Author Topic: Min days between a set of activities with weight=0%  (Read 104 times)

0 Members and 1 Guest are viewing this topic.

Omar Ben Ali

  • Level 3
  • ***
  • Posts: 246
    • View Profile
Min days between a set of activities with weight=0%
« on: September 10, 2017, 01:19:44 AM »
Hello,

If I take the weight=0% in the contraint "Min days between a set of activities ", and fet can't generate a timetables, Is that mean that the problem has no solution ?
Thanks in advance

Liviu Lalescu

  • Forum Administrator
  • Level 5
  • *****
  • Posts: 5094
  • FET author and forum moderator
    • View Profile
    • Homepage
Re: Min days between a set of activities with weight=0%
« Reply #1 on: September 10, 2017, 02:50:52 AM »
There might be a problem with the activities or the other constraints.

Note: even if the weight is 0% but consecutive if same day = true, then the activities must be consecutive if they are in the same day.

Omar Ben Ali

  • Level 3
  • ***
  • Posts: 246
    • View Profile
Re: Min days between a set of activities with weight=0%
« Reply #2 on: September 10, 2017, 03:23:42 AM »
thank you for the fast answer.

I also set min days between a set of activities=1  of all activities and delete all space contraintes , fet seems unable to find a solution.

Liviu Lalescu

  • Forum Administrator
  • Level 5
  • *****
  • Posts: 5094
  • FET author and forum moderator
    • View Profile
    • Homepage
Re: Min days between a set of activities with weight=0%
« Reply #3 on: September 10, 2017, 04:28:12 AM »
You need to lower your time constraints.

What FET are you using? Official, Algeria or Morocco?

You may post your file here or on the dedicated board Get Help on Your Input File (say if you use a FET different from the official), but I cannot guarantee I will see it myself. I hope other users can help as well.

Omar Ben Ali

  • Level 3
  • ***
  • Posts: 246
    • View Profile
Re: Min days between a set of activities with weight=0%
« Reply #4 on: September 10, 2017, 06:01:54 AM »
I use morocco version of Fet .
I have also lower almost time constraints ...
It seems dificult or impossible .
Thanks

Liviu Lalescu

  • Forum Administrator
  • Level 5
  • *****
  • Posts: 5094
  • FET author and forum moderator
    • View Profile
    • Homepage
Re: Min days between a set of activities with weight=0%
« Reply #5 on: September 10, 2017, 08:49:18 AM »
I removed all the time and the space constraints and it solves very easily. So the activities seem correct.

I saw you have teachers max 2 afternoons per week. You might want to allow 3 or more.

I think your data is simply too constrained. You need to relax the constraints.

Omar Ben Ali

  • Level 3
  • ***
  • Posts: 246
    • View Profile
Re: Min days between a set of activities with weight=0%
« Reply #6 on: September 11, 2017, 01:49:03 PM »
I am sorry. I made a mistake. Indeed i added the constraint "Two activities grouped id=118 and id=119" to two subgroups of a set of students  wich contradict the constraint "students  min hours daily  =2" and " students max gaps per week=0". This make the generation of the timetables impossible.
The question is why fet  remains four hours looking for an impossible timetable ?

Liviu Lalescu

  • Forum Administrator
  • Level 5
  • *****
  • Posts: 5094
  • FET author and forum moderator
    • View Profile
    • Homepage
Re: Min days between a set of activities with weight=0%
« Reply #7 on: September 11, 2017, 01:57:27 PM »
I am sorry. I made a mistake. Indeed i added the constraint "Two activities grouped id=118 and id=119" to two subgroups of a set of students  wich contradict the constraint "students  min hours daily  =2" and " students max gaps per week=0". This make the generation of the timetables impossible.
The question is why fet  remains four hours looking for an impossible timetable ?

Because I did not implement a test to check for a mistake like that. There are too many possible mistake cases to treat them all.

Omar Ben Ali

  • Level 3
  • ***
  • Posts: 246
    • View Profile
Re: Min days between a set of activities with weight=0%
« Reply #8 on: September 11, 2017, 02:21:14 PM »
Ok, I see there are many impossible cases.
Thank you Mr Liviu.

Liviu Lalescu

  • Forum Administrator
  • Level 5
  • *****
  • Posts: 5094
  • FET author and forum moderator
    • View Profile
    • Homepage
Re: Min days between a set of activities with weight=0%
« Reply #9 on: September 11, 2017, 02:24:06 PM »
And there are many timetables in which you cannot find a clear impossibility (like you found), but which cannot be solved (FET is too weak or even there might not exist at all a timetable).

rodolforg

  • Level 1
  • *
  • Posts: 38
    • View Profile
Re: Min days between a set of activities with weight=0%
« Reply #10 on: September 11, 2017, 09:32:31 PM »
Liviu, Is there any interest on implementing this kind of issue?

Liviu Lalescu

  • Forum Administrator
  • Level 5
  • *****
  • Posts: 5094
  • FET author and forum moderator
    • View Profile
    • Homepage
Re: Min days between a set of activities with weight=0%
« Reply #11 on: September 11, 2017, 11:41:42 PM »
Liviu, Is there any interest on implementing this kind of issue?

Unfortunately, I cannot take care of all the possibilities.

rodolforg

  • Level 1
  • *
  • Posts: 38
    • View Profile
Re: Min days between a set of activities with weight=0%
« Reply #12 on: September 13, 2017, 08:12:55 AM »
Liviu, Is there any interest on implementing this kind of issue?

Unfortunately, I cannot take care of all the possibilities.


So, what about the reported ones? :)

Liviu Lalescu

  • Forum Administrator
  • Level 5
  • *****
  • Posts: 5094
  • FET author and forum moderator
    • View Profile
    • Homepage
Re: Min days between a set of activities with weight=0%
« Reply #13 on: September 13, 2017, 08:21:14 AM »
Liviu, Is there any interest on implementing this kind of issue?

Unfortunately, I cannot take care of all the possibilities.


So, what about the reported ones? :)

I did not understand why Omar Ben Ali's file was infeasible (his justification). I just believed him.

I could implement some additional checks, but it would not be useful. There are too many possibilities. The biggest problem is that I may reject a feasible timetable. This is what I am very afraid of, so that is why I prefer to keep things stable as they are. It is critical.

I am sorry to contradict with you, rodolforg.

Maybe you could explain to me more detailed why Omar Ben Ali's file was impossible.
« Last Edit: September 13, 2017, 08:35:31 AM by Liviu Lalescu »

rodolforg

  • Level 1
  • *
  • Posts: 38
    • View Profile
Re: Min days between a set of activities with weight=0%
« Reply #14 on: September 13, 2017, 09:21:09 AM »
Liviu, not a problem at all.
Besides, I misunderstood him: I thought he has two activities that should be grouped and at same time min days between them... (i guess I was led by original post subject)
I tested here and FET does not check this case.

By reading the Omar Ben Ali's post once again, I guess he has two activities that must be grouped, and each activity belongs to a student set subgroup. The problem he had, I suppose, it is the constraint "student max gap per week=0". The student set group would have a gap, as the subgroups' activities would not be simultaneous. In fact, this case would be too difficult to detect/check.