Hello! Liviu
Why does FET run for 15 minutes successfully, sometimes for 30 minutes, but not for 10 hours?
tôi đã gửi file cho bạn qua gmail
There is no such rule. If you have a very very difficult timetables, then the timetable can be solved in less than 5 minutes if you have luck; but if you have bad luck it will be solved after 5 days or even 2 weeks. Please view my YouTube videos if you want to know more about this topic.
Hello, ngo,
Indeed, as Volker said, the generation is semi-randomized. I checked your file, could not get a timetable (generating on 8 threads) in 40 minutes. It is probably very difficult.
PS: If you want to upload your file here (if it is non-private), you can archive it if it is too large.
Why don't you report the error correctly at a certain time? I can't schedule it. For example, if after 15 minutes the scheduling fails, report the exact error. Let me reduce the weight or change that constraint
Sorry, but this is impossible.
https://www.mediafire.com/file/f7gutt1o5mxighz/TKB_dang_chay_ky_2_ep_sinh_ep_steam.fet/file
Can you try this file to see if you can get a schedule?
Because FET runs randomly, with unpredictable results, many users will have difficulty when encountering this situation
I can't download the file.
But i guess it is "only" a difficult file. In that case we don't know if it is solveable or not. Impossible to proof that by math. If you want to calculate all possibilities if the timetable is solveable or not, than current computers are much too slow. If you want to do a math poof of a difficult file, then a new computer needs more time to calculate all possible variants then you need much more years than the sun is old.
Your dataset is only difficult. I run it on multiple cores and the fasted so far solved it in 16 minutes. So it is correct that FET doesn't report an error, since it is possible to solve it.