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
I was exporting some models that were taking a long time to solve to CNF to check with an external solver. I was surprised when, for models I knew instances existed, the solver returned that they were unsatisfiable.
I then noticed that the exported CNF instances are identical independent of the number of steps I give them.
A simple example to reproduce:
sig B {}
sig A{ var b : set B }
fact init { no b }
pred step { some bval : B | bval not in A.b and A.b' = A.b + bval }
pred end { A.b = B}
fact stepsuntilend { step until end and eventually end }
run {} for exactly 3 A, exactly 2 B, 3 steps
run {} for exactly 3 A, exactly 2 B, 1 steps
The first one will find an instance just fine, while the second one won't (as 1 step is not enough). However, exporting to CNF will produce the same SAT instance (corresponding to one step, I presume).
The text was updated successfully, but these errors were encountered:
I was exporting some models that were taking a long time to solve to CNF to check with an external solver. I was surprised when, for models I knew instances existed, the solver returned that they were unsatisfiable.
I then noticed that the exported CNF instances are identical independent of the number of steps I give them.
A simple example to reproduce:
The first one will find an instance just fine, while the second one won't (as 1 step is not enough). However, exporting to CNF will produce the same SAT instance (corresponding to one step, I presume).
The text was updated successfully, but these errors were encountered: