Open JoD opened 5 years ago
Thanks for the bug report. We are looking into it and will let you know once we fix it.
Regarding your other observation. A Pseudo-Boolean constraint is encoded into CNF using an encoding that is exponential in the worst case. Therefore, if the number of added clauses reach a given threshold, we just report "Overflow in the Encoding". However, this is done because the formula will just become too large and will not be solved by the SAT solver. The latest version of Open-WBO has a linear encoding for Pseudo-Boolean constraints that would avoid this issue at the cost of a worst performance from the SAT solver. We will change the encoding of the Pseudo-Boolean to this method if we estimate the encoding will blow up. Do you have an example of a formula with the message "Overflow in the Encoding"?
I have about 15 instances, I've attached the smallest one (remove .txt extension). smallest.opb.txt
Thanks for looking into this!
(sorry for closing and reopening, pushed the wrong button apparently).
Hey @rubengmartins! Any progress on this issue? I'm evaluating different PB solvers and would definitely like to include Open-WBO in my experiment. Kind regards - Jo
Trying out open-wbo for pseudo-Boolean problems. Got a floating point exception for instance on pastebin when running the command
open-wbo_static -formula=1 <instance>
. Compiled open-wbo with the defaultmake rs
command, with the default makefile.On a related note, on some other instances, I get
Overflow in the Encoding
. What are the coefficient / degree limits for open-wbo?