I set room constraints in two ways:
- for younger students - a group has set of home rooms
- for older students - subject has preferred set of rooms
The problem arose when both younger and older students have the same subject. If I specify subject preferred set of rooms with weight 100% the timetable cannot be generated with this message:
QuoteCannot generate timetable, because for activity with id==nn you have no allowed preferred room (from preferred room(s) constraints).
If the weight is say 95% timetable can be generated but there is a lot of soft conflicts.
We have the similar case when it would be desirable to set for a subject e.g. two rooms out of ten with weight 100% and the rest eight with a little less weight (e.g. 95%) which would mean that it's preferable a subject should be in one of these two rooms and acceptable in other ones. When I set these two constraints we have again a lot of soft conflicts. Moreover, the statistics say that
QuoteYou have nn activities which might end up in an unspecified room
I think I'll have to set preferred rooms for activities unless there is another solution. The number of activities is much bigger than rooms and groups.
Looks like you have done 2 bugs.
about the 1st problem:
pleas go to FET -> Data -> Activities
Select the activity with id nn.
Check the right window.
There should be at least 2 space constraints that conflicts each other.
for example space constraint a) use room A with 100%.
and space constraint b) use room B with 100%.
That is of course impossible. If a) is fine - b) will have a conflict. Is b) is fine - a) will have a conflict.
The only exception from this is if there are exactly 2 space constraints. a) a home room and b) not a home room. In that case a) will be skiped.
About problem 2:
i think you mixed the weights. the constraint with 10 rooms must get the 100% weight while the constraint with the 2 rooms need an lower value.
if you have got a very difficult dataset you might think about deactivating that second constraints nearly all the time. So i search a solution without that constraint. As soon as you have got a fine table, open the _data_and_time solution from your results folder, activate the constraint with maybe 99,9999% and generate again. Now you will get a nice result pretty quick.
Quote from: liquid on May 12, 2018, 11:32:30 PM
We have the similar case when it would be desirable to set for a subject e.g. two rooms out of ten with weight 100% and the rest eight with a little less weight (e.g. 95%) which would mean that it's preferable a subject should be in one of these two rooms and acceptable in other ones. When I set these two constraints we have again a lot of soft conflicts. Moreover, the statistics say that QuoteYou have nn activities which might end up in an unspecified room
I think I'll have to set preferred rooms for activities unless there is another solution. The number of activities is much bigger than rooms and groups.
Please add 100% preferred rooms with all the 10 rooms (2+8) and 95% preferred rooms with only the 2 rooms, for this subject.