[Clp] Clp Digest, Vol 117, Issue 1

[Imre Pólik]

Hi Andy,

Interior-point methods give you exactly that: (apart from denegeracy and
maybe the presolver) the optimal solution returned by the algorithm is
unique and it changes continously with the constraints.

If you also want to include degeneracy, then I have bad news: the optimal
solution changes in a non-continous way, and the discontinuities happen
when when a constraint is degenerate.
Imre

On Sun, Mar 1, 2015 at 12:00 PM, <clp-request at list.coin-or.org> wrote:

> Send Clp mailing list submissions to
>         clp at list.coin-or.org
>
> To subscribe or unsubscribe via the World Wide Web, visit
>         http://list.coin-or.org/mailman/listinfo/clp
> or, via email, send a message with subject or body 'help' to
>         clp-request at list.coin-or.org
>
> You can reach the person managing the list at
>         clp-owner at list.coin-or.org
>
> When replying, please edit your Subject line so it is more specific
> than "Re: Contents of Clp digest..."
>
>
> Today's Topics:
>
>    1. Lexicographic Linear Optimization (Andy Somogyi)
>
>
> ----------------------------------------------------------------------
>
> Message: 1
> Date: Sat, 28 Feb 2015 14:37:00 -0500
> From: Andy Somogyi <andy.somogyi at gmail.com>
> To: clp at list.coin-or.org
> Subject: [Clp] Lexicographic Linear Optimization
> Message-ID: <39E324C5-9326-4E97-8590-56E23BC1D165 at gmail.com>
> Content-Type: text/plain; charset=windows-1252
>
> Hi All
>
> Would anyone here have some tips on how to efficiently implement a
> lexicographic linear optimization in CLP?,
>
> Basically something similar to the approach outlined in:
>
> K. H?ffner, S. M. Harwood, and P. I. Barton, ?A reliable simulator for
> dynamic flux balance analysis,? Biotechnology and bioengineering, vol. 110,
> no. 3, pp. 792?802, 2013.
>
>
> Currently, I?m doing a pretty cheese approach of performing an iteration
> over the optimization function coefficients, where I zero all of the
> coefficients, except for the present one, calculating the primal solution,
> then set this as a constraint and continue on to the next coefficient, i.e.
>
> for i,c in orderedObjectives:
>     zeroModelObjectives();
>     setModelObjective(i,c);
>     primal();
>     setModelConstraint(i, getPrimalSolution(i));
>
>
> Surely this is a very inefficient approach.
>
> The reason for doing this is that I need a unique solution to the LP
> problem, and I need that solution to vary continuously in response to
> smoothly varying constraints.
>
> Thanks
>
> ? Andy Somogyi, PhD
> Research Associate,
> Biocomplexity Institute,
> Indiana University Dept of Physics
>
>
> ------------------------------
>
> _______________________________________________
> Clp mailing list
> Clp at list.coin-or.org
> http://list.coin-or.org/mailman/listinfo/clp
>
>
> End of Clp Digest, Vol 117, Issue 1
> ***********************************
>



-- 
http://imre.polik.net
-------------- next part --------------
An HTML attachment was scrubbed...
URL: <http://list.coin-or.org/pipermail/clp/attachments/20150301/3ccc1b3e/attachment.html>