recycling we have been world leaders for years, but we need to
of a 170% increase in waste over the 20-year period – the implied catch up in others. According to the MRA consulting website we
rate is much less (about 3.95%).8 now recycle 58% of all waste.

We can run the spreadsheet multiple times, using different values What impact would this have on implications from the previous
for ‘w’. model? Make some assumptions about future improvements in
recycling, and follow their implications by including them in a
Try halving the rate of growth of waste production. How does that new model.
look? What would it mean in real-life terms? Is it manageable?
Similarly, for other choices.
B. Implications for landfill
Alternatively, set what are believed to be manageable levels of www.pc.gov.au/__data/assets/pdf_file/0016/21904/sub028.pdf
waste production and adjust ‘w’ in the spreadsheet to achieve
these outcomes. This gives an idea of what efforts need to be Australia has a strong dependence on landfill as a form of waste
made to contain the waste problem. management, since the majority of waste that is not recycled or
re-used in Australia is compacted and disposed of in landfills.
The website contains much information about land filling including
the following:

‘… the depth of the landfill is assumed as 20m, the waste density

is assumed as 750kg/m3 and the capacity is assumed to be 3
million tonnes.’

Build a model that incorporates landfill as a means of disposing

of waste that cannot be re-used or recycled, and consider future
implications of the outcomes it produces.

8
This example shows the importance of evaluating models in terms of their real-world implications. In this case, evaluation led to
the identification of an inconsistency in the data. In general, evaluating model outcomes means that the accuracy of claims made
in the media can often be tested, which is another reason why the ability to do mathematical modelling is important for informed
citizenry.

www.immchallenge.org.au
Text by Peter Galbraith for IM²C and ACER. Available under Creative Commons non-commercial licence CC BY-NC 4.0.

Australian Council for Educational Research 23

Example problem
Level: Junior or middle secondary
Intermediate modelling

Howzat!

Courtesy of Teaching Mathematics and its Applications

Describe the real-world problem

The picture above was taken during a cricket Test Match between Australia
and England at the Melbourne Cricket Ground. The photo shows the English
batsman Colin Cowdrey trying to make his ground while taking a sharp single,
against a run out attempt with Australian Wally Grout over the stumps. At the
time there was no third umpire so the players had to rely on the square leg
umpire for the decision. An umpire’s judgment is still called on in all forms of
cricket, except at the highest level where technology is available.
The umpire gave Cowdrey out. But would that have been the decision of a third
umpire with digital technology to fall back on?

Specify the mathematical problem before descent? The most favourable case for the batsman is
the former situation. (Students might discuss why.) This is the
Use information from the photo to argue whether the ‘out’
situation we will analyse first.
decision was correct.

Calculating time
Formulate the mathematical model
The bail can be seen in the photograph as the dark mark roughly
a quarter of the way down the left stump. An internet search gives
Assumptions
the standard height of a cricket stump as 28 inches (71.1 cm).
Firstly, we need assumptions that will enable us to introduce
The vertical distance, ‘h’, that the bail has fallen can be estimated
mathematics into the problem, based on variables we identify
as important. by direct measurement of the photograph.

We assume that the batsman achieves a speed (V) at the end of The time (t) taken to reach this point after the stumps were
the run that is consistent with a batsman running a sharp single broken can be estimated from the formula h = (1/2)gt2.
wearing batting gear.
(This is an application of the equation of motion
We also assume (supported by the photo) that the bat has been
h = ut + (1/2) at2
grounded properly.
to the vertical movement of the bail, where a, the acceleration
Most important is an assumption about the path of the bails. Did
is g = 9.8 m/s2, and u = 0 from the assumption of no upward
they fly horizontally on impact, or was there upward movement
movement of the bail.)

Australian Council for Educational Research 24

Calculating velocity For the purpose of this resource, an example data set is given
Students will need to decide on a realistic estimate of the below. Students conducting this exercise will likely find different
batsman’s speed. How fast can a batsman with pads and bat data, and should experiment with the effect a range of values has
run over a short distance? An internet search will give students on their model.
a range of estimates to discuss. For example, finely tuned
athletes run 100 metres at about 10 metres per second and Time since wicket was broken
1500 metres at about 7 metres per second. Students will need
h = (1/2)gt2
to make assumptions about the effect of cricket pads on speed, 1
the differences in the type of running between cricket and athletic therefore t = (2h/g) 5
1
sprinting and so on (a cricket pitch is only around 20 metres long, t = (0.3556/9.8) 2
sprinters take some distance to accelerate from a standing start
to full speed, etc.). t = 0.19 seconds

Calculating distance Distance travelled by batsman in this time

To estimate the distance that the bat is inside the batting crease s = Vt
as shown in the photo, we note that the distance on the photo s = 7 × 0.19 = 1.33 m
between the two creases, the white line of the popping (batting)
crease (in front of Grout) and the white line of the bowling crease Position of batsman when wicket fell
(that passes through the base of the stumps) must be scaled r−s
to represent the actual value of 4 feet (as per the standard 1.10 − 1.33 = −0.23 m
dimensions of a cricket pitch). A parallelogram can be drawn, using
the shadow of the bat as a guide, that enables the distance of the
bat inside the crease to be calculated as a fraction of this distance.
Interpret the solution
The batsman was out by about 23 centimetres.
We can then estimate the distance (d) travelled by the batsman in
this time at speed (V) using the formula d = Vt. Consider how outcomes are affected by changes in the values in
the model.
Now the distance ‘d’ travelled by Cowdrey since the wicket was
broken can be compared with the distance ‘r’ his bat is inside the Evaluate the model
crease. This decides the outcome. The umpire raised his finger –
This can be extended. What if the bails went up in the air after
do you agree?
the wickets were broken? Would that be better for the batsman
or worse? If the bails went upward on impact the time for the bail
Solve the mathematics to reach its position will be even longer than in this calculation.
The batsman will be even more ‘out’, which also means that there
Example data is no need to explore this assumption further for purposes of
It is best whenever possible for students to identify needed data deciding the question.
and source data themselves as the modelling proceeds. They
may find relevant sources alternative to those below. What if the batsman was running faster than the speed that was
assumed? Our estimation of running speed introduced numerical
We have drawn on references: data not provided in the photo. It can be reasonably argued (again
Laws of Cricket: https://www.lords.org/mcc/laws-of-cricket/ suited to discussion) that the value chosen is likely to be an
World Athletic Records: http://inglog.com/tools/world-records underestimate. Again, this confirms the decision as the faster the
batsman is running the more trouble he is in.

h distance the bail has moved 7 inches = 0.1778 m estimated from photo
t time for which the bail has been moving h = (1/2)gt 2
scientific formula
u initial velocity of the bail 0 ball was not moving
a acceleration of bail 9.8 m/s 2
gravity
conservative value based on speed of
v velocity of batsman 7 m/s professional athlete and accounting
for effect of cricket padding
s distance batsman has travelled since the wicket fell s = Vt scientific formula
90% of crease width =
0.9 × 4ft
r distance of the bat inside crease at time the photo was taken estimated from photo
0.9 × 1.2192 m
≈ 1.10 m
position of batsman when wicket fell, in relation to crease r−s solve for answer to problem

www.immchallenge.org.au

Text by Peter Galbraith and Derek Holton for IM²C and ACER. Available under Creative Commons non-commercial licence CC BY-NC 4.0.

Australian Council for Educational Research 25

Further testing of the robustness of the original solution can Finally, as a matter of empirical interest, this problem has been
be conducted by allowing the values of quantities subject to used with groups of students at both secondary and tertiary
measurements made from the photo, or estimated, to vary (say level. The overwhelming response in every case is that the
by 10%). Does this make any difference to the outcome? umpire got it right.

This is known as sensitivity testing and is important whenever

the purpose of modelling is to determine a ‘best’ outcome. Report the solution
Here the outcome sought is a correct decision. ‘Howzat’ is The modelling report could contain all the above components
an example of a prescriptive modelling problem - see the of the modelling problem and its solution. The report should
explanation of prescriptive modelling in the Introductory problem synthesise this data into a cohesive narrative, outlining decisions,
Adapt a Recipe. assumptions and conclusions.

www.immchallenge.org.au

Text by Peter Galbraith and Derek Holton for IM²C and ACER. Available under Creative Commons non-commercial licence CC BY-NC 4.0.

Australian Council for Educational Research 26

Example problem
Level: Middle secondary
Intermediate modelling

How many trees make a newspaper?

Describe the real-world problem

How many trees go into your morning read?

According to the environmental paper and cardboard annually.
conservation organisation Clean The environmental impact of
Up Australia, Australians’ per paper consumption is significant.
capita consumption of paper Although most used paper could
products is high, and increasing be recycled, around half of it ends
every year. Australians consume up in landfill.
more than four million tonnes of

Specify the mathematical problem

Estimate how many trees are needed to provide a year’s Identifying data required
circulation of a major newspaper. Data required include:

If the task is planned ahead of time, students can be asked to ■ how much paper is in a newspaper, possibly considering
track daily papers for a week (or several weeks), and to bring • dimensions of a page
newspapers to class when the problem is set.
• margin data, if it is decided to find how much paper is
unused for actual print

Formulate the mathematical model • daily page count or average pages over a week

This is an ideal team task. Form teams of four students who can • weight of a page of newsprint
then brainstorm, within their team, for a few minutes, a list of the ■ how many newspapers are produced in a year, i.e.,
data they will need to create a model, and the assumptions they circulation (number of copies printed)
will need to make. Outcomes can then be discussed and debated
by sharing with the whole group. (Typically, not all necessary data ■ weight of newsprint paper obtained from an average tree
and assumptions will be recognised at the start.) used for pulp; i.e., the number of trees required to produce a
given weight of paper.
Some examples are given below.

Australian Council for Educational Research 27

An ice-breaking very simple team activity is for each member to Situational assumptions
measure the dimensions of a page, both total and associated ■ The number of pages is the same each week day, and each
with print area. Averaging will give a measure to use in the model. weekend, throughout the year.
■ The pages have standard dimensions.
Collecting data ■ The paper quality (thickness, weight) remains constant.
How this proceeds depends on how much time is allotted to ■ The model excludes inserts and magazines, and assumes a
the task – for example spread over a week, or allotted a limited 52 week year and additional copies not sold.
amount of class time? If the former, after identification of needed
information, student teams can be assigned tasks of collecting
their required data, such as: Solve the mathematics
■ conducting internet research to find circulation figures of
Example data
publications
For the purpose of this resource, an example data set is given
■ conducting internet research to identify relevant ways to below for a fictitious daily regional newspaper.
convert quantities of paper into weight (kg). (If an electronic
balance is available, weighing some newspaper is easiest) Given this information and the assumptions, weekly calculations
can be completed and scaled up to give yearly estimates.
■ conducting internet research to estimate the amount of
paper provided by a tree.
Calculate paper usage
If time is limited, then websites can be provided to students. In Per week
that situation, the web links can be withheld until the students Single pages = (78 × 153 763 × 5) + (96 × 199 153)
have identified the need for the associated information. + (96 × 359 088) = 113 558 706
Single newsprint sheets (2 pages use 1 sheet of paper)
Many of the data requirements and ways of finding them = 0.5 × pages = 56 779 353
will generate discussion. Weight of paper, for example, is a
complicated measure. An internet search is likely to return some
information about paper weight measured in standard base Dimensions
weight, using pounds (lb). A search will also find, however, that in
Width (cm) 29
many countries using the metric system, the relevant measure is
grammage (grams per square metre, or g/m²). Height (cm) 40
Top margin (cm) 1.8
In Australia, the weight of paper and paperboard is most
commonly expressed as grammage, and it will be discovered Bottom margin (cm) 1.6
that since the 1970s, the grammage of newsprint has decreased Left margin (cm) 1.8
from a global standard of 52 g/m² to 48.8, 45 and 40 g/m2. This
is cheaper for newspaper producers, and increases the viability Right margin (cm) 1.8
of forest resources. Students can consider why publishers might Average number of pages
choose a certain weight of newsprint, and experiment with
Monday to Friday 78
various weights in their model.
Saturday and Sunday 96
Additionally, there is no standard agreement on how many trees
are required to create one paper’s worth (or one pound, one Circulation
kilogram, one tonne) of newsprint. Hence estimates are needed Monday to Friday 153 763
and these can vary with the source, because of variations in
Saturday 199 153
production processes.
Sunday 359 088
The following websites are useful references:
Weight
Conservatree Assume a representative standard newsprint
45 g/m2
http://conservatree.org/learn/EnviroIssues/TreeStats.shtml grammage
Number of trees to weight of paper
Australian Science
http://www.australianscience.com.au/environmental-science/ Internet research informs assumption that one
12 trees
paper-consumption-impact-in-australia/ tonne of paper consumes approximately 12 full-
per tonne
grown trees (Conserveatree)

www.immchallenge.org.au

Text by Peter Galbraith for IM²C and ACER. Available under Creative Commons non-commercial licence CC BY-NC 4.0.

Australian Council for Educational Research 28

Per year Interpret the solution
Single newsprint sheets = 52 × 56 779 353 = 2 952 526 356
In this problem, the building up of the solution has involved a
Area of sheet = 29 × 40 = 1160 sq cm = 0.116 sq m
consistent linking of all mathematical measures with their
Total area = 0.116 × 2 952 526 356 = 342 493 057sq m
real-world meanings. So interpretation has followed mathematical
calculations at all points.
Convert area of newsprint to weight
For this example, assume a representative value for grammage
45 g/m2. Other assumptions can be adopted to vary the Evaluate the model
calculations. Firstly, the calculations should be checked to see that all essential
variables have been built into the solution, and that arithmetic
Standard newsprint grammage = 45 g/m2 = 0.045 kg/m2
has been conducted accurately – noting that conversion between
Weight of newsprint = 0.045 × 342 493 057
different systems of units may be involved. The generalisability of
= 15 412 187.6 kg = 15 412.2 tonne
the model can be evaluated in terms of the formula developed for
number of trees used, and its usefulness also in terms of insights
Convert weight of newsprint to trees gained from sensitivity testing.
For this example, we assume a representative 12 trees per
tonne of paper. Other assumptions can be adopted to vary the Refinement
calculations.
The model can be refined by considering the impact of recycling.
1 tonne of newsprint uses 12 trees As an example for this problem, we have used information about
15 412.2 × 12 = 184 946 trees recycling from Clean Up Australia (http://www.cleanup.org.au/
files/clean_up_australia_paper_cardboard_factsheet.pdf).

Estimate area of trees needed

Australia leads the world in newsprint recycling
It is of interest to consider what this might look like in terms of
forest area. Australia’s newsprint recycling rate is almost 10 per cent
higher than in Europe, according to figures from the Publishers
Internet research can be used to discover how many trees on National Environment Bureau and media advocacy group The
average can be harvested/planted per hectare. This may result in Newspaper Works. Annual figures over more than a decade
different possible calculations. show that Australia consistently recovers around 78 per cent
of all newsprint. The average recycling rate for newsprint in
State-based government departments of primary industries, for Europe is 69 per cent. Australian newspapers only use tree
example, provide information on plantation management. pulp from plantation pine and use up to 40 per cent recycled
materials; but newspaper and magazines are the most
As an example calculation for this problem, using an average
abundant forms of paper waste.
value of 1000 stems per hectare, 184 946 trees are needed to
produce a year’s worth of the example newspaper; that is, How many trees are saved annually by using 40 per cent recycled
184.9 hectares or 1.85 sq km. materials in producing new newsprint, for the newspaper used in
this example question?
Generalise solution
0.40 × 184 946 = 74 000 (approximately)
Define symbols: page dimensions (l, w); number of pages weekly (n);
weekly circulation (c); grammage (g); trees/tonne of newsprint (t).

Create a formula that can be used to estimate the number of Report the solution
trees needed for the annual production of any daily newspaper. The modelling report could contain all the components of the
modelling problem and its solution, as developed in preceding
sections. The report should synthesise this data into a cohesive
Sensitivity testing
narrative, considering the implications of Australian plantation
Using the formula for convenience, it is useful to vary the inputs processes and use of recycled materials in producing new paper
from their values used in the example calculation and see how products.
the outcome is affected. This tests the sensitivity of the result to
changes in the inputs, and gives a sense of the robustness
of estimates.

www.immchallenge.org.au

Text by Peter Galbraith for IM²C and ACER. Available under Creative Commons non-commercial licence CC BY-NC 4.0.

Australian Council for Educational Research 29

Example problem
Level: Middle secondary
Intermediate modelling

Bushwalking

Describe the real-world problem

It is not uncommon for companions who enjoy bushwalking to differ in fitness and
energy. On tracks that lead out and back they will often walk together for a time at the
pace of the slower walker, until the slower walker indicates an intention to turn around
and return to base.

The faster walker has the choice of following the same action, but alternatively may
decide to carry on for a time at a faster pace before also returning. Especially if the
opportunity to travel further and faster is appreciated, the faster walker will want to go
as far as possible.

However, they will not want their companion to have to wait around too long at the end
of the walk for them to return.

Specify the mathematical problem Formulate the mathematical model

When the slower walker starts the return trip, for how much See the figure modelling the problem.
extra time should the faster walker travel on the outward path Let B = starting point and C = point that walkers reach together
before turning for home, so that both will arrive at the starting before separating.
point at the same time?
Suppose that S is the point reached by slower walker (s) on the
homeward path when faster walker (f) turns for home at F.
BF and FB are outward and return paths of the faster walker
along the same track.

B F

Australian Council for Educational Research 30

Assumptions Secondly, in terms of the real context, the outcome is amenable
While together, both proceed at the speed of the slower walker. to direct checking. Try it out having estimated a value for k. This
After parting, both walkers maintain their respective average helps to underline that real-world problem solving cannot live
walking speeds. entirely in a classroom.

Let t = time taken to reach C while walking together at the

speed V of the slower walker. Evaluate the model
During evaluation of the solution to an original problem, a
Slower walker (s) now starts to return while the faster walker
related problem at times suggests itself, and stimulates another
(f) continues on as far as F at speed (kV) for an additional time T
application of the modelling cycle or parts of it. In this case we
where k > 1.
might consider that it is more realistic in practical terms for ‘f’
For the walkers to arrive at B together: to aim to arrive back at base so that the slower walker doesn’t
have to wait ‘too long’ for the faster walker to return. That is, a
time taken for f to cover the distance FB = time taken for s to modified modelling question is set in terms of a time window
cover the distance SB. ‘w’, so that
0 < (difference in arrival times) < w.
So FB/kV = SB/V
Then in terms of formulating the model and solving the
mathematics we have the following.
Solve the mathematics After the walkers separate, ‘f’ travels a distance (2CF + CB) at
BC = Vt; CF = (kV)T; CS = VT speed (kV) to reach base B.
FB = BC + CF = Vt + kVT Time for ‘f’ to reach base = (2kVT + Vt)/kV = (2kT + t)/k.

SB = BC − CS = Vt − VT After the walkers separate, ‘s’ returns to base B at speed V

Hence V(t + kT)/kV = V(t − T)/V (same as outward trip).
So (t + kT)/k = (t − T) Time for ‘s’ to reach base = t.

2kT = (k − 1)t So ‘f’ reaches base at a time

T = [(k − 1)/2k]t [(2kT + t)/k -t] = [2kT − (k − 1) [(k − 1)t + kw)]/2k t]/k
after ‘s’ has arrived.
T can now be found by substituting different values for t, V, and
k. So we need 0 < [2kT − (k-1) t]/k < w
That is [2kT − (k-1) t]/k > 0 and [2kT − (k-1) t]/k < w
(It is common for the formulation of a model to merge
seamlessly into the mathematics of its solution.) [2kT − (k − 1) t]/k > 0
gives T > t(k − 1)/2k
and [2kT − (k − 1)t]/k < w
Interpret the solution gives T < [(k − 1)t + kw)]/2k
It is sufficient to know the time from the start of the walk to
the separation point (t), and the relative walking speeds (k) of Hence t(k − 1)/2k < T < [(k − 1)t + kw)]/2k
the two individuals. For the first we need to remember to look
at a watch at the start, and when point C is reached. (Note that For example, if w = 0.2 (12 minutes) and as before we consider
spending some time at C before resuming the walk will not the case where
affect the outcome.) And relative speeds are easily estimated by k = 2, t = 1, then we obtain 0.25 < T < 0.35.
comparing times taken to cover a chosen distance.
Interpreting, we have that if ‘f’ aims to have ‘s’ wait no longer
Firstly, check the mathematics for sensible outcomes. than 12 minutes at the starting point at the end of the walk,
f should continue for about 15 to 20 minutes from the point
k = 1 gives T = 0
where the walkers separate.
Both walkers turn together if they walk at the same pace, as
should happen. Report the solution
Suppose walkers stay together for an hour. The modelling report would contain all the above components
of the modelling problem and its solution synthesised into
t = 1 so T = (k − 1)/2k
a cohesive narrative. Additional working or explanations can
1 be added when judged to enhance the product, with its
Suppose k = 2: T = (‘f’ should continue on for 15 minutes).
4 completeness assessed in terms of, for example, the checklist
This seems a sensible figure. on report writing.

www.immchallenge.org.au

Text by Peter Galbraith for IM²C and ACER. Available under Creative Commons non-commercial licence CC BY-NC 4.0.

Australian Council for Educational Research 31

Example problem
Level: Middle secondary
Intermediate modelling

Farm dams l

Describe the real-world problem

Farmers get their feet wet

In a dry country like Australia, farm dams are part Dams lose water by evaporation from the water
of the lifeblood of rural life. Knowing the volume surface to the atmosphere. Annual average
of water at any time is important for planning evaporation rates are estimated using data
the numbers and distribution of livestock, and collected from locations throughout the country.
estimating when the supply is likely to run out They vary from about 100 cm in western
under drought conditions. Tasmania up to 400 cm in the desert regions of
Methods of estimating the volume of a partly northern Australia. Values in Victoria vary from
empty dam depend on interpreting physically about 140 cm in the south to 180 cm in the north,
observable signs. If depth markers are embedded according to the Bureau of Meteorology
when the dam is excavated, the volume can be (http://www.bom.gov.au/watl/evaporation).
estimated from the water level measured on the Loss of water by seepage is negligible.
marker. If dams do not have markers, some other
method is needed.

Livestock water requirements Table 1 Livestock water requirements

The table, containing data from a primary industries website,
shows water requirements for a variety of farm animals. It Stock Litres/animal/year
contains information that farmers would know for their own Sheep
stock, and is provided as a resource for this problem. nursing ewes on dry feed 3300
fat lambs on dry pasture 800
mature sheep — dry pasture 2500
fat lambs — irrigated pasture 400
mature sheep — irrigated pasture 1300
Cattle
dairy cows, dry 16 000
dairy cows, milking 25 000
beef cattle 16 000
calves 8000

Australian Council for Educational Research 32

Specify the mathematical problems Formulate the mathematical model (a)
Find a method of estimating the amount of water in a partly Discussion with students can be used to identify variables and
empty dam. measurements that will be needed to address the problem. Only
data that can be accessed are useful, and the depth of water
We have made an assumption that the dimensions of the is not one of these. One readily obtainable piece of data is the
dam when empty (as excavated) are known, and that they are distance the water surface has receded from its position when
as shown in Figures 1 and 2. The dam (Figure 1) has been the dam is full — a distance obtained simply by stepping out
constructed by excavating a horizontal square base ABTW. Sides the distance taken to walk from the edge of the dam, directly to
ABCD and WTSR are vertical, while the ends ADRW and BCST the current water level. A key insight is the recognition that this
have been sloped back to ground level so that the distances distance (x) is both measurable, and a key input to the model.
AD and BC are equal to half the length of AB (lengths are not
to scale). Livestock gain access to the water in the dam via the
sloping ends. Assumptions
Dams with rectangular surface areas and with access for animals
Figure 2 shows the vertical cross-section ABCD. The side length of can be represented by trapezoidal volumes of various kinds. The
the bottom square ABTW is ‘2a’, which is also the dam width. The dimensions of the dam when empty (as excavated) are known.
depth of the dam when full is ‘d’, and ‘h’ represents the depth when
it is partly empty. Sloping lengths AD and BC are of length ‘a’. For
this dam, a = 10 and d = 2 where the distances are in metres. Model for dam volume
We need to create a formula that enables the volume to be
estimated when the distance DP = x is known. (x is the distance
Problem a: Model for dam volume
found by stepping out or measuring the distance the water has
Find a way of estimating the volume of water in the dam when receded down the slope from its highest point when the dam is full.)
the (unknown) depth of water is ‘h’.
Volume of dam when full = area (ABCD) × width (2a)
Problem b: Model for water loss over time In trapezium ABCD: length CD = AB + 2x DN = 2a + 2acosθ where
sin θ = d/a
Suppose the dam is in northern Victoria. If the dam is full and no
more rain falls, how many days it would take for the dam to dry Area of ABCD = xa (2 + cos θ) × d
up? Now suppose that a herd of 100 beef cattle is sustained by Volume of dam (full) = 2 a2d (2 + cos θ)
the dam. If the dam is full and no more rain falls, how long will it Given a = 10, d = 2, sin θ ≈ 0.2 (θ ≈ 11.5˚ and cos θ ≈ 0.98)
be before other arrangements need to be made for the stock?
Volume of dam (full) = 1192m3 (approx.) = 1.192 megalitres

R
To find volume of partly filled dam consider cross-section ABQP
S in Figure 2.

W T PM/MA = DN/NA (similar triangles) so PM/h = a cos θ/d and

PM = ah cos θ/d.
1
D C Area (ABQP) = xh (4a + 2ah cos θ/d) so volume at height h is
2
given by:
A B V(h) = area (ABQP) × 2a = 2a2h[2 + (h/d) cos θ].
Hence V(h) = 200h(2 + 0.49h) − a quadratic in ‘h’.
Figure 1 Diagram of empty rectangular dam
(Note as a check that when h = d = 2, V = 1192 as above.)

Now we need to express ‘h’ in terms of ‘x’.

D
0 N C By similar triangles DNA and PMA:
x d
M Q (a − x)/h = a/d so h = (d/a)(a − x) = 0.2(10 − x).
P h
a Substitution in the formula for V(h) gives:
A B
2a V(x) = 40(10 − x)(2.98 − 0.098x).
Check: When x = 0, V = 1192 as above.
Figure 2 Vertical section through rectangular dam This formula can now be written more tidily as:
V(x) = 3.92(10 − x) (30.4 − x) which rounds to a value of 1192
when x = 0.

www.immchallenge.org.au

Text by Peter Galbraith for IM²C and ACER. Available under Creative Commons non-commercial licence CC BY-NC 4.0.

Australian Council for Educational Research 33

Solve the mathematics (a) Formulate the mathematical model (b) for water
Note that the constant checking of numerical values for loss over time
consistency, during formulation, can also be considered part of To estimate how long water in a dam will last, we need to make
the solution process. Volumes corresponding to different values further assumptions – about evaporation.
of ‘x’ calculated from the formula above are shown in Table 2
below. Volume columns show the respective percentages of the Assumptions
total capacity. Also shown (for comparison) are the values of the
Water in the dam is reduced by evaporation and in providing for
inaccessible depth measures (h) that correspond to the different
livestock. Seepage loss is negligible and can be ignored.
selected values of the accessible measure (x).
Internet research will typically show that the evaporation rate varies
Figure 3 is a graph of V = 3.92(10 − x) (30.4 − x) for values of x
with temperature, wind speed, sunshine, and relative humidity.
between 0 and 10.
It also varies throughout the year, but a rough daily evaporation rate
It could be used, for example, as a chart pinned on the kitchen
(average) can be found in (cm or mm) by dividing the average annual
wall from which the volume of water remaining can be read, given
value by 365 days. This will be sufficient accuracy for estimation
only the distance (x) to the water line from the high water mark
purposes, although in practice there will be seasonal variations.
when the dam is full.
Using the value for northern Victoria given in the problem
Table 2 Dam volume and depth against distance dam water has
description, we obtain the average amount of evaporation per day
receded from high water mark
= 180/365 ≈ 0.5(cm).
Volume This is the ‘depth’ of water that is lost across any exposed surface
x (metres) h (metres)
(% of whole)
area in a day. The volume lost will vary with the surface area.
0 100 2.0
Calculations will overestimate the number of days that suitable
1 87.0 1.8 water is available to animals. Near the end the dam will resemble
a bog and the water undrinkable.
2 74.7 1.6

3 63.1 1.4 Model for water loss over time

From Figure 2, the cross-sectional area when the depth of the
4 52.1 1.2 dam is ‘h’ [A(h)] is a rectangle with length PQ and width 2a.
5 41.8 1.0
A(h) = (2a + 2ahcosθ)/d × 2a = 4a2 (1 + hcosθ/d)
6 32.1 0.8 = 400(1 + 0.098h) when a = 10, d = 2.
When a = 0, A = 400 (area of the square base of the dam).
7 23.1 0.6
The area varies linearly with h, so the average cross-sectional
8 14.7 0.4 area occurs when h = 1.
So average value of the exposed area of water surface = 439.2 m2,
9 7.0 0.2
for which the daily evaporation loss would be
10 0 0 439.2 × (0.5)/100 ≈ 2.2 m3.

y
1100
1000
900
800
700
600
500
400
300
200
100
0
1 2 3 4 5 6 7 8 9 x
Figure 3 Graph of volume against x for rectangular dam

www.immchallenge.org.au

Text by Peter Galbraith for IM²C and ACER. Available under Creative Commons non-commercial licence CC BY-NC 4.0.

Australian Council for Educational Research 34

Solve the mathematics (b) for water loss over time For water loss over time, noting the observation about boggy
conditions when the dam is nearly empty, the figure obtained will
If the dam is full and no more rain falls, we can estimate how
overestimate the time drinkable water will be available. A safer
many days it would take for the dam to dry up:
estimate would be about 5 months.
1192/2.2 ≈ 541 days (approximately 18 months’ supply).

If the dam must sustain a herd of 100 beef cattle, then daily
consumption of water by 100 beef cattle (from the example data): Evaluate the model
16000 × 100/365 ≈ 4384 litres ≈ 4.38m3. Apart from the checking of working for possible errors
in mathematical calculations and/or in the application of
Total average water loss per day from (consumption +
technology, evaluation involves continuous checking against the
evaporation) ≈ 6.58 m3.
needs of the problem context.
Number of days of water available before other arrangements
Has the solution provided a sufficiently good answer to the
must be made to maintain livestock ≈ 1192/6.58 ≈ 181 days
problem posed, or do we need further work?
(about 26 weeks or 6 months).
Sometimes when the answer to the first question is ‘yes’ the
first answer obtained suggests a deeper exploration that only
Interpret the solution becomes obvious from the initial modelling effort. This then
Mathematical outcomes have been linked throughout to the dam stimulates another cycle of modelling with an amended purpose.
structure, its volume and dimensions, and practical implications –
A different perspective on evaluation was reported by a teacher
as for example the meaning of the graph in Figure 3. This is typical
who used a version of this problem with her Year 10 class. An
of problems involving a variety of mathematical calculations. Their
appreciative parent who happened to be a farmer told her that
meaning within the problem needs to be interpreted and assessed
stepping down the bank was the method he and others used in
as they arise, for testing numerical outcomes against the real
estimating the amount of water in a dam.
situation will often identify errors in calculation that need to be
addressed.

The formula obtained, translates the stepped out distance (x) into Report
a corresponding value for volume that gives estimates of volume The modelling report could contain all the above components
for any measured value of x. This would provide the basis for of the solution of the problem. It should summarise and
constructing a ready reckoner, or wall chart if desired. illustrate how the mathematical insights obtained advanced an
understanding of the problem — even if this sometimes means
The data provided, assumed a consistent evaporation rate
that the solution attempt in its present state is in need of further
based on an annual average. But the scenario posed — drought
development. All assumptions and choices of data values should
conditions — could be considered to be different from the
be explained and justified.
average. How will the outcome vary, if different values are
considered for the evaporation rate? In this case the report should develop as a systematic and
cohesive narrative: considering the implications of drought
The data provided for livestock water requirements are annual
conditions for livestock farmers; providing tools for farmers to use
averages. In a real-world scenario, are animals likely to need more
to estimate water supply, such as the data shown in Table 2 and
water in drought conditions than in average weather? How will
the graph in Figure 3; and recommending a timeframe within which
the outcome vary if different values are considered for the water
provisions should be made for alternative supplies for stock.
consumption rate?

www.immchallenge.org.au

Text by Peter Galbraith for IM²C and ACER. Available under Creative Commons non-commercial licence CC BY-NC 4.0.

Australian Council for Educational Research 35

 Suggestions for
student-generated modelling

At this stage, we encourage students to invent their own problems Catastrophic events
and to outline their approach to solutions. These problems should
■■ How can you predict the severity of the damage of a tsunami
be related to things that students find interesting and important.
from the Richter scale value based on energy?
Some topics with the potential to give rise to good modelling
projects are indicated below.
Environmental problems
Music ■■ Drought is causing water shortages in cities. How much
water do we have left in the Hinze Dam? Will it cater for the
■■ How many artists actually make any money from their music?
current and future population of the Gold Coast?
How much is piracy costing performers?
■■ Analyse the spread of introduced biological control agents.
Politics What is the culling rate needed to stabilise the cane toad
population in Australia?
■■ What is the benefit of minor parties? How are they managed
in other countries? ■■ What do models tell us about climate change?

Sport Disease (epidemics and pandemics)

■■ Is gambling good for the country? How well can winners be ■■ What do the past and current trends in HIV/AIDS diagnosis
predicted? and deaths suggest for the future number of people afflicted
and their odds for surviving?
■■ In archery, what is the optimum angle of elevation for the
release of an arrow in the sport of archery so that the ■■ Research Creutzfeldt–Jakob disease (sometimes described
arrow hits the perfect bullseye? How does air resistance as the human form of ‘mad cow’ disease). The Chief Medical
subsequently influence this optimum release angle? Officer in the United Kingdom has suggested fatalities from
the disease could number in the hundreds of thousands.
Games What are the rates of people who are infected but not
showing symptoms, compared to fatalities?
■■ What games can be analysed to find winning strategies?
A number of these were chosen and developed by students at a
modelling challenge sponsored on the Gold Coast over several
Manmade disasters
years by A B Paterson College and Griffith University.9
■■ When will the Aral Sea in Uzbekistan dry up completely? Can
it be saved?

9
See Galbraith, P., Stillman, G. & Brown, J. (2010). Turning ideas into modelling problems. In R. Lesh, P. Galbraith, C. Haines, & A.
Hurford (Eds.), Modeling students’ modeling competencies (pp. 133–144). New York: Springer.

Australian Council for Educational Research 36

Senior modelling problems for
middle or senior secondary

The problems at this level are designed to bridge from students’

previous learning into the more substantial and complex demands
of IM²C-style problems. These examples call upon initiative,
persistence, decisions about the type of mathematics to apply, use
of the web for data and solution methods, and how technology
might be utilised. Generally, more advanced mathematical
techniques will be needed. These problems contain scope for
individuals and groups to exercise initiative and demonstrate
attention
to detail.
As in the intermediate-level examples, we suggest that, provided
they have an understanding of the modelling process, students
work in groups of about four and divide the work between them.
They should start as a whole group and decide what needs to be
done, what assumptions need to be made; what data are needed;
which students will work on which sections; and how they will
construct a summary report. However, all students should have an
awareness of how the overall solution is proceeding, and feel free
to contribute to any part of the process.
At this level, we would expect that more time be spent on report
writing – again depending on total time availability. Students
should write summary reports and deliver verbal results using their
summaries. As much time should be spent on writing full reports
as possible. It might be useful for student groups to assess each
other’s work. Preferably this should be a blind assessment.
The need to communicate is hopefully already a part of students’
everyday classwork in maths. Students can develop skills in
presenting both verbally to the class and in writing as part of
their written answers. They should begin to treat their answers to
assigned problems/exercises in some detail. Rather than giving an
answer simply as ‘V = 3.67’, a more complete explanation can be
provided, such as: ‘I set up the volume, V, as an integral using the
given function. Then I did the definite integration. The answer was
3.67. So I know that the volume of the dam is 3.67 gigalitres.’
This approach to mathematics ought to help the students’
fluency in the subject (see the proficiency strand of the Australian
curriculum: Mathematics) and their fluency in English. In real life,
the first may not be as significant for them as the second. Being
able to communicate is an important skill in life. This is especially
true in the workplace, where increasingly, teams of people get
together to solve problems. Because members of these teams are
regularly drawn from various disciplines and backgrounds, they all
need to able to communicate their different perspectives in order
to maximise the value of their end product.

Australian Council for Educational Research 37

Example problem
Level: Senior secondary
Senior modelling

Nuclear blast

Berlyn Brixner/Los Alamos National Labora

Describe the real-world problem

The likely properties of the planned first atomic time intervals. Armed with a series of photos,
bomb were unknown until a trial detonation, he devised a question and conducted
the Trinity atomic test, was undertaken in the modelling to estimate the energy released in
New Mexico desert in 1945. the blast.

Among the invited witnesses to the trial was Taylor’s approach to the problem has been
a Cambridge Professor, Geoffrey Taylor, who discussed in various sources, including by
had been an advisor to the Manhattan project University of Cambridge Professor Timothy
team, the group responsible for developing the Pedley in the journal Mathematics Today in
nuclear device. Simply witnessing the trial did 2005. The paper describes (with illustrations)
not provide a measure of its strength. how to apply mathematics to real-world
problems, including a simplified working of the
Later, in 1947, photographs of the blast were nuclear blast problem.
made public in a variety of sources, including
Life magazine. Taylor was browsing through Sufficient information is provided to enable
one of these which contained a report of the the investigation to be illustrated through a
test, together with photos of the expanding modelling approach.
blast wave, taken over a succession of small

Specify the mathematical problem Mathematics Modeling Challenge program does not explicitly
address this aspect, as by its nature problems are specified as
Estimate the energy released in the bomb blast.
the starting point.) Secondly, this problem affirms the use of a
cyclic modelling process by a professional modeller.

Formulate the mathematical model The fundamental assumption is that the radius of the spherical
This problem illustrates two significant aspects about modelling blast wave (R) depends on a product of three factors: the time
activity. Firstly it illustrates an important attribute of a modeller elapsed since the explosion (t), the instantaneous energy released
– the identification by modellers of model-rich situations in the (E), and the density of air (ρ). Thus R = CtaEbρc, where C is a
first place. Curricular goals for students to use their mathematics dimensionless constant.
to solve problems of personal interest, in work contexts, and
as productive citizens require this ability. (The International (These assumptions are taken as a given for this development.
Their plausibility can be discussed if desired.)

Australian Council for Educational Research 38

Solve the mathematics photograph. Taylor used data from the series of photographs to
plot the graph of log R against log t.
Taylor then proceeded to apply dimensional analysis as elaborated
1
below, in order to find suitable values for the indices a, b and c, and (Note that R = (Et2 /ρ) 5 can be written as
to use the results to estimate the amount of energy released. 1
log R = log (E /ρ 5 ) + 2 log t, so that the graph of log R against
Using the standard notation used for dimensions dim R = [R] etc, log t is linear. The value of E can then be robustly estimated from
to express quantities in terms of the fundamental dimensions the intercept on the vertical axis). This is illustrated briefly in the
mass (M), length (L), and time (T) we have: Pedley article.

[R] = L, [t] = T, [E] = ML2 T-2, and [ρ] = ML-3 An internet search will identify many sources that discuss
aspects of the Taylor approach. Several of these provide a series
Thus dimensionally, using the formula above, we need: of photographs of the blast wave at successive time intervals.
L1 = M(b + c) L(2b − 3c) T(a − 2b) These can be used to give separate estimates of the energy
Equating dimensions on both sides of the equal sign we need : b released (as in the calculation above), or together to generate
+ c = 0; 2b − 3c = 1; a − 2b = 0 a log-log plot as used by Taylor. The sources do need vetting
before being given the students, as some include unnecessary
2 1 1
hence a = ,b= , and c = − . complications.
5 5 5
1
So R = C (Et2 /ρ) 5 A useful source (Codoban, 2004) is located at
http://www.atmosp.physics.utoronto.ca/people/codoban/
(A value of C ≈ 1 was assigned on the basis of knowledge of blast PHY138/Mechanics/dimensional.pdf
activity, and hence E = ρR5/t2.)

The above photograph of the blast contains a scale representing

100 m, and a label indicating that it was taken at t = 0.025 (s).
Report the solution
Expanding the photo from the link, taking measurements, and Despite post-war national security issues, Taylor published his
using the given scale to estimate the radius suggests a value for results showing the Trinity atomic bomb had a power equivalent
R of about 132 m. of about 17 kilotons of TNT. The US Army was annoyed – the
information was supposed to be classified!
Noting that density of air is 1.2 kg/m3 and substituting in the
above formula gives a value for E of about 7.7 × 1013 joule.
References
Codoban, S. (Ed.) (2004). Estimate of the energy released in the
Interpret the solution first atomic bomb explosion [tutorial notes for University of
The standard measure of strength of bomb blasts is given Toronto subject Physics for the Life Sciences].
as a comparison with the yield of the explosive material TNT, http://www.atmosp.physics.utoronto.ca/people/codoban/
trinitrotoluene. Internet research indicates that the energy PHY138/Mechanics/dimensional.pdf
released by 1 metric tonne of TNT is equivalent to about
4.148 gigajoules, where 1 gigajoule equals 1 billion joules. Pedley, T.J. (2005). Applying Mathematics. Mathematics Today,
Check that the value of E for the Trinity blast calculation (using 41(3), 79–83.
the single diagram above) converts to an energy equivalent of
Taylor, G.I. (1950). The formation of a blast wave by a very intense
about 18.4 kilotonnes of TNT.
explosion. II. The atomic explosion of 1945. Proceedings
of the Royal Society of London. Series A, Mathematical and
Physical Sciences, 201(1065), 175–86.
Evaluate the model http://www3.nd.edu/~powers/ame.90931/taylor.blast.wave.II.pdf
The method used by Taylor was slightly different from the above
which is based on measurement and calculation from one

www.immchallenge.org.au

Text by Peter Galbraith for IM²C and ACER. Available under Creative Commons non-commercial licence CC BY-NC 4.0.

Australian Council for Educational Research 39

Example problem
Level: Middle and senior secondary
Senior modelling

Population growth

Describe the real-world problem

Australia’s population to hit 23 million

Tuesday, 23 April, 2013.

Australia’s population will reach 23 million The ABS estimates that there is a birth
people overnight, and is on track to every one minute and 44 seconds,
surpass 40 million within 40 years. a death every three minutes and 32
seconds, and a new migrant arriving
Projections build on last year’s known every two minutes and 19 seconds.
population, and take into account fertility,
life expectancy and immigration figures. That means our population increases by
Figures from the Australian Bureau of one person every minute and 23 seconds
Statistics show that around 180 000 – more than 1000 people per day.
people move to Australia each year.

Specify the mathematical problem Formulate the mathematical model (a) for
daily growth
Problem a: Daily population growth
Increase in persons per day = no of births per day – no of deaths
Simple problem. Check and comment on the claim that ‘our per day + no of net migrants per day.
population increases by one person every minute and 23 seconds
– more than 1000 people per day.’
Solve the mathematics (a) for daily growth
This amounts to evaluating claims made by others on the basis As on 23 April 2013:
of someone else’s model. This is itself an important activity to
no of births per day = (24 × 60 × 60)/104 = 830.77
recognise and undertake when appropriate.
(one birth every 104 sec)
no of deaths per day = (24 × 60 × 60)/212 = 407.55
Problem b: Population growth over 40 years
(one death every 212 sec)
Advanced problem: Is it likely that Australia’s population will reach
no of (net) migrants per day = (24 × 60 × 60)/139 = 621.58
40 million in 40 years?
(one migrant arrives every 139 sec)

Australian Council for Educational Research 40

So, increase in persons per day = 830.77 − 407.55 + 621.58 Formulate the mathematical model (b) for
= 1044.8 (1045 persons)
growth over 40 years
Time interval between arrivals = (24 × 60 × 60)/1045 The advanced problem asks: Is it likely that Australia’s population
= 82.68 sec (1 min 23 sec approx) will reach 40 million in 40 years?

Given that births, deaths, and net migration have been established
Interpret the solution (a) for daily growth as the key variables, the emphasis changes to designing a model
The calculations verify the claims that on this day the population for long-term population forecasts – typically expressed in terms
increases by one every 1 min 23 sec, which is more than 1000 of years.
persons per day
It is useful to summarise what we know from the data given in the
report (see the table). The last column generalises the calculations
Evaluate the model (a) for daily growth to express the yearly change in a population (P) in terms of the
respective contributions from births, deaths, and migration.
The model is built in terms of births, deaths, and net immigration,
and gives results precisely in line with the claims. We can be The calculations assume that the data given for April 23 apply
confident it is a good model for the purpose it was created. across the year, which seems to be a suggestion within the
reporting. This is a point for discussion in its own right.
However, we can point to reservations about its wider use.
The model uses data for births, deaths, and net immigration that
are applicable on a particular day. For population predictions long- Assumptions
term, we need annual estimates of these quantities. For example, For our first model we will use the values given in the report, and
a migration rate of 621 per day (the figure that applies on 23 April included in the table. This assumes that the fractional birth rate,
2013) if used to calculate an annual figure gives a value over 226 000 the fractional death rate, and the rate of migration remain the
(many more than the average of 180 000, given by the ABS). same over the time scale of the model.
The challenge of estimating population size into the future is Note that in published statistics, migration rate refers to the net
considered in the next problem. rate of increase through migration.

These assumptions deserve discussion, as they provide a means

of making progress, but may also be a source of reservation in
accepting predictions. They will be revisited when evaluating
model outcomes.

Population in 2013 = 23 000 000 Population = P

Fractional birth rate (b)

births per day births per year
one per 104 = births per year ÷ total pop births per year
Births = (60 × 60 × 24) ÷ 104 = 830.77 × 365.25
seconds = 303 439 ÷ 23 000 000 = Pb
= 830.77 = 303 439
= 0.0132 per year (yr−1)

Fractional death rate (d)

deaths per day deaths per year
one per 212 = deaths per year ÷ total pop Deaths per year
Deaths = (60 × 60 × 24) ÷ 212 = 407.55 × 365.25
seconds = 148 856 ÷ 23 000 000 = Pd
= 407.55 = 148 856
= 0.00647 per year (yr−1)

Migrants per
year
migration per day migration per year = M/365.25
one per 139 Migration numbers (m)
Net migration = (60 × 60 × 24) ÷ 139 = 621.58 × 365.25 where
seconds = 227 032 persons/year
= 621.58 = 227 032 M = total
migrants in a
year

Note that net migration is independent of current population, while the number of annual births (Pb) and annual deaths (Pd) is directly
influenced by it.

www.immchallenge.org.au

Text by Peter Galbraith for IM²C and ACER. Available under Creative Commons non-commercial licence CC BY-NC 4.0.

Australian Council for Educational Research 41

Real-world material relevant to both modelling and discussion To solve by calculus
can be obtained from internet sources. Students will often do δP ≈ Pr δt + M δt
this on their own initiative, or can be prompted if relevant. (It is
important that such activity remains focused on the modelling = (Pr+ M) δt where r = b − d.
task – for example, is related to clarifying assumptions – and In the limit
does not become a diversion).
dP/dt = Pr + M
ln(Pr + M) = rt + constant, leading to
Solve the mathematics (b) for growth over 40 years rt rt
P(t) =P0e +M(e −1)/r where P(0) = P0
For this first model, we use the data provided in the stimulus text,
and summarised in the table. There are a number of ways this P40 = 40 526 191
problem can be approached mathematically, and three different
approaches are shown.
Interpret the solution (b) for growth over 40 years
The spreadsheet solution can be applied by anyone with
All methods confirm a predicted population of around 40 million
spreadsheet competence. The other solutions require knowledge
in 40 years’ time.
of geometric series and calculus respectively, which locates them
within senior mathematics.

Let P0 = initial population (in 2013) Evaluate the model (b) for growth over 40 years
Appropriate mathematical growth processes have been applied,
Let Pn = population in year n and similar results obtained using three different methods. It
seems reasonable to trust the methods.
Let r = natural population change rate
= birth rate (b) − death rate (d) i.e. (b – d) What about the assumptions leading to the values used for b, d,
and M which were all based on their values on a particular day:
Let M = average net annual immigration intake
April 23, 2013?

To solve by spreadsheet It is easy for death rates to have a ‘spike’ – for example during ’flu
P0 = 23 000 000; b = 0.0132; d = 0.00647; r = b − d = 0.00673; epidemics, or for some reason to be lower than average during a
M = 227 032 short interval. For predictions, we need stable average values. If
members of a species have an average life time of 10 years, then
P1 = P0 + rP0 + M = P0 (1 + r) + M = P0 R + M (where R = 1 + r) on average 1/10 of a population (the fractional death rate) will die
P2 = P1 R + M etc each year. Generalising, an average lifetime of ‘L’ means that on
average a fraction ‘1/L’ of the population die each year. In the fifth
• −1
column of the table we see that d = 0.00647 yr , which implies an
Copy down average life time of 154.6 years.

• Clearly the value on April 23 2013 was not typical and we need a
more representative value.
•
P40 = 40 459 252 Websites such as the Australian Institute of Health and
Welfare give figures for life expectancy for Australians
(http://www.aihw.gov.au/deaths/life-expectancy ).
To solve by geometric series
Proceeding as above (annual increments) Life expectancy is not quite the same as average lifetime but is
a close approximation (an interesting point of discussion if time
P1 = P0R + M is available). If the average life time in 2013 was 81.5 years, the
2
P2 = P1R + M = P0R + M (R + 1) average fractional death rate = 1/81.5 = 0.0122699 yr−1.

• Also, the net immigration rate is quoted officially at about 180 000

• per year, which is substantially less than the figure implied by the
specific value that applied on the special date.
n n−1 2
Pn = P0R + M (R + … R + R + 1)
n n−1 Leaving the value of ‘b’ alone, which is essentially determined by
Pn = P0R + M (R )/R − 1
personal family planning decisions, and adjusting the values of
P40 = 40 459 252 (annual increments) ‘d’ and ‘M’ as above, show that using the above three methods of
calculation, leads to a value of P40 = 31 200 000 (approximately).
P40 = 40 523 429 (monthly increments)
P40 = 40 523 960 (daily increments)

www.immchallenge.org.au

Text by Peter Galbraith for IM²C and ACER. Available under Creative Commons non-commercial licence CC BY-NC 4.0.

Australian Council for Educational Research 42

This is substantially less than the previous estimate. There An almost identical graph of a linear approximation has equation
are implications for jobs, housing, health provision, education… M = 430 652 – 31 500 000r (red). Now feasible solutions can be
if forecasts are relied upon. This adds a social context around considered by noting the real world meanings of r and M. For
the problem. example if r = 0 (so called replacement rate when deaths are just
matched by births) then M = 430 652, leading to a population of
40 226 080 in 2053. Is this realistic? Allowing ‘r’ and ‘M’ to vary
Refinement (advanced)
within the 40 year period provides for considering implications of
Over time, the natural growth rate (births – deaths) is monitored changes in family planning decisions (personal) and immigration
and published as statistics. Knowing these, immigration can be policy (national).
used as a policy tool to address population issues. We can use
this idea to pose a different question.

Suppose a population of 40 million is the goal for 2053 (40 years Report
from 2013). What combination of natural growth rates and The modelling report should contain all the components of the
immigration rates would achieve this? modelling problems, their solutions, interpretation and evaluation.
rt rt
The report could provide a cohesive narrative, including some
For this senior level refinement using P =P0e +M(e – 1)/r from discussion of implications of Australian population growth
previous work we need for policy, society and economy around matters such as jobs,
40r 40r housing, health, and education. When forecasts are astray where
40 000 000 = 23 000 000 e + (e -1)I/r
does responsibility lie? With demographers? With data problems?
This problem requires facility with CAS technology such as Maple Has the media been creative with facts or interpretations?
or Mathematica. Using the former and noting that MIG = M the
graph is shown on the axis below (blue in online format).

700000

600000

500000

MIG
400000

300000

200000

100000
-0.01 -0.005 0 0.005 0.01
r

www.immchallenge.org.au

Text by Peter Galbraith for IM²C and ACER. Available under Creative Commons non-commercial licence CC BY-NC 4.0.

Australian Council for Educational Research 43

Example problem
Level: Senior secondary
Senior modelling

Fifteen-forty

Describe the real-world problem

Results: The Championships, Wimbledon 2015

Ladies’ singles: Serena Williams def Garbine Muguruza: 6–4, 6–4
Gentlemen’s singles: Novak Djokovic def Roger Federer: 7–6 (7–1), 6–7
(10–12), 6–4, 6–3
Some years ago a tennis commentator made the following remark during
a Grand Slam tournament: ‘A top male player has a fifty-fifty chance of
winning a game from 15–40 on serve.’

Specify the mathematical problem Formulate the mathematical model

Evaluate this statement. That is: what probability can we assign
to the outcome of winning a service game from a score of Data
15–40? Students can be given the data table below or tasked with
collecting the required data. The official Wimbledon website
contains many years’ worth of results, including the most
up-to-date live scores. http://www.wimbledon.com/en_GB/scores.

Other Grand Slam tournament websites also contain similar data.

Finals match data

Wimbledon 2015 Williams Muguruza Djokovic Federer

 
37 of 68 43 of 61 95 of 145 94 of 141
1st serve in (r)
54% 70%   66% 67%

29 of 37 23 of 43 70 of 95 70 of 94
Winning 1st serve (f)  
78% 53% 74% 74%

11 of 31 6 of 18   30 of 50 23 of 47
Winning 2nd serve (s)
35% 33% 60% 49%

Australian Council for Educational Research 44

Situational assumptions Hence Pr(D) = p2 + 2pq × Pr(D)
Given the long run frequency of serving outcomes we make the Hence Pr(D) = p2/(1 − 2pq) = p2/(1 − 2p + 2 p2) since q = 1 − p.
following assumptions:
So Pr(G) = p4/(1 − 2p + 2p2)
■ The probability of the outcome of any particular serve is
estimated as the long run proportion for the outcome of Check:
serves as given in the table. If p = 0, Pr(G) = 0 and if p = 1, Pr(G) = 1
as should occur.
■ The outcome of any point is independent of the result of the
previous point. (Note that the reliability of these assumptions Interpret the solution
should increase with the quality of the players.)
If we take the case of Federer then p = 0.66
■ There are two mutually exclusive outcomes for each point Substituting this value for p gives Pr(G) = 0.34
(win or lose). That is, a 34 per cent chance of winning the game.

Verify that for Djokovic the figure is around 40 per cent.

win
We can find the value of ‘p’ that will meet the commentator’s
f
criterion as follows.

We need to solve p4/(1 − 2p + 2p2) = 0.5

1-f
r lose That is,
2p4 − 2p2 + 2p − 1 = 0 which we can solve with the help of
serve win Computer Algebra System (CAS) software.
p ≈ 0.75 is the single positive root
s Check:
1-r
When p = 0.75, Pr(G) = 0.506 (approximately 50 per cent chance
of winning).
1-s lose
It seems the commentator was a bit over optimistic!
Outcomes of a service point But, Wimbledon finals is the very best playing against the very
best. If the opponent is a lesser player (which happens most of
To win a point:
the time) then the opinion might be closer to the mark.
either the first serve is in (with probability given by r) and the
service point is won (with probability f) The Wimbledon website contains complete statistics for all seven
or the first serve is out (with probability 1 − r) and the second matches played by finalists. To obtain them click on the names of
serve is won (with probability s). players.

The collated data for Djokovic gives outcomes for 728 service
Hence, as seen in the figure using the notation from the table, the
points across seven opponents.
probability of winning a service point (p) is given by
p = rf + (1 − r)s Using the total data for the same variables shown in the table we
calculate:
r = 0.707 (0.71)
Solve the mathematics f = 0.77
Using the data in the table, we obtain: s = 0.636 (0.64)
p (Djokovic) = (0.66)(0.74) + (0.34)(0.60) = 0.69 giving p = 0.731 (0.73)
p (Federer) = (0.67)(0.74) + (0.33)(0.49) = 0.66
With p = 0.73 we obtain Pr(G) = 0.469 (around 47 per cent chance
Note that q = 1 − p is the probability of losing a service point, of winning).
and that the probability of winning a point as receiver is the So the estimate is looking quite a reasonable one, especially
complement of the opponent’s probability of winning as server. when we consider that many opponents on the circuit will be
lesser players than those accepted for a Grand Slam tournament.
Let Pr(G) be the probability that a player wins a service game
from 15–40.
Evaluate the model
The server must win the next two points and then win from deuce.
We can also take a statistical look at outcomes.
So our required probability is Pr(G) = p2 × Pr(D), where Pr(D) is the
probability of winning a service game from deuce. Assume that the choice of the seven opponents in the draw is
sufficiently random to support the estimation of confidence
To win from deuce the server must win the next two points intervals. (The sample is probably best regarded as representative
(probability p2) rather than random.)
or return to deuce and win from deuce.
Consider the 728 service points from Wimbledon 2015 as a
To return to deuce the server must win the first point and lose the sample from the population of many thousands that a quality
second (pq) or vice-versa (qp). player, such as a Grand Slam winner, is involved with while at the
top of his or her form over a period of years.

www.immchallenge.org.au
Text by Peter Galbraith and Alistair Carr for IM²C and ACER. Available under Creative Commons non-commercial
licence CC BY-NC 4.0.

Australian Council for Educational Research 45

Considering the probability of winning a point as given by the Report the solution
proportion of service success, then the standard error of the
This analysis has focused around Novak Djokovic as a case
sampling distribution of proportions is given by (√(p*(1-p*)/n)
study. There is opportunity to test the commentator’s claim using
where p* is the value (0.73) from the sample (n = 728).
a selection of top players, and ideally this should be pursued.
Then with (approximately) 95 per cent confidence we estimate
For example, what are the corresponding outcomes for top
the population value of p to lie in the interval:
women players? A similar approach can be used to investigate
p* − 2(√(P*(1 − p*)/n) ≤ p ≤ p* + (2√(P*(1 − p*)/n)
parallel outcomes for women players, using data from the
That is 0.698 ≤ p ≤ 0.764 which translates to website.
0.409 (41%) ≤ Pr(G) ≤ 0. 531 (53%)
Perhaps Serena Williams is just about the best of all time? Use
Given that some opponents will not be of Grand Slam quality, we available data to compare her serving performance with that of
might argue that this outcome is probably still conservative at the other top women players.
upper end.
Further problems could also be explored. For example, what are
It is interesting that the original Wimbledon data generated reasonable probabilities of winning a game from each of the
a p-value (0.69) that lay just outside the lower 95 per cent positions from 0–40 to 40 –0?
confidence limit. It is probably no surprise that such an extreme
Using the same notation, show that the probability of winning a
occurrence took place with Roger Federer as the opponent.
game from 40–0 can be expressed as:
The first evaluation suggested further avenues to explore, which Pr(G) = p(1+ q + q2) + q3 Pr(D)
were not envisaged at the start. This is an illustration of how Pr(G) = 0.998 when p = 0.73
interim results can often stimulate further modelling activity in and so on.
which aspects of the modelling cycle are re-activated, sometimes
a number of times.

www.immchallenge.org.au
Text by Peter Galbraith and Alistair Carr for IM²C and ACER. Available under Creative Commons non-commercial
licence CC BY-NC 4.0.

Australian Council for Educational Research 46

Example problem
Level: Senior secondary
Senior modelling

Farm dams ll

Describe the real-world problem

Farmers get their feet wet

In a dry country like Australia, farm dams are part Dams lose water by evaporation from the water
of the lifeblood of rural life. Knowing the volume surface to the atmosphere. Annual average
of water at any time is important for planning evaporation rates are estimated using data
the numbers and distribution of livestock, and collected from locations throughout the country.
estimating when the supply is likely to run out They vary from about 100 cm in western
under drought conditions. Tasmania up to 400 cm in the desert regions of
Methods of estimating the volume of a partly northern Australia. Values in Victoria vary from
empty dam depend on interpreting physically about 140 cm in the south to 180 cm in the north,
observable signs. If depth markers are embedded according to the Bureau of Meteorology
when the dam is excavated, the volume can be (http://www.bom.gov.au/watl/evaporation).
estimated from the water level measured on the Loss of water by seepage is negligible.
marker. If dams do not have markers, some other
method is needed.

Livestock water requirements Table 1 Livestock water requirements

The table, containing data from a primary industries website,
Stock Litres/animal/year
shows water requirements for a variety of farm animals. It
Sheep
contains information that farmers would know for their own
stock, and is provided as a resource for this problem. nursing ewes on dry feed 3300
fat lambs on dry pasture 800
mature sheep — dry pasture 2500
fat lambs — irrigated pasture 400
mature sheep — irrigated pasture 1300
Cattle
dairy cows, dry 16 000
dairy cows, milking 25 000
beef cattle 16 000
calves 8000

Australian Council for Educational Research 47

 S O T
R α θ
r C r
A B
R
83
s d
x Q x
P D h

Figure 1 Cross-sectional diagram of a partly full dam

Specify the mathematical problems Model for dam volume

The dam (Figure 1) has been excavated in the shape of an The aim is to create a formula that enables the volume to be
inverted spherical cap AMB – so that its cross-section is circular estimated in terms of the arc length AP(s) which we can measure
whatever the depth of water, and stock can drink from any point by stepping to the edge of the water, from the high water mark
on its perimeter. The diameter of the dam when full (AB) is 30 when the dam is full. (The distance ‘s’ needs to be related to the
metres, and its depth when full is 2 metres. The figure shows depth of water ‘h’ which cannot be measured directly.)
cross-sections of the dam, and of the hemisphere SMT of which
The volume of the cap of a sphere of depth ‘h’ is given by the
it is a part. PMD is a representative section of the partly filled dam
formula V = πh2(3R − h)/3, where R is the radius of the parent
of radius ‘x’ and depth ‘h’.
sphere. (This result can be provided after students have identified
a need for it, or can otherwise be obtained as an application of
Problem a: Model for dam volume
integral calculus.)
Find a way of estimating the volume of water in the dam when
the (unknown) depth of water is ‘h’. From Figure 1 we have from triangle OAC:
R2 = r2 + (R − d)2 so R = (r2 + d2)/2d
Problem b: Model for water loss over time
r = 15 and d = 2 gives R = 57.25.
Suppose the dam is in central Victoria. If the dam is full and no more
rain falls, how many days it would take for the dam to dry up? Then V (full dam) = πd2(3R − d)/3 = 710m3 (approx.) = 710 000 litres.
Now suppose that a flock of 300 mature sheep on dry pasture Similarly, from triangle OPQ:
is sustained by the dam. If the dam is full and no more rain falls,
R2 = x2 + (R − h)2 so that x2 = 2Rh − h2
how long will it be before other arrangements need to be made
for the stock? This links the radius of the water surface and depth, when the
dam is partly filled, that is, for various values of ‘h’.
Formulate the mathematical model (a) for
So we know how to calculate V when ‘h’ is known from
dam volume V = πh2 (3R − h)/3, but we need the volume in terms of a section
Discussion with students can be used to identify variables and of arc length ‘s’ that we can actually measure.
measurements that will be needed to address the problem. Only
data that can be accessed are useful, and the depth of water Hence we need to express ‘h’ in terms of ‘s’.
is not one of these. One readily obtainable piece of data is the
From Figure 1, we get h = OM − OQ = R − R cos θ.
distance the water surface has receded from its position when
Hence h = R (1 − cos θ) … (1)
the dam is full — a distance obtained simply by stepping out
Now θ = (θ +α) − α where θ +α = sin−1(r/R) ≈ 0.265 and
the distance taken to walk from the edge of the dam, directly to
α = s/R = s/57.25.
the current water level. A key insight is the recognition that this
So θ = 0.265 − s/57.25 … (2)
distance (s) is both measurable, and a key input to the model. Hence from (1) and (2) we have
h = 57.25[1 − cos(0.265 − s/57.25)].
Assumptions Check: When s = 0, h = 2 (rounded from 1.99)
■ Water in the dam is reduced by evaporation and in providing
for livestock. Seepage loss is negligible and can be ignored. To substitute this expression into the volume formula to obtain V
as a function of ‘s’ is routine but challenging and becomes messy:
■ Dams whose surfaces are bounded by curved perimeters
V = πR3(2 + cos(sin−1(r/R) − s/R))( 1 − cos(sin−1(r/R) − s/R))2.
and with access for animals can be represented by circular
or elliptical approximations. It is simpler to proceed in two stages: first calculate values of ‘h’
■ The dimensions of the dam when empty (as excavated) from chosen values of ‘s’.
are known. Then obtain corresponding volumes using V = πh2 (3R − h)/3.

www.immchallenge.org.au

Text by Peter Galbraith for IM²C and ACER. Available under Creative Commons non-commercial licence CC BY-NC 4.0.

Australian Council for Educational Research 48

 s (m) h (m) V (m3) V/capacity (%)
Formulate the mathematical model (b) for water
loss over time
0 2.00 710 100.00
To estimate how long water in a dam will last, we need to make
1 1.74 542 76.36 further assumptions – about evaporation.

2 1.51 406 57.14 Assumptions

Internet research will typically show that the evaporation rate varies
3 1.29 297 41.77
with temperature, wind speed, sunshine, and relative humidity.
4 1.09 211 29.72
It also varies throughout the year, but a rough daily evaporation rate
5 0.90 145 20.46 (average) can be found in (cm or mm) by dividing the average annual
value by 365 days. This will be sufficient accuracy for estimation
6 0.73 96 13.55 purposes, although in practice there will be seasonal variations.

8 0.45 36 5.08 So estimating the value for central Victoria, from data given
in the problem description, we obtain the average amount of
10 0.23 9.8 1.38 evaporation per day = 160/365 ≈ 0.44 (cm).
15 0.00 0.00 0.00 This is the ‘depth’ of water that is lost across any exposed surface
area in a day. The volume lost will vary with the surface area.
Table 2 Dam volume and depth against distance dam water has
receded from high water mark Calculations will overestimate the number of days that suitable
water is available to animals. Near the end the dam will resemble
a bog and the water undrinkable.

y
700 Solve the mathematics (b) for water loss over time
600
The daily amount lost by evaporation will vary with the area of
water surface, but by using the average value we can estimate
500 how long the water should last without further rain.
400
From the data provided for this problem, for central Victoria
300 we assume that the average amount of evaporation per day is
200 approximately 160/365 ≈ 0.44(cm). This is the ‘depth’ of water
that is lost across any exposed surface area in a day. The volume
100
x
lost will vary with the surface area.
0
1 2 3 4 5 6 7 8 9 From Figure 1, the cross-sectional area when the depth of the dam
is ‘h’ is a circle with radius ‘x’ and area πx2 where x2 = 2Rh − h2.
Figure 2 Graph of volume against ‘s’ for circular dam
Hence A(h) = π(2Rh − h2).

Solve the mathematics (a) for dam volume Note when h = 0, A = 0 and when h = 2,
A = 706.85 − the value of π (15)2.
(Note that ‘s’ takes values from 0 to the length of arc (AM)
= R(θ +α) ≈ 0.265 x 57.25 or about 15m.) Mean value of the cross-sectional surface area averaged over the
interval h = 0 to h = 2 is given by
Example: s = 1 gives h = 1.745 and thence V ≈ 542 (m3). This is 2
approximately 76% of the dam capacity of 710 m3. So when the
distance to the water’s edge has dropped by only 1 metre (approx
–
1
2−0
0
ʃ π (2Rh − h ) dh = –
2 2π
3
(3R − 2) ≈ 355.5, since R = 57.5.

6.67%), the volume has been reduced by almost one-quarter.

If the dam is full and no more rain falls, we can estimate how
Table 2 shows the results of calculating (using a spreadsheet) many days it would take for the dam to dry up:
these quantities for values of ‘s’ at (initially) one-metre intervals. 710/1.56 ≈ 455 (approximately 65 weeks or 15 months worth of
Notice that when the waterline has receded only three metres water supply).
from the top of the dam, the volume of water is much less than
50% of the capacity. If the dam must sustain a herd of 300 sheep on dry pasture, then
daily consumption of water by 300 sheep (from the example
To interpret the results, Figure 2 contains the graph of data) is 2500 × 300/365 ≈ 2055 litres ≈ 2.055m3.
V = πR3(2 + cos(sin-1(r/R) − s/R))( 1 − cos(sin-1(r/R) − s/R))2
Total average water loss per day from (consumption +
generated by using graphing technology. This could be used, for
evaporation) ≈ 3.62 m3.
example, as a chart pinned on the kitchen wall from which the
volume of water remaining can be read, given only the distance (s) Number of days of water available ≈ 710/3.62 ≈ 196 days (about
to the water line from the high water mark when the dam is full. 28 weeks or just over 6 months).

www.immchallenge.org.au

Text by Peter Galbraith for IM²C and ACER. Available under Creative Commons non-commercial licence CC BY-NC 4.0.

Australian Council for Educational Research 49

Interpret the solution Evaluate the model
Mathematical outcomes have been continually linked to the dam Apart from the checking of working for possible errors
structure, its volume and dimensions, and practical implications in mathematical calculations and/or in the application of
– as for example the meaning of the graph in Figure 2. This technology, evaluation involves continuous checking against the
is typical of problems involving a variety of mathematical needs of the problem context.
calculations. Their meaning within the problem needs to be
interpreted and assessed as they arise, for testing numerical Has the solution provided a sufficiently good answer to the
outcomes against the real situation will often identify errors in problem posed, or do we need further work? Why?
calculation that need to be addressed.
Sometimes when the answer to the first question is ‘yes’ the
The formula obtained translates the stepped out distance (s) into first answer obtained suggests a deeper exploration that only
a corresponding value for volume that gives estimates of volume becomes obvious from the initial modelling effort. This then
for any measured value of s. This would provide the basis for stimulates another cycle of modelling with an amended purpose.
constructing a ready reckoner or wall chart if desired.
A different perspective on evaluation was reported by a teacher
The data provided assumed a consistent evaporation rate who used a version of this problem with her year 10 class. An
based on an annual average. But the scenario posed — drought appreciative parent who happened to be a farmer told her that
conditions — could be considered to be different from the stepping down the bank was the method he and others used in
average. How will the outcome vary, if different values are estimating the amount of water in a dam.
considered for the evaporation rate?

The data provided for livestock water requirements are annual Report the solution
averages. In a real-world scenario, are animals likely to need more The modelling report could contain all the above components
water in drought conditions than in average weather? How will of the solution of the problem. It should summarise and
the outcome vary if different values are considered for the water illustrate how the mathematical insights obtained advanced an
consumption rate? understanding of the problem — even if this sometimes means
that the solution attempt in its present state is in need of further
For water loss over time, noting the observation about boggy
development. All assumptions and choices of data values should
conditions when the dam is nearly empty, the figure obtained will
be explained and justified.
overestimate the time drinkable water will be available. A safer
estimate would be about one month less. Perhaps? In this case the report should develop as a systematic and
cohesive narrative: considering the implications of drought
Students can contribute actively to these sorts of ideas, and
conditions for livestock farmers; providing tools for farmers to
resulting refinements, once they engage with properties of the
use to estimate water supply, such as the data shown in Table 2
real contextual setting.
and the graph in Figure 2; and recommending a timeframe
within which provisions should be made for alternative supplies
for stock.

www.immchallenge.org.au

Text by Peter Galbraith for IM²C and ACER. Available under Creative Commons non-commercial licence CC BY-NC 4.0.

Australian Council for Educational Research 50

Example problem
Level: Senior secondary
Senior modelling

Super Size Me

Disclaimer: This problem is about lifestyle

choices. It explores the impact of energy
intake (food selection) and exercise on
bodyweight, and the particular source of
food products is irrelevant.

Describe the real-world problem

Australian obesity a big problem

29 May, 2014. Worldwide adult obesity rates have jumped nearly 30 per cent in the last 30 years, according
to new analysis from the Global Burden of Disease Study 2013, published in medical journal The Lancet.
Adult obesity rates in Australia are climbing faster than anywhere else in the world, and are only slightly
less than those of the United States. In Australia, 63 per cent of adults are overweight or obese. In response
to the study, health experts have called on the governments in Australia to commit to a national strategy
to address overweight and obesity.
Commenting on the implications of the study, Professor Klim McPherson from Oxford University in the UK
stated that an appropriate global response to the worldwide obesity rates would focus on ‘curtailing many
aspects of production and marketing for food industries’.
This kind of public concern is not new. In America in 2004, concern over a perceived link between obesity and
food industries led to the making of the documentary film Super Size Me (2004). For 30 consecutive days,
the film’s creator Morgan Spurlock Morgan ate three meals daily consisting of nothing but McDonald’s food
and beverages (consuming approximately 5000 calories daily). If offered an ‘upsize’ he always took it, and
limited his daily exercise to that of the average American office worker. He ate everything on the McDonald’s
menu at least once. Spurlock, a 32-year-old male who was 188 cm tall and weighed 84.1 kg at the start of the
experiment, documented as his weight increased by 11.1 kg. There were other damaging effects to his body.

Specify the mathematical problem Formulate the mathematical model

Develop and evaluate a mathematical model to describe the
weight gain experienced by Morgan Spurlock. Use the model to Data
explore the respective effects of calorie intake and exercise. You ■ The unit of energy is the kilocalorie (Calorie) where 1 Calorie
may find useful information on energy and activity levels in the = 4.18 kilojoules.
appendix at the end of this question.
■ 7700 kcal ~1kg (biological data).

■ The average energy intake per day, I, is 5000 Calories (from

film description).

Australian Council for Educational Research 51

Assumptions Calculator
■ Energy processing is continuous, with no delays in w1 = 0.4904 + 0.9984 x 84.1 = 84.66
converting Calories into body functioning, and excess w2 = 0.4904 + 0.9984 x 84.46 = 84.82
Calories to additional weight. That is, individual daily effects and so on.
are averaged out over periods of time.
But this is tedious.
■ Weight change (per day) is determined by energy intake
(per day) less energy used (per day). Spreadsheet
■ Energy used per day = the rate at which a body burns energy, See table below.
that is, the basal metabolic rate (BMR) + energy used in
Row 3 contains the initial values (day zero).
activity (conduct web search to collect data, or see the
appendix to this resource). Formula for BMR is in column E. For example, E4 =
(10*B4+6.25*$D$3-5*$C$3 + 5)
Basal metabolic rate
Web search identifies the Mifflin formula, circa 1990, specifying that: Formula for energy used in activity is in column G. For example,
BMR (Calories) = 10w + 6.25h − 5a + 5 for males, where w is G4 = ($F$3*E4)
weight in kilograms, h is height in centimetres and a is age in
Formula for weight is in column B. For example, B4 = (H3 − E3 −
years.(For females BMR = 10w + 6.25h − 5a − 161.)
G3)/7700 + B3
Since the experiment lasts 30 days, we assume that age and
height can be treated as constants. The film description specifies Geometric series
that a = 32, h = 188, and starting weight = 84.1 kg. w1 = a + bw0
Hence take BMR = 10 × weight + 1020
w2 = a + bw1 = a + b(a + bw0) = a(1 + b) + b2 w0
Energy used in activity w3 = a + bw2 = a + b(a + bw0) = a(1 + b + b2) + b3 w0 … leading to:
From the film description we assume a sedentary lifestyle over w30 = a(1 + b + b2 + … + b29) + b30w0 = a(1 − b30)/(1 − b) + b30w0
the course of the experiment, and from the information in the
appendix this implies an additional energy usage per day of about w30 = 94.53.
0.2 × BMR.
Hence, energy used per day is 1.2 BMR = 12w + 1224. Calculus
If we treat time as continuous then we can approach a solution
Converting energy to weight through calculus.
Weight today = weight yesterday + (Intake in calories yesterday −
Let weight at time t (days) after the start of the ‘diet’ be w (kg)
calories used yesterday) converted to weight.
and consider the change in weight from w to w + δw that occurs
Let wn be weight after day ‘n’ (on day 0 original weight between t and t + δt.
is w0 = 84.1 kg).
I = average daily intake of Calories (I = 5000 from film data). If the average daily energy intake is I, then during time interval
Let En be the energy used on day ‘n’ (En = 12w + 1224). from t to t + δt intake is ≈ Iδt (Calories).

w1 = w0 + (I − E0)/7700 Similarly, the energy used between t and t + δt ≈ 12w + 1224

w1 = w0 + (3776 − 12w0)/7700 = 3776/7700 + (1 − 12/7700)w0 = (Calories).
0.4904 + 0.9984 w0
Hence δw ≈ [I − (12w +1224)]) δt/7700 so that in the limit dw/dt =
Hence w1 = a + bw0 where a = 0.4904, and b = 0.9984. (3776 − 12w)/7800.
Similarly w2 = a + bw1 and so on… w30 = a + bw29. Thus,

Solve the mathematics # dw # dt

^ 3776 - 12w h
= 7700
There are a number of methods that could be used to solve the
mathematics.

A B C D E F G H
1 day weight age height BMR Activity factor Activity energy Energy Intake (I)
2   (kg) (yr) (cm) (kcal/day) (BMR multiplier) (kcal/day) (kcal/day)
3 0 84.10 32 188 1861 0.2 367 5000
4 1 84.46     1865   368 5000
5 2 copy — — copy — copy copy
33 30 94.51     1965   393 5000

www.immchallenge.org.au

Text by Peter Galbraith for IM²C and ACER. Available under Creative Commons non-commercial licence CC BY-NC 4.0.

Australian Council for Educational Research 52

We then have ln(3776 − 12w) = −0.00156t + c. Other similar calculations that represent lifestyle decisions drive
home the information that diet is the predominant problem, and
Hence (3776 − 12w) = Ae−0.00156t (where A = ec). exercise can only go so far if that is not addressed.
When t = 0, w = 84.1 so A = 2767 and Refinement of the model could be achieved through conducting
w = (3776 − 2767 e−0.00156t )/12 another scenario.
When t = 30, w = 94.63 Sixteen-year-old Sarah weighs a steady 55 kg and had an activity
level that is ‘lightly active’. She receives a skateboard for her
The latter approach would be unlikely to be within the scope of
birthday and now skateboards for an hour a day, adjusting her
Year 10 students, for whom a spreadsheet solution would be
food intake so that her weight remains at 55 kg. Recently some
accessible. In general, of course, if a problem can be approached
fast food outlets opened near her home, and after skateboarding
in more than one way then initial insights are confirmed or
Sarah now has a daily routine in which she eats a cheeseburger
challenged, and sometimes new insights obtained.
and drinks a 375 ml coke. The rest of her diet remains the same.
What is the effect of this habit on Sarah’s weight after two
months (60 days)?
Interpret the solution
Before leaping to interpretation it pays to reflect back on the (Alternatively, replace Sarah’s information with a personal choice
context of the modelling. Here a condition was that Morgan of weight, food and exercise).
consumed approximately 5000 Calories each day. So in reflecting
on our solution it would be reasonable to allow the average Further information about food content and exercise is provided
amount to vary say by 5 per cent from this figure; that is, to carry in the appendix.
out calculations assuming average intakes of between 4800 and
5200 calories. What does considering a range of values tell us?
Report the solution
We find that for I = 4800, the predicted weight gain is 9.7 kg After investigating these problems, write a letter to yourself
(approximately), while for I = 5200 the predicted gain is 11.2 kg summarising the understanding you have obtained about the
(approximately). effects of food and physical activity on fitness. Include advice for
maintaining a healthy lifestyle.
Activity level is a more tenuous measure: allowing a 20 per cent
tolerance above (0.24) and below (0.16) the value used, with I = The letter should summarise briefly the main findings from the
5000 gives a range of weight after 30 days of between 94.2 and two problems, based on Morgan and Sarah. It should highlight
94.8 kg. insights obtained about the impacts of exercise and diet, and
perhaps contain self-directed advice for a healthy lifestyle.
It is reasonable to say that the model, built on the basis of daily
energy intake (food) and daily energy use (BMR + additional
activity) tracks the increase in weight in a logical and coherent
way. Given the assumptions it is this property, rather than precise Appendix
numerical values, that are important for evaluation purposes. Basal metabolic rate (BMR) is the rate at which energy is used
(The calculated weight gain for the standard conditions is about 6 when the body is at complete rest. Use an internet search to find
per cent lower that the reported figure.) a way of estimating BMR (Mifflin formula).

To find the total energy used it is necessary to estimate the

contribution of additional physical activity. There are various ways
Evaluate the model
to do this but a useful one is to express it as a multiple of BMR,
We should not think of this modelling exercise primarily as an which varies with lifestyle.
attempt to predict precisely the weight gain of Morgan Spurlock
over 30 days. Its purpose, rather, is to develop a model that is Examples might include:
applicable in estimating the effects of energy intake and energy
use on body weight. The film provides specific data against which ■ sedentary, little or no exercise, desk job = BMR × 1.2
to test the model.
■ lightly active, exercise or sports one to three days per week =
Given that it appears sensitive to the ‘right’ data, and gives results BMR × 1.375
that are consistent with the outcomes reported in the film, we
■ moderately active, exercise or sports six days per week =
now consider what further insights might be obtained. We can
BMR × 1
consider some hypothetical circumstances that give insight into
the respective impact of food intake and exercise. (This represents ■ very active, hard exercise every day, or exercising twice per
a move from the evaluation phase back into modelling activity.) day = BMR × 1.725

Suppose we hold energy intake at 5000 and double (increase ■ extra active, hard exercise two or more times per day, or
by 100 per cent) the activity level from 0.2 BMR to 0.4 BMR, training for marathon or other intensive event = BMR × 1.9
assuming that such a diet is compatible with such good
intentions. Then we obtain a weight after 30 days of 93.1kg When more energy (calories) are taken in by way of food than are
compared with 94. 5. A modest reduction for hard work! needed by the body, the excess calories are converted to extra
kilograms of weight at a rate of 7700 Calories = 1 kg.
Suppose instead we reduce the daily Calorie intake to 4000
(20 per cent reduction) and the original sedentary lifestyle is
maintained. Now after 30 days we obtain a weight value of 90.75
kg, almost 2.5 kg less than the previous value.

Australian Council for Educational Research 53

 Energy used Sample fast food items Energy (kcal)
Activity
(kcal/kg/h)
Plain hamburger: ground beef patty,
Sitting quietly 0.4 grilled (40 g), hamburger bun (65 g) and 328
salad
Writing 0.4
Plain hamburger and medium chips:
Standing relaxed 0.5
ground beef patty, grilled (40 g), 497
Driving a car 0.9 hamburger bun (65 g) and salad

Vacuuming 2.7 Cheeseburger: ground beef patty, grilled

(40 g), hamburger bun (65 g), lettuce,
Walking rapidly 3.4 391
sliced tomato and onion, tomato sauce,
Skateboarding 5.0 cheese (16 g)

Running 7.0 Cheeseburger and large chips: ground

beef patty, grilled (40 g), hamburger bun 728
Tennis 7.0 (65 g) and salad, large portion of chips
Swimming 7.9 Double hamburger and medium chips:
2 ground beef patties, grilled (80 g),
606
hamburger bun (65 g) and salad,
medium portion chips
Pizza with cheese, tomato and olives:
223
medium portion (90 g)
Pizza with cheese, tomato and olives:
844
large portion (340 g)
Lemonade (375 ml) 158
Dry ginger (375 ml) 124
Bundaberg ginger beer (375 ml) 188
Coca Cola (375 ml) 158
Pepsi cola (375 ml) 169
Fanta (375 ml) 210

www.immchallenge.org.au

Text by Peter Galbraith for IM²C and ACER. Available under Creative Commons non-commercial licence CC BY-NC 4.0.

Australian Council for Educational Research 54

Example problem
Level: Senior secondary
Senior modelling

Temporary traffic lights

Describe the real-world problem

Lane closure between Newtown and Highbury

1 February, 2015. Traffic restrictions will apply on the Kings Highway from
3 March, to allow for the construction of a new eastbound overtaking lane
approximately one kilometre east of Highbury. To facilitate this work, the existing
eastbound traffic lane will be temporarily closed for approximately three weeks.
For safety reasons, traffic will be restricted to one lane, on a 600-metre section
of the Kings Highway, through the work site where 55- and 25-kilometre-per-
hour speed limits will be in place. Temporary traffic signals will be in use to
manage traffic flow and delays can be expected.

Specify the mathematical problem ■ Last car through on a green light at A will take (d/V + safety
margin(s)) sec to clear the road, before the light at B turns
Two-way traffic along a 600m single lane section of roadway is
green
controlled by temporary traffic lights at each end. A speed limit
of 55 km/h applies. How should the timing of the lights be set to ■ Let T (same for both) be the green light time at A and B.
achieve an efficient flow of traffic in both directions?

Formulate the mathematical model

We first need to understand how the light cycle operates.
A B
d
■ Traffic lights are at A and B, where d = 600, and V (max
speed) = 55km/h.

■ Assume that the road is a main highway – no pushbikes or Complete cycle of light changes at A (same for B)
farm machinery allowed.
Time 0 T T + d/V + s 2T + d/V + s 2T + 2d/V +2s
■ Assume that the traffic is similar in both directions and
sufficiently dense to build up while the light is red. Light at A G R R R G

Light at B R R G R R

Australian Council for Educational Research 55

Solve the mathematics For cars to just clear during a cycle of light change we need n = N.
The length of cycle is 2(T + d/V + s) Thus 15T/(6 + 15a) = 2(T + 45)/b
so, the proportion of green light time = T/2(T + d/V + 2s). So T = 90(2 + 5a)/(5b − 10a − 4)
Using V = 15 m/s (54km/r) and and a safety margin of 5s gives We have noted above a reasonable (minimum) estimate for
proportion of green light time = T/2(T+45). b: b ≈ 2.
For example: T = 60 gives a value of 0.29 (29%) Drivers are ‘rearing to go’ when vehicle in front moves so estimate
T= 180 gives a value of 0.4 (40%) a ≈ 0.5.

This gives T ≈ 405.

Interpret the solution
Does this mean that longer green light times (at both ends) are
more efficient? Evaluate the model
We don’t know. Nobody is going to sit comfortably for 6 to 7 minutes in a queue
with no sign of activity; so, review the outcome.
What are the implications for traffic flow?
Our model so far is not wrong, but we evaluate it as inadequate Do we need a new model or to amend some aspects of the
to answer this question, as yet we have no basis for assigning approach?
values to T.
Suppose T = 240 (4 min): n = 267 and N = 285 (18 vehicles wait a turn)
How do we obtain estimates of the respective numbers of T = 180 (3 min): n = 200 and N = 225 (25 vehicles wait a turn)
vehicles arriving at the lights, and leaving while they are green? T = 60 (1 min): n = 67 and N = 105 (38 vehicles wait a turn)

First, consider vehicles arriving at A during a cycle of light But now the percentage missing out is approaching 40 per
changes. cent, and the build-up will be rapid. So balance driver frustration
with volume of traffic movement. How many light changes are
Defensive driving advice says to leave 2s between vehicles tolerable on a busy road: Two? Three? Four?
passing a roadside marker.
Conduct a retrospective estimate of omitting acceleration phase:
On a busy approach road this gives an estimate of the rate at If T = 180, the nth vehicle (number 200) will have travelled
which vehicles arrive, and typically join a waiting queue. Keeping 200 × 6 = 1200 metres when it just makes the light, and will have
things general, we denote this parameter by b. been travelling for about 80 s. The initial phase is a very small
fraction of this.
So, the number of vehicles arriving during a cycle of length
2(T + d/V + s) is given by Revisit parameter values: l, s, a, b.
N = 2(T + d/V + s)/b
that is N = 2(T + 45)/b Note from the formula for T that we must have: 5b − 10a > 4.

A model for vehicles leaving A during green light time is shown Consider different times of day: the effect of b.
below.
Consider the influence of heavy truck presence.

Investigate the effect of including an acceleration phase, for

A example:
n 3 2 1 T = 240: n = 263 versus 267 (change 1.4%)
T = 180: n = 197 versus 200 (change 1.5%)
x T = 60: n = 63 versus 67 (change 5%)

Next, consider vehicles waiting for green light.

Average vehicle length (including space to vehicle in front) = l. Report the solution
Vehicle n is last through the light when it turns green. The modelling report should contain all the above components
a is the average delay time between successive vehicles of the modelling problem and its solution. The report should
taking off. synthesise this data into a cohesive narrative, considering the
So available time in motion for nth vehicle to reach lights = T − an. implications for safety and driver frustration. The report could
recommend a particular option that best balances the factors of
Assume that we can ignore short period of acceleration to wait time and traffic build up.
speed V.
Then ln = V (T − an) and so n = VT/(l + aV) = 15T/(6 + 15a).

www.immchallenge.org.au

Text by Peter Galbraith for IM²C and ACER. Available under Creative Commons non-commercial licence CC BY-NC 4.0.

Australian Council for Educational Research 56

Teamwork
and report writing

Substantial mathematical modelling projects (such as IM2C both of the correctness of the mathematics and of its relevance to
projects) are best done in groups. However individual group the aspects of the modelling task for which it has been proposed.
members need to be competent in all stages of the modelling
Teams of students can collaborate by bringing complementary
process, with expertise developed through prior experience with
perspectives to paired problem-solving (one an initiator and the
a range of modelling problems. In terms of the IM2C, we suggest
other a checker).
that teams of four are valuable in view of the number of tasks that
have to be covered in producing answers and a report. However, The evaluation stage in the modelling process is crucial: strong
this and later advice, is relevant to consider for all modelling opinion, if not underpinned or resisted appropriately with
contexts, whether IM2C related or not. mathematical argument, can derail the best of intentions.
It can be worthwhile assigning different members of a team to A team will have more avenues to pursue if its members know
different tasks – again this depends on the expertise of members. more mathematics – but advanced mathematics alone will
There are the ‘mathematicians’ who go from the assumptions not produce a successful report. The team should also include
to the mathematics; the ‘project managers’ who look after and members who know what is required of the modelling process,
keep track of overall progress and make sure that a summary and who realise the importance of mathematics in industry,
or report is written, that the assumptions are reasonable and science, and life generally.
relevant, that the mathematics is correct, and that the conclusions
follow from the mathematics and answer the posed problem.
Different members of a team may do more than one task. But it Using technology
is also important that role assignment remains flexible, as often It will be helpful for some members of a team to have an
members of a team can contribute at times to areas that may not awareness and facility with technology. The ability to write
be their prime responsibility, and this needs to be provided for. computer programs is also valuable.
Report writing is fundamental to modelling. For teams in the early Technological tools do not simply amplify cognitive processes
stages of modelling, the report may be as simple as a series of – they can fundamentally change the nature of a task and the
bullet points that lead to the problem solution. As teams mature, requirements to complete it.
they should be encouraged to write a full summary that will state
the problem, indicate all assumptions, give some idea of the Graphics calculators or software such as Matlab, Maple and
maths used and provide a conclusion. Gradually students should Mathematica lead to less tedious and more efficient execution of
work up to a full report with details of all the mathematics and all calculus and other types of problems, and generate tables, graphs
the sources that were used to complete the problem. and so on.

To help students in their report writing, we suggest that regular Mathematical expertise should not be considered simply as the
classwork should include pertinent explanations as well as accumulation of internal mental processes and structures, but
symbols and equations. It will be easier to see how students reason also as a process of appropriating tools that transform tasks, and
and their general understanding and fluency will be enhanced. the relationship of individuals to them.
The most powerful use of technology for problem-solving is in
Selecting a team
enabling individuals, including students, to explore ideas and
Successful teamwork in mathematical modelling depends both tackle solution paths that would otherwise be beyond them.
on team composition and on the quality of collaboration between
For the IM2C, each team should plan to have access to at least
team members.
two computers loaded with software such as a computer algebra
Modelling is a cyclic process involving different tasks in different system and spreadsheet program, and team members who are
stages (see the mathematical modelling framework, p. 6–7). proficient in these tools.

It is important that all team members have knowledge that

Collaborative competence
enables them to contribute (perhaps in different ways) to the
substance of the mathematical modelling involved. It is possible to consider teamwork from a Vygotskian notion of a
‘zone of proximal development’ (ZPD).
For example, if one or two individuals provide and argue for a
particular mathematical approach, others need to supply critique, The zone of proximal development is the gap between what
a student can achieve working alone and when aided.

Australian Council for Educational Research 57

The theory posits that education is most effective when it is ■■ Times are nearly always underestimated, so set an early
neither too easy nor too hard; when tasks challenge an individual deadline for the first tasks.
to stretch past existing competence while support or scaffolding
■■ Plan sufficient time for report writing, which is likely to take at
the learner to advance.
least a day.
The Vygotskian notion of a zone of proximal development (ZPD),
■■ Meet or video conference once every day to discuss progress
through which individual learning is scaffolded by utilising
and next steps.
differential levels of expertise existing between expert and novice
participants, has been extended to apply within collaborative
group activity (see e.g., Goos, Galbraith and Renshaw).10 Unlike Documentation
the traditional scaffolding notion, this view of the ZPD involves ■■ Maintain a record of all sources, websites and pieces
egalitarian relationships. It refers to activity in peer groups where of mathematics. Be careful and systematic, and update
each member brings some relevant knowledge and/or skills, continually, so information can be easily retrieved if required
but requires others’ contributions to build on or refine these in the final report.
attributes for effective group progress to be made. Such situations
■■ Back up documentation and drafts frequently.
approximate practices within mathematical teams that are striving
to go beyond the established boundaries of their knowledge, and ■■ Don’t throw anything away, even apparently redundant copies
so provide authentic experiences of doing mathematics under of electronic files – you never know when something might
conditions of uncertainty, as modelling indeed sets out to do. come in handy. It is always possible that your latest good
idea may not be as good as something that you might have
Importantly within such teams, when the direct influence of the
discarded earlier.
teacher is removed, students must take personal responsibility
for the ideas that they are constructing and testing, so that
authorship of mathematical knowledge is then vested in Order of tasks
individuals and their partners. ■■ To begin, the team should come to a mutual understanding of
If recognition of peer contribution to a common task doesn’t meaning of the first mathematical question to be addressed.
happen, the team cannot function properly. It therefore follows that ■■ Decide and allocate initial internet searches to be undertaken,
social factors involving relationships and respect between group but allow time for additional searches as you get deeper into
members are also factors in constructing effective teams. the problem and the report.
■■ Consider relevant technology. Search online for specific
Timelines and tasks
programs, and for data to support that provided in the
In preparing for tasks such as those provided by the IM2C, and problem statement.
modelling projects in general, a routine should be rehearsed in
■■ Assign sub-tasks that can be carried out in parallel by
practice situations, using models obtained from other sources well
different team members, and ensure all individual work is
before the challenge problem itself is undertaken.
brought together for critique and integration.
Team members should understand the routine so that it can be
■■ Midway through the challenge period, the team should begin
put into action automatically over the days of the challenge.
to collate members’ work, draft a summary and outline the
maths used.
Timelines
■■ Begin to consider the advice the team will give in response to
■■ Plan an overall approach and timeline. Make it flexible the challenge question being posed.
within limits.
■■ Everyone in a team should be responsible for critically
■■ Take note of the total time available for the project. assessing what is written as it is written.
■■ Schedule meeting times to share and review progress. It ■■ Errors and approaches need to be discussed. Were the
may not be feasible to schedule too far in advance: students assumptions that were made useful or too strong? Were
typically schedule meetings at unusual times outside school all the assumptions used? Were there implicit assumptions
hours to suit individual availability. that should be stated explicitly? Could the problem be
■■ Team members should work together for at least the first and approached in a different way?
final days of the challenge period, because of the importance of Are more data needed?
the initial planning and the final report writing. ■■ Collect all references together as a list and think about the
■■ Agree on a time by which all raw modelling activity needs to need for one or more appendices.
be completed.

10
Goos, M., Galbraith, P., & Renshaw, P. (2002). Socially mediated metacognition: Creating collaborative zones of proximal development in
small group problem solving. Education Studies in Mathematics, 49(2), 193–223.

Australian Council for Educational Research 58

■■ Identify what work still needs to be done to answer the ■■ Showing and justifying all mathematical working.
challenge problem.
■■ Setting out all mathematical working, graphs, tables, etc.
■■ At least one full day should be devoted to report writing.
■■ Interpreting mathematical results in terms of the
real-world problem.
Teacher/mentor assistance
■■ Evaluating the result – does your answer make sense? Does it
The role of the teacher or mentor is to guide students to become help to answer the problem?
proficient modellers.
■■ Dealing with refinements to the original problem.
The teacher should scaffold of task implementation, without
providing specific advice on the problem. ■■ Qualifying the solution.

Teachers might direct students to the framework for modelling, or ■■ Recommending the solutions arising from the work.
give suggestions on time management. What further work is needed?

For formative modelling tasks outside the IM2C, a teacher can In practice, several of these activities can occur at the same time.
exercise discretion in deciding how direct hints should be. For example, obtaining a mathematical result, interpreting it, and
evaluating its correctness or relevance,
The aim is to help students overcome blockages to are aspects that are often dealt with together.
progress, but not compromise critical decision-making
and learning processes. Evaluating the solution(s) in terms of the problem requirements
is crucial. It is the final crescendo in this orchestral work. It is here
that reporting to the client is to be found. In the heart of a report
Writing a modelling report
a team may use all the complicated mathematics and modelling
Models are designed to address particular problems in specific that they know or that they can find out, but it is necessary to
situations. So, models must be capable of being evaluated and communicate the results to the client. If this is not done clearly the
used by others, including non-mathematicians. whole point of the exercise is lost.
Information needs to be communicated clearly and fully.
Report summary checklist
While they should always contain information that gives a complete
picture of what the modelling has achieved – and not achieved! Writing reports is not easy. It takes a great deal of practice to
produce reports of value. Students should be guided to practice
The IM2C requires that a team report be produced in which the writing reports before the challenge.
contributions of individual members are merged. For modelling The mathematics and the modelling should not, in themselves,
undertaken within other contexts (e.g. school based activity) other be the focus of the report. The focus of the report should be a
reporting styles can be employed. For example, when assessment solution to the real-world problem, and the mathematics and the
of individual expertise is required, an effective approach has modelling are the means to this end.
been for students to work collaboratively, but be responsible for
providing an individual report. The checklist below is relevant to any The 2016 IM2C required participants to submit a one-page
modelling project. summary sheet and a solution of up to 20 pages. A well-written
summary is vital. Consider the summary as a direct statement to
While reports on mathematical models vary in style and detail, they the people who needed to solve the real-world problem. Be aware
should always contain information that gives a complete picture of that the end users’ mathematical ability may not be high, so the
what the modelling has achieved – and not achieved! practical suggestions may need to be in everyday language.

Report writing checklist The summary should:

A checklist for a good report may include: ■■ state the problem

■■ Describing the real-world problem being addressed. ■■ state the assumptions made

■■ Specifying the resulting mathematical questions precisely. ■■ give a brief description of the mathematics used

■■ Listing all assumptions and their justification. ■■ provide practical suggestions to solve the
real-world problem.
■■ Indicating sources of imported information
(for example, websites). The summary might also discuss what might have been
done to develop the solution further on another occasion.
■■ Explaining how numerical values used in calculations were
decided on.

Australian Council for Educational Research 59

Modelling report
illustrating use of criteria

The example in this section illustrates a modelling report from compiling a report and justifying results or recommendations). It
an Australian competition (not related to the International differed in that the students selected their own problem context
Mathematics Modeling Challenge (IM²C)). The purpose of this rather than having it prescribed. Students also had substantially
example is to illustrate the general applicability of the criteria used less time (1.5 days plus overnight if students chose to use that
for the Australian component of the IM²C. opportunity) for completion and report construction. This meant,
for example, that only one mathematically based question could
The AB Paterson Gold Coast Modelling Challenge was devised and
be generally addressed, and that evaluation of outcomes could
hosted annually by Trevor Redmond for a period of approximately
be foreshadowed and discussed but not always pursued, and
10 years. Its senior section involved Year 10 and Year 11 students
the outcomes could not always be fed into a second round of
from secondary schools, almost all new to modelling, working
modelling.
in mixed-school teams during a two-day modelling challenge.
Student teams (most commonly comprising four members) were What follows is an example from the Gold Coast Modelling
assigned from different participating schools. The challenge Challenge, annotated with mentor comments. This single example
commenced with an introduction to modelling by mentors who illustrates, in a targeted way, the attributes that are important
worked with five or six teams in a class setting for approximately in developing and reporting on any modelling problem. Thus, it
two hours on a common modelling task chosen by the mentor. contains features that are consistent with, but less extensive than,
The purpose here was to provide a common understanding of those that are expected within a complex context such as the
what is involved in modelling and reporting. The introduction IM²C. The performance of the team is discussed in terms of the
class covered the cyclic modelling process by applying it to the criteria used for judging within the 2016 Australian component of
problem chosen for this purpose. Following this introduction, each the IM²C, which were not known to student teams involved in the
student team chose their own real situation to investigate, devised IM²C, nor of course the students here.
a mathematical problem to address, and worked collaboratively
This example can be viewed as providing a bridge to the more
until lunchtime on the second day to complete their modelling
extensive demands of the international challenge. Alternatively,
and construct a poster or report describing their work. Each team
for those not intending to go the full journey to the latter, it
gave an oral presentation within the class group in the post-lunch
provides a self-contained illustration essential components of the
session, followed by a public display of posters from all groups
development and reporting of a modelling problem.
later in the afternoon.
The material that follows is illustrative. In using selected excerpts
Mentors facilitated students’ pathways in modelling, rather than
from student work, it does not set out to demonstrate a perfect
specifying approaches or choosing situations or questions for
solution. The focus is on how the students systematically
groups to model. The mentors intervened as little as possible.
introduced and addressed essential phases of the modelling
Their role was to enable students to develop and evaluate their
process for a problem selected by them. It thereby demonstrates
own ideas in a productive manner. For example, mentors would
that the Australian criteria developed for grading modelling
answer questions asked of them by students, draw attention to
projects are workable.
where teams were in relation to their modelling progress, and
what this might mean going forward. They did not provide specific In the excerpts that follow ‘student’ is used to identify material
advice as to how to proceed mathematically towards a solution or included in the team report, while ‘mentor comment’ indicates
direct students down a particular path. feedback from a competition mentor or the authors. Note, as
indicated above, that the headings and associated text outline the
The Gold Coast Challenge has both similarities to and differences
major modelling steps.
from the IM²C. The Gold Coast Challenge was similar to the IM²C
in the enactment of a cyclic modelling process (e.g., making and
justifying assumptions; choosing the mathematical approach;
interpreting mathematical outcomes in terms of the real-world
context; evaluating progress in terms of problem requirements;

Australian Council for Educational Research 60

Australian judging criteria for the International Mathematics Modeling Challenge 2016

Criteria Sub-criteria
1 Problem definition ■■ specification of precise mathematical questions from the general problem statement

2 Model formulation ■■ identification of assumptions with justification

■■ choice of variables
■■ identification and gathering of relevant (needed) data
■■ choice and justification of parameter values
■■ development of mathematical representations

3 Mathematical processing ■■ application of relevant mathematics

■■ invocation and use of appropriate technology
■■ checking of mathematical outcomes for procedural accuracy
■■ interpretation of outcomes in terms of the problem situation

4 Model evaluation ■■ adequacy and relevance of findings in relation to problem situation

■■ further elaboration or refinement of problem
■■ relevance of revised solution(s) following revisiting and further work within earlier criteria
■■ quality of answers to specific questions posed in problem statement

5 Report quality ■■ summary page quality: succinctness, power to attract reader

■■ overall organisation of the report: logical presentation including
� description of the real-world problem being addressed
� specification of the mathematical questions
� listing of all assumptions wherever they are made
� indication of how numerical parameter values used in calculations were decided on
� setting out of all mathematical working: graphs, tables; technology output and so on
� interpreting the meaning of mathematical results in terms of the real world problem
� evaluating the solution(s) in terms of the problem requirements

Gold Coast inundated Students: Mathematical question

At what point in time will the Q1 (or a later building) lobby, which is
Problem definition 3m above sea level, be submerged due to sea level rises, and what
will be the mean maximum temperature of that month in Surfers
Students: Rationale for choice of modelling topic Paradise at that time?
Climate change is the term used to describe the changing nature (Note that this is slightly different in nature to the IM²C in that this
of the world’s weather patterns. Many meteorologists have stated problem, with the above wording, was determined by the students
that climate change could potentially result in detrimental impacts themselves.)
upon our current way of life, through rising global temperatures
and sea levels. The increasing temperatures due to climate Mentor comment
change have been reported to cause rising sea levels due to The students have identified an issue they deem significant, and
the expansion of water around the world due to heat (including the mathematical question is personally relevant given they live on
the melting of ice). The concept that the climate is changing is the Gold Coast and are familiar with the Q1 building. The question
strongly supported by evidence obtained by experts. is very specific given the tools available, as becomes apparent as
the solution attempt proceeds, and revisiting becomes necessary
as part of the cyclic modelling process.

Australian Council for Educational Research 61

Model formulation γ = 0.001χ + 24.64. This equation shows that the equilibrium line
of the periodic function for the mean maximum temperature over
Students: Assumptions time graph is sloping upwards.
It is assumed that the influence of rainfall and evaporation in the This linear equation can be substituted into the general periodic
ocean is negligible due to the fact that this would only contribute to equation instead of d.
the water cycle which would in turn feed back into the oceans. If this
The a value in a periodic equation is the amplitude of the wave;
is not the case, the process used to determine the increase in sea
essentially half the range of values for the γ axis.
level from the temperature increase would not be able to be used to
produce accurate results. The b value in a periodic function is the period of the function
being the average distance between two crests (or troughs) on
It is assumed that the whole surface of Surfers Paradise is a flat
a wave. Calculating the period as 12, the b value for the periodic
plane 3 metres above sea level that contains no obstructions
equation is 2π ÷ b = 0.5235.
to the path of the ocean as it rises. If this was not the case, the
calculations for the rise in sea level would be too simplistic to be To determine the value of c, the equation can temporarily exclude
valid, and there could be a significant difference between them the c value and be used to predict values for γ. Because these
and the actual sea level increase necessary to reach the lobby of values will have been translated incorrectly due to the lack of a c
the Q1. value, they can be matched up with corresponding γ values and
the χ distance between these selected corresponding points can
It is assumed that the trend demonstrated in the data set
be used to determine the value of c.
used continues into the future. If the trend does not continue
into the future, it would not be possible to accurately predict data One pair of crests and one pair of troughs have been identified
points. for analysis. The adjacent troughs have χ-values of 605 and
609, respectively – indicative of a horizontal phase shift of four
It is assumed that there is a correlation between temperature and
months. This indicates that predicted γ values have been shifted
rising sea levels; if there is no correlation, it would be illogical to
four months forward, and thus the c value is −4.
attempt to predict the change in sea level using the temperature of
Surfers Paradise. Thus the equation to predict the mean maximum temperature
over a month in Surfers Paradise becomes
It is assumed that the melting of the polar ice caps does not
contribute to the rise of the sea level and that rising sea levels γ ≈ 5.75 sin(0.5235(χ+4)) + (0.001χ + 24.64).
are entirely caused by expansion of water due to heat. If this
assumption proves invalid, the predicted point in time at which Actual mean maximum temperature (degrees) and predicted mean
the Q1 lobby would be at sea level would be too far in the future, maximum temperature (degrees) without a c value over the number of
as melting polar ice could increase the rate at which the sea level months since January 1938
rises; making the predicted time frame for the Q1 lobby to be at
sea level inaccurate.

Students: Defining the variables

The independent variable, χ, is the time in months, in one month Mean
intervals, since January 1938. Therefore: χ = 0 = January 1938; the maximum
temperature
first month that temperature records are available.
(degrees)
The dependent variable, γ, is the mean maximum temperature of
Surfers Paradise over one month measured in degrees Celsius.

Students: Building a model

The general form for a periodic function is γ = asin2π/b(χ − c) + d,
where a is the amplitude, b is the period, c is the phase shift, and Number of months since January 1983
d is the vertical translation of the function. Thus it would not be
possible to predict a change in the mean maximum temperature Actual mean maximum temperature (degrees)
of Surfers Paradise unless an equation representing the general
change in the climate was used in place of the constant d value. Predicated mean maximum temperature (degrees)

To obtain the equation for d, one must determine the average rate
at which the temperature increases. A linear regression performed
upon the data gave an equation of

Australian Council for Educational Research 62

Mentor comment a result that was incorrect by less than one tenth of a degree in
this particular case would indicate that it is quite accurate. Based
The assumptions are careful and relevant, foreshadowing
upon this, it would be reasonable to assume that the model
potential limitations as well as introducing necessary
could accurately predict the future trend of the mean maximum
simplifications to enable mathematical formulation.
temperature of Surfers Paradise.
Relevant data were identified as those available from official
Mentor comment
records kept from 1938.
The students first sought to test the accuracy of the periodic
The independent and dependent variables are appropriate –
function constructed to predict temperatures into the future.
other subsidiary variables and parameters were introduced as
An application of mathematics which preceded a direct attack
the modelling progressed. An annual periodic relationship was
on the problem question itself. This amounted to checking the
assumed between temperature and time of year based on real
quality of a tool before using it for its designed purpose – a wise
world seasonal cycles.
thing to do. The students identified an error in a hand calculation
Arguments supporting choice of parameters for the periodic that was ascribed to a rounding error – in fact a mentor check
function were presented carefully. Some innovative thinking showed that it was because the calculator mode was set to
is apparent – in particular, the replacement of ‘d’ by a linear degrees instead of radians! However, the team appropriately
expression to capture the slowly increasing base temperature. discarded it in favour of a computer-generated calculation.
There is some loose description that entangles the period with the The students used calculators and computers appropriately
coefficient, but the students have the mathematics right. for generating functions, displaying them graphically, and
undertaking messy computations that would have been
Their incorporation of the translation constant ‘c’ is clever, even impossible using real data without technology to assist.
though subject to error. They inferred a value of 4 empirically from
the graph of data – rather than a value of 3 (1/4 of a period). The The students recognised that output from a model of this type
numerical impact on outcomes is minimal. should be checked against real data, but missed that the checking
should be over a cycle rather than for a single point. If the point
The final criterion in this set (development of mathematical
happens to be where the cosine/sine has value 0, then that term
representations) merges with the first criterion of the following
will not be evaluated – this happened here.
group below (application of relevant mathematics). There is
interplay between these activities throughout the remainder of
Mathematical processing
the solution process, and it is natural that this should occur. The
important aspect is that both are identifiably present. Students: Solving the mathematical problem
Students: Verification of the periodic model According to Baltimore City College (internet source provided by
students) the volume of water will increase by 0.0088% when
This perception of the model being fairly accurate (e.g., graph),
heated 10 degrees. The Q1 lobby is only approximately 3 metres
can be verified by independently comparing a single predicted
above sea level, and it is unlikely that the world temperature
data point with a single actual data point. Below, the model is used
will rise by 10 degrees due to global warming, so this value
to generate the mean maximum temperature of Surfers Paradise
must be made smaller. To determine the expansion of water at
over the 80th month from January 1938.
temperatures lower than 10 degrees, a ratio must be used as
γ ≈ 5.75 sin(0.5235(χ + 4 )) + ( 0.001χ + 24.67) seen on the poster. Therefore, for each degree of temperature
γ ≈ 5 .75 sin(0.5235(80 + 4 )) + ( 0.001 × 80 + 24.67) increase, the world’s water will expand by 0.00088%. The volume
of the Earth’s water is approximately 1.3 billion cubic kilometres,
γ ≈ 5.75 sin(0.5235 × 84) + (0.08 + 24.67) and its surface area is 361 million square kilometres (reference
γ ≈ 5.75 sin(43.974) + 24.75 given by students). Because the volume of a rectangular prism is
known, and the world’s oceans have already been put in terms of
γ ≈ (5.75 × 0.6943) + 24.75 a rectangular prism, it is possible to calculate the increase in the
γ ≈ 3.992 + 24.75 height of the ocean by using the formula below (see continuation
of solution).
γ ≈ 28.74
Mentor comment
This calculation, however, produces a value of 24.702291 when
performed automatically by a calculator due to the fact that it has The Baltimore reference is one of many that can be used for this
not compounded errors due to rounding. The actual value of the purpose. The students did not list in among references which
mean maximum temperature in the 80th month from January was an omission on their part. In addressing the modelling
1938 was 24.8 degrees. Thus, the calculated value was off by question, the team developed further assumptions, for example
approximately 0.0977 degrees, while the value calculated by hand collecting the oceans into a cuboid to make the calculation of
with rounding errors using the model was off by 3.94 degrees. estimates tractable, and researching internet sources to provide
The fact that the model, when used with a computer, produced key parameter information on water expansion to feed into the

Australian Council for Educational Research 63

calculation. A clever first choice but a crude one. It enabled an A linear regression performed upon the results of the
order of magnitude estimate. It was suggested in later discussion mathematical model produced an equation of γ = 0.001χ + 24.64,
with the team that there were other ways that might be used where γ is the mean of the mean maximum temperatures, and
to approximate local water volumes to the same purpose more χ is the number of months since January 1938. Using this linear
appropriately. They saw the point but there was not time to pursue equation, the number of months since January 1938 at which the
the implications. sea level would rise by 3 metres can be calculated.

Students: Solving the mathematical problem (continued) The current number of months since January 1938 is

v = lwh ((2009 × 12) + 11) − (1938 × 12) = 24119 − 23256 = 863 months,
(temperature increase × 0.00088) × (1 300 000 000) = 361 000 000 and the most recent mean temperature is 25.275 degrees.
000h With a temperature increase of 0.946671 the temperature would
(temperature increase × 0.00088) × (1 300 000 000) ÷ 361 000 000 need to be
000 = h 25.275 + 0.946671 = 26.2217 degrees. Substituting that value into
0.003169 × temperature increase = h the equation the formula becomes
Mentor comment 26.2217 = 0.001χ + 24.64.

Here the students have collapsed some calculations and scaling Below, the equation is algebraically rearranged to solve for χ.
in the one procedure. They have obtained a correct outcome but 26.2217 = 0.001χ + 24.64
without adequate explanation. Their approach was as follows: 26.2217 - 24.64 = 0.001χ
expansion volume = temperature increase (degrees) × (expansion/ 1.5817/0.001= χ
degree) × present volume 1581.7 = χ

expansion volume = area of water surface × expansion height (h) Rounding χ up to 1582 months since January 1938, the number of
years since January 1938 is 131.833, or approximately 132 years.
So (temperature increase) × 0.00088 × 1 300 000 000 = 361 000 Adding that on to 1938, the result is that in the year 1938 + 132 =
000 × h (km) 2070, the Q1 lobby will be at sea level, with the mean of the mean
Leads to h = 0.003169 × temperature increase (metres) maximum temperatures of that year at 26.2217 degrees, and the
The mathematics is correct, but its communication leaves mean maximum temperature of that month at
something to be desired. γ = 5.75 sin (0.5235(1587.1+ 4)) + ((0.001 × 1581.7) + 24.67) =
24.19 degrees.
Students: Solving the mathematical problem (continued) This will occur in 2070 − 2009 = 61 years’ time.
Therefore, one can calculate the increase in sea level when given Mentor comment
the increase in temperature in degrees Celsius. Assuming a height
As noted above, in addressing the modelling question, the team
increase of at least 3 metres to reach the lobby of the Q1, the
developed further assumptions, for example collecting the oceans
equation becomes:
into a cuboid to make the calculation of estimates tractable, and
0.003169 × temperature increase = 0.003 researching internet sources to provide key parameter information
temperature increase = 0.003 ÷ 0.003169 on water expansion to feed into the calculation.
temperature increase = 0.946671
Using the formula previously generated, predictions were
Therefore, the temperature increase, necessary to create a calculated and interpreted. The interpretation of the outcome
rise in sea level of 0.003 kilometres, or 3 metres, is 0.946671 triggers the next part of the analysis.
degrees Celsius. Of course, for this change to actually take effect
the change in temperature must be substantial and sustained;
it should be the mean of the mean maximum temperatures
generated by the model. Due to the periodic nature of the model,
the equation should have a linear regression performed upon it
to determine an equation for its equilibrium line, thereby allowing
one to determine the point in time at which the mean of the mean
maximum temperatures will have increased to the point at which
the sea level will have risen by 3 metres.

Australian Council for Educational Research 64

Model evaluation Mentor comment

Students: Review of solution As an evaluation method the students looked up a prediction from
a meteorologist (Kurt Wayne) to provide an external ‘real world’
After reading over the process previously applied to determine the estimate against which to test their outcome (they again omitted
amount of time taken for the lobby of the Q 1 to be at sea level, the web reference from their report). They identified an arithmetic
it became apparent that the equation to calculate the sea level error, caused by misreading expansion data, realised their
increase when given the temperature increase was incorrect, as predicted date was far too soon, and subsequently addressed the
the percentage of water expansion was not correctly placed inside error. They revisited their original caveat concerning the possible
the equation; it should have been decreased by a factor of 10 to impact of melting ice, coming up with a much longer timeframe
allow one to multiply the percentage by the temperature increase. – seemingly stretched by another arithmetic error. Calculations
This is taken into account in revised calculations: temperature by the mentor based on data from the same website used by the
increase = 9.46671 leading to 10101.7 = χ. team gave an answer of about 350 years from 1938 – interestingly
Rounding χ up to 10102 months since January 1938, the within cooee of the Kurt Wayne prediction. Apart from the
number of years since January 1938 is 10102 ÷ 12 = 841.833, or question of accuracy, the variations in the predicted timeframes
approximately 842 years. Adding that on to 1938, the result is that served to illustrate why there is so much debate about the impact
in the year 1938 + 842 = 2780, the Q1 lobby will be at sea level, of climate change – itself a useful outcome. In various ways, the
with the mean of the mean maximum temperatures of that year at team can be seen to have considered all of the evaluation sub-
34.7414 degrees, and the predicted mean maximum temperature criteria in their efforts to solve the problem chosen.
of that month will be Summary comment
γ = 5.75 sin (0.5235(10102 + 4)) + ((0.001 × 10102) + 24.67) = In considering the quality of the overall modelling report it is noted
35.053 degrees. that the summary page is an IM²C requirement that did not apply
This will occur in 2780 − 2009 = 771 years’ time. in the present case. From the foregoing discussion and illustration,
The model predicts that there will be a sea level rise of 3 metres it is reasonable to say that all the characteristics listed above were
over 771 years. According to meteorologist Kurt Wayne, the sea displayed, and displayed consistently. Notwithstanding some
level will rise ‘by 15cm by 2030’. Assuming a constant rate of sea procedural errors, and incomplete referencing, the reader has been
level increase, this could be interpreted as given impressive evidence concerning modelling qualities of the
team.
15cm ÷ 10 years = 0.75cm/year
300cm ÷ 0.75cm/year = 400 years Finally, the purpose of this forensic examination is reiterated.
The goal has not been to provide an assessment of a specific
So Kurt Wayne predicts that in 400 years, the sea level would rise project as such, nor to display an ‘ideal’ solution, but to test the
by 3 metres, assuming a constant rate of sea level rise. The model applicability and robustness of a set of criteria to a modelling
created does not however, take into account the melting of polar project with no specific links to the IM²C. That is important
ice caps; something that could contribute significantly to the rising for providing a basis for the evaluation of modelling projects,
of the sea level. The difference of 311 years between Kurt Wayne’s irrespective of their specific educational context.
prediction and the model created could be reasonably suggested
to be due to the fact that the model generated only takes into In the following section, the criteria are applied with the specific
account the expansion of water due to a temperature increase, intention of providing advice for attaining quality within the IM²C
and not other factors like polar ice caps melting, moving water into project.
the ocean, which would logically only increase the rate at which
the sea level rises, effectively decreasing the number of years
required to rise the sea level by 3 metres.

Australian Council for Educational Research 65

 Relating modelling qualities
to judging criteria

In this section, we draw on specific comments from IM²C judges ■■ a reiteration of the Australian IM²C judging criteria
and examples from various submitted IM²C solutions. These
examples are taken from the teams that gained outstanding and ■■ the IM²C problem for 2015, along with an edited extract of the
meritorious awards in the International Mathematical Modeling international judges’ commentary
Challenge. The reader can find the IM²C entries via https://www.
immchallenge.org.au/supporting-resources/previous-immc- ■■ the IM²C problem for 2016, along with edited extracts of the
problems international and Australian judges’ commentary, and further
author commentary
We discuss and illustrate qualities required in a modelling report
to demonstrate how student reports can address the IM²C judging ■■ the IM²C problem for 2017, along with edited extracts of the
criteria. The main aim of this advice is to help future Australian international and Australian judges’ commentary.
teams successfully tackle future IM²C problems.
Full instructions, rules for entry and judges’ commentary can be
Note that these cover all the points raised in this guidebook on found on the website www.immchallenge.org.au.
the mathematical framework, the modelling process and writing a
report; and that were used in the previous section on a modelling
report illustrating use of criteria, to look at the structure of the
team report on the ‘Gold Coast inundated’ problem.

The IM²C began in 2015, and Australian teams have participated

in. This discussion gives:

Australian Council for Educational Research 66

The International
Mathematical Modeling
2
Challenge (IM C)
The IM2C is an international modelling schools, the IM2C seeks to develop and with a deeper experience both of how
competition involving teams of secondary enhance students’ ability to visualise, mathematics can explain our world and
students from a number of countries. understand and apply mathematics what working with mathematics looks
The IM2C poses a number of real-world in real-world contexts, providing a like. By mobilising students in teams, the
mathematical scenarios, and each team valuable opportunity for the practical IM2C replicates real-world conditions;
works for several days using freely demonstration of in-school learning and requiring collaboration and contribution
available material (from the web and other application of theory. from different skill sets, perspectives and
sources). At the end of this time, each methodologies to achieve overall success.
The point of involving teams of students is
team presents a report on their solution.
that difficult problems in society are almost Providing an opportunity for peer-based
The challenge awards prizes to the top always tackled by groups of people with learning, this aspect of the IM2C also
teams. The challenge has two levels: a different areas of expertise. The Challenge helps to incorporate and reinforce the
national and an international level. On provides opportunities for students to Australian Curriculum proficiency strands
the basis of the results of the national get together to pool their knowledge – understanding, fluency, problem-
competitions, the top two teams will be and understanding for a common goal, solving and reasoning – as students work
entered for judging at the international level. to communicate, to reason and to solve together, communicate with one another
problems. These are important factors of and employ creativity, reason and logic to
The main aim of the IM2C is to promote
the Proficiency Strand of the Australian successfully solve the defined problem.
mathematical modelling, encouraging
Curriculum and important skills in life.
participants to explore the application of
mathematics in real situations to solve Real-world problems require a mix of
problems of importance. Encouraging an different kinds of mathematics for their
extension of experience in mathematical analysis and solution, and take time and
modelling for students in secondary teamwork. The IM2C provides students

A fuller explanation of the instructions and rules for entry to the Australian
challenge can be found on the website www.immchallenge.org.au

Australian Council for Educational Research 67

Past IM2C questions

2015 IM2C: Movie scheduling

Describe the real-world problem

A great deal of preparation must take place before a movie can be filmed. Important sets and scenes need
to be identified, resource needs must be calculated, and schedules must be arranged. The issue of the
schedule is the focus of the modeling activities. A large studio has contacted your firm, and they wish to have
a model to allow for scheduling a movie. You are asked to answer the questions below. You should provide
examples and test cases to convince the movie executives that your model is effective and robust.

Question 1: Question 2:
Develop a model that will produce a filming schedule given the Develop a model that will take the information and schedule
following constraints: generated from the first question and can adjust them in the
event that some delay in one aspect or the availability of some
■ The availability dates of the stars of the film. asset changes. For example, if one of the stars has an accident
and cannot film for a certain period of time, you should be able to
■ The time required to film at a list of specific sites.
adjust the schedule.
■ The time required to construct and film on a list of sets.

■ The availability dates for specific resources. For example a Question 3:

war movie might require helicopters which are available only
Use the model developed in the first question to develop a way
at specific times.
to determine the most important constraints. That is, identify the
■ Some scenes cannot be shot until after certain computer constraints that will cause the longest delays if a problem occurs.
generated content is defined and other physical items are
constructed. Your schedule must include extra time to allow
for redoing some shots if they turn out to be inadequate
after editing and review.

Australian Council for Educational Research 68

 (see solution 2015015) used both a heuristics-based algorithm
and Kuhn’s algorithm (also known as the Hungarian method) that
guarantees a globally optimal solution and compared the results
of the two methodologies. The better papers generalised to larger
problems and tested their models using a case study or a practical
example, and carefully presented their conclusions.

Mathematical processing
A good example of sensitivity analysis can be found in solution
2015005, in which a great amount of random data was used
to identify which variables were most sensitive. Another good
example is in solution 2015004, in which the team provided a
methodology for reacting to changes by choosing the objective
of ‘changing the current schedule as little as possible’. Both
papers 2015004 and 2015005 stood out above the rest and were
2015 international judges’ commentary recognised as outstanding by the judges.
(edited extract) Use graphs, charts, networks and other appropriate visualisations
where possible to aid understanding. A good example of using
What characteristics distinguished the better papers? The
visualisations is provided by solution 2015017. The pictures of
better teams developed and presented their models in a very
graphs included in their work clarify the complexity of the making
logical manner. They moved from the very vague scenario with
of a film, thereby helping the reader understand the process that
which they were provided to identifying a problem they could
will be modelled in the following paragraphs.
model mathematically. They explained their assumptions very
clearly and discussed how well the assumptions were met by Model evaluation
the situation they had identified. After analysing their model for
A good example of testing a model can be found in solution
solutions, they tested the model’s conclusions against test cases
2015007. This team extracted data by hand from an actual movie,
they constructed or found in their research. They performed a
taking the data from the screen. This down-to-earth approach
sensitivity test to determine how the conclusions changed based
links the model with real-world data and gives a basis for testing
upon changes in their data, thereby identifying the most important
and improving the modelling process. For these reasons, paper
variables.
2015007 was awarded the meritorious award.
Problem definition
Report writing
Team 2015010 succeeded in profoundly developing the
The better papers excelled at scientific writing. The papers
building of their model. Step-by-step, this team introduces the
had a structure that was easy to follow. A very distinguishing
various elements of their model in a very clear manner that is
discriminator among all papers was the quality of the summary.
understandable to a wide audience. Due to time restrictions, they
A good summary provides a very clear overview of what is
did not succeed in extending this approach through the remainder
accomplished in the paper.
of the paper, but clearly demonstrated the ability to present their
ideas in a clear logical manner. For these reasons, paper 2015010 An excellent example of a well-constructed summary can be
was awarded the meritorious award. found in solution 2015016. The summary gives an excellent
overview of the work in a well-structured manner that invites the
Model formulation reader to read the complete work.
The better papers typically developed an optimisation problem Our advice to future contest participants is to allow plenty of
that generalised to larger data sets: more scenes, more filming time to construct a report. Build a structure which allows you to
sites, more actors with greater restrictions on their availability, and present the development of your model in a logical fashion. Test
so forth. Typically, the algorithms for solving their optimisation your model in as realistic a fashion as possible, and clearly state
models were computationally complex. The better papers used your conclusions. Similarly, present the analysis and conclusions
heuristics for finding good solutions in reasonable amounts of of your model in a manner that can be easily understood by a wide
time, or adding constraints to their models to reduce the number audience. Finally, write a summary that gives an overview of your
of possible solutions. The team awarded the meritorious award work that excites the reader to study your paper.

Full judges’ commentary and team solutions can be found on the website
https://www.immchallenge.org.au/supporting-resources/previous-immc-problems

Australian Council for Educational Research 69

Past IM2C questions

2016 IM2C: Record insurance

Ethiopia’s Haile Gebrselassie first won in 1994 and won for a third time in 2011. (Haile Gebrselassie at Vienna City Marathon 2011,
Alexxx86/Wikimedia Commons)

Describe the real-world problem

In athletics, one of the possible distances to run is 15 000 meters 1. For the 15k run described above with a 25 000 euro bonus
or 15k. For this type of run, 15k on a street track, there is a world what is the average cost of the bonus?
record, as there are records for all other distances that are run The insurance company will add an amount to the computed
in athletics (e.g., the marathon). In such a race, the organising average cost. The amount of the addition may be very reasonable
committee will usually pay a significant amount of money as a or not. The insurance company expects to cover their costs and
bonus to the winner if he or she succeeds in setting a new world realise a profit over a long time period with multiple subscribers.
record. These amounts of money can get quite large in order to The organising committee can decide to purchase the insurance
attract top runners: in the race shown in the picture there was or not (that is, ‘self insure’).
a 25 000 euro bonus if the winner succeeded in improving the
15k world record – which, by the way, he (un)fortunately did not 2. What criteria should the insurance company use in
achieve. Had he done so, there would have been a major financial determining the amount to add to the average cost for
problem for the organising committee, since they had not the above race? Specifically, how do they weight each
purchased any insurance. factor in determining their decision? For example, begin by
considering the case where the insurer will add 20% to cover
Usually, insurance will be purchased by the organising committee his operating costs, time value of money, and realise a profit
for such a running event, since the financial risks can be quite over a period of time.
large. The fee they will have to pay for such insurance will be, of
3. (a) What criteria should the organising committee use
course, significantly lower than the bonus they would have to pay
to determine whether or not they should purchase the
for a world record. Let’s define the average cost of the bonus as
insurance? Assume that they intend to sponsor this race
the ratio of the amount of bonus divided by the expected number
many times in the near future. By self insuring, they expect
of times the event is replicated before the current record is
to save the insurance company’s added cost over a period
broken. For example, if based on our analysis, we currently expect
of time.
the record to be broken every 25 repetitions under conditions
(b) But should they take the risk?
prevailing for a specified event, then the average cost of the
bonus is 1000 euro per race. The first question is:

Australian Council for Educational Research 70

Now consider that you are a member of the organising 5. Develop a general decision-scheme for the organising
committee of a major track meet with 20 men’s and 20 women’s committees to determine for each event whether they
athletic events, including field events (long jump, high jump, etc.). should purchase insurance or self insure. This scheme
should be written in a form easily understood and
4. Assume the organising committee can purchase the implemented by a typical organising committee.
insurance or not for each of the 40 events. For example,
they may choose to insure 10 of the 40 events. What factors Your submission should consist of a one-page Summary Sheet
should the organising committee consider in their decision and your solution cannot exceed 20 pages for a maximum of 21
to purchase insurance or not for each of the events at pages. (The appendices and references should appear at the end
the meet? Specifically, how do they weight each factor in of the paper and do not count toward the 20 page limit.)
determining their decision?

Zevenheuvelenloop

Zevenheuvelenloop (Seven Hills Run in English) is an annual 15 kilometres road

running race held in Nijmegen, Netherlands. It was first organised in 1984 and
has grown to be one of the largest road races in the Netherlands; it attracted
over 30,000 runners in 2008. The race has attained IAAF Label Road Race status.
The inaugural edition of the race in 1984 featured only an 11.9 kilometre course
as the Dutch athletics federation (Koninklijke Nederlandse Atletiek Unie) would
not allow new races to be longer than 12 km. The current undulating, hilly course
begins in Nijmegen, follows a path to Groesbeek and then loops back towards
Nijmegen to the finish line. Zevenheuvelenloop lends itself to fast times: Felix
Limo broke the men’s world record in 2001 and, at the 2009 edition, Tirunesh
Dibaba broke the women’s world record over 15 km. In 2010 Leonard Komon
improved Limo’s still standing World Record.

A number of athletes have achieved victory at the Zevenheuvelenloop on

multiple occasions; Tonnie Dirks, Tegla Loroupe, Mestawet Tufa, Sileshi Sihine
and Haile Gebrselassie have each won the race three times. The 2002 winner,
South African Irvette Van Blerk, won the race at the age of fifteen, having entered
the race while holidaying in the Netherlands. The race was used as the test event
for the development of the ChampionChip personal RFID timing system.

Source: Zevenheuvelenloop. (2017, September 19). Wikipedia, The Free Encyclopedia. https://
en.wikipedia.org/wiki/Zevenheuvelenloop

Australian Council for Educational Research 71

Winners
Key: Course record

Time Time
Edition Year Men’s winner Women’s winner
(m:s) (m:s)
32nd 2015 Joshua Cheptegei (ETH) 42:39 Yenenesh Tilahun (ETH) 50:05
31st 2014 Abera Kuma (ETH) 42:18 Priscah Jeptoo (KEN) 46:56
30th 2013 Leonard Komon (KEN) 42:15 Tirunesh Dibaba (ETH) 48:43
29th 2012 Nicholas Kipkemboi (KEN) 42:01 Tirunesh Dibaba (ETH) 47:08
28th 2011 Haile Gebrselassie (ETH) 42:44 Waganesh Mekasha (ETH) 48:33
27th 2010 Leonard Komon (KEN) 41:13 WR Genet Getaneh (ETH) 47:53
26th 2009 Sileshi Sihine (ETH) 42:14 Tirunesh Dibaba (ETH) 46:29 WR
25th 2008 Ayele Abshero (ETH) 42:17 Mestawet Tufa (ETH) 46:57
24th 2007 Sileshi Sihine (ETH) 42:24 Bezunesh Bekele (ETH) 47:36
23rd 2006 Micah Kogo (KEN) 42:42 Mestawet Tufa (ETH) 47:22
22nd 2005 Haile Gebrselassie (ETH) 41:56 Berhane Adere (ETH) 47:46
21st 2004 Sileshi Sihine (ETH) 41:38 Lydia Cheromei (KEN) 47:02
20th 2003 Richard Yatich (KEN) 42:43 Mestawet Tufa (ETH) 49:06
19th 2002 Kamiel Maase (NED) 43:41 Irvette van Blerk (RSA) 51:06
18th 2001 Felix Limo (KEN) 41:29 WR Rose Cheruiyot (KEN) 48:40
17th 2000 Felix Limo (KEN) 42:53 Berhane Adere (ETH) 48:06
16th 1999 Mohammed Mourhit (BEL) 43:30 Lyubov Morgunova (RUS) 49:45
15th 1998 Worku Bikila (ETH) 42:24 Tegla Loroupe (KEN) 50:06
14th 1997 Worku Bikila (ETH) 42:20 Catherina McKiernan (IRL) 48:30
13th 1996 Josephat Machuka (KEN) 43:06 Marleen Renders (BEL) 50:09
12th 1995 Josephat Machuka (KEN) 42:23 Hellen Kimaiyo (KEN) 49:44
11th 1994 Haile Gebrselassie (ETH) 43:00 Liz McColgan (GBR) 49:56
10th 1993 Khalid Skah (MAR) 43:35 Tegla Loroupe (KEN) 50:06
9th 1992 Carl Thackery (GBR) 43:54 Tegla Loroupe (KEN) 50:53
8th 1991 Tonnie Dirks (NED) 44:09 Ingrid Kristiansen (NOR) 48:46
7th 1990 Tonnie Dirks (NED) 44:53 Carla Beurskens (NED) 52:06
6th 1989 Tonnie Dirks (NED) 43:31 Carla Beurskens (NED) 50:36
5th 1988 Robin Bergstrand (GBR) 46:20 Marianne van de Linde (NED) 52:53
4th 1987 Marti ten Kate (NED) 45:11 Gerrie Timmermans (NED) 57:16
3rd 1986 Sam Carey (GBR) 46:2 Denise Verhaert (BEL) 53:33
2nd 1985 Klaas Lok (NED) 45:28 Joke Menkveld (NED) 57:28
1st 1984 Leon Wijers (NED) 36:55 Anne Rindt (NED) 45:48

Winners by country

Country Men’s race Women’s race Total

Ethiopia 11 12 23
Netherlands 7 6 13
Kenya 9 7 16
United Kingdom 3 1 4
Belgium 1 2 3
Ireland 0 1 1
Morocco 1 0 1
Norway 0 1 1
Russia 0 1 1
South Africa 0 1 1

Australian Council for Educational Research 72

2016 Australian judges’ commentary We saw many approaches to the decision of the organising
committees. What were the short- term and long-term goals of the
(edited extract) committee? What long-term and short-term risks were involved in
Problem definition accomplishing their goals if they did not buy the insurance? What
does ‘bankruptcy’ mean for a particular organising committee?
The better papers resulted as teams developed and presented
their models in a very logical manner. We strongly suggest that Papers 2016021, 2016026, and 2016033 were all awarded the
teams clearly answer each specific question in the problem. The designation of outstanding and employ distinct objectives for the
use of headers helps the judges understand which aspect of the weighting of short-term and long-term risks and the weight placed
given problem the team is currently addressing. on avoiding bankruptcy. They then develop decision models to
achieve their stated objectives.
Meritorious paper 2016036 is an example of presenting a model in
a very logical and reasoned manner. Each question is addressed Model evaluation
by reference to the model they developed from their assumptions.
It is important to address the strengths and weaknesses of
Mathematical processing the model you construct. All models have limitations and it is
important that you recognise and state those limitations.
Meritorious paper 2016004 is an example of excellent
mathematics but needed greater attention by the authors to their We strongly recommend that you read paper 2016026, which
explanation of why the mathematics is appropriate to a model was awarded the designation of outstanding, and demonstrates
developed from their assumptions, even though their examples the use of mathematics appropriate to the team’s assumptions,
are quite good. combined with some very good writing skills.

Model formation Report quality

There was great variety in modelling the probability of setting a Our advice to future contest participants is to allow plenty of time
world record at a particular event in a given track meet. Some to construct a report. In fact, consider working on the report as
papers used a simple approach estimating the trend in time soon as you begin work on the problem, as communicating your
intervals between world records. Others considered the historical ideas and approach is critical in this challenge.
records of specific athletes and determined the probability of a The better papers demonstrated excellent writing, particularly in
specific athlete breaking the current world record. Those athletes the quality of the summary. An example of an excellent summary
who had a reasonable chance of breaking a record were then can found in the meritorious paper 2016025. The summary gives
placed in a group of ‘top athletes’. They then considered the an excellent overview of the work in a well-structured manner that
number of top athletes attending an event and the probability that invites the reader to read the details of the complete work.
none of these athletes would break the record. Some weighted
recent data more heavily and examined the recorded times of the The best papers presented the analysis and conclusions of the
athlete versus their age and other variables. model in a manner that could be easily understood by a wide
audience. Consider who will use the model you have built and
Paper 2016021, which was awarded the designation of explain your model to that audience as well as to the judges. Use
outstanding, uses a simple approach to the probability by graphs, charts, networks and other appropriate visualisations
estimating the number of races before a record is broken. Paper where possible to aid understanding. An excellent example of
2016026, which was also awarded the designation of outstanding, incorporating graphs intelligently and conveniently for the reader is
examines particular athletes and computes the probability of that contained in meritorious paper 2016025.
athlete breaking the record based upon his recent performances
in the event and the current world record. They then compute the Consider who will use the model you have built and explain your
probability that no top performer attending the race will break model to that audience as well as to the judges.
the record. Paper 2016033, which was awarded the designation Remember that the judges are all from different countries. Explain
of outstanding, tested two approaches. The first is based upon your work using universally understood language. Also, the judges
assuming that participants have equal probability to break the are not familiar with the curricula of each school district. Thus,
record each year. The second is based upon assuming the world you should build a structure which allows you to present the
record will be broken at an interval within a range. After testing, development of your model in a logical and easily understood
they chose the second approach. fashion.
A wide variety of approaches were used to model the decisions The judges are not looking for the papers using the most
of the insurance company: how would they price the insurance sophisticated mathematics – this is a mathematical modelling
premium to cover their operating expenses, make a profit competition. Rather, they expect you to use the mathematics you
exceeding alternative investments, yet at a price attractive to already know in a logical manner to reach conclusions from your
organising committees? assumptions. Typically, simpler is better.
Meritorious paper 2016015 presents a unique method that is
reasonable to employ. This team also excels at communicating
ideas and justifying assumptions.

Full judges’ commentary and team solutions can be found on the website
https://www.immchallenge.org.au/supporting-resources/previous-immc-problems

Australian Council for Educational Research 73

2016 international judges’ commentary The nature and level of mathematical content of good modelling
reports can vary, and the relationship between the level and
(edited extract) complexity of mathematics used to the overall quality of the report
Problem definition is also not direct. Use of simple mathematics presented as part
of a well-argued and clearly described solution can be far more
An essential early step in working on a modelling problem is to effective than attempts to use more advanced techniques that are
decide what the problem is really about. Sometimes this is not poorly explained or justified.
obvious.
Whatever the level of mathematics used, though, it should be used
For the 2016 problem, scaffolding was provided in the form of correctly, and this feature of a mathematical modelling report
specific questions – these were presented to guide teams through will stand out very clearly to reviewers, particularly because they
the stimulus material and to help them make worthwhile progress will have a mathematical background themselves. For example,
towards the desired end result – as well as a dataset. However, if a report refers to ‘equations’ that are missing one or more of
this scaffolding may have made some teams think that the main a left-hand-side, an equals sign, or a right-hand-side (and so
problem was to answer these questions. In fact, a better approach are expressions rather than equations), this will not be viewed
would have been to see those as stepping stones towards the positively. Similarly, if a formula from a book or article is used, it
problem as articulated in the final question, which was about must be transcribed and applied correctly.
formulating well-argued and well-supported advice to the race
organising committee about how it should handle the risks The use of mathematical tools, especially technological tools such
associated with offering incentive payments to competing athletes as programmable calculators, or the use of statistical software,
across a number of events. can also be a help as well as a hindrance. Wise decisions about
the use of technology will stand out in a well-considered piece
Model formulation of modelling, as will poor decisions. Graphing tools can be very
powerful aids to investigating relationships among variables, and
The best models were clearly identified, and the variables used
statistical techniques can be very helpful in establishing best fit
in the models were fully explained. The assumptions underlying
relationship models. But including many colourful graphs that
the models used were also identified, explained and justified. In
serve no real purpose will generally not contribute positively to
solving real-world modelling problems, it is usually necessary to
a modelling report. Wherever possible, predictions made from
impose assumptions in order to simplify the situation sufficiently
the models applied should be checked against available (or
to make progress towards a solution.
constructed) data; and all mathematical work undertaken should
In addition, the choice of data values to test, demonstrate or be interpreted in relation to the problem under study.
to justify the models used is a significant part of the modelling
It should be obvious to a reader why a particular mathematical
process, and therefore also warrants solid argument and
tool or technique was chosen as the most useful in the
justification as to the choices made.
circumstances. This will usually warrant explanation and
Testing the sensitivity of a model is also an important step in justification, and results from using that tool should be interpreted
developing a model that will function as desired in the situation for with clear reference to its relevance to the purpose of the work.
which it is built. Using a range of parameter values to evaluate this
sensitivity is often a very useful activity. Model evaluation
An essential element of a good modelling effort involves thorough
Mathematical processing
evaluation of the model. Does it give information needed to make
The particular mathematical tools that could be applied to any decisions about the questions at issue? To what extent would
modelling task can vary enormously. Sometimes a simple piece of changes in the assumptions made, and changes to the settings
mathematics can be all that is needed to analyse some aspect of assumed in the way the problem is specified, affect the usefulness
the problem situation. The level of complexity of the mathematics of the model? Sensitivity analysis and other forms of model
used is not necessarily a critical factor, rather it is the relevance, evaluation can be very challenging, but they must be pursued.
appropriateness and usefulness of the mathematics that is most
Model evaluation can be undertaken by varying the numeric values
important, and its interpretation in relation to the situation under
that could apply in the situation under study, including by finding
study. The key factor is the fitness for purpose of the mathematics
suitable data from relevant sources. In some circumstances, this
used.
can be done systematically through simulating data, for example
For example, a graph of data that shows a relationship that where historical data cannot easily be obtained.
is not linear can provide a clue that a non-linear model for the
relationship should be explored. It would not matter how well a
linear regression model fits if it is not at all appropriate for this
particular problem.

Full judges’ commentary and team solutions can be found on the website
https://www.immchallenge.org.au/supporting-resources/previous-immc-problems

Australian Council for Educational Research 74

Report quality A good modelling report does not look like a mathematics
assignment. It is not a set of results as if it were a response to
A good quality report of the outcomes of a modelling task can
a test or a structured assignment. It’s a piece of analysis and
take a variety of forms, and the most important feature of a high
discussion that meets the needs of the particular situation under
quality report is that it meets the need at hand. In the case of
study. Detailed calculations and other mathematical content
the 2016 IM²C problem, for example, the context of the problem
would often best be presented as an attachment or appendix to
statement provided the perfect opportunity to imagine the
the main report.
organising committee as the key audience for a piece of written
advice as to what they should do regarding managing their event, In writing a summary, the goal should be to engage the reader
attracting high-level performers, and managing the associated and to draw them in to the analysis provided. In the 2016 case, an
risks through their insurance decisions. Such a report may have ‘executive summary’ could have been a useful model for the report
contained some mathematics, but only to the extent that it would summary.
help the organising committee understand the advice and reasons
for it.

Further commentary on 2016 solutions

In this table, we look at examples to illustrate the points that the IM²C Australian judges discussed. These examples are taken from the
teams that gained outstanding and meritorious awards in the 2016 IM²C.

Australian judges’ notes on approaches to the 2017 IM²C problem, by judging criteria

Criterion 1: Problem definition

Specification of precise mathematical questions from the general problem statement
See 2016026 as well as 2016021 in, for example, the start of part 2.

Criterion 2: Model formulation

Identification of assumptions with justification
Teams 2016021, 2016025 and 2016026 provide good descriptions and justifications of their assumptions. The first two teams do this
in prose, while the third team uses a table.

Choice of variables
See 2016036 for a good example.
Team 2016021 lists the main variables used at the very start of the report. The reasons for these choices appear as the report unfolds.

Identification and gathering of relevant (needed) data

Team 2016025, in Graph 2, shows technology output. It shows additional data represented as a graph showing as points the marathon
records since 1908. (Note that the am times for the vertical axis should be hour times.) This graph appears to have been produced by
the use of a commonly available piece of software.
Team 2016026 collects data on past world records (see the solution’s section 1.7).
Team 2016036 produces a frequency distribution table from accessing the Zevenheuvelenloop data from the 26th to the 32nd race
data.
Team 2016044 uses the 100 metres, 10 000 metres and high jump world records. This was so that it could predict various
world records and then apply their model to the problem in hand. This enabled the team to minimise the effect of things such as
‘technological advancement, development in training practices and anti-doping laws’.

Choice and justification of parameter values

See the section on ‘the amount of the insurance premium’ in solution 2016025.

Development of mathematical representations

See, for example, the section on ‘the amount of the insurance premium’ in 2016025, and solution 2016021, part 2.

Australian Council for Educational Research 75

 Criterion 3: Mathematical processing
Application of relevant mathematics
A variety of mathematical techniques and ideas are used by different teams in the attempts to find solutions to the questions. These
include:
■■ graphs (2016025)
■■ probability (2016026)
■■ regression (2016026)
■■ curves of best fit (2016025)
■■ benchmark percentage (2016026)
■■ product and summation notation (2016026)
■■ calculus (2016026, 2016036 - integration)
■■ labelled graphs or networks (2016026)
■■ expected values (2016026)
■■ matrices (2016026 – eigenvalues)
■■ statistics (2016004, 2016021, 2016036 – Pareto distribution, extreme value theory, binomial distribution, confidence intervals)
■■ Excel spreadsheets (2016004).

Invocation and use of appropriate technology

This is done by 2016021 by using Monte Carlo Simulation in part 3 of the report. Given the importance of technology in modelling,
it would be good to have a few more varied examples of technology use.

Checking of mathematical outcomes for procedural accuracy

This was consistently done for each of the models produced by team 2016036 (see sections 2.4, 3.4 and 4.4).
See also the ‘example case’ in parts 2 and 3 of 2016021.

Interpretation of outcomes in terms of the problem situation

Page 4 of team 201604’s work uses consecutive years that were and were not followed by a world record in the 10 000 metres race.
On the basis of this, the team was able to set limits for the range of probabilities that a world record would or would not succeed a
given period of years. This was then fed into their decision process for the organising committee.

Criterion 4: Model evaluation

Adequacy and relevance of findings in relation to problem situation
See team 2016021 in ‘strength and weaknesses’ in part 5 of their work; team 2016015 in ‘strengths and weaknesses’ on page 5; and
team 2016036, in, for example, section 4.5.

Further elaboration or refinement of problem

Team 2016021 approaches the first question in two ways: first through direct computation and then using confidence intervals.
Team 2016033 uses two approaches to Question 3.

Relevance of revised solution(s) following revisiting and further work within earlier criteria
2016026 follows two models through together for the first section. These were produced by different approaches and compared at the
end of the first section.

Quality of answers to specific questions posed in problem statement

201604 discusses the advantages and disadvantages of the team’s world-record determining model at the end of stage 4.
2016021 notes on page 5 of 20 that the error increases as the number of estimations increases.

Australian Council for Educational Research 76

 Criterion 5: Report quality
Summary page quality: succinctness, power to attract reader
We strongly recommend reading the summary of team 2016025 in full because the international judges’ make specific reference to it as
an excellent example. This report is clearly well written and provides material of interest that is easy to read and able to be understood by
the recipients, the organising committee of the 25km race.

Overall organisation of the report

Solution 0216036 was praised by the judges for it being ‘an example of presenting a model in a very logical and reasoned manner.
Each question is addressed by reference to the model they developed from their assumptions.’

Logical presentation including:

■■ description of the real-world problem being addressed

Team 2016026 provides a ‘problem restatement’ almost immediately after the summary. This appears as a list.
Team 2016021 includes an introduction that reorganises and expands the ideas inherent in the various questions.

■■ specification of the mathematical questions

Team 2016021 specifies mathematical questions throughout its work.

■■ listing of all assumptions wherever they are made

2016026, for example, introduces new assumptions as they are needed at the start of answering each question. The team though
does provide ‘overall assumptions and justifications’ immediately after the ‘problem restatement’. In all cases the assumptions and
justifications are presented in table form.

■■ indication of how numerical parameter values used in calculations were decided on

This can be found in all reports.

■■ setting out of all mathematical working: graphs, tables; technology output and so on

Graph 2 from team 2016025 is technology output. The curves are curves of best fit, one (blue) modelling the marathon record times
up to the late 1960s and the other (red) for the last 40 years. Before this graph there is a clear explanation of the graph and its potential
use in the Project.
Team 2016033 produces specifically designed programs to facilitate calculations (see, for example, Approaches 2.1 and 3.2.2). Team
2016021 also does this in their appendix on MATLAB programs.
Excel spreadsheets were used by 2016044 to fit data distributions and make calculations easy.

■■ interpreting the meaning of mathematical results in terms of the real world problem

This is done well by all teams.

■■ evaluating the solution(s) in terms of the problem requirements

2016026 discusses the limitations of its model, including the large number of variables that the team has needed to use. The team is
also concerned about the usability of its model without a computer.
2016044 uses spreadsheets do calculations and so makes the work within the capabilities of the organising committee.
It is worth noting that, of the seven team’s work considered here, only one provided simple instructions for the organising committee
in their conclusion section. The instructions were clear to the teams but generally required some quite sophisticated mathematical
knowledge to be able to calculate the unknowns and make decisions as to whether to insure or not. Team 2016015 produces a
‘decision-making scheme’, but it requires the organising committee to be able to find the probability of a world record, as well as a
geometric mean. The closest any team appears to have got was 2016021 who provide computation device they call a ‘risk calculator’
for Question 3.

Australian Council for Educational Research 77

Past IM2C questions

2017 IM2C: Jet lag

Describe the real-world problem

Scenario 1: Small meeting
Time: Mid-June
Organising international meetings is not easy in many ways,
including the problem that some of the participants may Participants: 6 individuals from:
experience the effects of jet lag after recent travel from their home ■ Monterey CA, USA
country to the meeting location which may be in a different time
zone, or in a different climate and time of year, and so on. All these ■ Zutphen, Netherlands
things may dramatically affect the productivity of the meeting. ■ Melbourne, Australia
The International Meeting Management Corporation (IMMC) has ■ Shanghai, China
asked your expert group (your team) to help solve the problem
■ Hong Kong (SAR), China
by creating an algorithm that suggests the best place(s) to hold
a meeting given the number of participants, their home cities, ■ Moscow, Russia
approximate dates of the meeting and other information that the
meeting management company may request from its clients.
Scenario 2: Big meeting
The participants are usually from all corners of the Earth, and Time: January
the business or scientific meeting implies doing hard intellectual Participants: 11 individuals from:
team work for three intensive days, with the participants
■ Boston MA, USA (2 people)
contributing approximately equally to the end result. Assume
that there are no visa problems or political limitations, and so any ■ Singapore
country or city can be a potential meeting location. ■ Beijing, China
The output of the algorithm should be a list of recommended ■ Hong Kong (SAR), China (2 people)
places (regions, zones, or specific cities) that maximise the overall
■ Moscow, Russia
productivity of the meeting. The questions of costs are not of
primary importance, but the IMMC, just as any other company, has ■ Utrecht, Netherlands
a limited budget. So the costs may be considered as a secondary
■ Warsaw, Poland
criterion. And the IMMC definitely cannot afford bringing the
participants in a week before the meeting to acclimatise or give ■ Copenhagen, Denmark
them the time to rest after a long exhausting journey.
■ Melbourne, Australia
Test your algorithm at least on the two following datasets. Your submission should consist of a one-page summary sheet. The
solution cannot exceed 20 pages for a maximum of 21 pages. (The
appendices and references should appear at the end of the paper
and are not included in the 20-page limit.)

Australian Council for Educational Research 78

2017 international judges’ commentary time. What is the effect on the six-person group productivity of one
or more members with low productivity as defined by your team?
(edited extract) Would a better solution be to minimise the maximum number of
Problem definition time zones crossed by a member of the group, or minimise the
maximum distance travelled, or minimise the maximum time of
The better papers moved from the rather vague scenario that travel of any member of the group?
required maximising productivity to identifying a problem they
could model mathematically. Paper 2017045 shows a very Mathematical processing
consistent, clear and pleasantly readable approach of the problem,
This is a mathematical modelling challenge, so solutions should
even though it may to be a rather simple one. This paper shows
not force the mathematics upon the given scenario. Rather, a
that even without using complicated formulae it is possible to
team should begin with the simplest mathematics that solves the
describe a good model. Furthermore, in both paper 2017019 and
problem the team has identified. Later, if you feel it is appropriate,
paper 2017026, a nice use has been made of various types of
you can refine your model to increase the accuracy, or change
pictures to explain their method and thinking, thus avoiding heaps
your assumptions to find a more appropriate solution.
of text and formulae, thereby making the reading much more easy
and pleasant.

Model formulation
Teams developed a wide range of definitions of productivity. First,
they identified variables, such as ‘jet lag’ caused by traveling in
an east-west direction (typically approximated by the number of
times zones crossed), the distance travelled by each attendee, the
total travel time (including layovers) of each member, the ability
of the destination city to host the meeting, changes in sunset
time for each attendee, climate conditions especially temperature
and altitude, and other factors. Next, there was great variation on
forming an objective function.
Some teams picked a single factor such as jet lag to optimise.
Some teams picked several factors and weighted the relative
importance of each factor to form the productivity function.
Several teams picked the variables they wished to consider
and addressed them sequentially. For example, Team 2017045
outlined a process of first determining a time zone in which to
meet; then, within that time zone, the general area that minimised
the average travel distance and travel time; then, within that
area, an acceptable climate zone; then, within that zone, the
cities appropriate to host the meeting. By considering the most
important variables first and allowing an acceptable range of
satisfaction, they developed a process for converging to a solution
they liked.
Another distinct approach was to research a list of ideal cities
to host meetings. Students found different organisations which
ranked the desirability of the best host cities based upon different
criteria. They then began eliminating cities based upon one or
more of the variables listed, such as jet lag, distance travelled, time
of travel, and so on.
After defining productivity using a principal variable or a subset of
variables, another important decision needed to be addressed. For
example, suppose jet lag were the only variable considered.
Should we minimise the average amount of jet lag? For example,
if Shanghai were chosen, the average time zones crossed would
be minimised. But one of the six members (Monterey, USA) would
have to cross nine time zones. And the member from Melbourne,
Australia would have a great distance to travel and a long travel

Australian Council for Educational Research 79

Model evaluation nice description why some aspects were not taken into account)
and 2017054 (very ‘to the point’).
There were some common shortcomings among solutions
that could have been addressed through more rigorous model Outstanding paper 2017057 presents a comprehensive and
evaluation. appropriate mathematical model, clearly communicates
the team’s modelling process, and articulately explains their
Some teams did not consider travel time in realistic settings. Many
results, as well their model’s strengths and weaknesses. These
teams used distance between cities as an approximation, which
characteristics, in general, distinguish better papers.
is not a good assumption. As a result, those teams computed
absurd answers such as the middle of Siberia and North Korea. Pictures, especially graphs, tables and schedules, can explain
In reality, considering actual travel time makes the problem quite efficiently your ideas and by using them you can sometimes
much easier. The meeting will preferably be held in cities with avoid a lot of text. Also the use of relevant pictures and graphs
international connections or towns that are easily accessible from makes a report clearer and more pleasant to read. But please
such cities. realise sometimes tables might better be in an appendix. The
use of formulae is of course quite essential in a mathematical
Many students used temperature difference as a penalty. But
modelling assignment. But the use of unexplained formulae will
in reality, even people from very cold areas in winter would not
not make the report more convincing. So always make sure the
mind being in a place like San Diego in winter, even though the
reader believes the formulae used are at least understood by the
temperature difference is huge.
writers themselves. Also realise the readers of your report are,
Some teams over-emphasised the difference between east-west though experienced mathematicians, not experts in all parts of the
vs west-east travel. Actually, east-west long distance travel is great world of math!
typically faster because of the jet stream. Hong Kong to Seattle is
Appendices, especially references, are very useful. But do not
11 hours while Seattle to Hong Kong is 13.
expect the judges to read your appendices. We may refer to an
Report writing appendix to check a reference or a step in your computer code,
but otherwise may not examine or read an appendix. So, do not
The summaries of the better papers were excellent. Examples place anything important to the development of your model in an
of good summaries can be found in papers 2017007 (a clear appendix.
explanation of the process followed by the team), 2017020 (with a

Australian Council for Educational Research 80

2017 Australian judges’ commentary zones needed to be crossed to get to the meeting’s location. After
this team established that mathematical objective, team members
(edited extract) then systematically tested each potential location time zone to
Problem definition determine the total number of time zone changes that would
occur. This enabled them to narrow their solution to the time zone
An essential starting point is to clarify exactly what the IM2C that created the fewest time zone changes for all participants,
problem requires. The problem statement asks teams to develop which is where their investigation stopped. Another iteration of the
an algorithm, and to test that algorithm on at least two given modelling process for that team might have been to then apply
scenarios. The question does not ask where those two meetings other factors, such as existing flight routes, or climate factors, to
should be held, but asks for a systematic process (an algorithm) further narrow their recommendation.
that could be followed to determine a location for a meeting
of participants from different home locations that maximises Model evaluation
productivity of the participants, especially in relation to the effects
No modelling process is complete without an evaluation of the
of jet lag.
solution proposed. Does it answer the question? How would
Another essential step in an effective modelling activity is to a change in the assumptions or starting conditions affect it?
transform the statement of what is required into a mathematical What additional factors could be taken into account to make
objective. As well as developing a clear understanding of what the solution work in a wider variety of circumstances? Very few
would constitute an answer to the question in its context, of the IM2C 2017 reports considered the extent to which their
the goals of the exercise must also be expressed in clear solution was ‘best’ or what other possible solutions might have
mathematical terms. For example, the mathematical objective been equally or almost as good. Very few teams showed that their
could be to minimise total distance travelled by the meeting solution would apply to completely different scenarios from the
participants, or time spent travelling, and so on. two cases given as part of the problem statement.

Model formulation and mathematical processing Report writing

In between the beginning process of defining the goals of the The end-point of the modelling process is to communicate
task and defining that in mathematical terms, and the end the results in a form that can be understood and used by its
process of writing a report, the processes of model formulation, audience. The report required for the IM2C comprises three
mathematical processing, and model evaluation take place. parts: a one-page summary sheet, a report of the solution,
Those processes would often occur multiple times, since the first and appendices, which includes references. However, the
attempt to solve the problem might expose other issues to be exact way a report is constructed should be determined in
taken more into account to provide the best possible solution. A light of its purpose and audience. A modelling report is not the
better mathematical formulation might be needed, which might same as a school mathematics assignment. It might take the
require an adjusted model that a different kind of mathematical form of a recommendation or set of recommendations to a
processing could assist, and an updated interpretation of the committee, together with an explanation and justification of the
results and evaluation of the outcomes would then be needed. recommendations.
An example taken from the 2017 IM2C reports was from a team
that did an excellent job of defining the cumulative effects of jet
lag across the meeting participants as the total number of time

Australian Council for Educational Research 81

Australian judges’ notes on approaches to the 2017 IM²C problem, by judging criteria

Better approaches Problematic treatments

Model formulation
Identifying relevant variables

■■ Recognise that meetings are usually held in climate- ■■ Look for an ‘average’ climate – ignores that people react
controlled buildings. differently to their usual climate, to variations in climate, and
to shorter-term weather changes.
■■ Seek to incorporate climate in recognition of out-of-meeting
activities. ■■ Fail to distinguish between weather and climate.

■■ Use scientific data to optimise working conditions (choose

‘best’ latitude with defensible definition of ‘best’).

Identifying assumptions and other factors

■■ State assumptions clearly and explain why they are made ■■ Making pointless, unrealistic, or unfair assumptions (e.g. no
(e.g. to simplify the problem). flight delays, no crying babies will be on the flight, all meeting
participants are in good health to minimise health-related
■■ Show how the assumptions contribute to the solution path
exacerbation of jet lag).
followed.
■■ Factors simply stated with no justification or evidence, and
■■ Consider the possible impact of changing assumptions.
no link to solution.
■■ Factors used are justified (e.g. evidence cited) and linked to
■■ Extensive exploration of flight costs, hotel costs, meeting
solution.
room hire costs etc.
■■ Impact on cost of using home location of a participant.

Mathematical processing
How ‘distance travelled’ was treated
■■ Consider actual flight arrangements (such as proximity to ■■ Treat all distances ‘as crow flies’ rather than actual journeys
airport, existence of direct flights, total travel time) required, including great circle calculations that don’t take
actual flight routes into account, or using three-dimensional
coordinates only.
■■ Ignore multiple participants from particular origins

How time zones were treated

■■ Seek to minimise the total number of time zone changes for ■■ Only time zone considered without taking account of actual
participants. journeys.
■■ Note that the average UTC offset (or equivalent) does not ■■ Time zone calculations performed without checking the total
necessarily minimise total time zone changes. time zone changes that result (e.g. note that the ‘average
time zone’ calculation does not necessarily yield the location
■■ Take account of changes in time zone differences at different
with the least number of time zone changes)
times of the year.
■■ Ignore multiple participants from particular origins.
■■ Quantify the daily period of alertness (and its overlap) for
participants (and therefore spell out the impact of jet lag). ■■ Ignore the effect of daylight saving on time zone data.
■■ Recognise jet lag diminishes each day. ■■ Assume the days required for jet lag recovery can be added
to the pre-meeting time (contradicts problem statement).
■■ One report defined a productivity function, applied it to
participants according to their ‘normal alert hours’, and
integrated it to find the total work achieved.

Model evaluation
■■ Consider the possible existence of multiple ideal locations. ■■ Find just one proposed city for each scenario.
■■ Consider the range of different locations that could provide ■■ Propose a solution that does not pass the laugh test (eg,
more or less equivalent solutions. clearly looks wrong from inspection of maps provided; is in
the middle of no-where)
■■ Consider the applicability of the algorithm for other scenarios
(eg, different kinds of configurations of origin locations –
such as several participants coming from particular region,
with only one or two coming from different region)

Australian Council for Educational Research 82

Past IM2C questions

2018 IM2C: The best hospital

Describe the real-world problem

Almost everyone will seek health care at some point in life. In an

emergency, an individual will most likely go to the closest hospital,
but for non-emergencies, we may have a choice of where to
seek treatment.

Suppose there are 4 or 5 hospitals reasonably accessible to your

residence. You wish to pick the “best” hospital. How would you
measure and choose the “best” of these local hospitals? Suppose
the gravity of your situation is such that you are willing to travel
for your health care and you want to pick the “best” of 50 or
so hospitals. What variables do you use and how well can you
measure each?

Certainly, mortality is an important variable. Measuring death

rates has the advantage that death is a definite unique event.
The total number of deaths may not be a good measure of the
quality of the hospital at all, but the number of evitable deaths
could be a very good measure. How do we decide whether a
death is evitable or inevitable? Each death case can be coded
with data to include, for example, primary diagnosis, age, gender,
the urgency of admission, comorbidity, length of stay, social Your submission should consist of:
deprivation, and other factors. With large sample sizes, the
performance of different hospitals could possibly be measured ■ one-page summary sheet
by comparing cases with similar characteristics.
■ two-page memo
In addition to mortality there are other factors that one might
■ your solution of no more than 20 pages, for a maximum
want to use in measuring the overall quality of a hospital. A few
of 23 pages with your summary and memo
possible variables include:
note: the reference list and any appendices do not count toward
■ The experience of the doctors
the 23-page limit and should appear after your completed solution.
■ The amount of attention one expects to receive from the
staff and attending physician

Your team is tasked to use mathematical modelling to address Glossary

three requirements. Evitable: capable of being avoided.

1 Develop a model that uses mortality to measure the quality Comorbidity: the presence of one or more medical
of a hospital. conditions co-occurring with a primary condition.

2 Develop a model that uses other factors, in addition to User-friendly: easy to learn, use, understand, or deal with.
mortality, to measure the quality of a hospital. Based on the
factors you include from particular hospitals, your model Social deprivation: hardship caused by a lack of the
must result in information to make a decision of which ordinary material benefits of life in society.
hospital is the best.

3 In addition to the mathematical analysis, you provide in

your report, include a two-page “user-friendly” memo that a
person without much mathematical expertise or computing
ability can use to choose a hospital.

View the 2018 Australian and international judges’ commentary at

https://www.immchallenge.org.au/supporting-resources/previous-immc-problems/previous-problem-2018

Australian Council for Educational Research 83

Additional example problems
from the IM 2C

The International Mathematics Modeling Challenge has been run annually since 2015. A number of other contests have led up to the
establishment of IM²C. The Consortium for Mathematics and its Applications has long run the High School Mathematical Contest in
Modeling, in the United States of America. Several sample prompts from this contest are outlined below.

 Bank service
The bank manager is trying to improve customer satisfaction by offering
better service. Management wants the average customer to wait less than
two minutes for service and the average length of the queue (length of the
waiting line) to be two persons or fewer. The bank estimates it serves about
150 customers per day. The existing arrival and service times are provided.
Determine if the current customer service is satisfactory according to the
manager’s guidelines. If not, determine, through modelling, the minimal changes
for servers required to accomplish the manager’s goal.

 Emergency medical response

The Emergency Service Coordinator for an area is interested in stationing the three
available ambulances to provide the best likely medical response. Area maps,
population data and average travel times are provided. Determine the best locations
for the three ambulances to maximise the number of people who can be reached
within 8 minutes of an emergency call.

American elk
Prior to the arrival of European colonisation on the North American continent, the
ecological bio-diversity was much richer than we currently know in the 21st century.
The elk native to the eastern areas of the United States was hunted to extinction
sometime in the early 1800s. A similar species, the Manitoba Elk, adapted to living
in the western US and Canadian prairie, developing tolerances to certain diseases,
foods, and environmental differences. How would these adaptations affect an
introduction of Manitoba Elk into the eastern states? Population data from a
reintroduction trial are provided. Build a mathematical model to determine whether
the elk survive or die out. Come up with a plan to improve the growth of the elk
population over time.

 Fuel prices
Fuel prices fluctuate significantly from week to week. Consumers would like to
know whether to fill up the tank (fuel price is likely to go up in the coming week)
or buy a half tank (fuel price is likely to go down in the coming week). Develop a
model that a consumer could use each week to determine how much fuel – full
tank or half tank – to purchase.

Australian Council for Educational Research 84

 Space shuttles
On July 21, 2011, the 135th and final United States Space Shuttle landed
in Florida after its 13-day mission into orbit, complete with a docking at
the International Space Station. The National Aeronautics and Space
Administration will now have to rely on other nations or commercial
endeavours to travel into space until a replacement vehicle is developed and
constructed. Develop a comprehensive 10-year plan complete with costs,
payloads, and flight schedules to maintain the International Space Station.

 Search and find

Finding lost objects is not always an easy task, even when you have
knowledge of a general location. Consider the following scenario: you have
lost a small object, such as a class ring, in a small park. A map of the park
is to be provided. It is getting dark and you have your pen light flashlight
available. If your light shines on the ring, you assume that you see it. You
cannot possibly search 100 per cent of the region. Determine how you will
search the park in minimum time. An average person walks approximately 4
miles per hour. You have about two hours to search. Determine the chance
you will find the lost object.

 Bicycle club
Several cities are starting bike share programs. Riders can pick up and drop
off a bicycle at any rental station. These bicycles are typically used for short
trips within the city centre, either one-way or roundtrip. The idea is to help
people get around town on a bike instead of a car. Those making longer trips
(such as commuting to work) are likely to use their own bikes. Some of the
challenges are how to determine where to locate the rental stations, how
many bikes to have at each station, how/where to add new locations as the
program grows, how many bikes to move to another location and when (time
of day, day of week). The downtown city maps, the bike rental locations and
the number of bikes at each location for several are available online. Develop
an efficient bike rental program for these cities.

Curbing city violence

A regional city has had lots of problems with gangs and violence over the
years. The mayor, chief of police, and city council need your help. Data are
available for the following: incidents of violence, homicides, assaults, regional
population (census data), unemployment, high school enrolment, high school
attrition, graduation rate, prison population (adult and juvenile), released on
parole, parole violations. Analyse and model these data to give the city a plan
to reduce violence. Prepare a report for the mayor outlining your proposals.

Full problems with example data and possible solutions can be found on the website
https://www.immchallenge.org.au/supporting-resources/additional-example-problems

Australian Council for Educational Research 85

View publication stats