Within a university setting you’d typically find a consensus amongst most staff and students that timetabling examinations is a reasonably straightforward process. Primarily, a simple matter of ensuring students are not timetabled to sit more than one examination at the same time. It would be normal as a starting point to make sure no clashes exist in the timetable. This is known as a potential or ‘feasible’ solution and is, pretty much, where you end up with when using a basic piece of timetabling software.
With more advanced software tools we can, at this point, look to the power of optimisation to help with the provision of quality. By this we mean ensuring that the overall schedule is the best it can be for every student – that it’s optimal. But what does this actually mean? Well, in this context we have equated optimal with ‘best’. Best usually means students have the maximum ‘spread’ of examinations across the designated examination period. This can be measured by factors such as the number of students with back to back examinations, the number of students with overnight examinations or the number of students with a minimum of X hours between examinations (where X can be defined by the timetabler). Now, of course, each institution wants the best overall schedule – but how can we be sure this is ‘fair’ for each individual student. Can we guarantee the equitable nature in terms of quality for each student?
In providing the optimal or best timetable we have to quantify and measure spread and fairness, and provide transparency and accessibility. Then, of course, we have to balance this with the optimal number of rooms, their optimal sizes, the optimal number of periods per day, their length and the optimal number of days and weeks. And there are usually a host of constraints to deal with related to ordering and fixed dates or weeks. And, then, there are further factors around special arrangements for individual students and scheduling invigilation staff optimally to oversee the entire process cost effectively. (It may be obvious to you at this point that examination timetabling, which initially looked like a pretty clear-cut process, is really not quite so straightforward.) There really are a multitude of factors to consider when developing a high quality timetable for each student – which, of course, requires software with some very complex computation going on under the hood.
EventMAP’s exam timetabling software builds in a body of research on optimisation algorithms that stretches back almost 20 years and now includes the more recent concept of fairness that’s been the subject of recent research that we had input into. Over the years we’ve used this body of research to continually refine our exam timetabling products, to the point where a product like EXAM lets timetablers harness all the complexity inherent in optimal solutions to create timetables that work beautifully for everyone involved.
