Dissertation Topics
This page contains details for the topics available for final year dissertations for MMath students, and for projects for BSc students. For full information on the BSc and MMath Final Year Projects, please see this page.
These topics are also offered to students in MSc Mathematics.
For more information on any of these projects, please contact the project supervisor.
 Filippo Cagnetti

For more information, please email Dr Filippo Cagnetti or visit his staff profile
 Project 1: Isoperimetric Inequalities (Dr F. Cagnetti)

The isoperimetric inequality [1,2] states that the ball minimizes the perimeter among all the sets with the same volume:
!$$\begin{equation} \tag*{(I)} P (B_r) \leq P (E), \quad \text{ for any set }E \subset \mathbb{R}^n \quad (r > 0 \text{ chosen such that } E = B_r). \end{equation}$$! Here, !$P (E)$! and !$E$! denote the perimeter and the volume of the set !$E$!, respectively, while !$B_r$! is a ball with radius !$r$!. In order to give a rigorous meaning to inequality (I), one has to clarify what we mean by perimeter, if !$E$! is a set whose boundary is not regular. This is achieved by introducing the notion of set of finite perimeter [3], which is a very useful tool in geometric variational problems. We will give a rigorous proof of inequality (I), and we will consider interesting related problems.
Key words: Isoperimetric Inequality, sets of finite perimeter.
Recommended modules: Measure and Integration, Perturbation theory and calculus of variations, Functional Analysis.
References:
!$[1]$! E. De Giorgi: Sulla proprietà isoperimetrica dell'ipersfera, nella classe degli insiemi aventi frontiera orientata di misura finita. Atti Accad. Naz. Lincei Mem. Cl. Sci. Fis. Mat., Sez. I 8 (1958), 3344.
!$[2]$! E. De Giorgi, Selected Papers, in L. Ambrosio, G. Dal Maso, M. Forti, S. Spagnolo (Eds.), Springer Verlag (2005).
!$[3]$! F. Maggi: Sets of Finite Perimeter and Geometric Variational Problems: an Introduction to Geometric Measure Theory. Cambridge University Press, 2012.
 Project 2: Steiner Symmetrization (Dr F. Cagnetti)

Steiner Symmetrization is a very simple and powerful technique in analysis, and has several remarkable applications to problems of geometric and functional nature. The Steiner symmetral of a set !$E \subset \mathbb{R}^n$! about a hyperplane !$H \subset \mathbb{R}^n$! is the set !$E^S$! with the following property: the intersection of !$E^S$! with any straight line !$L$! orthogonal to !$H$! is a segment, symmetric about !$H,$! whose length equals the 1dimensional measure of !$L \cap E$!. One can prove that for any set !$E \subset \mathbb{R}^n$! there holds !$E^S=E$!, and Steiner inequality (see, for instance, Theorem 14.4 in [3]) is satisfied: !$$\begin{equation} \tag*{(S)} P (E^S) \leq P(E) \quad \text{ for any set }E \subset \mathbb{R}^n. \end{equation}$$! Here !$P (E)$! and !$E$! denote the perimeter and the volume of the set !$E$!, respectively. We will first give a rigorous proof of (S), and then investigate the family of sets such that equality holds true in Steiner inequality.
Key words: Isoperimetric Inequality, sets of finite perimeter.
Recommended modules: Measure and Integration, Perturbation theory and calculus of variations, Functional Analysis.
References:
!$[1]$! E. De Giorgi: Sulla proprietà isoperimetrica dell'ipersfera, nella classe degli insiemi aventi frontiera orientata di misura finita. Atti Accad. Naz. Lincei Mem. Cl. Sci. Fis. Mat., Sez. I 8 (1958), 3344.
!$[2]$! E. De Giorgi, Selected Papers, in L. Ambrosio, G. Dal Maso, M. Forti, S. Spagnolo (Eds.), Springer Verlag (2005).
!$[3]$! F. Maggi: Sets of Finite Perimeter and Geometric Variational Problems: an Introduction to Geometric Measure Theory. Cambridge University Press, 2012.
 Miroslav Chlebik

For more information, please email Dr Miroslav Chlebík or visit his staff profile
 Project 1: Absolutely Minimizing Lipschitz Extensions and Infinity Laplacian (Dr M. Chlebík)

A continuous realvalued function !$u$! defined on a domain !$U\subseteq \mathbb{R}^n$! (!$n\geq 2$!) is called absolutely minimizing, if for any open set !$V\subset U$! and any Lipschitz function !$v$! on !$\overline{V}$! !$$ v\bigm_{\partial V}=u\bigm_{\partial V} \qquad \implies \qquad \\nabla u\_{L^\infty(V)}\leq \\nabla v\_{L^\infty(V)}.$$! It is wellknown that !$u$! is absolutely minimizing if and only if it is the solution of the infinity Laplacian, which is the (highly degenerate) EulerLagrange equation for the prototypical problem in the calculus of variations in !$L^\infty$!. The problem of regularity of these functions is widely open, at this time it is unknown whether they are differentiable everywhere if !$n>2$!. We examine various techniques to study pointwise behaviour of these functions.
Miroslav Chlebik Presentation [PDF 309.98KB]
Key words: Lipschitz mappings, optimal Lipschitz extension,degenerate elliptic PDEs, infinity harmonic functions.
Recommended modules: Functional Analysis, Partial Differential Equations
References:
!$[1]$! Aronsson, G., Crandall, M. G. and Juutinen, P., A tour of the theory of absolutely minimizing functions, Bull. Amer. Math. Soc. 41(2004), no. 4, 439505
!$[2]$! Crandall, M. G., Evans, L. C. and Gariepy, R. F., Optimal Lipschitz extensions and the infinity Laplacian, Calc. Var. Partial Diff. Equations 13(2001), no. 2, 123139
 Project 2: Metric Dimension (Dr M. Chlebík)

Hausdorff dimension is the principal notion of dimension in the context of fractal sets in !$\mathbb{R}^n$!, or even for general metric spaces. However, other definitions are in widespread use, for example, packing dimension, upper and lower boxcounting dimension, upper and lower Minkowski dimension, ... We will examine some of these and their interrelationship.
Miroslav Chlebik Presentation [PDF 309.98KB]
Key words: Hausdorff dimension, Lipschitz mappings, rectifiable sets, fractals
Recommended modules: Measure and Integration, Functional Analysis
References:
!$[1]$! Falconer, K., Fractal geometry: Mathematical Foundations and Applications, John Wiley & Sons Ltd., 1990
 Project 3: Arcs with Increasing Chords (Dr M. Chlebík)

A curve !$C$! in the plane has the increasing chord property if !$\x_2x_3\\leq \x_1x_4\$! whenever !$x_1$!, !$x_2$!, !$x_3$! and !$x_4$! lie in that order on !$C$!. Larman & Mc Mullen showed that !$$ L\leq 2\sqrt 3ab, $$! where !$C$! is a plane curve with the increasing chord property with length !$L$! and endpoints !$a$! and !$b$!. We will examine how to improve the above constant "!$2\sqrt 3$!". (It is conjectured that !$L\leq \frac23\piab$!, with equality if !$C$! consists of two sides of a Reuleaux triangle.)
Miroslav Chlebik Presentation [PDF 309.98KB]
Key words: curve length, Lipschitz curve, calculus of variations
Recommended modules: Functional Analysis, Partial Differential Equations
References:
!$[1]$! Larman, D. G. and McMullen P., Arcs with increasing chords, Proc. Cambridge Philos. Soc. 72(1972), 205207
 Antoine Dahlqvist

For more information, please email Dr Antoine Dahlqvist or visit his staff profile
 Project 1: RANDOM MATRICES AND FREE PROBABILITY

See PDF for full description
Antoine Dahlqvist  Random matrices and Free Probability [PDF 345.10KB]
 Project 2: FROM M/M/1 TO BROWNIAN QUEUES, SYMMETRIES OF THE BROWNIAN PATH.

See PDF for full description
 Masoumeh Dashti

For more information, please email Dr Masoumeh Dashti or visit her staff profile
 Project 1: Metrics on the space of probability measures (Dr M. Dashti)

Studying the convergence properties of sequences of probability measures comes up in many applications (for example in the study of approximations of probability measures and stochastic inverse problems). In such problems, it is of course important to choose an appropriate metric on the space of the probability measures. This project consists of learning about some of the important metrics on the space of probability measures (for example: Hellinger, Prokhorov and Wasserstein), and studying the relationship between them. We also look at convergence properties of some sampling techniques.
Key words: probability metrics, rates of convergence, Bayesian inverse problems
Recommended modules: Introduction to Probability, Measure and Integration.
References:
!$[1]$! Gibbs A. L. and Su F. E. (2002) On choosing and bounding probability metrics.
!$[2]$! Robert, C. P. and Casella, G. (2004) Monte Carlo statistical methods. Second edition. Springer Texts in Statistics. SpringerVerlag, New York.
 Project 2: Inverse problems: classical and Bayesian approach (Dr M. Dashti)

Consider the problem of finding the initial temperature field of a one dimensional heat equation from (noisy) measurements of the temperature function at a positive time. This is an example of an inverse problem (considering the underlying heat equation, given initial temperature field, as the direct problem). Such problems where the function of interest cannot be observed directly, and has to be obtained from other observable quantities and through the mathematical model relating them, appears in many practical situations. Inverse problems in general do not satisfy Hadamard's conditions of wellposedness: for example in the case of the above inverse heat problem, the solution (here the initial field) does not depends continuously on the temperature function at a positive time. We can, however, obtain a reasonable approximation of the solution in a stable way by regularizing the problem using a priori information about the solution. In this project, we will study classical regularization methods, and also the Bayesian approach to regularization in the case of statistical noise.
Key words: Inverse problems, Tikhonov regularization, Bayesian regularization
Recommended modules: Partial differential equations, Functional analysis, Probability and statistics, Measure and Integration.
References:
!$[1]$! Engl H. W., Hanke M. and Neubauer A. (2000) Regularization of inverse problems, Kluwer Academic Publishers.
!$[2]$! Stuart A. (2010) Inverse problems: a Bayesian perspective, 19, 451559.
 Project 3: Conditional regularity of the NavierStokes equations in terms of pressure (Dr M. Dashti)

We start by studying LerayHopf weak solutions of the three dimensional NavierStokes equations which are known to exit globally (for all positive times). The strong solutions are only known to exist locally. There are, however, results which show the global existence of strong solutions under extra conditions on the velocity field or pressure (conditional regularity results). In this direction, we will study Serrin's conditional regularity result and then examine similar conditions in terms of the pressure field.
Key words: NavierStokes equations, Regularity theory
Recommended modules: Partial differential equations, Functional analysis, Measure and Integration.
References:
!$[1]$! Chae L. and Lee J. (2001) Regularity criterion in terms of pressure for the NavierStokes equations, Nonlinear Analysis 46. 727735
!$[2]$! Serrin J. (1962) On the interior regularity of weak solutions of the NavierStokes equations. Archive for Rational Mechanics and Analysis, 9, 187195.
!$[3]$! Temam R. (2001) NavierStokes Equations: Theory and Numerical Analysis, American Mathematical Society.
 Nicos Georgiou

For more information, please email Dr Nicos Georgiou or visit his staff profile
 Project 1: Optimal sequence alignments (Dr N. Georgiou)

Consider three words in diﬀerent languages, for example Henry, Enrico and Heinrich. The alphabet used to construct these words is the Latin alphabet and just by looking at them we are convinced that the words look “the same”. One way to make this rigorous is to “align” the words so that similar letters line up. For example,
H e  n r y   H e i n r i c h  E  n r i c o and then introduce a score function for which aligned letters gain a score +1, mismatched letters (like h and o in the last two words) get penalised by −a and gaps (denoted by underscores) are penalised by −b.
An alignment is called optimal if it achieves the highest possible score from all possible alignments between the words and a high score indicates a higher probability that the words are “similar”. This can be used for example when comparing DNA sequences of two species.
While the example above is deterministic, one can add randomness to this by creating two words over any ﬁnite alphabet with random letters, and then trying to ﬁnd the optimal score and alignment, as well as the behaviour of these scores when the alphabet size k, costs a, b and size n change.
The project has several aspects suitable for various forms of an MSc thesis:
(1) Theoretical aspects: There exists a vast literature on the longest common subsequence (LCS) of words from a ﬁnite alphabet. This is only the case that the gap penalty is 0, but already the project can be only on the LCS.
(2) Numerical aspects. Eﬃcient algorithms computing optimal alignments and optimality regions (what are those?:)) are scarce and buried in mathematical biology books and journals.
(3) The problem is a window to an area of mathematics called “Algebraic Statistics” that can serve as an umbrella to the thesis. Several other problems are analysed in the area with similar techniques.
Key words: Sequence alignment, global alignment, optimality regions, multiple sequence alignments, algebraic statistics.
 Project 2: Corner growth processes;(Dr N. Georgiou)

Consider a collection of independent Bernoulli random variables !$\{ X_v\}_{v\in \mathbb Z^2}$! with !$\mathbb P(X_v=1)=p=1q$! and interpret the event that !$X_v=1$! as the event of having site !$v$! as marked. For any rectangle !$[m]\times[n] =\{ 1,2,...,m \}\times\{1,2,...,n\}$! we can define the random variable !$L(m,n)$! that denotes the maximum possible number of marked sites that one can collect along a path from !$(1,1)$! to !$(m,n)$! that is strictly increasing in both coordinates. It is possible that there is more than one optimal path, and any such path is called a `Bernoulli longest increasing path (BLIP).'
The random variables !$L(m,n)$! satisfy a certain property, called subadditivity. By Kingman's Subadditive Ergodic Theorem one can prove !$n^{1}L([nx],[ny])\rightarrow \Psi(x,y)$! a.s. and in !$L^1$!. Part of the project will be to prove the closed formula for !$\Psi(x,y)$! given by
\begin{equation} \Psi(x,y) =\left\{ \begin{array}{lll} x, & \textrm {if } x < py \\ \displaystyle \frac{2\sqrt{pxy}p(x+y)}{q}, & \textrm {if } p^{1}y\geq x\geq py \\ y, &\textrm {if } y < px \end {array} \right. \end{equation}
There is a vast literature in statistical physics that studies this model as a simplified alternative to the hard longest common subsequence (LCS) model (see other projects).
Key words: Longest increasing path, Hammersley process, totally asymmetric simple exclusion process, corner growth model, last passage percolation, subadditive ergodic theorem
 Project 3: Traffic flow models via totally asymmetric simple exclusion processes;(Dr N. Georgiou)

The totally asymmetric simple exclusion process (TASEP) is a stochastic particle system in which particles move only in one direction, without being able to overtake each other. The model has been used a few times to model traffic flow in narrow highways, for which there are rigorous mathematical results, or even implemented to make predictions about traffic in city grids but in this case without the mathematical rigour.
The goal of this project is threefold. First there is the theoretical component of understanding the mathematics behind the hydrodynamic limits of the particle system and find the limiting PDE. Second, we will use free traffic data and develop statistical tests to identify and estimate relevant parameters that appear in the hydrodynamic limit above. The third is to develop Monte Carlo algorithms that take the estimated parameters, build the stochastic model, and show us the traffic progress in a given road network.
Supervisor: Dr. Nicos Georgiou
Helpful mathematical background: Random processes, Monte Carlo Simulations, Statistical Inference.
Some Bibliography:
[1] N. Georgiou, R. Kumar and T. Seppäläinen TASEP with discontinuous jump rates https://arxiv.org/pdf/1003.3218.pdf
[2] H.J. Hilhorst and C. AppertRolland, A multilane TASEP model for crossing pedestrian traffic flows https://arxiv.org/pdf/1205.1653.pdf
[3] J.G. Brankov, N.C. Pesheva and N. Zh. Bunzarova, Onedimensional traffic flow models: Theory and computer simulations. Proceedings of the X Jubilee National Congress on Theoretical and Applied Mechanics, Varna, 1316 September, 2005(1), 442–456.
 Peter Giesl

For more information, please email Dr Peter Giesl or visit his staff profile
 Project 1: Computational analysis of periodic orbits in nonsmooth differential

See PDF for full description
Peter Giesl  Computational analysis of periodic orbits in nonsmooth differential [PDF 11.57KB]
 Project 2: Calculation of Contraction Metrics

See PDF for full description
Peter Giesl  Calculation of Contraction Metrics [PDF 16.77KB]
 Project 3: Dimension of Attractors in Dynamical Systems (MMath and PGT only)

See PDF for full description
 James Hirschfeld

For more information, please email Prof. James Hirschfeld or visit his staff profile
 Project 1: Algebraic Geometry (Professor J. W. P. Hirschfeld)

Given one or more polynomials in several indeterminates, what do their set of common zeros look like? Curves and surfaces are typical examples. This topic examines the basic theory of such objects. It can be done both at an elementary level and at a more sophisticated level. The material of the Term 7 course on Ring Theory would be handy.
James Hirschfeld Presentation 1 [PDF 36.89KB]
Key words: polynomial, algebraic geometry
Recommended modules: Coding Theory
References:
!$[1]$! Reid, M. Undergraduate Algebraic Geometry, University Press, 1988.
!$[2]$! Semple, J. G. and Roth, L. Introduction to Algebraic Geometry, Oxford University Press, 1949
 Project 2: Cubic Curves (Professor J. W. P. Hirschfeld)

Cubic curves in the plane may have a singular point or be nonsingular. The nonsingular points on a cubic form an abelian group, which leads to many remarkable properties such as the theory of the nine associated points, from which many other results can be deduced. A nonsingular (elliptic) cubic is one of the most beautiful structures in mathematics.
James Hirschfeld Presentation 2 [PDF 25.58KB]
Key words: algebraic curve, cubic, group
Recommended modules: Coding Theory
References:
!$[1]$! Seidenberg, A. Elements of the Theory of Algebraic Curves AddisonWelsley 1968
!$[2]$! Clemens, C.H. A scrapbook for Complex Curve Theory Plenum Press 1980
 Project 3: Finite Geometry (Professor J. W. P. Hirschfeld)

In defining a vector space, the scalars belong to a field, which can also be finite, such as the integers modulo a prime. Many combinatorial problems reduce to the study of geometrical configurations, which in turn can be analysed in a geometry over a finite field.
James Hirschfeld Presentation 3 [PDF 26.96KB]
Key words: geometry, projective plane, finite field
Recommended modules: Coding Theory
References:
!$[1]$! Dembowski, P. Finite Geometries, Springer Verlag, 1968
!$[2]$! Hirschfeld, J.W.P. Projective Geometries over a Finite Field Oxford University Press, 1998.
 Project 4: Coding Theory (Professor J. W. P. Hirschfeld)

Error correction codes are used to correct errors when messages are transmitted through a noisy communication channel.
Here is the basic idea.To send just the two messages YES and NO, the following encoding suffices:
YES = 1, NO = 0:If there is an error, say 1 is sent and 0 arrives, this will go undetected. So, add some redundancy:
YES = 11, NO = 00:Now, if 11 is sent and 01 arrives, then an error has been detected, but not corrected, since the original messages 11 and 00 are equally plausible.
So, add further redundancy:
YES = 111, NO = 000:Now, if 010 arrives, and it is supposed that there was at most one error, we know that 000 was sent: the original message was NO.
Most of the theory depends on vector spaces over a finnite field.References
1. R. Hill, A First Course on Coding Theory, Oxford, 1986; QE 1302 Hil. The course
is mostly based on this book.
2. V.S. Pless, Introduction to the Theory of ErrorCorrecting Codes, Wiley, 1982, 1989;
QE 1302 Ple.
3. S. Ling and C.P. Xing, Coding Theory, a First Course, Cambridge, 2004; QE 1302
Lin.
4. https://www.maths.sussex.ac.uk/Staff/JWPH/TEACH/CODING21/index.html
 Istvan Kiss

For more information, please email Prof Istvan Kiss or visit his staff profile
Mathematical epidemiology is the study of the spread of diseases, in space and time, with the objective to trace factors that are responsible for, or contribute to, their occurrence [16]. Mathematical models are frequently used in real applications (e.g. control of Childhood disease, FootandMouth disease and Pandemic Influenza outbreak) with the aim to predict the time course of an epidemic and to determine the efficacy of various control strategies such as vaccination and contact tracing. Many such models assume that individuals can either be susceptible (S), infected and infectious (I), and recovered or removed (R). In these basic but fundamental models, susceptible or healthy individuals can become infected upon contact with infected individuals and these can then recover and become susceptible again or become immune or removed with no further impact on the epidemic. In this context the following projects are proposed:
 Project 1: Adaptive networks and their impact on epidemic models (Prof I. Z. Kiss)

We will consider a pairwise model that allows to capturing epidemic dynamics on a network that evolves in time [8]. Namely, the network has a fixed set of edges that can become deactivated and reactivated, as a possible response by individuals who try to avoid infection. The aim of this project is to formulate an SIS (susceptibleinfectedsusceptible) pairwise model and to analyse this both analytically and numerically in order to determine the epidemic threshold, disease prevalence and to characterise the interaction between disease and network dynamics. The project will involve model formulation and the derivation of differential equations to construct the pairwise model, as well as analytical and numerical analysis of the resulting model.
Key words: networks, ordinary differential equations, dynamical systems, simulation, stochastic processes, Matlab/Python
Recommended modules: Mathematics in Everyday Life, Introduction to Mathematical Biology, Differential Equations, Computing with Matlab, Probability Models, Probability and Statistics, Random Processes, Statistical Inference
 Project 2: Model reduction techniques for epidemic dynamics on networks (Prof I. Z. Kiss)

This project will focus on various types of stochastic epidemic models [7, 9] based on the classic SIS (susceptibleinfectedsusceptible) and SIR (susceptibleinfectedremoved) models. In this project we aim to derive exact models by exploring the symmetries of the network in term of the networks automorphism group. We will start from simple toy networks, with extension to more realistic networks, and we will formulate ordinary differential equation models that are related to the Kolmogorov forward equations corresponding to a continuous time Markov Chain. The project will involve model formulation, numerical solution to the formulated model, as well as comparison to simulation results.
Key words: probability, stochastic processes, Markov Chain, Kolmogorov equations, automorphism, ODEs, Matlab/Python
Recommended modules: Mathematics in Everyday Life, Introduction to Mathematical Biology, Differential Equations, Computing with Matlab, Probability Models, Probability and Statistics, Random Processes, Statistical Inference
 Project 3: Inference methods for epidemic models using meanfield and network models (Prof I. Z. Kiss)

The application of inference methods (maximum likelihood and Bayesian methods)will be explored in the context of deterministic and stochastic epidemic models. The susceptibleinfectedsusceptible (SIS) and susceptibleinfectedrecovered (SIR)epidemic models will be considered using pairwise or other meanfield models,as well as the full stochastic counterpart of the explicit stochastic epidemic simulation on networks. Possible questions include: (a) can we recover the network or its properties from epidemic data, (b) can we identify the source of infection from infection data, and (c) how can these methods be extended to realworld networks and spreading processes such as tweets on twitter and spread of memes. Excellent introductory reading for this topic can be found in [1113].
Key words: inference, networks, simulation, probability, stochastic processes, Matlab/Python
Recommended modules: Mathematics in Everyday Life, Introduction to Mathematical Biology, Differential Equations, Computing with Matlab, Probability Models, Probability and Statistics, Random Processes, Statistical Inference
 Project 4: Working with and analysing realworld networks (Prof I. Z. Kiss)

We will consider a number of realworld networks, such as the network of global cargo ship movements [14] or other technological or social networks, and apply tools from network sciences to uncover their properties in terms of degree distribution, clustering, community structure, path length [15,16] and by simulating various spreading processes on them. We will also aim to combine network analysis with the processes unfolding on these networks to better understand how such networks emerged and continue to evolve.
Key words: networks, ordinary differential equations, dynamical systems, simulation, stochastic processes, Matlab/Python
Recommended modules: Mathematical Modelling, Mathematics in Everyday Life, Computing with Matlab, Applied Mathematics, Probability and Statistics, Introduction to Mathematical Biology, Random Processes
References for all three projects:
!$[1]$! Roberts, M. & Heesterbeek, H. (1993) Bluff your way in epidemic models. Trends Microbiol. 1, 343348.
!$[2]$! Keeling, M. J. & Rohani, P. (2008) Modeling infectious diseases in animals and humans. USA: Princeton University Press.
!$[3]$! Britton, N. F. (2003) Essential Mathematical Biology: Infectious Diseases. London: Springer.
!$[4]$! Diekmann, O. & Heesterbeek, J. (2000) Mathematical Epidemiology of Infectious Diseases: Model Building, Analysis and Interpretation. Mathematical and Computational Biology. New York: John Wiley & Sons.
!$[5]$! Daley , D.J. & Gani, J. (2001) Epidemic Modelling: An Introduction. Cambridge Studies in Mathematical Biology, Cambridge University Press.
!$[6]$! Brauer, F., van den Driessche, P, &Wu, J. (2008) Mathematical epidemiology. Lecture Notes in Mathematics series. Berlin Heidelberg: Springer.
!$[7]$! Simon, P.L., Taylor, M. & Kiss, I.Z. (2011) Exact epidemic models on graphs using graphautomorphism driven lumping. J. Math. Biol. 62, 479508.
!$[8]$! I.Z. Kiss, L. Berthouze, T.J. Taylor and P.L. Simon (2012) Modelling approaches for simple dynamic networks and applications to disease transmission models. Proc. Roy. Soc. A. 18, 14712946.
!$[9]$! I.Z. Kiss, C.G. Morris, F. Selley, P.L. Simon & R.R. Wilkinson (2013) Exact deterministic representation of Markovian SIR epidemics on networks with and without loops. Submitted to J. of Math. Biol. (http://arxiv.org/abs/1307.7737).
!$[10]$! P. van Mieghem & R. van de Bovenkamp (2013) NonMarkovian infection spread dramatically alters the susceptibleinfectedsusceptible epidemic threshold in networks. Phys. Rev. Lett. 110(10):108701.
!$[11]$! T. Britton and P. D. O'Neill (2002) Bayesian inference for stochastic epidemics in populations with random social structure. Scandinavian Journal of Statistics, 29(3):375390.
!$[12]$! I. Brugere, B. Gallagher, and T. Y. BergerWolf (2018) Network structure inference, a survey: Motivations, methods, and applications. ACM Comput. Surv., 51(2):24:124:39.
!$[13]$! M. Gomez Rodriguez, J. Leskovec, and A. Krause (2010) Inferring networks of diffusion and influence. In Proceedings of the 16th ACM SIGKDD International Conference on Knowledge Discovery and Data Mining, KDD '10, pages 10191028, New York, NY, USA, ACM.
!$[14]$! P. Kaluza, A. Kölzsch, M.T. Gastner and M.T. and B. Blasius (2010) The complex network of global cargo ship movements. Journal of the Royal Society Interface, 7(48), pp.10931103.
!$[15]$! M. E. J. Newman (2003) The structure and function of complex networks, SIAM Review 45, 167–256.
!$[16]$! A.L. Barabási (2016) Network science. Cambridge University Press.
 Konstantinos Koumatos

For more information, please email Dr Konstantinos Koumatos or visit his staff profile
 Project 1: A variational approach to microstructure formation in materials: from theory to design of smart materials (Dr K. Koumatos)

From the prototypical example of steel to modern day shapememory alloys, materials undergoing martensitic transformations exhibit remarkable properties and are used in a wide range of applications, e.g. as thermal actuators, in medical devices, in automotive engineering and robotics.
The properties of these materials, such as the toughness of steel or Nitinol being able to remember its original shape, are related to what happens at small length scales and the ability of these materials to form complex microstructures. Hence, understanding how microstructures form and how they give rise to these properties is key, not only to find new applications, but also to design new materials.
A mathematical model, developed primarily in the last 30 years [1,2,3], views microstructures as minimizers of an energy associated to the material and has been very successful in explaining many observables. In fact, it has been successful even in contributing to the design of new smart materials which exhibit enhanced reversibility and low hysteresis, properties which are crucial in applications.
In this project, we will review the mathematical theory  based on nonlinear elasticity and the calculus of variations  and how it has been able to give rise to new materials with improved properties. Depending upon preferences, the project can be more or less technical.
Key words: microstructure, energy minimisation, elasticity, calculus of variations, nonconvex variational problems
Recommended modules: Continuum Mechanics, Partial Differential Equations, Functional Analysis, Measure and Integration
References:
!$[1]$! J. M. Ball, Mathematical models of martensitic microstructure, Materials Science and Engineering A 378, 6169, 2004
!$[2]$! J. M. Ball and R. D. James, Fine phase mixtures as minimizers of energy, Archive for Rational Mechanics and Analysis 100 (1), 1352, 1987
!$[3]$! K. Bhattacharya, Microstructure of martensite: why it forms and how it gives rise to the shapememory effect, Oxford University Press, 2003
!$[4]$! X. Chen, V. Srivastava, V. Dabade R. D. James, Study of the cofactor conditions: conditions of supercompatibility between phases, Journal of the Mechanics and Physics of Solids 61 (12), 25662587, 2013
!$[5]$! S. Muller, Variational models for microstructure and phase transitions, in Calculus of variations and geometric evolution problems, 85210, 1999
 Project 2: Polyconvexity and existence theorems in elasticity (Dr K. Koumatos)

The equilibrium problem of nonlinear elasticity can be formulated as that of minimising an energy functional of the form !$$ \mathcal E(u) = \int_\Omega W(\nabla u(x))\,dx, $$! subject to appropriate boundary conditions on !$\partial\Omega$!, where !$\Omega\subset \mathbb{R}^n$! represents the elastic body at its reference configuration and !$u:\Omega\to \mathbb{R}^n$! is a deformation of the body mapping a material point !$x\in \Omega$! to its deformed configuration !$u(x)\in \mathbb{R}^n$!. The function !$W$! is the energy density associated to the material and physical requirements force one to assume that !$$ W(F) \to \infty, \mbox{ as }\det F\to0^+ \mbox{ and } W(F) = \infty, \,\det F \leq 0. \tag{$\ast$} $$! As the determinant of the gradient expresses local change of volume, the conditions above translate to the requirement of infinite energy to compress a body to zero volume as well as the requirement that admissible deformations be orientationpreserving. It turns out that (!$\ast$!) is incompatible with standard conditions required on !$W$! to establish the existence of minimisers in the vectorial calculus of variations. In this project, we will review classical existence theorems as well as the seminal work of J. Ball [1] proving existence of minimisers for !$\mathcal E$! and energy densities !$W$! that are !${\it polyconvex}$! and fulfil condition (!$\ast$!). Such energies cover many of the standard models used in elasticity.
Key words: nonlinear elasticity, polyconvexity, quasiconvexity, existence theories, determinant constraints
Recommended modules: Continuum Mechanics, Partial Differential Equations (essential), Functional Analysis (essential), Measure and Integration
References:
!$[1]$! J. M. Ball, Convexity conditions and existence theorems in elasticity, Archive for Rational Mechanics and Analysis 63 (4), 337403, 1977
!$[2]$! B. Dacorogna, Direct methods in the calculus of variations, volume 78, Springer, 2007
 Project 3: Compensated compactness and existence theory in conservation laws via the vanishing viscosity method (Dr K. Koumatos)

Existence of solutions to nonlinear PDEs often relies in the following strategy: construct a suitable sequence of approximate solutions and prove that, up to a subsequence, the approximations converge to an appropriate solution of the PDE. A priori estimates coming from the PDE itself typically allow for convergence of the approximation to be established in some weak topology which, however, does not suffice to pass to the limit under a nonlinear quantity. This loss of continuity with respect to the weak topology is a great obstacle in nonlinear problems. In a series of papers in the 1970's, L. Tartar and F. Murat (see [3] for a review) introduced a remarkable method, referred to as compensated compactness, which gives conditions on nonlinearities !$Q$! that allow one to establish the implication: !$$ V_j \rightharpoonup V \Longrightarrow Q(V_j) \rightharpoonup Q(V)\tag{$\ast$} $$! under the additional information that the sequence !$V_j$! satisfies some differential constraint, e.g. the !$V_j$!'s could be gradients, thus satisfying the constraint !${\rm curl}\, V_j = 0$!. Note that (!$\ast$!) is not true in general and it is the additional information on !$V_j$! that ``compensates'' for the loss of compactness. In this project, we will review the compensated compactness theory and investigate its consequences on the existence theory for scalar conservation laws in dimension 1 via the vanishing viscosity method. In particular, we will use the socalled divcurl lemma to prove that a sequence !$u^\varepsilon$! verifying \begin{align*} \partial_t u^\varepsilon + \partial_x f(u^\varepsilon) & = \varepsilon \partial_{xx} u^\varepsilon\\ u(\cdot,t = 0) & = u_0 \end{align*} converges in an appropriate sense as $\varepsilon\to0$ to a function $u$ solving the conservation law \begin{align*} \partial_t u + \partial_x f(u) & = 0\\ u(\cdot,t = 0) & = u_0. \end{align*}
Key words: compensated compactness, divcurl lemma, weak convergence, oscillations, convexity, wave cone, conservation laws, vanishing viscosity limit
Recommended modules: Continuum Mechanics, Partial Differential Equations (essential), Functional Analysis (essential), Measure and Integration (essential)
References:
!$[1]$! C. M. Dafermos, Hyperbolic conservation laws in continuum physics, Springer, 2010
!$[1]$! L. C. Evans, Weak convergence methods for nonlinear partial differential equations, American Mathematical Society, 1990
!$[1]$! L. Tartar, Compensated compactness and applications to partial differential equations, in Nonlinear analysis and mechanics: HeriotWatt symposium, 136212, 1979
 Project 4: On the Di PernaLions theory for transport equations and ODEs with Sobolev coefficients (Dr K. Koumatos)

For !$t\in \mathbb{R}$!, consider the system of ordinary differential equations !$$ \frac{d}{dt}X(t) = b(X(t)),\quad X(0) = x\in \mathbb{R}^n. \tag{!$\ast$!} $$! The classical CauchyLipschitz theorem (aka PicardLindel\"of or Picard's existence theorem) provides global existence and uniqueness results for (!$\ast$!) under the assumption that the vector field !$b$! is Lipschitz. However, in many cases (e.g. fluid mechanics, kinetic theory) the Lipschitz condition on !$b$! cannot be assumed as a mere Sobolev regularity seems to be available.
In pioneering work, Di Perna and Lions [2] established existence and uniqueness of appropriate solutions to (!$\ast$!) under the assumption that !$b\in W^{1,1}_{{\tiny\rm loc}}$!, a control on its divergence is given and some additional integrability holds. In this project, we will review the elegant work of Di Perna and Lions.
Remarkably, their proof of a statement concerning ODEs is based on the transport equation (a partial differential equation) !$$ \partial_t u(x,t) + b(x)\cdot {\rm div}\, u(x,t) = 0, \quad u(x,0) = u_0(x) $$! and the concept of renormalised solutions introduced by the same authors. The relation between (!$\ast$!) and the transport equation lies in the method of characteristics which states that smooth solutions of the transport equation are constant along solutions of the ODE, i.e. !$$ u(X(t),t) = u(X(0),0) = u_0(X(0)) = u_0(x). $$!
Key words: ODEs with Sobolev coefficients, DiPernaLions, transport equation, renormalised solutions, continuity equation
Recommended modules: Continuum Mechanics, Partial Differential Equations (essential), Functional Analysis (essential), Measure and Integration (desirable)
References:
!$[1]$! C. De Lellis, Ordinary differential equations with rough coefficients and the renormalization theorem of Ambrosio, Seminaire Bourbaki 972, 2007
!$[2]$! R. DiPerna and P. L. Lions, Ordinary differential equations, transport theory and Sobolev spaces, Inventiones Mathematicae, 98, 511517, 1989
!$[3]$! L. C. Evans, Partial Differential Equations, American Mathematical Society, 1998
 Yuliya Kyrychko

For more information, please email Y.Kyrychko@sussex.ac.uk or visit her staff profile
 Project 1: Mathematical modelling of infectious diseases

This project is devoted to the development and analysis of the mathematical models for the spread of infectious diseases. The models developed in this project will be analysed analytically using methods of nonlinear dynamics and numerically in MATLAB to compare the analytical findings on the disease progression with numerical observations. A number of various scenarios will be considered as the models are developed to account for the effects of vaccination, quarantine, delayed onset of symptoms and similar.Keywords: mathematical modelling, epidemics, nonlinear dynamics.Recommended modules: Introduction to Mathematical Biology, Dynamical Systems
 Project 2: Timedelayed models of couples systems

This project aims to identify and analyse models of coupled elements, which are connected with timedelays. These types of systems arise in various different disciplines, such as engineering, physics, biology etc. The interesting feature where the current state of the system depends on the state of the system some time ago makes such models much more realistic and leads to various potential scenarios of dynamical behaviour. The models in this project will be analysed analytically to understand their stability properties and find critical time delays as well as numerically using MATLAB.
Keywords: mathematical modelling; delay differential equations; stability analysisRecommended modules: Introduction to Mathematical Biology, Dynamical Systems
 Omar Lakkis

For more information, please email Dr Omar Lakkis or visit his staff profile
 Project 1: Geometric Motions and their Applications (Dr O. Lakkis)

Geometric constructs such as curves, surfaces, and more generally (immersed) manifolds, are traditionally thought as static objects lying in a surrounding space. In this project we view them instead as moving within the surrounding space. While Differential Geometry, which on of the basis of Geometric Motions, is a mature theory, the study of Geometric Motions themselves has only really pickedup in the late seventies of the past century. This is quite surprising given the huge importance that geometric motions play in applications which range from phase transition to crystal growth and from fluid dynamics to image processing. Here, following the socalled classical approach, we learn first about some basic differential geometric tools such as the mean and Gaussian curvature of surfaces in usual 3dimensional space. We then use these tools to explore a fundamental model of geometric motions: the Mean Curvature Flow. We review the properties of this motion and some of its generalisations. We look at the use of this motion in applications such as phase transition. This project has the potential to extend into a research direction, depending on the students will and ability to pursue this. Extra references will be given in that case. One way of performing this extension would be to implement computer code simulating geometric motions and analysing the algorithms.
Omar Lakkis Presentations [PDF 358.53KB]
Key words: Parabolic Partial Differential Equations, Surface Tension, Geometric Measure Theory, Fluiddynamics, Growth Processes, Mean Curvature Flow, Ricci Flow, Differential Geometry, Phasefield, Levelset, Numerical Analysis
Recommended modules: Finite Element Methods, Measure and Integration, Numerical Linear Algebra, Numerical Differential Equations, Intro to Math Bio, Applied Whatever Modelling.
References:
!$[1]$! Gurtin, Morton E., Thermomechanics of evolving phase boundaries in the plane. Oxford Mathematical Monographs. The Clarendon Press, Oxford University Press, New York, 1993. ISBN 0198536941
!$[2]$! Huisken, Gerhard, Evolution Equations in Geometry, in Mathematics unlimited2001 and beyond, 593604, Springer, Berlin, 2001.
!$[3]$! Spivak, Michael, A Comprehensive Introduction to Differential Geometry. Vol. III. Second edition. Publish or Perish, 1979. ISBN 0914098837
!$[4]$! Struwe, Michael, Geometric Evolution Problems. Nonlinear Partial Differential Equations in Differential Geometry (Park City, UT, 1992), 257339, IAS/Park City Math. Ser., 2, Amer. Math. Soc., Providence, RI, 1996.
 Project 2: Stochastic differential equations: computation, analysis and modelling (Dr O. Lakkis)

Stochastic Differential Equations (SDEs) have become a fundamental tool in many applications ranging from environmental risk management to mechanical failure control and from neurobiology to financial analysis. While the need for effective numerical solutions of SDEs, which are differential equations with a probabilistic (uncertain) data, closed form solutions are seldom available.
This project can be specialised, according to the student's tastes and skills into 3 different flavours: (1) Analysis/Theory, (2) Analysis/Computation, (3) Computational/Modelling.
(1) We explore the rich theory of stochastic processes, stochastic integration and theory (existence, uniqueness, stability) of stochastic differential equations and their relationship to other fields such as the KacFeynman Formula (related to quantum mechanics and particle physics), or Partial Differential Equations and Potential Theory (related to the work of Einstein on Brownian Motion), stochastic dynamical systems (large deviation) or Kolmogorov's approach to turbulence in fluiddynamics. Prerequisites for this direction are some knowledge of probability, stochastic processes, partial differential equations, measure and integration and functional analysis.
(2) We review the basics of SDEs and then look at a practical way of implementing algorithms, using any one of Octave/Matlab/C/C++, that give us a numerical solution. In particular, we learn about pseudorandom numbers, MonteCarlo methods, filtering and the interpretation of those numbers that our computer produces. Although not a strict prerequisite, some knowledge of probability, ordinary differential equations and their numerical solution will be useful.
(3) We look at practical models in environmental sciences, medicine or engineering involving uncertainty (for example, the ideal installation of solar panels in a region where weather variability can affect their performance). We study these models both from a theoretical point of view (connecting to their Physics) and we run simulations using computational techniques for stochastic differential equations. The application field will be emphasised and must be clearly to the student's liking. (Although very interesting as a topic, I prefer not to deal with financial applications.) The prerequisites are probability, random processes, numerical differential equations and some of the applied/modelling courses.
Omar Lakkis Presentations [PDF 358.53KB]
Key words: Stochastic Differential Equations, Scientific Computing, Random Processes, Probability, Numerical Differential Equations, Environmental Modelling, Stochastic Modelling, FeynmanKac Formula, Ito's Integral, Stratonovich's Integral, Stochastic Calculus, Malliavin Calculus, Filtering.
Recommended modules: Probability Models, Random Processes, Numerical Differential Equations, Partial Differential Equations, Introduction to Math Biology, Fluiddynamics, Statistics.
References:
!$[1]$! L.C. Evans, An Introduction to stochastic differential equations. Lecture notes on authors website (google: Lawrence C Evans). University of California Berkley.
!$[2]$! C. W. Gardiner, Handbook of stochastic methods for physics, chemistry and the natural sciences. 3rd ed., Springer Series in Synergetics, vol. 13, SpringerVerlag, Berlin, 2004. ISBN 3540208828
!$[3]$! P.E. Kloeden; E. Platen; H. Schurz, Numerical solution of SDE through computer experiments. Universitext. SpringerVerlag, Berlin, 1994. xiv+292 pp. ISBN 3540570748
!$[4]$! A. Beskos and A. Stuart, MCMC methods for sampling function space, ICIAM2007 Invited Lectures (R. Jeltsch and G. Wanner, eds.), 2008.
!$[5]$! Joseph L. Doob, Classical potential theory and its probabilistic counterpart, Classics in Mathematics, SpringerVerlag, Berlin, 2001, Reprint of the 1984 edition. ISBN 3540412069
 Michael Melgaard

For more information, please email Prof. Michael Melgaard or visit his staff profile
 Project 1: Spectral and scattering properties of Quantum Hamiltonians (Prof. M. Melgaard)

Quantum Operator Theory concerns the analytic properties of mathematical models of quantum systems. Its achievements are among the most profound and most fascinating in Quantum Theory, e.g., the calculation of the energy levels of atoms and molecules which lies at the core of Computational Quantum Chemistry.
Among the many problems one can study, we give a short list:
 The atomic Schrödinger operator (Kato's theorem and all that);
 The periodic Schrödinger operator (describing crystals);
 Scattering properties of Schrödinger operators (describing collisions etc);
 Spectral and scattering properties of mesoscopic systems (quantum wires, dots etc);
 Phase space bounds (say, upper bounds on the number of energy levels) with applications, e.g., the Stability of Matter or Turbulence.
Key words: differential operators, spectral theory, scattering theory.
Recommended modules: Functional Analysis, Measure and Integration theory, Partial Differential Equations.
References:
!$[1]$! M. Melgaard, G. Rozenblum, Schrödinger operators with singular potentials, in: Stationary partial differential equations Vol. II, 407517, Handb. Differ. Equ., Elsevier/NorthHolland, Amsterdam, 2005.
!$[2]$! Reed, M., Simon, B., Methods of modern mathematical physics. Vol. IIV. Academic Press, Inc., New York, 1975, 1978,1979,1980.
 Project 2: Variational approach to KohnSham models with magnetic fields (Prof. M. Melgaard)

Quantum Mechanics (QM) has its origin in an effort to understand the properties of atoms and molecules. Its first achievement was to establish the Schrödinger equation by explaining the stability of the hydrogen atom; but hydrogen is special because it is exactly solvable. When we proceed to a molecule, however, the QM problem cannot be solved in its full generality. In particular, we cannot determine the solution (i.e., the ground state) to !$HΨ=EΨ$!, where !$H$! denotes the Hamiltonian of the molecular system, !$Ψ$! is the wavefunction of the system, and !$E$! is the lowest possible energy. This problem corresponds to finding the minimum of the spectrum of !$H$! or, equivalently, !$$E= \inf \{ \, \mathcal{E}^{\rm QM}(Ψ) \, : \, Ψ \in \mathcal{H}, \:\: \ Ψ \_{L^{2}} =1 \, \}, where \ \mathcal{E}^{\rm QM}(Ψ):= \langle Ψ, H Ψ \rangle_{L^{2}}$$! and !$\mathcal{H}$! is the variational (Hilbert) space. For systems involving a few (say today six or seven) electrons, a direct Galerkin discretization is possible, which is known as Full CI in Computational Chemistry. For larger systems, with !$N$! electrons, say, this direct approach is out of reach due to the excessive dimension of the space !$ℜ^{3N}$! on which the wavefunctions are defined and the problem has to be approximated. Quantum Chemistry (QC), as pioneered by Fermi, Hartree, Löwdin, Slater, and Thomas, emerged in an attempt to develop various ab initio approximations to the full QM problem. The approximations can be divided into wavefunction methods and density functional theory (DFT). For both, the fundamental questions include minimizing configuration, divided into Question I (i) necessary and sufficient conditions for existence of a ground state (=a minimizer), and Question I (ii) uniqueness of a minimizer, and Question II, necessary and sufficient conditions for multiple (nonminimal) solutions (i.e., excited states).
A magnetic field has two effects on a system of electrons: (i) it tends to align their spins, and (ii) it alters their translational motion. The first effect appears when one adds a term of the form !$eħm^{1} {s} \cdot \mathcal{B}$! to the Hamiltonian, while the second, diamagnetic effect arises from the usual kinetic energy !$(2m)^{1}  {\mathbf p} ^{2}$! being replaced by !$(2m)^{1}  \mathbf {p} (e/c) \mathcal{A}^{2}$!. Here !${\mathbf p}$! is the momentum operator, !$\mathcal{A}$! is the vector potential, !$\mathcal{B}$! is the magnetic field associated with !$\mathcal{A}$!, and !${s}$! is the angular momentum vector. Within the numerical practice, one approach is to apply a perturbation method to compute the variations of the characteristic parameters of, say, a molecule, with respect to the outside perturbation. It is interesting to go beyond and consider the full minimization problem of the perturbed energy. In HartreeFock theory, one only takes into account the effect (ii), whereas in nonrelativistic DFT it is common to include the spindependent term and to ignore (ii) and to study the minimization of the resulting nonlinear functional, which depends upon two densities, one for spin "up" electrons and the other for spin "down" electrons. Each density satisfies a normalisation constraint which can be interpreted as the total number of spin "up" or "down" electrons.
The proposed project concerns the abovementioned problems within the context of DFT in the presence of an external magnetic field. More specifically, molecular KohnSham (KS) models, which turned DFT into a useful tool for doing calculations, are studied for the following settings:
Recent results on rigorous QC are found in the references.
 As a first step towards systems subject to a magnetic field, Question I(i) is addressed for the unrestricted KS model, which is suited for the study of open shell molecular systems (i.e., systems with a odd number of electrons such as radicals, and systems with an even number of electrons whose ground state is not a spin singlet state). The aim is to consider the (standard and extended) local density approximation (LDA) to DFT.
 The spinpolarized KS models in the presence of an external magnetic field with constant direction are studied while taking into account a realistic local spindensity approximation, in short LSDA.
Key words: differential operators, spectral theory, scattering theory.
Recommended modules: Functional Analysis, Measure and Integration theory, Partial Differential Equations.
References:
!$[1]$! M. Melgaard, G. Rozenblum, Schrödinger operators with singular potentials, in: Stationary partial differential equations Vol. II, 407517, Handb. Differ. Equ., Elsevier/NorthHolland, Amsterdam, 2005.
!$[2]$! Reed, M., Simon, B., Methods of modern mathematical physics. Vol. IIV. Academic Press, Inc., New York, 1975, 1978,1979,1980.
 Project 3: Resonances for Schrödinger and Dirac operators (Prof. M. Melgaard)

Resonances play an important role in Chemistry and Molecular Physics. They appear in many dynamical processes, e.g. in reactive scattering, statetostate transition probabilities and photodissociation, and give rise to longlived states well above scattering thresholds. The aim of the project is carry out a rigorous mathematical study on the use of Complex Absorbing Potentials (CAP) to compute resonances in Quantum Dynamics.
In a typical quantum scattering scenario particles with positive energy arrive from infinity, interact with a localized potential !$V(x)$! whereafter they leave to infinity. The absolutely continuous spectrum of the the corresponding Schrödinger operator !$T(\hbar)=\hbar^{2}D+V(x)$! coincides with the positive semiaxis. Nevertheless, the Green function !$G(x,x'; z)= \langle x  (T(\hbar)z)^{1} x \rangle$! admits a meromorphic continuation from the upper halfplane !$\{ \, {\rm Im}\, z >0 \,\}$! to (some part of) the lower halfplane !$\{ \, {\rm Im}\, z < 0 \,\}$!. Generally, this continuation has poles !$z_{k} =E_{k}i Γ_{k}/2$!, !$Γ_{k}>0$!, which are called resonances of the scattering system. The probability density of the corresponding "eigenfunction" !$Ψ_{k}(x)$! decays in time like !$e^{t Γ_{k}/ \hbar}$!, thus physically !$Ψ_{k}$! represents a metastable state with a decay rate !$Γ_{k}/ \hbar$! or, rephrased, a lifetime !$\tau_{k}=\hbar / Γ_{k}$!. In the semiclassical limit !$\hbar \to 0$!, resonances !$z_{k}$! satisfying !$Γ_{k}=\mathcal{O}(\hbar)$! (equivalently, with lifetimes bounded away from zero) are called "longlived".
Physically, the eigenfunction !$Ψ_{k}(x)$! only make sense near the interaction region, whereas its behaviour away from that region is evidently nonphysical (Outgoing waves of exponential growth). As a consequence, a much used approach to compute resonances approximately is to perturb the operator !$T(\hbar)$! by a smooth absorbing potential !$iW(x)$! which is supposed to vanish in the interaction region and to be positive outside. The resulting Hamiltonian !$T_{W}(\hbar):=T(\hbar)iW(x)$! is a nonselfadjoint operator and the effect of the potential !$W(x)$! is to absorb outgoing waves; on the contrary, a realvalued positive potential would reflect the waves back into the interaction region. In some neighborhood of the positive axis, the spectrum of !$T_{W}(\hbar)$! consists of discrete eigenvalues !$\tilde{z}_{k}$! corresponding to !$L^{2}$!eigenfunctions !$\widetilde{Ψ}_{k}$!.
As mentioned above, the CAP method has been widely used in Quantum Chemistry and numerical results obtained by CAP are very good. The drawback with the use of CAP is that there are no proof that the correct resonances are obtained. (This is in stark contrast to the mathematically rigorous method of complex scaling). In applications it is assumed implicitly that the eigenvalues !$\tilde{z}_{k}$! near to the real axis are small perturbations of the resonances !$z_{k}$! and, likewise, the associated eigenfunctions !$\widetilde{Ψ}_{k}$!, !$Ψ_{k}(x)$! are close to each other in the interaction region. Stefanov (2005) proved that very close to the real axis (namely, for !$ {\rm Im}\, \tilde{z}_{k} =\mathcal{O}(\hbar^{n})$! provided !$n$! is large enough), this is in fact true. Stefanov's proof relies on a series of ingenious developments by several people, most notably Helffer (1986), Sjöstrand (1986, 1991, 1997, 2001, 2002), and Zworski (1991, 2001).
The first part of the project would be to understand in details Stefanov's work [2] and, subsequently, several open problems await.
Key words: operator and spectral theory, semiclassical analysis, micro local analysis.
Recommended modules:Functional Analysis, Measure and Integration theory, Partial Differential Equations.
References:
!$[1]$! J. Kungsman, M. Melgaard, Complex absorbing potential method for Dirac operators. Clusters of resonances, J. Ope. Th., to appear.
!$[2]$! P. Stefanov, Approximating resonances with the complex absorbing potential method, Comm. Part. Diff. Eq. 30 (2005), 18431862.
 Project 4: Critical point approach to solutions of nonlinear, nonlocal elliptic equations arising in Astrophysics (Prof. M. Melgaard)

The Choquard equation in three dimensions reads:
!$$\begin{equation} \tag*{(0.1)} Δ u  \left( \int_{ℜ^{3}} u^{2}(y) W(xy) \, dy \right) u(x) = l u , \end{equation}$$! where !$W$! is a positive function. It comes from the functional:
!$$\mathcal{E}^{\rm NR}(u) = \int_{ℜ^{3}}  \nabla u ^{2} \, dx \int \int  u(x) ^{2} W(xy) u(y)^{2} \, dx dy,$$!
which, in turn, arises from an approximation to the HartreeFock theory of a onecomponent plasma when !$W(y) =1/  y  $! (Coulomb case). Lieb (1977) proved that there exists a unique minimizer to the constrained problem !$E^{\rm NR}(\nu) = \inf \{ \, \mathcal{E}(u) \, : \, u \in \mathcal{H}^{1}(ℜ^{3}), \ u \_{L^{2}} \leq \nu \, \}$!.
The mathematical difficulty of the functional is caused by the minus sign in !$\mathcal{E}^{\rm NR}$!, which makes it impossible to apply standard arguments for convex functionals. Lieb overcame the lack of convexity by using the theory of symmetric decreasing functions. Later Lions (1980) proved that the unconstrained problem (0.1) possesses infinitely many solutions. For the constrained problem, seeking radially symmetric, normalized functions !$\ u \_{L^{2}} =+1$!, or more generally, seeking solutions belonging to:
!$$\mathcal{C}_{N}= \{ \, φ \in \mathcal{H}_{\rm r}^{1} (ℜ^{3}) \, : \, \ φ \_{L^{2}} =N \, \} ,$$! the situation is much more complicated and conditions on !$W$! are necessary. In the Coulomb case, Lions proves that there exists a sequence !$(l_{j}, u_{j})$!, with !$l_{j} > 0$!, and !$u_{j}$! satisfies !${(0.1)}$! (with !$l=l_{j}$!) and belongs to !$\mathcal{C}_{1}$!
We may replace the negative Laplace operator by the socalled quasirelativistic operator, i.e., the pseudodifferential operator !$\sqrt{ δ +m^{2} } m$!; this is the kinetic energy operator of a relativistic particle of mass !$m \geq 0$!. It is defined via multiplication in the Fourier space with the symbol !$\sqrt{k^{2} +m^{2}} m$!, which is frequently used in relativistic quantum physics models as a suitable replacement of the full (matrix valued) Dirac operator. The associated timedependent equation arises as an effective dynamical description for an !$N$!body quantum system of relativistic bosons with twobody interaction given by Newtonian gravity, as recently shown by Elgart and Schlein (2007). This system models a Boson star.
Several questions arise for the quasirelativistic Choquard equation (existence, uniqueness, positive solutions etc) and the first part of the project would be to get acquainted with recent (related) results, e.g., [1] and [2].
Key words: operator and spectral theory, semiclassical analysis, micro local analysis.
Recommended modules:Functional Analysis, Measure and Integration theory, Partial Differential Equations.
References:
!$[1]$! S. Cingolani, M. Clapp, S. Secchi, Multiple solutions to a magnetic nonlinear Choquard equation, Zeitschrift fr Angewandte Mathematik und Physik (ZAMP) , vol. 63 (2012), 233248.
!$[2]$! M. Melgaard, F. D. Zongo, Multiple solutions of the quasi relativistic Choquard equation, J. Math. Phys. !${53}$!(2012), 033709 (12 pp).
 Veronica Sanz (This project only available to MSc Data Science students)

For more information, please email Prof Veronica Sanz or visit her staff profile
PLEASE NOTE THAT PROF SANZ IS NOT AVIALABLE FOR PROJECT SUPERVISION IN 19/20
 Project 1: Searching for fundamental theories of Nature with unsupervised machine learning (Prof Veronica Sanz)

In High Energy Particle Physics we contrast data with new theories of Nature. Those theories are proposed to solve mysteries such as 1.) what is the Dark Universe made of, 2.) why there is so much more matter than antimatter in the Universe, and 3.) how can a light Higgs particle exist.
To answer these questions, we propose mathematical models and compare with observations. Sources of data are quite varied and include complex measurements from the Large Hadron Collider, underground Dark Matter detection experiments and satellite information on the Cosmic Microwave Background. We need to incorporate all this data in a framework which allows us to test hypotheses, and this is usually done via a statistical analysis, e.g. Bayesian, which provides a measure of how well a hypothesis can explain current observations. Alas, this approach has so far been unfruitful and is driving the field of Particle Physics to an impasse.
In this project, we will take a different and novel approach to search for new physics. We will assume that our inability to discover new physics stems from strong theoretical biases which have so far guided analyses. We will instead develop unsupervised searching techniques, mining on data for new phenomena, avoiding as much theoretical prejudices as possible. The project has a strong theoretical component, as the candidate will learn the mathematical/physical basis of new physics theories including Dark Matter, the Higgs particle and Inflation. The candidate will also learn about current unsupervisedlearning techniques and the interpretation of data in HighEnergy Physics.
The strategy adopted for this project holds the potential to open a new avenue of research in High Energy Physics. We are convinced that this departure from conventional statistical analyses mentioned above is the most effective way to discover new physics from the huge amount of data produced in the Large Hadron Collider and other experiments of similar scale.
Reaching the scientific goals outlined here would require modelling huge amounts of data at different levels of purity (raw measurements, pseudoobservables, reinterpreted data), and finding patterns which had not been detected due to a focus on smaller sets of information. Hence, we believe that research into unsupervised learning in this context will have far reaching applications beyond academic pursuits. As the world becomes increasingly dataorientated, so does our reliance on novel algorithms to make sense of the information we have in our possession. To give some examples, we can easily expect the development of unsupervised learning integrated into facial recognition software and assist in the discovery of new drugs, which provides a boost in the security and medical sector respectively.
 Nick Simm

For more information, please email Dr Nick Simm or visit his staff profile
 Project 1: Random matrix theory and the Riemann zeta function

See PDF for full description
Simm: Random matrix theory and the Riemann zeta function [PDF 156.59KB]
 Project 2: Asymptotic analysis of integrals and applications

See PDF for full description
Simm: Asymptotic analysis of integrals and applications [PDF 124.34KB]
 Enrico Scalas

For more information, please email Prof. Enrico Scalas or visit his staff profile
 Project 1: The replication crisis in science (Prof. E. Scalas)

In the last decade, several results published even by selfimportant journals such as Nature and Science had to be retracted because they could not be reproduced. In this project, we will not focus on deliberate scientific fraud, but rather on errors made by applied scientists because of the misuse and abuse of probability and statistics. The socalled replication crisis virtually affects all the natural sciences [1]. The situation is so serious that, recently, 72 applied statisticians called for a revision of the significance level to be used for pvalues in hypothesis testing [2]. However, there are serious doubts that this will cure the problem [3].
In this dissertation, you will review the main factors leading to the current dire situation in science and you will work out possible solutions based on probabilistic and statistical techniques.
References:
!$[1]$! Baker, M. (26 May 2016). "1,500 scientists lift the lid on reproducibility". Nature. 533 (7604): 452–454. doi:10.1038/533452a.
!$[2]$! Benjamin, D. et al. (2017). "Redefine statistical significance". Available at: https://osf.io/preprints/psyarxiv/mky9j.
!$[3]$! Colquhoun, D. (2017), "The reproducibility of research and the misinterpretation of P values". Available at: http://www.biorxiv.org/content/biorxiv/early/2017/08/07/144337.full.pdf
 Ali Taheri

For more information, please email Dr Ali Taheri or visit his staff profile
 Project 1: Hardy Spaces !$H^p$! and Boundary Behaviour of Holomorphic Functions (Dr A. Taheri)

The study of boundary behaviour of holomorphic functions in the unit disc is a classical subject which has been revived and generalised to higher dimensions as well as other geometries due to recent developments in the theory of ellipic PDEs, e.g., one such development being the H1 and BMO duality.
The aim of this project is more modest and lies in understanding the interplay between holomorphic functions in the disc on the one hand and the Poisson integral of Borel measures on the boundary circle. The results here lead to surprising qualitative properties of holomorphic functions.
Key words: Poisson integrals, Nevanlinna class, Nontangential convergence, M&F Riesz theorem
Recommended modules: Complex Analysis, Functional Analysis, Measure Theory
References:
!$[1]$! Real and Complex Analysis by Rudin
!$[2]$! Introduction to !$H^p$! spaces by Koosis
!$[3]$! Theory of !$H^p$! spaces by Duren
!$[4]$! Bounded Analytic Functions by Garnett.
 Project 2: Fourier Series in !$L^p$! Spaces and Kolmogorov's Theorem (Dr A. Taheri)

Fourier analysis has been one of the major sources of interesting and fundamental problems in analysis. It alone plays one of the most significant roles in the development of mathematical analysis in the past 2 centuries.
The aim of this project is to study Fourier series, specifically in the context of: !$L^2$!  the Hilbert space approach, continuous functions, and !$L^p$! with !$1 < p < ∞$!.
Particular emphasis goes towards the convergence/divergence properties using Functional analytic tools, Baire category arguments, singular integrals.
Key words: !$L^p$! spaces, Summability kernels, Baire category, Singular integrals, Hilbert transform
Recommended modules: Complex Analysis, Functional Analysis, Measure and Integration
References:
!$[1]$! Fourier Analysis, T.W. Koner, Cambridge University Press, 1986
!$[2]$! Real and Complex Analysis, W. Rudin, McGraw Hill, 1987
!$[3]$! Real Variable Methods in Harmonic Analysis, A. Torchinsky, Dover, 1986.
 Project 3: Oscillation and Concentration Effects in Nonlinear PDEs (Dr A. Taheri)

In the theory of nonlinear partial differential equations the study of the oscillation and concentration phenomenon plays a key role in settling the question of the existence of solutions. Here the aim is to understand the basics of weak versus strong convergence for sequences of functions and to introduce a tool known as Young measures for detecting the mechanisms that could prevent strong convergence.
Key words: Young measures, Weak convergence, DivCurl lemma
Recommended modules: Partial Differential Equations, Functional Analysis, Measure Theory
References:
!$[1]$! Parameterised Measures and Variational Principles, P.Pedregal, Birkhäuser, 1997.
!$[2]$! Partial Differential Equations, L.C. Evans, AMS, 2010.
!$[3]$! Weak Convergence Methods in PDEs, L.C. Evans, AMS, 1988.
 Project 4: Singularities in Harmonic Maps Between Manifolds (Dr A. Taheri)

Harmonic maps between manifolds are extremals of the Dirichlet energy. It is wellknown that depending on the topology and global geometry of the domain and target manifolds these harmonic maps can develop singularities in all forms and shapes. The aim of this project is to introduce the student to the theory and some of the basic ideas and important tools involved.
Key words: Harmonic maps, Dirichlet energy, Minimal connections, Singular cones.
Recommended modules: Partial Differential Equations, Introduction to Topology, Algebraic Topology, Functional Analysis
References:
!$[1]$! Infinite dimensional Morse theory by Chang
!$[2]$! Two reports on Harmonic maps by Eells and Lemaire
!$[3]$! Cartesian Currents in the Calculus of Variations by Giaquinta, Modica and Soucek.
 Chandrasekhar Venkataraman

For more information, please email Dr Chandrasekhar Venkataraman or visit his staff profile
 Project 1: Modelling, analysis and simulation of biological pattern formation (Dr C Venkataraman)

The formation of structure or patterns from homogeneity is ubiquitous in biological systems such as the intricate markings on sea shells, pigment patterns on the wings of butterflies and the regular structures made by populations of cells. Their is a rich theory for mathematical modelling of these phenomena that typically involves systems of PDEs. In this project we will understand and analyse some classical models for pattern formation and then extend them to take into account phenomena such as nonlocal interactions or growth and curvature. Dependent on the interests of the student we will either focus on the approximation of the models or their analysis.
Recommended modules: Introduction to Mathematical Biology, Advanced Numerical Analysis, Numerical Solution of Partial Differential Equations, Partial Differential Equations, Programming in C++
References:
!$[1]$! Turing, A. M. (1952). The chemical basis of morphogenesis. Phil. Trans. R. Soc. Lond. B
!$[2]$! Murray JD (2013) Mathematical Biology II: Spatial Models and Biomedical Applications. Springer New York
!$[3]$! Kondo, S.,and Miura, T. (2010). Reactiondiffusion model as a framework for understanding biological pattern formation. Science.
!$[4]$! Plaza, R. G., SanchezGarduno, F., Padilla, P., Barrio, R. A., & Maini, P. K. (2004). The effect of growth and curvature on pattern formation. Journal of Dynamics and Differential Equations
 Project 2: Mathematical problems from cell biology

Mathematical modelling, analysis and simulation can help us understand a number of cell biological questions such as, How do cells move? How do they interact with their environment and each other? How do cell scale interactions influence tissue level phenomena? In this project we will review and extend models for either cell migration, receptorligand interactions or cell signalling. The models typically involve geometric PDE with coupled systems of equations posed in different domains, cell interior, cellsurface, extracellular space. Dependent on the interests of the student we will either focus on the derivation, the approximation, or the analysis of the models.
Recommended modules: Introduction to Mathematical Biology, Advanced Numerical Analysis, Numerical Solution of Partial Differential Equations, Partial Differential Equations, Programming in C++
References:
!$[1]$! Elliott, C. M., Stinner, B., and Venkataraman, C. (2012). Modelling cell motility and chemotaxis with evolving surface finite elements. Journal of The Royal Society Interface
!$[2]$! Croft, W., Elliott, C. M., Ladds, G., Stinner, B., Venkataraman, C., and Weston, C. (2015). Parameter identification problems in the modelling of cell motility. Journal of mathematical biology
!$[3]$! Elliott, C. M., Ranner, T., and Venkataraman, C. (2017). Coupled bulksurface free boundary problems arising from a mathematical model of receptorligand dynamics. SIAM Journal on Mathematical Analysis
!$[4]$! Ptashnyk, M., and Venkataraman, C. (2018). Multiscale analysis and simulation of a signalling process with surface diffusion. arXiv preprint
 Minmin Wang

For more information, please email Dr Minmin Wang or visit her staff profile
 Project 1: Probabilistic and combinatorial analysis of coalescence

See PDF for full description
Minmin Wang  Probabilistic and combinatorial analysis of coalescence [PDF 64.35KB]