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 encountered a peculiar issue with Z3 incorrectly returning unsat, for a formula in the 'QF_NRA' logic. Funny enough, if I add an additional constraint, Z3 will realize that it is indeed sat. This problem occurs with version 4.13.0 of Z3, I tried an older version (4.8.6), which doesn't seem to suffer from this problem.
I tried to introduce named assertions and use the (get-unsat-core) feature to investigate this a bit further, however then z3 doesn't terminate (at least not after 20 minutes, where it previously took 2 seconds without the (get-unsat-core)).
Attached is also the zipped version of the (get-unsat-core) file.
Hope you can help figure out what is going wrong here.
Hi Z3 team,
I encountered a peculiar issue with Z3 incorrectly returning unsat, for a formula in the 'QF_NRA' logic. Funny enough, if I add an additional constraint, Z3 will realize that it is indeed sat. This problem occurs with version 4.13.0 of Z3, I tried an older version (4.8.6), which doesn't seem to suffer from this problem.
In more detail:
We have some
Real
variables:that once each is assigned
8
should give sat. And indeed asserting:returns sat. HOWEVER, removing all 3 assertions and just leaving the last line:
makes the formula unsatisfiable. Leaving any of the 3 assertions on the individual variables is also sufficient to make z3 realize it is sat.
The full
.smt2
file can be seen below:I tried to introduce named assertions and use the (get-unsat-core) feature to investigate this a bit further, however then z3 doesn't terminate (at least not after 20 minutes, where it previously took 2 seconds without the (get-unsat-core)).
Attached is also the zipped version of the (get-unsat-core) file.
Hope you can help figure out what is going wrong here.
Cheers, Aleksander
files.zip
The text was updated successfully, but these errors were encountered: