Skip to main content

Research topics and projects

This is a full list of the research projects that supervisors have identified for potential PhD or MSc students. Self-funded students and students with external scholarships are welcome to apply for any research project: our funding page has full details on funding options for UK and international students. There are a limited number of PhD scholarships for outstanding UK students to work on any project identified on the full project list. International students can apply for support via the Chancellor's International Scholarships and other schemes.

Condensed Matter Physics
Filter PhD projects: Funded / all.
Theoretical and Computational Physics
Filter PhD projects: Funded / all.
Medical and Biophysics
Filter PhD projects: Funded / all.
Astronomy and Astrophysics 
Filter: All projects.
Centre for Fusion, Space and Astrophysics
Filter PhD projects: Funded / all.
Elementary Particle Physics
Filter: All projects.

Show all news items

Statistical Mechanics of Decision Making

Supervisor: Prof. Robin Ball, Theory Group

Optimisation under Uncertainty is a pervasive problem in real life: decisions have to be made (for example about a design) before one knows the full detail of how it will be challenged to perform. Even in model problems, this can mean the figure of merit is performance averaged over many input scenarios, or it can be determined by some other statistic of the performance distribution.

The first strand of your PhD will be where performance is assessed through simulations, with scope for collaboration with Warwick Business School. We recently published what is the most efficient algorithm which tends to an exactly thermal equilibrium [1]: the question is to what extent one can beat this by algorithms which are only approximately thermal. In the background here is a big physics question: if the microstate dynamics is only approximately thermal, what can you say about the evolution of macrostates?

A second strand concerns theoretical studies of model transport and other service networks, with scope for collaboration with University of Utrecht and their partnership with Dutch Railways. We have already shown that key resource limitations can drive spatial instability and near collapse of service in a simple rail network model [2]. The questions are then: Can we understand the vulnerability to instability in more general and less model specific terms? Can we diagnose when the system is close to instability? Can we design to improve its performance under uncertain perturbations?

The research will entail a mixture of model simulations and analytic approaches. In addition to an EPSRC studentship, it is anticipated that visits to Utrecht can be funded through the Dutch collaboration.

[1] Ball, Robin, Branke, Juergen and Meisel, Stephan. (2017) Optimal sampling for simulated annealing under noise.INFORMS Journal on Computing . ISSN 1091-9856 (In Press) Preprint:

[2] Ball, Robin, Panja, Debabrata and Barkema, Gerard T. (2016) A two component railway model exhibiting service collapse. Preprint:

Tue 27 November 2018, 14:00 | Tags: Theoretical and computational (funded) PhD