Suggestion about a new set of constraints

[choko]

Hello, choko,

It seems you are a mathematician who has a good grasp of FET? You identified that FET will have difficulties moving activities in this situation. I would have hoped that FET would be clever here. I did my best to avoid cycling.

Really thank you so much for your appreciation. That really means a lot to me. And again, your work is very very much appreciated.

I am terrible sorry If I made you feel offended. I can assure you that it totally was not my intention at all.

I myself, as a graduate in Mathematics and computer science, actually tried to use linear programming to write an excel program that tried to generate a timetable. But it turned out that the pivot table was too large that I could never handle. And then I found FET very much useful and I have been using it and admiring you since then. Thank you so much for your help and reply.

Maybe you could force FET to place firstly D6..D50 and F6..F50, by using the option 'group activities in the initial order'. You can see the normal order of activities, see what activity is first, and group D6..D50 and F6..F50 with it (so you will have 91 activities in this group activities in the initial order item), or group D6..D50 and F6..F50 with another activity which is near them and before D1..D5 and F1..F5.

In my situation, those E1 to E4 are actually very large blocks. Simply speaking, each of E1 to E4 has to be held at the same time with, 7 more activities. And indeed, I have like 20 more activities like E1, making a totally of 8x5x3, 120, activities to be held at 21 time slots. Even worse, some of them have different duration. So I am not sure if it is a good idea to allocate these activities in the last steps. But I will try to see if it works.

Another solution is to use space constraints. Maybe you could share your file with us/me (here/my email) and I'll try to see it.

If you categorically don't want to use space constraints, then Vangelis' suggestion is exactly what you need? It is a nice constraint, but difficult to implement. I am not sure its algorithm of timetable generation is feasible for FET. But, as I said, if we find sponsors for it, I might give it a try. Until then, it is in the TODO list.

The space constraints have been used in my fet file. And I am still figuring out if there is anything I misunderstood. Will definitely try it again.

I am not sure I'll answer so fast in the next interval, I am preparing a new FET version tonight (very minor changes). But if your problem is urgent, please mention this and I'll see it sooner.

My problem is totally not urgent. Still trying to figure out if there is any other method that help.

Again, your work and help are very very much appreciated! Thank you so much.

[Liviu Lalescu]

Thank you very much for the appreciation and kind words!

I would like to check the bottleneck you are reporting (FET not being able to swap the necessary activities in a short time), but I cannot do anything without your file.

I hope a good selection of space constraints will solve the problem. I could try to help you here, but again I need your file for this.

[choko]

Thank you very much for the appreciation and kind words!

I would like to check the bottleneck you are reporting (FET not being able to swap the necessary activities in a short time), but I cannot do anything without your file.

I hope a good selection of space constraints will solve the problem. I could try to help you here, but again I need your file for this.

Just sent an email. Thank you so much again.

[Liviu Lalescu]

You are welcome! Thank you, too!

I hope I found the problem and improved the file. Probably the constraint suggested by choko and Vangelis is not needed for this case (but it might be useful in other cases).

choko's file is now solvable much faster and easier. I am waiting for his feedback to make sure it is correct.

[Liviu Lalescu]

Hmm... the problem is not solved if we need to take care of all the practical requirements in choko's file. Maybe the problems would be solved by the suggested constraint, ""a set of activities share min common time slots with any of another set of activities" (and maybe in selected time slots)".

[choko]

You are welcome! Thank you, too!

I hope I found the problem and improved the file. Probably the constraint suggested by choko and Vangelis is not needed for this case (but it might be useful in other cases).

choko's file is now solvable much faster and easier. I am waiting for his feedback to make sure it is correct.

Yes. your suggestion and modification boosted the flexibility by removing redundant constraints in my situation. Despite some of the unsolved practical problems, the generation takes only a very acceptable time and the result is quite satisfactory.

Million thanks again