This paper (click the title to get a postscript copy) was presented at The First International Conference on the Practice and Theory of Automated Timetabling in Edinburgh in 1995 and is based upon a survey of British Universities carried out around Spring 1995. The actual survey used is available (in postscript format) here or from the Papers section of our ftp server under questionnaire.ps.
The authors are quite happy for anyone who wishes to use this survey as a basis for their own survey of either another country's Universities or for another similar problem (eg. course timetabling).
To help in this respect I have included an annotated version of the questionnaire below and hopefully, in the near future, will be able to invite comments over the web either on the significance of the results or on how the survey may be made even more clear or altered for a different domain. Discusion is very welcome on TTP - The Timetabling Mailgroup.
Back to the Timetabling Page.
The survey was split into four sections. These were:
In this section, the timetablers were asked to describe various features of their particular timetabling problem. In particular:
Obviously the more detailed information available, the more useful the results of the survey. It is, however, difficult to elicit such information without precise questions and precise questions often miss the mark when there is so much variation between problems. We found, for example, that a number of Universities did not use the concept of periods at all and that, in some cases, the amount of space available varies according to how many outside buildings are available.
Two questions that should have been included in the original survey are:
Some strange results came with the question "How many different lengths of exams are there?" It appears that some universities will allow extra time of a variable length to special needs students. This is the best explanation so far.
The survey asked whether a computer was used, as a scheduler or merely to hold data, and how much time was spent on each part of the process.
Very few universities were forecoming with any documentation for their system and we are very grateful to those who did. Few reported using a commercial package although they may have been more reticent in answering the survey. I suspect that several universities have misunderstood the survey and claimed to have an automated scheduler whereas what they really have is a program that helps them do it themselves.
This section asked about constraints used. Fourteen constraints were given and another eleven found that had not been previously thought of. Three questions were asked for each constraint:
Click here to see the constraints found by the survey.
This section simply asked for other relevent information not found on the rest of the questionnaire.