SCIP, GLOP in MPSolver: other possibilities ?

[Emmanuel Seve]

Hi ,

I'm starting to use OR-Tools and il have a programme running with MPSolver. 

For now, it's working perfectly with SCIP but for a question of license i can't continue to use it. I need an apache v2 license. GLOP is an option but GLOP can not handle integer variables and i have a mix of double and integer variables. 

I could use CP-SAT by dividing my double variables to "transform" them into integer variables but if possible i would like to continue with the MPSolver option.

Thanks,

Emmanuel

[Emmanuel Seve]

correction: dividing my new double variables with a sufficiently high number of course

[Laurent Perron]

You can use the CP-SAT backend with mp-solver. 

It will try to scale continuous variables to discrete ones. 

It works better with 8 or more threads. 

--
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 on the web visit https://groups.google.com/d/msgid/or-tools-discuss/024ddb62-c655-4be9-9c2d-98332723d487n%40googlegroups.com.

[Emmanuel Seve]

Thanks Laurent for your fast answer

What do you mean by "try": not sure about the answer ?

I'm novice with OR-tools , so can you explain about thissentence " It works better with 8 or more threads". 

So no possilities with MPSolver ?

[Emmanuel Seve]

No sure to understand this neither :  "CP-SAT backend with mp-solver"

Le samedi 15 janvier 2022 à 17:13:46 UTC+1, Laurent Perron a écrit :

[Laurent Perron]

1) replace "SCIP" by "SAT" when creating the MPSolver.

2) solver.SetNumThreads(8)  # Does not work with SCIP

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

To view this discussion on the web visit https://groups.google.com/d/msgid/or-tools-discuss/20cecc53-235a-409d-bc8d-48ba0ac64c6en%40googlegroups.com.

[Emmanuel Seve]

Merci beaucoup Laurent,

It works perfectly. To be sure i changed the unity of my data so i can safely declare them as integer.

You confirm that a code using the MPSolver with the CP-SAT backend is well under apache v2 

[Laurent Perron]

CP-SAT is part of or-tools, thus apache 2.

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

To view this discussion on the web visit https://groups.google.com/d/msgid/or-tools-discuss/64d0ebf9-56f4-451e-800d-f1a9689a490an%40googlegroups.com.