There are many opportunities to do a PhD in the School of Computing Science at the University of Glasgow. Some current opportunities and projects are listed below; contact the School for further information.
The School of Computing Science has a number of Scottish Graduate Academy prize studentships allocated to it which are available to excellent candidates from across the world. There is no residence qualification and candidates from any country may apply.
We are willing to consider applicants in any area of computer science and informatics but may give preference to students who are working in SICSA theme areas:
Specific research topics suggested by supervisors at the School are shown on the SGA website here: http://www.sicsa.ac.uk/graduate-academy/prize-studentships/phd-projects-2010
Applicants must apply first to the University of Glasgow for a PhD place and also to SICSA for a studentship.
The University of Glasgow School of Computer Science, has a CASE studentship along with Sciencesoft Ltd for a PhD student to work on the problem of visualisation of large oil reservoirs.
The student will have their fees paid, a standard EPSRC stipend, plus a top up provided by the commercial partner.
Applications should be directed to Dr Paul Cockshott (wpc@dcs.gla.ac.uk)
Project Background
The accurate and efficient three-dimensional visu- alisation of reservoir simulation models is vital to the petroleum industry, where the production from oil and gas fields is simulated using complex 3D models of the reservoirs. There is increasing demand that the uncer- tainty in the results of these computer simulations be reduced through the use of finer-scale models of the reservoir and the trend is to build higher-resolution 3D models to simulate advanced processes such as water injection into reservoirs and carbon capture by storing CO2 in spent oil fields.
The largest models generated are limited by the avail- able hardware, but the increasing availability of lost- cost, multi-core computer systems has facilitated the growth in the size of reservoir model that can be sim- ulated, with multi-million cells models now common- place. Finer-scale models are being designed to allow more accurate modelling of physical processes. At the current rate of increase 10–50 million cell models will be used within the next two-to-three years, which will overwhelm the existing generation of graphics cards and displays. For models with even a few million cells, individual cells will be rendered below the scale of a single pixel of the display when viewing the full reser- voir, and all sense of detail is lost. Even allowing for the development of graphics and display hardware, the expected growth in performance will be siggnificantly slower than the rate of increase in model size.
This generates a need for new algorithmic techniques and more economical forms of data representation that could be use to give visually acceptable rendering of very large datasets.
The computer vision and graphics group at the School of Computer science has for some 15 years now been researching the capture and visualisation of 3D data. We have past experience in developing 3D capture tech- nologies; 3D modeling techniques for human bodies for medical and entertainment purposes; 3D rendering techniques for surfaces and point clouds; 3D volumet- ric compression techniques for medical cross sectional data. In addition we have worked on several image compression mechanisms. A common feature of much of our previous work has been the use of scale space techiques, which allow as much work as possible to be done on a reduced scale model of the problem.
We believe that this general approach is likely to be fruitful in the work that Sciencesoft are engaged upon, so the development of appropriate scale space techniques for geological modelling will provide the foundation paradigm of the research.
Page maintained by John O'Donnell, jtod@dcs.gla.ac.uk, www.dcs.gla.ac.uk/~jtod/