Reading MPS files

190 views
Skip to first unread message

Mark

unread,
Jul 12, 2018, 8:32:54 PM7/12/18
to Pyomo Forum
It's been a while since I've used pyomo and I see there's been a lot of development progress - great work!

Is it now possible to read a MIP in .mps format, modify it using the normal pyomo api and solve it using CBC?

Bynum, Michael Lee

unread,
Jul 12, 2018, 10:38:57 PM7/12/18
to pyomo...@googlegroups.com
I could be wrong, but I don’t think this is quite possible.

Michael
On Jul 12, 2018, at 6:31 PM, Mark <slash...@gmail.com> wrote:

It's been a while since I've used pyomo and I see there's been a lot of development progress - great work!

Is it now possible to read a MIP in .mps format, modify it using the normal pyomo api and solve it using CBC?

--
You received this message because you are subscribed to the Google Groups "Pyomo Forum" group.
To unsubscribe from this group and stop receiving emails from it, send an email to pyomo-forum...@googlegroups.com.
For more options, visit https://groups.google.com/d/optout.

Watson, Jean-Paul

unread,
Jul 12, 2018, 11:01:15 PM7/12/18
to pyomo...@googlegroups.com
I concur - to the best of my knowledge, there is no such capability. Although, to contradict Michael, it is “possible” :)

One of the reasons we haven’t developed this capability is that the model would be “flat” and therefore almost impossible to manipulate - basically, of the form Ax>=b, lacking all symbolics. Which, perhaps arguably, moves it outside the realm of an algebraic modeling language. So we can think about this more, can you provide a sketch of your use case?

Jpw

Mark

unread,
Jul 13, 2018, 9:16:30 AM7/13/18
to Pyomo Forum
It's very common for me to have model instances stored in either .lp or .mps format from project to project.
I was hoping that it would be possible to work with these in pyomo similar to how one can load them into CPLEX or Gurobi, but with using an open-source solver such as CBC.
Honestly, 90+% of the time I wouldn't need to do much in the way of manipulation; I'd mostly just be looking to solve the model using different solver hyper-parameters and runtimes.
It would be useful if one could preserve the same variable names and constraint names though.  It seems to me that shouldn't be too difficult a thing to do.
Reply all
Reply to author
Forward
0 new messages