Projects for prospective PhD or MSc students in the Statistics & Probability group
If you are a graduate student interested in any of the projects listed below, please contact us or apply directly for admission via the Graduate school. As an additional source of information, you may also want to have a look at the slides from the Postgraduate Opportunities Meeting from 26/11/2014.
| PhD project: Open topic |
We have just advertised an `open topic' PhD studentship (3.5 years). The project topic could be chosen within the scope of any of our research interests, in accordance with a potential supervisor. If you are interested in such an open topic project, please get in touch either directly with the relevant supervisor, or enquire here for general information.
Supervisor: Flexible.
Funding:
- For UK students, funding for one studentship is likely to be available via EPSRC.
- For EU students, EPSRC provides tuition fee waivers (without subsistence etc.)
Deadline for Application: 05/01/2015, by e-mail including a cover letter (addressed to Dr Ian Jermyn) and CV. Interviews will be held by phone or in person from 12-16 Jan 2015. Formal application through the Graduate school would be required following a successful interview.
Start date: 01/10/2015.
| PhD project: Electricity generation adequacy and investment projection under uncertainty |
Electricity systems worldwide are in a period of rapid change, due to the need to integrate a high capacity of renewable generation. In particular, it is necessary to modify existing analysis methods and market structures to accommodate renewables. In Great Britain, there are also specific concerns over whether there will be sufficient generating capacity connected to the system, and whether sufficient new investment will be forthcoming, due to the imminent retirement of large capacities of older coal plant. As with any study of rare events, directly relevant data for assessing the risk of generating capacity shortfalls are very sparse; periods of truly extreme demand happen very infrequently. The first phase of this project will study how one can perform meaningful adequacy assessments under these circumstances. Issues to be considered include: limited data on extreme demands and in particular on any statistical wind-demand relationship which might exist, realism of standard conventional plant models, the almost complete lack of relevant data on the contribution of interconnectors to other systems, and the relevance of historic generation and demand data for studying adequacy of future generation systems (looking 5, 10 or more years ahead). The second phase of the project will study uncertainty in outputs of economic projection models which are used to assess the consequences for installed generating capacity of different electricity market design options. The initial task here will be to understand systematically how these model outputs depend on the wide range of inputs, both numerical (e.g. projections of underlying demand growth) and structural (quantifications of how generating companies take investment decisions). The ultimate aim is to understand better the relationship between these models and real systems, and hence to propose improved ways of carrying out such modelling. This project will build on the expertise of Prof. Goldstein in Bayesian methods and analysing uncertainty in complex computer models, and of Dr. Dent in power systems analysis (in particular his work with National Grid on the GB statutory generation capacity adequacy study). We anticipate that in addition to opportunities for the development of novel statistical methodology to meet the aims of the project, this work will also contain substantial interaction with National Grid on important applied problems.
Supervisors: Prof. Michael Goldstein (Mathematical Sciences) and Dr. Chris Dent (Engineering)
Funding: Please enquire.
Start date: Please enquire.
| PhD project: Bayes linear theory and computation for standard statistical problems |
The purpose of this topic is extend Bayes linear methodology, in particular computationally. There already exists research software which implements some of the methodology. However, while the software works as a research tool - so we use it for their own calculations - it has never been suited to wider usage by statisticians and other researchers. This is partly for technical reasons (the language is constrained by limitations placed on computing power
available in the previous decade), partly for semantic reasons (the language is hard to learn and unfamiliar
to this generation of scientists) and partly because all model construction starts from scratch. That is,
there has been no attempt to provide users with ways to generate standard analyses such as regression,
analysis of variance, dynamic linear models, and so forth. One intention of the project will be to extend
Bayes linear methodology to such common scenarios, and to implement them in the statistical programming language R. This topic will suit someone interested in mixing statistics and programming.
For more information, please click here.
Supervisor: Professor David Wooff.
Funding: Please enquire.
Start date: Flexible.
| PhD project: Others |
There are several other routes to apply for a PhD studentship. The project topic could be chosen within the scope of any of our research interests, in accordance with a potential supervisor. If you are interested in such an open topic project, please get in touch either directly with the relevant supervisor, or enquire here for general information.
Supervisor: Flexible.
Funding:
- For UK students, funding for one studentship is likely to be available via EPSRC. Please apply as soon as possible via the Graduate School.
- For EU students, EPSRC provides tuition fee waivers (without subsistence etc.). Please apply as soon as possible via the Graduate School.
- Chinese Students may consider applying for a CSC award. Deadline for applications to the Graduate school: Usually in January.
- Outstanding students (from anywhere in the world, including UK/EU) may apply for a Durham Doctoral Studentship (DDS). Deadline for applications to the Graduate school: Usually in February.
- If you come with your own funding (for instance, from your home Government or University), simply apply anytime via the Graduate School, stating your funding source in the application form.
Start date: Please enquire.
Back to the Statistics and Probability Postgraduate Page.
Last updated: 27/11/2014, JE.