|
CBR for Timetabling:
Additional information
|
|
|
Objectives: This project will seek to
investigate a variety of meta-heuristic, heuristic and hybrid methods
applied to university timetabling to gain a better understanding of which
methods work best on which problems and also to study why a certain
technique is better on some problems than others. Resulting from this will
be a study of the key features of the problems involved with regards to
measuring similarity between these problems. This knowledge will then be put
into a case-based reasoning system which will attempt to match new
timetabling problems with those in its case-base according to how similar
they are. The case-base will contain a range of university timetabling problems together with the technique which was successfully applied to obtain high quality solutions. The technique used for the most similar problem in the case-base can then be retrieved and applied to the new problem with a high degree of confidence that it will produce good quality results. The overall aim of our case-based reasoning system will be to provide a method of solving new timetabling problems to a high standard by using past experience of problem solving. Whilst this technique is unlikely to beat the results of specially designed problem-specific techniques, it should provide competitive results saving a large amount of time and work in discovering the best technique to implement. The project has resulted in a large number of papers having been published and submitted for publication.
|
