Hi
When you set the constraints of the times in which the teachers are not available, maybe you don't know how many teachers wil be finally available for each time slot (unless you count them manually). So, my suggestion is to include a new statistic to show that information, in order to correct impossible constraints before start the timetable generation. Of course, an authomatic detection (less teachers available than needed) would be better!.
Cheers
Sorry, but I do not understand. I already do a check on the number of activities for each teacher, versus his not available times.
Your suggestion isn't that easy as you might think now. It's easy if teacher/students have full timetables (all timeslots are used), but it's complicated if the tables are not full.
Liviu already include common checks like that. The checks also need to care about timetables that are not full (not all timeslots are used). Maybe the checks can be improved if the timetables are full, but that mean: a) you must code a check that check if the timetables are full. b) it will help only a few datasets, because most datasets (compare examples) don't have full timetables.
Excuse me if my previous post wasn't clear!
Of course, Volker is right: it will be a weak information know how many teachers are available for each time slot if the timetable isn't full. Furthermore, the fact to have for each of all time slots more teachers than groups don't will guarantee that the whole problem will be solvable. But if the timetable is full, and this condition fails at some time slot, it will be certain that the problem will be impossible to solve.
The idea arose when I was entering the preferences of my workmates and I suspected that too many of them were requesting be free at some times, i.e. last class of friday (of course!). Then I missed to have a simple account as an alternative in order to avoid passing sheet after sheet, looking for the petition and count them.
Anyway, thank you very much!
Now I understood - I'll add this in the TODO.