I made a fet file
the same file i opened in 4 computers in the school lab using the same windows (7)
i used the same fet Algeria41 in the 4 computers
one of them solved at 5 hours
BUT THE OTHERS AFTER 20 hours these are the results:
the second one reached 687 out of 693
the third one 689
the fourth one 678
by the way all computers have the same matrial
because fet always start with an other seed.
if you place the same seed, you will get similar results.
please go to fet -> settings -> seed
and just enter 2 numbers.
repeat the test with the same 2 numbers and you will get exactly the same result in the same time.
if you do it on an other computer, then you will also get exactly the same result. Time might be of course higher or lower, depending on your computer.
maybe have a look at this:
https://lalescu.ro/liviu/fet/forum/index.php?topic=166.0
thanks mr VOLKER
Can I always put the random seed at 1 for both x and y?
if you want to do a benchmark? yes.
for a default user it won't be clever to do that.
YES SIR
BUT IF I HAVE A DIFFICULT TIMETABLE AND I WANT FET ENGINE TO TRY ALL THE POSSIBILTIES NO MATTER HOW OFTEN IT TAKES IT ONLY MATTERS THAT IT SOLVES AT THE END, WHAT TWO NUMBERS I SHOULD TRY FOR X AND Y
a random one. so in best case you should fet randomly select a number, since a human is not random enough. there is no "rule" behind that. it just should be random. it is bad to set always the same.
i have very very difficult timetables and with my dataset i notice that it is better to run multiple timetables with a 20 minutes time limit (so fet shoose random numbers).
with my dataset it is bad to generate only a single one, since in worst case i even got after 2 weeks no solution.
maybe it is a bit like playing tennis and asking if you should do a long hit or a short hit. If you should play it to the right side or to the left side. It is bad if you always do it the same. You must play randomly if you want to win.
thank you sir
i will try multiple
i have a question:
if I reach 691 out of 693 activities done
can I take the random seed of the file + omit just one constraint?
in nearly all cases:
if you modify, add or remove a single constraint it mean: If you take the same seed the results can be totally different and also the needed time might be totally different.
thanks sir :)