[Couenne] Smallish optimization problem

[Ted Ralphs]

Hi Juhani,

Yes, the entire Couenne stack is under the EPL. I'm not too familiar with
C#, but it looks like you can easily call C++ from it and the API of
Couenne is in C++ (you could also check out the API of the Optimization
Services project: https://github.com/coin-or/OS/wiki). Both Couenne itself
and Optimization Services produce .nl files, which can be passed directly
to Couenne (that is the only way to get a problem in to Couenne at the
moment). If you want a modeling language, the easiest way to produce a .nl
file without AMPL (using open source tools) for this type of model would be
to use Pyomo (https://github.com/pyomo). You could also check out JuMP (
https://github.com/JuliaOpt/JuMP.jl). By the way, for convex quadratic
problems, Cbc should be all you need.

Hope this helps!

Cheers,

Ted

On Mon, Aug 12, 2019 at 8:33 AM Juhani Miettunen <
juhani.miettunen at piimega.fi> wrote:

> Hello all,
>
>
>
> We have a sawmill ERP SW product done with Microsoft tools. Programming
> language is C#. A new feature for the product requires mixer integer
> optimization with quadrative objective (least squares). Problem is quite
> small, some data from the NEOS result report below.
>
>
>
> I made an AMPL model about the problem and Couenne works fine with it. I
> used NEOS web service for the testing. Other solvers seem to work fine as
> well, including CBC.
>
>
>
> A few questions:
>
>    - Is Couenne and all the libraries used by it under EPL license (or
>    other business friendly license)
>    - Is there a suitable API available for Microsoft .NET and C#
>    - Is there an easy way to convert the AMPL model to .nl or directly to
>    suitable matrices without AMPL license
>
>
>
> Any good advice how to get Couenne embedded to our SW package?
>
>
>
> Kind Regards,
>
> Juhani
>
>
>
>
>
>
>
> You are using the solver couenne.
>
> Checking ampl.mod for couenne_options...
>
> Checking ampl.com for couenne_options...
>
> Executing AMPL.
>
> processing data.
>
> processing commands.
>
> Executing on prod-exec-3.neos-server.org
>
>
>
> Presolve eliminates 53 constraints and 1 variable.
>
> Adjusted problem:
>
> 21 variables:
>
>         10 binary variables
>
>         11 nonlinear variables
>
> 19 constraints, all linear; 64 nonzeros
>
>         19 inequality constraints
>
> 1 nonlinear objective; 11 nonzeros.
>
>
> _______________________________________________
> Couenne mailing list
> Couenne at list.coin-or.org
> https://list.coin-or.org/mailman/listinfo/couenne



-- 
Dr. Ted Ralphs
Professor, Industrial and Systems Engineering
Lehigh University
(610) 628-1280
ted 'at' lehigh 'dot' edu
coral.ie.lehigh.edu/~ted
-------------- next part --------------
An HTML attachment was scrubbed...
URL: <http://list.coin-or.org/pipermail/couenne/attachments/20190812/66c846f2/attachment.html>