List of Possible Research Projects in the Applied and Interdisciplinary Informatics Research Group
These projects are only suggestions and examples of areas that the research group is interested in, and not all members of staff who are interested in supervising students are listed. Potential research students are encouraged and welcome to produce their own suggestions in these research areas or other research areas in the field. Moreover, it is unlikely that a research project will remain fixed throughout the lifetime of a studentship as students develop a greater appreciation of the subareas that are of most interest to them.
Please email cgj@kent.ac.uk Colin Johnson, the Head of the Research Group, telephone 01227 827562 or visit the group home page: http://www.cs.kent.ac.uk/research/groups/aii/ for further information.
| Genetic Programming with Estimation of Distributions and Principal Component Analysis | ||
|---|---|---|
| Genetic Programming (GP) is a novel way of
generating computer program code, based on an evolutionary process
where a population of programs changes over time via a process of
selection, crossover and mutation. This has proven effective at finding
programs that solve problems in a number of areas. Recently a new GP-type process has been developed, called Tree-Swarm Optimization (Veenhuis et al., 2005). This is based on encoding programs as points in a high-dimensional space, then exploring this space by the particle swarm optimization algorithm. This opens up a number of new opportunities for exploring this space in new ways. One example would be Estimation of Distribution Algorithms, which uses statistical summaries of the population as a way of sharing information across the population. A second example would be Principal Component Analysis (PCA), where we can extract the most important directions in a highdimensional space. This could be used, for example, to extract useful subprograms from the population. The aims of this project would be to investigate these new techniques, and apply them to a number of case studies. |
||
| Contact: Colin Johnson | Email: cgj@kent.ac.uk | |
 |
||
| Area Proportional Euler Diagrams | ||
| Many data sets are visualized effectively with
area proportional Venn and Euler diagrams, where the area of the
regions formed by intersecting curves is in proportion to a defined
specification. Some work has already been performed on area
proportional Venn diagrams, where all possible intersection of curves
are present. However, Euler diagrams are arguably a more effective
visualization method, because some of the regions can be empty, and so
the diagram can be less complex. No significant research has yet been
performed on area proportional Euler diagrams, and so this PhD project
would be breaking new ground. This project would develop novel techniques for generating area proportional Euler diagrams. An effective approach would be to initially look at diagrams of limited size, for instance three and four curve diagrams. Classes of embeddings would be derived and geometric methods for ensuring the areas of regions are correct would be developed. Following this, more general embedding methods for arbitary curve sizes would be produced, based on non-area proportional generation techniques. |
||
| Contact: Peter Rodgers | Email: pjr@kent.ac.uk | |
|
||
| Bio-molecular memory elements | ||
|
This project will look at possibilities to design, model and
(mathematically) analyse distributed memory systems based on realistic
biomolecular interactions. The questions to be addressed during this
research is (i) how can data be efficiently stored in biomolecular
systems (ii) how can stored data be retrieved and processed (iii) what
are the fundamental properties of bio-molecular systems, in particular,
what are the theoretical limitations on storing data in biological
systems. This |
||
| Contact: Dominique Chu | Email: dfc@kent.ac.uk | |
|
||