Same model: infeasibility using presolve and optimality without

[Niklas Terés Leute]

Hi,

I’m using CP-SAT (latest version) and seeing unexpected behavior. With the exact same model, the solver returns “infeasible” when presolve=True, but “optimal” when presolve=False. I have the exported model file (pb.txt) but it seems I can't attach it here. Let me know if I can send it to you somehow.  

Additionally, the infeasibility with presolve only occurs when I add the cumulative constraints in a certain order; if I change the order (even though the constraints are identical), the solver returns “optimal”.

Could this be a bug related to presolve?

Thanks,

Niklas

[Laurent Perron]

Definitely. 

Please send the proto to me. 

--Laurent

--
You received this message because you are subscribed to the Google Groups "or-tools-discuss" group.
To unsubscribe from this group and stop receiving emails from it, send an email to or-tools-discu...@googlegroups.com.
To view this discussion visit https://groups.google.com/d/msgid/or-tools-discuss/f61175cb-e0b2-4e7f-aaea-5388384f7011n%40googlegroups.com.

[Niklas Terés Leute]

Sent. Thanks!

[Laurent Perron]

Have you tried with 9.15 ?

Laurent Perron | Operations Research | lpe...@google.com | (33) 1 42 68 53 00

To view this discussion visit https://groups.google.com/d/msgid/or-tools-discuss/2051f5bc-e116-435e-bf07-b20ef08c3573n%40googlegroups.com.

[Niklas Terés Leute]

I used 9.14.6206. I just realised 9.15 had been released. It still hasn't been released on pypi, has it?

[Laurent Perron]

Not yet, but the wheels should be on github

To view this discussion visit https://groups.google.com/d/msgid/or-tools-discuss/63b5b8a1-576f-4eb4-8cd0-f7e41d95ca3en%40googlegroups.com.

--

--Laurent