Accurately predicting the number of students likely to enrol in particular units is important for universities to construct timetables and determine workload allocations for academic staff.
By the time that student numbers are confirmed for a unit, timetabling has been close to finalised and workload allocations already determined. Therefore, mismatch between predicted and allocated unit enrolments leads to sub-optimal timetabling and workload allocation.
This project aims to develop and evaluate a methodology to provide some analytical rigor to a process that has otherwise relied on ad-hoc spreadsheets and educated guess work.
The methodology will involve formulating a graph theoretic framework to describe possible course plans and unit selections and applying monte-carlo simulation over a large number of replications to determine confidence intervals of enrolment numbers for each unit. Historical data will be used to calibrate and validate the approach.
Throughout the research project, you will be involved in:
- literature review
- data gathering
- simulation model formulation
- computational analyses
- scientific writing.
We expect a thesis and/or journal publication and a proof-of-concept level end-user tool to be produced as part of the research project.
Skills and experience
We expect you to have some experience with simulation tools and good programming capabilities in C#, C++, Java, Python or MATLAB.
Contact the supervisor for more information.