APPLIED MATHEMATICS 4

2020

SCHOOL OF MATHEMATICS AND STATISTICS

Contents
1 Entry Requirements 1

2 Course Administration 2

3 Structure of Applied Mathematics 4 3

3.1 Course work . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 3
3.2 Essay . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 3
3.3 Course descriptions for the Core Applied Maths Units: . . . . . . . . . . . 4
3.4 Course descriptions for Selective Applied Maths Units: . . . . . . . . . . . 4

4 Assessment Procedures 6

5 The Essay/Project 7
5.1 Assessment of the Essay/Project . . . . . . . . . . . . . . . . . . . . . . . 8
5.2 Important Dates . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 8
5.3 Essay/Project Guidelines . . . . . . . . . . . . . . . . . . . . . . . . . . . . 9
5.4 Suggested Topics for the Essay/Project . . . . . . . . . . . . . . . . . . . . 10

6 Rights and Responsibilities 32

7 Scholarships, Prizes and Awards 33

8 Life After Fourth Year 34

 1 Entry Requirements

Entry requirements vary slightly depending on whether the candidate has completed a
Bachelor of Science (BSc) degree or a Bachelor of Advance Studies (BAS) degree. Entry
to the Honours Programme in Applied Mathematics is usually based on satisfying the
following criteria (and subject to approval by the Head of the School of Mathematics &
Statistics):
1. Faculty requirement:
BSc: The candidate must have qualified for the Pass Degree in Mathematics (or
discipline cognate with Mathematics) with a SciWam of at least 65, or
BAS: Completed the Bachelor degree with two majors, at least one of which must
be in Mathematics (or a discipline cognate with Mathematics) with a WAM of
at least 65, and
2. Mathematics requirement:
BSc The candidate must have completed 24 credit points of 3000 level or above
mathematics units with an average mark of at least 65 % for advanced level
units and an average mark of at least 75 % for mainstream units.
BAS The candidate must have completed a mathematics major with an average
mark of at least 65 % for advanced level mathematics units and an average
mark of at least 75 % for mainstream units.
3. Essay / Project supervision:
BSc/BAdvS The candidate is expected to find a prospective supervisor from among the
Applied Mathematics staff, who is agreeable to supervise the candidate’s essay
or project in the candidate’s chosen topic.

Students from institutions other than the University of Sydney must possess qualifications
which are deemed equivalent to the above. There is some flexibility in these conditions;
students not quite meeting them, but keen to pursue Honours in Applied Mathematics are
invited to contact the Course Coordinator for advice. Students are expected to meet with
prospective supervisors to discuss the potential for essays or projects before submitting
the honours application.
Applications must be submitted to the Faculty of Science.
Application and enrolment information should also be obtained from the Faculty of Sci-
ence, either in person of from their website. The Faculty will also provide AAM compu-
tations once the final third-year results are in.
Graduate Diploma in Science and MSc(Qualifying) applicants should see the Director of
Postgraduate Studies before enrolling.

1
2 Course Administration

The Course Coordinator for Applied Mathematics 4 in 2020 is:

Prof. Marek Rutkowski
Room 814, Carslaw Building
Phone: (02) 9351 1923
Email: marek.rutkowski@sydney.edu.au

2
3 Structure of Applied Mathematics 4

3.1 Course work

Full-time students normally will enrol in 48 credit points, 24 of which will be coursework.
Student pursuing honours under the BSc degree will enrol in the ‘shell units’ MATH4401
A-D, while students in the BAS degree will enrol directly in the units themselves. Course-
work requirements are the following:

1. Core Units:
Students are required to enrol in the following two 6 credit point core units

– Semester 1 - MATH 4411 - Applied Computational Mathematics

– Semester 2 - MATH 4412 - Advanced Methods in Applied Mathematics

2. Selective Units:
In addition to the Core Units, students should choose their remaining coursework
from units offered by the School of Maths listed as 4000 level or higher which have
not already been taken for credit.

In addition to the courses listed above, students may be allowed to take one course
from a related discipline. Approval from the Essay/Project Supervisor and the Course
Coordinator needs to be obtained prior to enrolment. The choice of course work has to
ensure that the student covers the proper material required for the student’s project or
essay. (Further details should be obtained from the Fourth-Year Course Coordinators in
Pure Mathematics and Statistics.)

3.2 Essay
In addition to the courses, each student is also required to write an essay or project on
an Applied Mathematics topic, under the supervision of a member of staff of the School
of Mathematics and Statistics. This is considered to be the major aspect of the Honours
Programme, and is detailed in Section 5.

3
3.3 Course descriptions for the Core Applied Maths Units:
MATH4411 Applied Computational Mathematics (Sem-1) Prof G. Gottwald

The computational approach to study scientific problems serves a twofold pur-

pose in mathematics. Numerical formulations and analysis help scientists in
exploring complex systems, often giving rise to hitherto unknown phenomena
which deserve further analytical treatment and understanding, or they can be
used to verify existing mathematical theories. On the other hand, the need for
numerical simulations, for example in studying complex systems such as the
ocean-atmosphere systems, requires new mathematics to allow for an efficient
simulation. This course will highlight examples of this fruitful interplay of
analytical mathematics and numerical computations.
The course will present a variety of numerical methods together with some of
the underlying theory covering the numerical integration of partial differential
equations, geometric integration, stochastic differential equations as well as
some basic machine learning. When solving continuous time systems such
as ordinary, partial or stochastic differential equations, we need to discretize
time and space. How this affects the accuracy of our computations will be
discussed, for example. The lectures will prepare the theory and background.
MATH2x63 and/or MATH3x76 or similar courses (such as COSC) provide a
useful background for this course, but are not required.

MATH4412 Advanced Methods in Applied Mathematics (Sem-2) TBD

Course description to appear.

3.4 Course descriptions for Selective Applied Maths Units:

MATH4413 Applied Mathematica Modelling (Sem-1) TBD

Course description to appear.

MATH4414 Advanced Dynamical Systems (Sem-2) TBD

Course description to appear.

MATH541A/B Special Topics in Applied Mathematics (Sem-1) TBD

Course description to appear.

MATH542A/B Deterministic and Stochastic Systems (Sem-2) TBD

4
 Course description to appear.

MATH543A/B Mathematical Models for Natural Phenomena (Sem-2) TBD

Course description to appear.

Units previously labeled as 39XX and now labelled as 40XX are also available to be taken
as selective courses. Course descriptions can be found on the School’s website.
http://www.maths.usyd.edu.au/u/UG/SM

5
4 Assessment Procedures

The Honours mark for each student is computed based on the following:

• 50 % for the Project/Essay assessment;

• 50 % for 4 courses (12.5 % for each).

The marking scale for Honours is significantly different from the undergraduate marking
scale at the University of Sydney. The Essay/Project, in addition to all the fourth-year
courses, will be marked with this scale in mind. This scale appears below.

Grade of Honours Faculty-Scale

First Class, with Medal 95–100
First Class (possibly with Medal) 90–94
First Class 80–89
Second Class, First Division 75–79
Second Class, Second Division 70–74
Third Class 65–69
Fail 00–64

6
5 The Essay/Project

A significant part of the Honours year is the completion of an Honours Essay or Project
by each student. Each student must choose an Essay/Project supervisor who is willing
to supervise the student’s chosen topic for the Essay or Project.
A list of available topics appears in Section 5.4. Students are welcome to choose differ-
ent topics, provided that they are able to obtain a supervisor for that topic from within
the School. Essay/project topics and supervisors should be finalised by the beginning
of the First Semester, so that students can commence work immediately on their Es-
says/Projects.
The following list shows the main applied mathematics resarch areas:

• Dynamical Systems

• Financial Mathematics and Mathematical Economics

• Geophysical and Astrophysical Fluid Dynamics

• Industrial and Biomedical Modelling

• Integrable Systems

• Mathematical Biology

For detailed information about these areas and the corresponding staff, please have a look
at the webpage
http://www.maths.usyd.edu.au/res/AppMaths.html

7
5.1 Assessment of the Essay/Project

The Essay or Project will be marked according to the following.

• 90 % for the final written report

This will be marked by 3 different markers, one of whom is the supervisor, and each
marking will therefore constitute 30 % of the final Essay/Project mark. Note that
the assessment also includes a one page report submitted at the end of the first
semester (see section 6.3).

• 10 % for a seminar presentation on the Essay/Project

Three typed and bound copies of the final Essay/Project should be submitted to the
Applied Mathematics Honours Course Coordinator, who will then distribute these copies
to three markers (one of which is the supervisor) for marking. Due dates for submission
appear in Section 5.2.
The seminar is an opportunity for each student to present the material of his or her
Essay or Project to a mathematically literate audience. The seminar talk will usually
be of 25 minutes duration, with an additional 5 minutes set aside for questions. The
Course Coordinator will provide additional information and help Honours students in
their preparation for the seminar. The presenter of the best AM4 seminar will be awarded
the Chris Cannon Prize.

5.2 Important Dates

The following are important dates for all students intending to complete their essay/projects
by the end of the first or the second semester of 2020 .

Semester 1:

• Seminar: Week of April 20-24th, 2020 (week 8)

• Essay Submission: Monday, 25th May 2020 (week 13)

An electronic file (pdf format) and three typed and bound copies of the Essay/Project
are to be handed in to the Applied Mathematics Honours Course Coordinator by
this date and time. Note the Electronic pdf file must be submitted to the Honours
Coordinator and uploaded via the Turnitin system to the University’s LMS website
by this deadline.

Semester 2:

• Seminar: Week of 21-25 September, 2020 (week 8)

8
 • Essay Submission: Monday, 2nd November 2020 (week 13)
An electronic file (pdf format) and three typed and bound copies of the Essay/Project
are to be handed in to the Applied Mathematics Honours Course Coordinator by
this date and time. Note the Electronic pdf file must be submitted to the Honours
Coordinator and uploaded via the Turnitin system to the University’s LMS website
by this deadline.

5.3 Essay/Project Guidelines

• The student should consult the supervisor on a regular basis, preferably at least
once a week. This is the student’s responsibility.
• A realistic schedule for work on the essay or project should be drawn up at an
early stage, and adhered to as closely as possible. If it proves necessary to modify
the original plans, a revised schedule should be drawn up after discussion with the
supervisor.
• At the end of Semester 1, a one page report is to be submitted to the Honours coor-
dinator. This report includes a half page description about the students aim/scope
of the project/essay and a half page description about what the student has achieved
in semester 1 and what the student wants to achieve in semester 2. This report has
to be approved by the supervisor before submission.
• The essay/project should be both a discursive and a critical account of the selected
topic. It should be written at a level that an expert Applied Mathematician can be
expected to understand, though he/she need not be an expert in the field covered.
The work must contain substantial mathematical content.
• The essay/project should be based on some four to six original primary source arti-
cles, which themselves represent a substantial contribution to the topic. Secondary
sources, such as books, review papers, etc., should also be consulted and cited.
• Original research is not essential.
• The length of the essay/project should be between 40 to 60 typed A4 pages. Only
in exceptional circumstances, and after consultation with the supervisor, should the
essay exceed 60 pages. This number includes all figures, contents pages, tables,
appendices, etc. Computer programs essential to the work should be included (with
adequate commentary) as additional material in appendices.
• Students should be careful to provide full and correct referencing to all material
used in the preparation of essays and projects. Be explicit in stating what is your
contribution and what is someone else’s contribution. Avoid quoting verbatim unless
reinforcing an important point.

9
 • Three examiners will be appointed to assess each essay/project. One of these exam-
iners will be the student’s supervisor. Although marking schemes may differ, marks
will generally awarded for:

(i) selection and synthesis of source material;

(ii) evidence of understanding;
(iii) evidence of critical ability;
(iv) clarity, style and presentation;
(v) mathematical and/or modelling expertise.

• Students are advised to read the pamphlet entitled “Guide to Essay Writing for
Science Students” available from the Science Faculty Office.
• The preferred method of typesetting mathematical documents these days is using
LATEX. This is available from the computers at the School. Students are recom-
mended to use LATEX in typesetting their Essays/Projects. Additional information
on LATEX is available from the Course Coordinator.
• Students who have worked on their essays or project topics as Vacation Scholars are
required to make a declaration to that effect in the preface of their essay/project.

5.4 Suggested Topics for the Essay/Project

The following is a list of possible essay/project topics for Applied Mathematics 4 students
in 2020 . Prospective students interested in any of these topics are encouraged to discuss
them with the named supervisors as early as possible.
However, this list is not exhaustive. Students may wish to suggest their own topics for
essays or projects. Before commencing work, however, each student must find a member
of staff who will agree to supervise the essay/project. For topics other than those listed
below, the student and supervisor must submit a brief written outline of the proposed
project or essay for approval by the Course Coordinator.

Asymmetric information in modelling of financial markets (Essay or Project)

Supervisor: Dr Anna Aksamit (Carslaw room 606; phone 9351 2979)
Differing levels of information available to agents are an essential aspect of decision-
making problems. It is natural to conjecture that an agent who possesses more information
in a financial market typically gains advantage over less-informed agents.
In this project, you will investigate mathematical framework used to formalise the idea
of different levels of information available to agents in a stochastoc model of a financial
market and quantify the role of information in various decision-making problems. The
essential tool to address these problems is the enlargement of filtration theory, which is
an actively researched area of stochastic analysis. You will learn the recent results from
that theory and you will apply them to solve utility maximisation problems incorporating
two distinct information flows. The related literature includes the recent monograph by
Aksamit and Jeanblanc (2017) and the paper by Chau et al. (2018).
10
Optimisation problems for stochastic stopping games (Essay or Project)
Supervisor: Dr Anna Aksamit (Carslaw room 606; phone 9351 2979)
Game theory, which focuses on solutions to optimisation problems with interactions
between agents, has undoubtedly plenty of applications in various fields, in particular,
economics and finance.
In this project, you will study a particular class of two-player zero-sum stopping games,
which are also known as Dynkin games, where the players choose their respective optimal
stopping rules. The main theoretical goal of the essay/project is to investigate the most
essential notions with optimal stopping rules, such as the value of the game (notice that
even the existence of the value of the game is by no means ensured a priori) and the
classical concept of the Nash equilibrium. In the next step, it is also possible to analyse
applications to arbitrage-free pricing of some classes of game options. You may pursue to
investigate several variations of the game, including asymmetric information or the case
of agents with heterogeneous beliefs. The related literature includes the recent papers by
Ekström et al. (2017), Esmaeeli et al. (2018), Gensbittel and Rainer (2018), and Grün
(2013).

Statistical laws, information theory, and natural language (Project)

Supervisor: A/Prof. E. Altmann (Carslaw room 615; phone 9351-2448)
How dissimilar are texts from different authors, epochs, or literary genres? Information
theory provides a quantitative and mathematically well-defined answer to these questions.
For instance, suppose a single (random) word is shown to you and your task is to guess
from which of two texts this word was extracted from. The dissimilarity between the two
texts is then quantified as the (expected) amount of (Shannon) information you obtain
from the word to solve your task. While these and other natural-language questions
have long been a main motivation for information theory, the recent availability of large
amounts of texts (e.g., on the Internet and due to the digitization of information) opens
new possibilities of investigations and applications. In this project we will explore the
consequences of statistical regularities observed in large datasets of written language (e.g.,
Zipf’s law) to the computation of similarity measures. Example of questions we will
address from an information-theoretic perspective are: how similar are spoken and written
English? Is Australian English more similar to British or to American English? Given
a series of observations of two sources, how can we quantify the extent into which one
source is affected by the other? This project requires programming skills and the interest
in applications of statistical methods.

Fractals in transiently chaotic systems (Project)

Supervisor: A/Prof. E. Altmann (Carslaw room 615; phone 9351-2448)
The most interesting dynamics of dynamical systems often happens during the tran-
sient time it needs to reach an asymptotic state which is often trivial (e.g., the rest state of
autonomous systems with friction). This project will investigate how chaos and fractality
are defined and quantified during such transients. After a revision of the main concepts

11
of transient chaos theory, it will focus on one of two areas of current research: (i) the
appearance of fractal basin boundaries in undriven systems showing ”double transient
chaos”; or (ii) optical billiards in which the intensity of light is partially reflected and
partially transmitted. The challenge in both cases is to to generalize and apply measures
of fractal dimension used in traditionally transiently-chaotic systems (in particular, the
uncertainty dimension). The project involves a combination of analytical calculations and
numerical simulations.

Data and Models in Complex Networks (Project)

Supervisor: A/Prof. E. Altmann (Carslaw room 615; phone 9351-2448)
This project will be co-supervised with Dr. Lamiae Azizi. The field of complex net-
works combines mathematical models for the generation of networks with the empirical
analysis of network data. Crucial in this approach is the identification of statistical regu-
larities that appear ”universally” in various types of networks and that serve as motivation
to consider generative models that explain these regularities. Paradigmatic examples are
the small world properties and the scale-free distribution of node degrees. The modern
availability of large amounts of data that can be represented in form of networks bring
new opportunities and challenges for the mathematical modeling of complex networks.
In particular, a more rigorous connection between the statistical analysis of the data and
the mathematical generative models is crucial. The goal of this project is to re-frame
previously established connections between statistical analysis and generative models as
a (Bayesian) inference problem on suitably defined probabilistic network models.

Analysis and Simulation of Simpler Mechanical Systems (Project)

Supervisor: Prof H. Dullin (Carslaw room 714; phone 9351 4083)
This project is about the dynamics of mechanical system with few degrees of freedom
(for example the double or triple pendulum, the Foucault pendulum, a spinning top, the
Levitron, ... I am open to your suggestions, please come talk to me if you have further
ideas/suggestions). The goal of the project is to get a good understanding of the dynamics
of the chosen systems. Depending on the example different techniques will be applied,
ranging from analytical computations to the the numerical integration of the problem.

Simulating a Closed Chain of Planar Rigid Bodies (Project)

Supervisor: Prof H. Dullin (Carslaw room 714; phone 9351 4083)
A chain of planar rigid bodies is a simple mechanical system with n segments connected
by joints that allow free rotation. Connecting the first segment to the last by another
joint gives a closed chain. Since the distance between the joints is fixed the closed chain
has n degrees of freedom. Reduction by translations and rotations leaves n − 3 degrees
of freedom specifying the shape. For certain parameters the dynamics of this system is
chaotic in the sense of Anosov. The goal of the project is the numerical solution of the
equations of motion for the case of the five-gon, with the ultimate goal of finding many
periodic orbits.

12
Rotation number near the 1:3 resonance (Project)
Supervisor: Prof H. Dullin (Carslaw room 714; phone 9351 4083)
Consider a Hamiltonian system with two degrees of freedom near an equilibrium point
with frequencies (of the harmonic oscillator approximation) in 1:3 resonance. Every such
system can be approximated by the so called resonant Birkhoff normal form. This resonant
Birkhoff normal form is an integrable system, and the goal of the project is to compute the
rotation number (i.e. the ratio of frequencies) of the motion for general initial condition
near the equilibrium point. These are given by complete elliptic integrals, which need to
be analysed.

Symplectic integration of the regularised planar ciruclar restricted 3 body problem (Project)
Supervisor: Prof H. Dullin (Carslaw room 714; phone 9351 4083)
The restricted three body problem describes the motion of a test particle in the field
of two heavy masses rotating around each other in circular orbits. The problem has a sin-
gularity when the test particle collides with either of the other masses. The collision can
be regularised, such that the solutions are defined for all times. The goal of the project
is to construct and implement a symplectic integration method for the regularised prob-
lem. This can then be used to study periodic orbits in this chaotic system, in particular
including collision orbits.

Designing optimal contract (Project)

Supervisor: Prof B. Goldys (Carslaw room 709; phone 9351-2976)
Contract theory is part of economics focused on designing optimal contracts between
principals and agents. Only recently, mathematical methods have become available in this
area. The breakthrough has been made by Sannikov in his seminal paper: A continuous-
time version of the principal-agent problem. Rev. Econom. Stud. 75 (2008) 957-984.
We will study the Sannikov paper and will try to extend it to more realistic situations.
Familiarity with stochastic analysis is necessary for this project.

Stochastic PDEs in credit modelling (Project)

Supervisor: Prof B. Goldys (Carslaw room 709; phone 9351-2976)
Pricing and hedging of credit derivatives is one of the most challenging problems of
Mathematical Finance. A standard tool to build mathematical models in this area is the
theory of copula functions. A major deficiency of this method is that it is difficult to
incorporate dynamics in the model. This problem has been recently addressed by using
the mean field method well known in Physics, see Bush, N.; Hambly, B. M.; Haworth, H.;
Jin, .; Reisinger, C. Stochastic evolution equations in portfolio credit modelling. SIAM J.
Financial Math. 2 (2011), no. 1, 627-664. Using this approach we can derive a stochastic
partial differential equation that describes time evolution of the fraction f(t,x) of agents
that did not default before a given time and are in a distance x to the default. The aim
of this project is to study this approach and develop more realistic models.

13
Stochastic differential equations with memory and applications to stochastic volatility mod-
els (Project)
Supervisor: Prof B. Goldys (Carslaw room 709; phone 9351-2976)
IA standard assumption of the Black-Scholes model for pricing of options is that the
volatility of the asset, if random, is driven by Wiener process. However, recent care-
ful analysis of market data show that the volatility is a much rougher process than
Wiener process, see for example Gatheral, Jaisson, Rosenbaum. Volatility is rough,
https://arxiv.org/abs/1410.3394 , In this project we will investigate a simple model of
stock prices in the case when volatility is modelled as a solution to a stochastic differ-
ential equation with memory. The main question will be how to derive (approximate)
option prices and hedges.

Reaction-diffusion equation on a half-line driven by the boundary noise (Project)

Supervisor: Prof B. Goldys (Carslaw room 709; phone 9351-2976)
Partial differential equations with random boundary conditions arise in many problems
of Science and Engineering. In this project we will study an important nonlinear PDE
with random boundary conditions, for details see from Brze?niak, Zdzis?aw; Goldys, Ben;
Peszat, Szymon; Russo, Francesco Second order PDEs with Dirichlet white noise boundary
conditions. J. Evol. Equ. 15 (2015), no. 1, 1?26. We will focus on the existence and
uniqueness of stationary states and the rate of convergence to equilibrium. The project
will extend some results

Linear processes driven by space-time homogeneous noise (Project)

Supervisor: Prof B. Goldys (Carslaw room 709; phone 9351-2976)
Processes evolving randomly in space and time are often described using partial dif-
ferential equations driven by external (independent of the solution) noise. Such PDEs
are nontrivial even in the linear case and they are frequently applied to model random
phenomena in physics, fluid dynamics, biology and engineering, see for example Sturm,
Anja On convergence of population processes in random environments to the stochastic
heat equation with colored noise. Electron. J. Probab. 8 (2003), no. 6.
In this project we will focus on linear PDEs driven by the so-called space-time homo-
geneous noise. Such an assumption leads to a rich class of examples useful in applications
but still rather poorly understood. The project requires knowledge of Fourier analysis
and some basic functional analysis.

Data-driven modelling: Finding models for observations in finance and climate (Project)
Supervisor: Prof G. Gottwald (Carslaw room 625; phone 9351-5784)
When given data, which may come from observations of some natural process or data
collected form the stock market, it is a formidable challenge to find a model describing
those data. If the data were generated by some complex dynamical system one may try
and model them as some diffusion process. The challenge is that even if we know that the
data can be diffusive, it is by no means clear on what manifold the diffusion takes place.

14
This project aims at applying novel state-of-the-art methods such as diffusion maps and
nonlinear Laplacian spectral analysis to determine probabilistic models. You will be using
data from ice cores encoding the global climate of the past 800kyrs as well as financial
data. In the latter case you might be able o recover the famous Black-Scholes formula
(but probably not).

This project requires new creative ideas and good programming skills.

Optimal power grid networks and synchronisation (Project)

Supervisor: Prof G. Gottwald (Carslaw room 625; phone 9351-5784)
Complex networks of coupled oscillators are used to model systems from pacemaker
cells to power grids. Given their sheer size we need methods to reduce the complexity
while retaining the essential dynamical information. Recent new mathematical method-
ology was developped to describe the collective behaviour of large networks of oscillators
with only a few parameters which we call ?collective coordinates?. This allows for the
quantitative description of finite-size networks as well as chaotic dynamics, which are both
out of reach for the usually employed model reduction methods.
You will apply this methodology to understand causes of and ways to prevent glitches
and failure in the emerging modern decentralised power grids. As modern societies in-
crease the share of renewable energies in power generation the resulting power grid be-
comes increasingly decentralised. Rather than providing a power supply constant in time,
the modern decentralised grid generates fluctuating and intermittent supply. It is of
paramount importance for a reliable supply of electric power to understand the dynamic
stability of these power grids and how instabilities might emerge. A reliable power-grid
consists of well-synchronised power generators. Failing to assure the synchronised state
results in large power outages as, for example, in North America in 2003, Europe in 2006,
Brazil in 2009 and India in 2012 where initially localised outages cascade through the grid
on a nation-wide scale. Such cascading effects are tightly linked to the network topology.
Modern power-grids face an intrinsic challenge: on the one hand decentralisation was
shown to favour synchronisation in power grids, on the other hand decentralised grids are
more susceptible to dynamic perturbations such as intermittent power supply or overload.
The project uses analytical methods as well as computational simulation of models for
power grids. You will start with a simple network topology and then, if progress is made,
use actual power grid topologies.

Data assimilation in numerical weather forecasting and climate science (Project)

Supervisor: Prof G. Gottwald (Carslaw room 625; phone 9351-5784)
Data assimilation is the procedure in nuemrical weather forecasting whereby the infor-
mation of noisy observations and of an imperfect model forecast with chaotic dynamics,
we cannot trust, is combined to find the optimal estimate of the current state of the
atmosphere (and ocean). Data assimilation is arguably the most computationally costly
step in producing modern weatehr forecasts and has been topic of intense research in the
last decade. There exists several approaches, each of which with their own advantage and
disadvantages. Recently a method was introduced to ada[tively ?pick? the best method

15
to perform data assimilation. This method employs a switch which, although it seems to
work, has not been linked to any theoretical nor physical properties of the actual flow.
This project will be using toy models for the atmosphere to understand the witch with
the aim of improving the choice of the switching parameter.

Networks of coupled oscillators (Project)

Supervisor: Prof G. Gottwald (Carslaw room 625; phone 9351-5784)
Many biological systems are structured as a network. Examples range from micro-
scopic systems such as genes and cells, to macroscopic systems such as fireflies or even an
applauding audience at a concert.
Of paramount importance is the topography of such a network, ie how the nodes, let’s
say the fireflies, are connected and how they couple. Can they only see their nearest
neighbours, or all of them. Are some fireflies brighter than others, and how would that
affect the overall behaviour of a whole swarm of fireflies? For example, the famous ’only
6 degrees of separation’-law for the connectivity of human relationships is important in
this context.
In this project we aim to understand the influence of the topography of such a net-
work. Question such as: How should a network be constructed to allow for maximal
synchronization will be addressed.
This project requires new creative ideas and good programming skills.

Weather derivatives (Project)

Supervisor: Prof G. Gottwald (Carslaw room 625; phone 9351-5784)
Weather derivatives were introduced to safe guard economies which severely depend
on the weather. The question we propose in this project is whether the value of weather
derivatives has any information about the actual weather which will occur; in other words:
Can they be used to do weather forecasting. The trader probably uses different forecast
models from different weather centres across the world to determine the value; can the
value of a weather derivative be understood as a multi-model forecast. We will use real
data.
This project requires new creative ideas and good programming skills. Some exposure
to statistical ideas and Bayesian methodology would be desirable but is not essential.

Complexity analysis of bi-rational maps (Project)

Supervisor: Dr G. Gubbiotti (Carslaw room 630; phone 9351-3879)
In this project we aim to study the behaviour of some bi-rational maps of the complex
projective space into itself. These maps can be derived in various ways, e.g. as stationary
form of some known differential-difference equations. Indeed, it is known that this kind of
maps can have only three possible behaviours: chaotic, integrable or linearisable. These
beahaviours are characterised by the rate of growth of the iterates of these equations, i.e.
by measuring their complexity. With this analysis we want to uncover the underlying
structures of these maps which is the first step to insert those equations in the general
theory.
The project is an hands-on one: the student will be taught the basic theory on algebraic
entropy and will be introduced to the Julia programming language as fundamental tool
to make explicit computations.

16
Symmetries and integrability of discrete and semi-discrete equations (Essay or Project)
Supervisor: Dr G. Gubbiotti (Carslaw room 630; phone 9351-3879)
Symmetry is one of the most proficient concepts in mathematical physics. In some
sense, we expect the reality to follow some symmetry rules, so symmetric systems have a
special place beneath the general ones. The concept of of symmetry for differential was
introduced in the XIX Century by the Norvegian mathematician Sophus Lie. Nowadays
symmetries have been defined almost for all kind of functional equations, especially for
semi-discrete and completely discrete equations which are my main area of expertise.
A specific class of symmetries, known as generalised symmetries, also proved to be
one of the possible keys to define integrability both in the continuous and in the discrete
setting. An essay in this field would provide a critical survey of the ways of computing
generalised symmetries and on the possible use of them in proving integrability. A project
in this field may require to use symmetries to prove the integrability of some equation or
to derive some particular solutions.

Periodic orbits in classical mechanics (Project)

Supervisor: Dr G. Gubbiotti (Carslaw room 630; phone 9351-3879)
A famous theorem proved by Bertrand in the XIX states that the only radial forces
admitting closed and periodic orbits are the Kepler problem and the isotropic harmonic
oscillator. These two have the common property of being maximally superintegrable, i.e.
they admit the maximum number possible of first integrals. Nekhoroshev in the 70ies
proved that this property is shared by every maximally superintegrable system.
Based on this result a simple perturtibative test to check weather or not a system is
maximally superintegrable has been proposed. In this project we plan to apply this test
as a sieve test to some relevant examples to uncover their properties.

Hidden symmetries, linearisation and superintegrability (Essay)

Supervisor: Dr G. Gubbiotti (Carslaw room 630; phone 9351-3879)
It is well known that the Kepler problem can be brought in the form of a linear harmonic
oscillator. It is less known that this property is shared by many superintegrable systems.
This kind of linearisation are usually linked to the existence of hidden symmetries, i.e.
non-local symmetries that allow to linearise the system.
An essay on this subject will cover all the relevant known cases and give a critical
survey of the state of the art on both linearisation and on the methods to find non-linear
symmetries.

Discrete soliton equations (Project)

Supervisor: Prof N. Joshi (Carslaw room 629; phone 9351-2172)
Famous PDEs such as the Korteweg-de Vries equation (which have soliton solutions)
have discrete versions (which also have soliton solutions). These discrete versions are equa-
tions fitted together in a self-consistent way on a square, a 3-cube or an N-dimensional
cube. These have simple, beautiful geometric structures that provide information about
many properties: solutions, reductions to discrete versions of famous ODEs, and deeper
aspects such as Lagrangians. This project would consider generalisations of such struc-
tures and/or properties of the solutions, such as finding their zeroes or poles.

17
Integrable discrete or difference equations (Essay or Project)
Supervisor: Prof N. Joshi (Carslaw room 629; phone 9351-2172)
The field of integrable difference equations is only about 20 years old, but has already
caused great interest amongst physicists (in the theory of random matrices, string the-
ory, or quantum gravity) and mathematicians (in the theory of orthogonal polynomials
and soliton theory). For each integrable differential equation there are, in principle, an
infinite number of discrete versions. An essay in this area would provide a critical survey
of the many known difference versions of the classical Painlevé equations, comparisons
between them, and analyse differing evidence for their integrability. Project topics would
include the derivation of new evidence for integrability. The field is so new that many
achievable calculations remain to be done: including derivations of exact solutions and
transformations for the discrete Painlevé equations.

Exponential asymptotics (Project)

Supervisor: Prof N. Joshi (Carslaw room 629; phone 9351-2172)
Near an irregular singular point of a differential equation, the solutions usually have
divergent series expansions. Although these can be summed in some way to make sense
as approximations to the solutions, they do not provide a unique way of identifying a
solution. There is a hidden free parameter which has an effect like the butterfly in chaos
theory. This problem has been well studied for many classes of nonlinear ODEs but
almost nothing is known for PDEs and not much more is known for difference equations.
This project would include studies of a model PDE, like the famous Korteweg-de Vries
equation near infinity, or a difference equation like the string equation that arises in 2D
quantum gravity.

Cellular Automata (Essay or Project)

Supervisor: Prof N. Joshi (Carslaw room 629; phone 9351-2172)
Cellular automata are mathematical models based on very simple rules, which have
an ability to reproduce very complicated phenomena. (If you have played the Game of
Life on a computer, then you have already seen automata with complicated behaviours.)
This project is concerned with the mathematical analysis of their solutions, which lags
far behind corresponding develop- ments for differential or difference equations.
In this project, we will consider a family of cellular automata called parity filter rules,
for which initial data are given on an infinite set. For example, consider an infinitely long
train of boxes, a finite number of which have a ball inside, whilst the remainder are empty.
At each time step, there is a simple rule for moving the leftmost ball in a box to the next
empty box on the right. Continue until you have finished updating all nonempty boxes in
the initial train. (Try this out for yourself with adjacent boxes with three balls, followed
by two empty boxes and then two boxes with balls inside. What do you see after one
update? Two updates?) It turns out that these box-and-ball systems replicate solitons,
observed in solutions of integrable nonlinear PDEs. In this project, we will consider how
to derive parity filter rules from nonlinear difference equations, and how to analyse their
solutions. One direction for the project is to analyse the solutions as functions of initial
data. Another direction is to develop ways to describe long-term behaviours.

18
Other Possible Research Topics (Essay or Project)
Supervisor: Prof N. Joshi (Carslaw room 629; phone 9351-2172)
Additional projects under the supervision of Prof Joshi include research topics in the
areas of dynamical systems and mathematical physics. Please contact Prof Joshi for
further details.

Modelling the evolution of human post-menopausal longevity and pari bonding (Project)
Supervisor: A/Prof P. Kim (Carslaw room 621; phone 9351-2970)
A striking contrast between humans and primates is that human lifespans extend well
beyond the end of the female reproductive years. Natural selection favours individuals
with the greatest number of offspring, so the presence of a long female post-fertile period
presents a challenge for understanding human evolution.
One prevailing theory that attempts to explain this paradox proposes that increased
longevity resulted from the advent of grandmother care of grandchildren. We have de-
veloped preliminary age-structured PDE models and agent-based models to consider the
intergenerational care of young proposed by this Grandmother Hypothesis. The project
will involve extending the models to consider whether the presence of grandmothering
could increase the optimum human longevity while simultaneously maintaining a rela-
tively early end of fertility as seen in humans (and killer whales).
Analytical approaches will involve developing numerical schemes for the PDEs and
analytically and numerically studying the steady state age distributions and growth rates
of the populations with and without grandmothering and under different life history pa-
rameters, e.g. longevity and end of fertility.
We have now also begun to explore mating strategies, especially pair bonding, yet
another unique human characteristic among mammals. Speculations about how pair
bonding developed from our ancestral roots abound and are open to being quantified,
modelled, and analysed. Like the grandmothering models, these investigations will involve
PDEs or agent- based models.

Modelling cancer immunotherapy (Project)

Supervisor: A/Prof P. Kim (Carslaw room 621; phone 9351-2970)
A next generation approach to treating cancer focuses on cancer immunology, specif-
ically directing a person’s immune system to fight tumours. Recent directions in cancer
immunotherapy include

• Oncolytic virotherapy: infecting tumours with genetically-engineered viruses that

preferentially destroy tumour cells and induce a local anti-tumour immune response,

• Preventative or therapeutic cancer vaccines: stimulating a person’s immune system

to attack tumour colonies to prevent or hinder tumour development,

• Cytokine therapy: using immunostimulatory cytokines to recruit immune cells and

enhance existing anti-tumour immune responses.

19
 These treatments can be used alone or in combination with each other or with other
forms of treatment such as chemotherapy. Since immunotherapy often involves immune
responses against small tumours, often close to inception, they are highly spatially de-
pendent and often probabilistic.
The goal of the will be to develop differential equation and possibly probabilistic agent-
based models to understand the tumour-virus-immune dynamics around a small, devel-
oping tumour and determine conditions that could lead to effective tumour reduction or
complete elimination. The project will involve developing the models and schemes for
numerically simulating the ODE and PDE systems, and if possible, performing a stability
analysis of the ODE system.

The Schrödinger Equation on Quantum Graphs (Essay or Project)

Supervisor: Dr R. Marangell (Carslaw room720; phone 9351-5795)
A metric graph is a (usually planar) set of vertices and edges where the edges have a
defined length. A “quantum graph” is a metric graph equipped with a differential operator
(called a Hamiltonian ) and some boundary conditions on the vertices of the graph.
The study of quantum graphs involves aspects from a wide variety of mathematical
areas, including graph theory, combinatorics, spectral theory, mathematical physics, PDEs
and spectral theory.
This is essay or project will focus on the study of the Schrödinger operator

H = −∂xx + V (x)

on Quantum Graphs. We will be interested in the spectrum (eigenvalues and eigenfunc-

tions) of the Schrödinger operator for various potentials V(x). This will depend heavily
on the shape of the graph - and in general the spectrum will be very different for trees
than it will be for graphs with cycles. It will also be different depending on whether the
graph is bounded or not. This project will study as many of these cases as possible.
The spectrum of the Schrödinger equation on a quantum graph will share spectral prop-
erties of operators in both one and two dimensions, and as such some pretty interesting
things can happen. In particular compactly supported eigenfunctions of the Schrödinger
operator can exist on a quantum graph which is not a thing that can happen for (a lot
of) ODEs (whenever there is uniqueness of solutions).

PDEs on Quantum Graphs (Essay or Project)

Supervisor: Dr R. Marangell (Carslaw room720; phone 9351-5795)
See the previous entry for the definition of a Quantum Graph. This is actually a series
of projects. For this type of project we’ll pick one of the following (depending on your
tastes/random chance):
• The Heat Equation (this will be similar to the Schrödinger equation project above)

• The Wave Equation

• Reaction-diffusion equations (with or without advection - this is a huge number of

PDEs)

20
 • The Korteweg de Vries equation
• The Nonlinear Schrödinger Equation
• The sine-Gordon equation
• Any other interesting PDE
The project will focus on the study of such a PDE on graphs. We will investigate
application, steady states of such a graph and their stability/instability, other interesting
solutions that might arise, how to construct numerical simulations on Graphs, and how
the geometry/topology of the graph plays a role in all of this. Much of this project will
depend on your taste, so if you are interested in this type of thing, please email me and
we can meet to discuss more details.

Periodic wave trains in the nonlinear Klein Gordon Equation (Essay or Project)
Supervisor: Dr R. Marangell (Carslaw room720; phone 9351-5795)
The 1+1 nonlinear Klein Gordon equation is

utt − uxx + f 0 (u) = 0

Where f 0 (u) is some (nice enough) function called the potential. For a wide class of
potentials, the solutions exhibit periodic travelling wave trains.
Passing to a moving frame and linearising leads to periodic travelling wave equations
in which the (temporal) spectral parameter is now part of an advective term.
A type of dynamic instability, called dynamical Hamiltonian-Hopf instabilities is known
for wavetrains travelling fast enough. This project will search for dynamical Hamiltonian-
Hopf instabilities and attempt to numerically compute the spectrum of the periodic trav-
elling wavetrains. Then we will numerically simulate the temporal dynamics of such
travelling waves. Because the spectrum of the periodic wavetrain enters the right half
plane away from the origin, it is not entirely understood how Hamiltonian-Hopf instabil-
ities will manifest themselves dynamically, and this work will provide useful information
for linking dynamic and spectral behaviour.

Symmetry Breaking in the defective Nonlinear Schrödinger equation with a potential (Es-
say or Project)
Supervisor: Dr R. Marangell (Carslaw room720; phone 9351-5795) This project will exam-
ine how solutions to the Nonlinear Schrödinger (NLS) equation with two defects bifurcate
with the addition of a potential. Solutions to the NLS equation with a pair of defects can
be constructed via phase portrait techniques, and one can see that a family of symmet-
ric states can be found, some of which can be shown to be stable. Similarly, the linear
Schrödinger equation with double-well potential also admits stable symmetric solutions.
There are other, unstable solutions in both cases. The purpose of this project is to ex-
amine what happens when we use both of these equations to perturb off of each other. If
the magnitude of the potential and defects are chosen carefully, then we should be able to
observe that each of these cases can be thought of as a bifurcation of the other. Hopefully
some new and interesting solutions will emerge.

21
The vortex filament equation and solitons in firenados (Essay or Project)
Supervisor: Dr R. Marangell (Carslaw room720; phone 9351-5795)
The vortex filament equation is the relatively simple-to-state PDE for a curve in 3-D

γt = γx × γxx .

This equation models a thin rotating vortex, like a water sprite, a tornado, or even a
firenado. It turns out that this equation is transformable into the Nonlinear Schrödinger
(NLS) equation. What this means is that solitons exist (and indeed have even been found)
within this equation. Moreover, because the NLS has exact solutions, there is a range of
closed solutions to this equation. These closed solutions turn out to be torus knots. This
project will explore all of these features, as well as compute, simulate and plot closed form
solutions and solitons of thin rotating vortices. Differential Geometry would be a plus for
this project as well as familiarity with at least some form of computational software tool,
though these are not explicitly necessary.

Chemotaxis in models with zero diffusivity (Essay or Project)

Supervisor: Dr R. Marangell (Carslaw room720; phone 9351-5795) Chemotaxis is the
movement of a cell via advection either towards or away from a chemical source. It has
been used in many biological models, from slime-moulds to motile bacteria, to roadway
construction by humans. Typically linear diffusivity has been studied, but lately models
where the diffusivity is allowed to go to zero have become of interest. This project
will examine the existence of travelling wave solutions in such models, as well as some
elementary stability properties of such solutions.

Stability in a model of herd grazing and chemotaxis (Essay or Project)

Supervisor: Dr R. Marangell (Carslaw room720; phone 9351-5795) This project will ex-
amine a model of the formation of a herd of grazing animals. The model will focus on
two major factors, how the animal seeks food and how the the animals interact with each
other. Remarkably, the model shares many properties with another, well studied model,
that of so-called bacterial chemotaxis. The aim of this project will be to analyse, both
numerically and analytically, such a model, and to understand certain special solutions
in the model, called travelling waves, as well as their stability.

Absolute and convective instabilities (Essay or Project)

Supervisor: Dr R. Marangell (Carslaw room720; phone 9351-5795) Absolute and convec-
tive instabilities are instabilities that result from points in the essential spectrum crossing
into the real axis. This project will focus on several toy problems which explore what hap-
pens as such instabilities are present. The aim of this project is twofold - first to explore
the mechanisms that lead to such spectral instabilities, and then secondly to understand
the dynamic implications of such instabilities.

22
Absolute spectrum of St. Venant roll waves (Essay or Project)
Roll waves are a phenomenon that occurs when shallow water flows down an inclined
ramp. Mathematically they can be modelled by the St. Venant equations. Typically
roll waves occur as periodic solutions, however if they are far enough apart, they can be
treated as solitary waves. In this case, the spectrum of the linearised operator governs
their dynamics, and in particular, their stability properties. This project will focus on
computing the absolute and essential spectrum of these solitary waves. Medium computa-
tional skills are required for this project. Supervisor: Dr R. Marangell (Carslaw room720;
phone 9351-5795)
Perron-Frobenius methods in PDEs (Essay or Project)
Supervisor: Dr R. Marangell (Carslaw room720; phone 9351-5795) The Perron-Frobenius
theorem is a cute theorem from linear algebra, which states that a matrix with only
positive entries, has a simple largest eigenvalue, and that the eigenvector for this eigenvalue
contains only positive entries. Moreover, this is the only eigenvalue whose eigenvector has
this property. This theorem can be extended to linear operators, and as such can be used
to gain information about the spectral stability of solutions to PDEs. This project will
begin by investigating applications of this to simple, scalar PDEs, and will then move up
to systems. In particular, investigation of stability in population models which involve an
allee effect will be particularly amenable to these methods.
A discrete and continuous dictionary (Essay or Project)
Supervisor: Dr R. Marangell (Carslaw room720; phone 9351-5795) There are many results
in the field of second order linear difference equations, and second order linear ODEs that
are the same in ‘spirit’. One such example is the Abel/Liouville/Jacobi identity, which
states that the determinant of a fundamental set of solutions obeys a straightforward first
order linear differential equation. The analogue is that the determinant of a fundamental
set of solutions to a second order difference equation obeys a first order difference rule.
This project will attempt to highlight which results should be connected, and for theorems
with certain hypotheses, how to translate one into the other.
The history of continued fractions (Essay)
Supervisor: Dr R. Marangell (Carslaw room720; phone 9351-5795) The origin of continued
fractions is not well known, though it is traditionally placed at around the time of Euclid’s
algorithm. By manipulating the Euclidean algorithm, one can derive a simple continued
fraction of the rational number pq . Continued fractions gained much interest after John
Wallis published his famous identity in 1655
4 3 × 3 × 5 × 5 × 7 × ···
= ,
π 2 × 4 × 4 × 6 × 6 × ···
and it was observed that the expression could be transformed into a continued fraction.
Continued fractions continued to be explored through the 19th century with incredibly
prominent mathematicians of the age such as Hermite, Jacobi, Gauss, Stieltjes and Cauchy
making contributions. Continued fractions (in some form) are still of interest today in the
fields of orthogonal polynomials as well as special functions, and integrable systems. This
essay will explore the use and theory behind continued fractions from their ‘inception’
nearly 2000 years ago up through the current time.

23
The history of the integral (Essay)
Supervisor: Dr R. Marangell (Carslaw room720; phone 9351-5795)
This essay will follow the history of the integral from the method of exhaustion of the
ancient greeks, up through the modern notions of the Lebesgue integral. From here the
project can go in a couple of different ways - either into modern notions of stochastic
and probablistic integration, into modern notions of integral transforms, or into computer
algorithms for symbolic and numeric integration. Perhaps unsurprisingly, there aren’t
many books available on this subject, so the student will be expected to find most of
their own sources for this project, but the Wikipedia articles ‘Integral’ and ‘History of
Calculus’ and the references therein are a good place to start. If this project is chosen,
some care must be taken to ensure that enough mathematics in included in the essay.

Other Possible Research Topics (Essay or Project)

Supervisor: Dr R. Marangell (Carslaw room720; phone 9351-5795)
Other projects under the supervision of Dr Marangell include topics in the areas of
nonlinear standing or travelling waves, topics in the application of geometric and topolog-
ical methods in dynamical systems and PDEs, and other research topics in the history of
mathematics and science in general. Examples of nonlinear standing and travelling waves
come from models in a wide range of areas which include mathematical biology, chemistry
and physics. More specific examples would be standing/travelling waves in population
dynamics, combustion models, and quantum computing, but really there are many, many
examples, so please contact Dr Marangell for further details.

PDE models for the distribution of ingested lipids in macrophages in atherosclerotic plaques.
(Project)
Supervisor: Prof M.R. Myerscough (Carslaw room 626; phone 9351-3724)
Atherosclerotic plaques are accumulations of lipid (fat) loaded cells and necrotic (dead)
cellular debris in artery walls. They are caused by LDL (which carries ‘bad cholesterol)
penetrating the blood vessel wall, becoming chemically modified (usually oxidised) and
setting off an immune reaction. In response to this immune reaction, macrophages (a
type of white blood cell) enter the artery wall and consume the modified LDL. In this
way macrophages accumulate lipids and as more LDL and more cells enter the vessel walls
the population of cells also grows. Other processes can affect the growth or regression of
the plaque, such as cell death, cells leaving the tissue and lipid export from inside cells to
HDL (which carries good cholesterol which is good because its been carried away from the
plaque). When atherosclerotic plaques grow very large and rupture they can cause heart
attacks and strokes which are one of the two leading causes of death in the developed
world. (The other is cancer.)
We have written a partial differential equation model for the accumulation of cells and
lipids in plaques. In this model, the number of macrophages in the plaque is a function of
both time t and accumulated lipid a. The primary equation is an advection equation with
nonlinear source and sink terms, including a term with an integral convolution that models

24
what happens when macrophages phagocytose (=eat) other macrophages that are dead
or dying. We have done an analysis of this model at steady state when all the processes
(lipid ingestion, macrophages leaving the plaque, the action of HDL) occur at a constant
rate. This project will build on this analysis and has the aim of producing numerical
solutions to the model when model processes are functions of a, the accumulated lipid
inside the cell.
This project is particularly suitable for students who are interested in applications of
mathematics to biomedical problems, have completed a third year unit on PDEs and have
at least some experience in coding in Matlab, C, Python or similar.

Poncelet Porisms (Essay)

Supervisor: Dr Milena Radnovic (Room 633; phone 9351-4543)
Suppose that two conics are given in the plane, together with a closed polygonal line
inscribed in one of them and circumscribed about the other one. The Poncelet porism
states that then infinitely many such closed polygonal lines exist and all of them with the
same number of sides. That statement is one of most beautiful and deepest contributions
of the XIX-th century geometry and has many generalisations and interpretations in
various branches of mathematics. In this essay, the student will present rich history and
current developments of the Poncelet porism.

Elliptical billiards and their periodic trajectories (Essay or Project)

Supervisor: Dr Milena Radnovic (Room 633; phone 9351-4543)
We consider billiards in a domain bounded by arcs of several conics belonging to a
confocal family. When the boundary of such a billiard does not contain reflex angles, the
system turns out to be integrable. Geometrically, the integrability has the following man-
ifestation - for each billiard trajectory, there is a curve, called caustic, which is touching
each segment of the trajectory. For elliptical billiards, the caustics are conics from the
same confocal family.
Integrability implies that the trajectories sharing the same caustic are either all periodic
with the same period or all non-periodic.
On the other hand, if there is at least one reflex angle on the boundary, the integrability
will be broken, although the caustics still exist. Such billiards are thus called pseudo-
integrable and there may exist trajectories which are non-periodic and periodic with
different periods sharing the same caustic.
An essay on this topic would provide a review of classical and modern results related to
the elliptical billiards. In a project, the student would explore examples of billiard desks.

Other Possible Research Topics (Essay or Project)

Supervisor: Dr Milena Radnovic (Room 633; phone 9351-4543)
Additional projects under the supervision of Dr Radnovic include research topics in
the areas of dynamical systems and mathematical physics. Please contact Dr Radnovic
for further details.

25
Representations of American options through European claims (Essay or Project)
Supervisor: Prof M. Rutkowski (Carslaw room 814; phone 9351-1923)
A challenging mathematical problem of great importance for the Finance industry
was formulated by Jourdain and Martini (2001): they were interested in pricing of path-
independent European claims in the Black-Scholes model and they discovered that, for a
large class of European payoffs, in the region where the European price increases with the
time to maturity, this price is equal to the American price of another claim. They also
furnished particular examples where the corresponding American claims can be computed
explicitly. However, the full characterisation of American claims obtained in this way
was, and still remains, an open question, although a first step towards verifying the
representability of certain American claims was recently done by Lenga (2017).
As shown by Jourdain and Martini (2002), some preliminary theoretical results con-
cerning the problem of representation of an American option through a European claim
can also be used to obtain new approximation schemes for prices of American options.
Recall that in the Black-Scholes model, no closed-form expression is available even for the
price of the standard American put and thus numerous numerical approximation methods
have been specifically designed for this problem. This essay/project combines an intrigu-
ing theoretical problem in the field of Financial Mathematics with numerical aspects of
arbitrage-free pricing of American options and thus it is particularly suitable for a student
with a working knowledge of computational methods.

Analytical valuation under funding costs and credit risk (Essay or Project)
Supervisor: Prof M. Rutkowski (Carslaw room 814; phone 9351-1923)
Prior to the global financial crisis of 2007-2008, financial institutions tended to ignore
the credit risk of highly-rated counterparties in valuing and hedging contingent claims
traded over-the-counter, that is, bilateral contracts negotiated between two default-prone
entities. Then, in just the short span of one month of 2008, several mainstream financial
institutions experienced critical credit events in a painful reminder of the default-riskiness
of even large names.
The credit valuation adjustment (CVA) and the funding valuation adjustment (FVA)
are the price adjustments due to the default risk and the cost of funding the trade. Trading
desks back the deal with a client by hedging it with other dealers in the market, and this
may involve maintaining a number of hedging accounts in the underlying assets, in cash,
or in other correlated assets when proxy hedging. Typically, the funds needed for these
operations are raised from the internal treasury of the dealer and, ultimately, they come
from external funders. Interest charges on all borrowing and lending activities need to
be covered and this affects the contract’s valuation. The causes of these adjustments are
accounted for at the level of the contract payoffs and the resulting all-inclusive price is
written as a solution to an advanced mathematical problems, such as semilinear PDEs or
BSDEs (see Bielecki et al. (2018)).
The challenging theoretical goal, which underpins this project, is to identify classes of
simple contracts and market models where this all-inclusive price of an uncollateralized
contract can be computed analytically yielding a closed-form expression. Some examples

26
of such contracts are given in the existing literature (see, for instance, Brigo et al. (2017)
or Bichuch et al. (2018)) but a lot of further research is needed to fully understand both
linear and nonlinear valuation adjustments.

Control of boundary-layer flows (Essay or Project)

Supervisor: A/Prof S. Stephen (Carslaw room 525; phone 9351-3048)
This project is in the field of hydrodynamic stability of boundary-layer flows where
viscous effects are important. The aim is towards understanding more fully the transition
process from a laminar flow to a turbulent one.
We will consider rotating flows which are relevant to the flow over a swept wing and
to rotor-stator systems in a turbine engine. Experiments show that the boundary layer
becomes unstable to stationary or travelling spiral vortices.
The project will investigate the effect of different surface boundary conditions on
boundary-layer flows over rotating bodies. Effects such as suction, partial-slip, com-
pliance and wall shape can be modelled. Suction, for example, is used to achieve laminar
flow control on swept wings.
The resulting system of governing ordinary differential equations will be solved nu-
merically for the basic flow, determining important values such as the wall shear. The
linear stability of these flows to crossflow instabilities will be investigated. These take the
form of co-rotating vortices, observed in experiments, and only occur in three-dimensional
boundary layers.
The flow for large Reynolds number, corresponding to large values of rotation, will be
considered. In this case the boundary layer thickness will be very small so asymptotic
methods of solution will be used. Different asymptotic regimes will need to be considered
and solutions obtained in each region. Matching the solutions between the regimes and
satisfying the boundary conditions will lead to an eigenrelation. Inviscid and viscous
instability modes will be considered.
The effect of the surface boundary conditions on the disturbance wave number and
wave angle will be determined. This will have applications in possible control of boundary
layers as boundaries causing stabilisation of the instabilities could lead to a delay in the
transition process from a laminar flow to a turbulent flow.

Inverse Problems for Partial Diffeential Equations - A Geometric Analysis Perspective

(Essay or Project)
Supervisor: Dr L. Tzou (Carslaw room 615; phone 9351-1917)
The analysis of mathematical models arising from various imaging techniques has be-
come a prominent research topic due to its potential applications in many fields including
oil exploration and early detection of malignant breast tumor. In many cases one asks
the question of whether one can deduce interior information about an object by mak-
ing measurements on its surface. Mathematically this amounts to recovering the model
parameters for a partial differential equation from boundary behavior of its solutions.
This project will explore how these mathematical models behave in various geometric
settings. We will see how this applied mathematics problem can motivate us to ask
interesting questions in both geometry and analysis.

27
Pattern Formation Through Geometric Microlocal Analysis (Essay or Project)
Supervisor: Dr L. Tzou (Carslaw room 615; phone 9351-1917)
This essay is under the supervisor’s ARC Discovery Project Probing the Earth and the
Universe with Microlocal Analysis and has the potential to continue to a PhD position
partially funded by ARC and jointly supervised by applied mathematicians at Macquarie
University. There can also be a computational component to the project.
The student will have opportunities to attend, subject to availability, the workshop or-
ganized by the supervisor at the Banff International Research Station and Fudan Confer-
ence on Microlocal Analysis. During the first semester student will also have opportunity
to interact with prominent mathematicians in related fields who are visiting through the
School’s International Visitor’s Program.
We model the effect of drought on vegetation patterns on variable terrains. Mathemat-
ically these are described by the dynamics and stability of localized patterns in reaction-
diffusion systems on surfaces of non-constant curvature. In systems such as the (rescaled)
Schnakenberg model which often serves as a prototype for such studies, the behaviour of
localized patterns depend on the asymptotic of the singularity for the Green’s function
of the Helmholtz operator near the singularity. In the general geometric setting, it is
often difficult to derive an explicit formula for the Green’s function (or even to compute
it numerically), which is why existing theory can only treat simple geometries such as the
disk and sphere.
In the language of microlocal analysis, however, this difficulty becomes a standard
parametrix construction. This recent breakthrough is allowing us to study pattern for-
mation on terrains which were previously inaccessible.

Non-equilibrium statistical mechanics of a chain: the birth of temperature. (Essay or

Project)
Supervisor: Dr G. Vasil (Carslaw room 627; phone 9351-4163)
If you release a simple, frictionless pendulum from a given height, it will oscillate
forever in periodic motion. If you attach a second pendulum at the bottom of the first,
the result becomes chaotic with all the associated rich behaviour. But the system will
still more-or-less swing back and forth forever.
If you attach 100, 1,000, or 1,000,000 more frictionless elements to the bottom of the
configuration something even more interesting happens. The swinging motion begins to go
away. The initial coherent kinetic energy transfers into the small-scale random fluctuations
of the individual elements. The transfer of energy is a process called ”thermalisation” (or
the birth of temperature), and the damping of the motion is the emergence of friction in
an otherwise dissipation-free system.
Even more fascinating, the system can generate configurational entropy, and this gives
rise to emergent entropic forces. This is a fancy way of saying that the initial inextensible
pendulum chain configuration turns into a hot bouncing bungee cord.
How does this happen exactly? How fast does temperature, entropy, friction, and
tension emerge? What can we predict about the statistics? What happens if we drive the
system externally? What does this say about hydrodynamical turbulence? This project

28
will investigate all any and all of these question and more. The project will start by
directly simulating the pendulum system numerically. After analysing the results, we can
use theoretical techniques from equilibrium and non-equilibrium statistical mechanics to
better understand these kinds of systems.
I also have a number of similar and related problems on non-equilibrium statistical
systems. You should come talk to me if this area interests you in general.
Computational projects with Dedalus. (Essay or Project)
Supervisor: Dr G. Vasil (Carslaw room 627; phone 9351-4163)
Dedalus is a multi-purpose solver than can simulate a broad array of nonlinear partial
differential equations (PDEs).
You can find out about Dedalus here:
http://dedalus-project.org
I have been leading the development of this project for the past few years. The idea
behind Dedalus is that the user can enter their equations in text-based format and experi-
ment with different behaviour much more easily than anything that came before. The code
runs on computers ranging from laptops to the world’s most powerful supercomputers.
And you can see some examples of past work done here:
http://vimeo.com/dedlaus
In the past, my collaborators and I have solved problems ranging from solar magneto-
hydrodynamics, the dynamics of boats in Arctic seas, buckling elastic beams, nonlinear
optics on topologically complex graphs, turbulent thermal convection, nuclear burning
flames in white dwarf stars, cancer modelling, and more.
I have a range of potential projects that students can investigate with Dedalus. Some
of these are extensions of past problems, and some are entirely new. I’m also happy to
talk to students who have a particular application in mind that they would like to try.
The idea is to simulate something interesting and model the behaviour after you know
the answer!
Improved numerical algorithms for multidimensional spectral methods. (Essay or Project)
Supervisor: Dr G. Vasil (Carslaw room 627; phone 9351-4163)
Simulating important physical, biological, finical, and mathematical dynamics usually
involves solving a large system of linear equations. An important key to many numerical
methods is finding a way to solve linear systems in as few operations as possible. The
first way to do this is to use a basis that produces matrices with as few non-zero elements
as possible. But even with very sparse matrices, then actual time to find the solution can
still differ by many orders of magnitude with different methods.
Two general classes of methods exist for solving large, sparse, linear systems. The first
class is ”direct methods”. These are things like Gaussian elimination that you commonly
learn in undergraduate courses. The second class are ”iterative methods”. Iteration
involves applying a matrix to a vector over and over again in a clever way that converges
to the desired solution. Both of these methods have their pros and cons.
The idea of this project is to study iterative methods in a new context. This project
has the potential to accelerate important computational algorithms by many orders of
magnitude. The results of this would contribute back to the code base of the Dedalus
project and apply to many different areas.

29
Possible Research Topics (Essay or Project)
Supervisor: Prof M. Wechselberger (Carslaw room 628; phone 9351-3860)
Projects under the supervision of Prof Wechselberger include research topics in the
areas of relaxation oscillators, return maps, physiological rhythms, mathematical neuro-
science, and more generally, dynamical systems.
These areas of research are concerned with the study of oscillatory patterns of so
called ‘slow/fast systems’. These systems are ubiquitous in nature and control most of
our physiological rhythms. E.g. one cycle of a heart beat consists of a long interval
of quasi steady state interspersed by a very fast change of state, the beat itself. This
very fast relaxation of energy leads to the notion of a relaxation oscillator and shows the
appearance of multiple time-scales (slow/fast) in the system. To study periodic solutions
in such systems one can analyse an associated return map. It is well known that already
1D maps can have a rich variety of dynamics, like periodic, quasiperiodic and chaotic
solutions, the most famous example is the 1D circle map. Plotting the winding number
versus a control parameter in this case yields the famous ‘devil’s staircase’.
For more information on possible topics, please see the research link on Prof Wechsel-
berger’s webpage http://www.maths.usyd.edu.au/u/wm/

Portfolio optimisation with liquid American options. (Essay or Project)

Supervisor: Dr Zhou Zhou (Carslaw room 527; phone 9351 2108)
As a classical problem in mathematical finance, portfolio optimisation has been stud-
ied extensively in various setups. American-style options, with huge trading volumes, are
perhaps the most popular financial derivatives in the financial market. However, so far
there have been very limited studies on portfolio optimisation involving liquidly traded
American options. Among the few works, Bayraktar and Zhou (2016) analyses a gen-
eral utility maximisation problem with the short-selling constraint on liquid American
options. They establish a duality result for the value function associated with the utility
maximisation problem.
In this essay/project, you will consider the portfolio optimisation in which stocks are
traded dynamically and American options are traded statically. Different from Bayraktar
and Zhou (2016), here we assume that American options can be sold. Due to the nature
of American options, investors who short the options will face the uncertainty of the
exercise times. How to deal with this uncertainty and what are the corresponding trading
strategies will be the core questions for this essay/project. In addition to probability,
stochastic processes and control, the project will also involve game theory. Hence it
will be particularly suitable for students who are interested in game theory and/or its
applications. For some relevant works for the project, see Bayraktar and Zhou (2014,
2017).

Mean-field stopping games (Essay or Project)

Supervisor: Dr Zhou Zhou (Carslaw room 527; phone 9351 2108)
Companies or individuals often need to decide the best times (stopping strategies) to
take certain actions in order to maximise their profits/utilities. When there is more than

30
one investor (player) involved, players will interact with each other. In this situation, how
the players choose stopping strategies is of particular interest (e.g., competing companies
choose times to enter/quit the market or update products). Such real-world problems can
be modelled as stopping games.
There have been a lot of works on stopping games with different applications. Tra-
ditional methods of game theory assume that each player keeps an eye on every other
player. As a result, roughly speaking, the dynamics of an n-player game can be repre-
sented by a system of (at least) n coupled differential equations (if such representation
exists). Unfortunately, when n is relatively large (say, n ≥ 10), the system of n equations
are usually computationally intractable. On the one hand, as can be seen from Zhou
(2015), the construction of a three-player stopping game is already very complicated. On
the other hand, in practice when there are a lot of players, each player may not keep
track of every other player’s behaviour. More likely, she may only pay attention to the
aggregate behaviour of the cohort of other players.
The above intractability and practical situation motivates us to seek an effective tool
to study stopping games with a large number of players, where the payoff of each player
depends on her own strategy and some relevant features of the aggregate behaviour of
other players. The tool of mean-field games serves this purpose. So far, there are only a
few works on mean-field stopping games (see, e.g., Carmona et al. (2017), Nutz (2018),
Nutz et al. (2018)), and a lot of further research needs to be done in order to deepen the
understanding of this topic. In this essay/project, you will consider mean-field stopping
games and their applications in finance and economics. You will also investigate some
particular models (e.g., bank-run models) and examine the existence of equilibria. Finally,
you will aim to construct/characterise the equilibria when they exist.

31
6 Rights and Responsibilities

Applied Mathematics 4 students will have access to the following.

• Office space and a desk in the Carslaw building.

• A computer account with access to e-mail and the World-Wide Web, as well as TEX
and laser printing facilities for the preparation of essays and projects.
• A photocopying account paid by the School for assembling essay/project source
material.
• After-hours access key to the Carslaw building. (A deposit is payable.)
• A pigeon-hole in room 728 — please inspect it regularly as lecturers often use it to
hand out relevant material.
• Participation in the School’s social events.
• Class representative at School meetings.

Applied Mathematics 4 students have the following obligations.

• Regular attendance at the regular weekly seminars in applied mathematics.

• Have regular meetings with project/essay supervisors, and meet all deadlines.

• Utilise all School resources in an ethical manner.

• Contribute towards the academic life in Applied Mathematics at the School of Math-
ematics and Statistics.

32
7 Scholarships, Prizes and Awards
The following prizes may be awarded to Applied Mathematics 4 students of sufficient
merit. Students do not need to apply for these prizes, which are awarded automatically.

Joye Prize in Mathematics Value: $5300, with medal and shield

Awarded to the most outstanding student completing Honours in the School of Mathe-
matics and Statistics.
University Medal
Awarded to Honours students who perform outstandingly. The award is subject to Faculty
rules, which require a Faculty mark of 90 or more in Applied Mathematics 4 and a Third
Year WAM of 80 or higher. A medal is always awarded when the Faculty mark is 95 or
higher. More than one medal may be awarded in any year.
K.E. Bullen Memorial Prize Value: $1000
Awarded annually on the recommendation of the Head of the School of Mathematics and
Statistics in consultation with the professors of Applied Mathematics to the most profi-
cient student in Applied Mathematics 4, provided that the student’s work is of sufficient
merit.
Barker Prize Value: $550
Awarded at the Fourth (Honours) Year examination for proficiency in Pure Mathematics,
Applied Mathematics or Mathematical Statistics.
M.J. and M. Ashby Prize Value: $360
Offered annually for the best essay, submitted by a student in the Faculty of Science,
that forms part of the requirements of Pure Mathematics 4, Applied Mathematics 4 or
Mathematical Statistics 4.
Norbert Quirk Prize No IV Value: $250
Awarded annually for the best essay on a given mathematical subject by a student enrolled
in a Fourth Year course in mathematics (Pure Mathematics, Applied Mathematics or
Mathematical Statistics) provided that the essay is of sufficient merit.
Australian Federation of Graduate Women Prize in Mathematics. Value: $175
Awarded annually, on the recommendation of the Head of the School of Mathematics
and Statistics, to the most distinguished woman candidate for the degree of BA or BSc
who graduates with first class Honours in Applied Mathematics, Pure Mathematics or
Mathematical Statistics.
Chris Cannon Prize Value: $100
For the best adjudged essay/project seminar presentation of an Applied Mathematics 4
student.

Note: some of the prize values may change. A complete list of the prizes offered by
the School of Mathematics and Statistics, as well as any changes to current prize val-

33
ues can be found at http: // www. maths. usyd. edu. au/ u/ About/ prizes. html on the
school’s website.

8 Life After Fourth Year

Postgraduate Studies:
Many students completing the Honours programme have in the past gone on to pursue
postgraduate studies at the University of Sydney, at other Australian universities, and
at overseas universities. Please see the Director of Postgraduate Studies if interested
in enrolling for a MSc or PhD at the School of Mathematics & Statistics. Students
who do well in Applied Mathematics 4 may be eligible for postgraduate scholarships,
which provide financial support during subsequent study for higher degrees at Australian
universities. The honours coordinator is available to discuss options and provide advice
to students interested in pursuing studies at other universities.
Careers:
Students seeking assistance with post-grad opportunities and job applications should feel
free to ask lecturers most familiar with their work for advice and written references.
The Director of the Applied Mathematics Teaching Programme, the Course Coordinator
and the course lecturers may also provide advice and personal references for interested
students.

34