You signed in with another tab or window. Reload to refresh your session.You signed out in another tab or window. Reload to refresh your session.You switched accounts on another tab or window. Reload to refresh your session.Dismiss alert
{{ message }}
This repository has been archived by the owner on Jul 2, 2020. It is now read-only.
Using the following program to solve a very easy instance of n-queens problem leads to an infinite solving time. This problem only occurs when we request to print all the solution (-a) found using an ActivityBased strategy (-bb 3).
And thank you for the bug report.
The thing is: looking for all solutions of a satisfaction problem with a black-box search strategy may lead to find the same solution more than once.
In your case, there is a combination of unlucky parameters that produces a never ending loop...
Indeed, ABS restarts on each solution and during the probing phase, on each failure.
In addition, a fast restart strategy is added to force restarting every 500 failure.
Finally, a nogood-from-restart option is plugged-in, that should prevent the same search tree to be rediscovered again and again.
But the latter is not able to extract nogood when no decision are refuted, which is the case here due to ABS that is good at building path to solutions.
Long story short, you must add a nogood-from-solution option, which is not available as a CLI option...
I'll add it and let you know.
Best
# for freeto subscribe to this conversation on GitHub.
Already have an account?
#.
Using the following program to solve a very easy instance of n-queens problem leads to an infinite solving time. This problem only occurs when we request to print all the solution (-a) found using an ActivityBased strategy (-bb 3).
queens.xcsp3:
The text was updated successfully, but these errors were encountered: