[Top][All Lists]

[Date Prev][Date Next][Thread Prev][Thread Next][Date Index][Thread Index]

Re: [Help-glpk] want to contribute

From: Andrew Makhorin
Subject: Re: [Help-glpk] want to contribute
Date: Wed, 18 Apr 2001 01:53:31 +0400

Dear Brady,

Thank you very much for your interest in GLPK.

>     In particular, I'm interested to know:
>     - Given the presence of some other free software LP/MIP solvers, like
>lp_solve and bonsaiG, what was the motivation for starting GLPK? (I was
>attracted just because it's a GNU project, so maybe that's reason enough)

Most probably that's reason enough, because the GNU project is a great
thing :+) Besides, there is really not so much free software for LP/MIP.

>     - What is the current long-term vision for the project?  Improving
>current components and/or adding new ones? 

In the future I see the following GLPK features (ranged by importance):

- GLPK input language (currently I'm working on it);

- advanced branch-and-bound heuristics for MIP;

- LP and MIP preprocessing, decomposition, analysis, and probing;

- advanced implementation of the interior point solver;

- heuristic (inexact) search methods for MIP;

- LP parametric analysis;

- quadratic programming based on the simplex method.

>     - What would be most helpful in the short-term? 

MIP is one of the most important and interesting things. However, the
current version of GLPK MIP solver implements the only heuristic (by
Driebeck and Tomlin) due to that it is not able to solve many MIP
problems. Therefore if you wish to contribute to GLPK development,
implementation of advanced branch-and-bound heuristics would be most
helpful. What d'you think about that? Or you have other ideae? Please
inform me.

Best regards,

Andrew Makhorin

Andrew O. Makhorin, Ph.D.,
Department for Applied Informatics,
School of Aerospace Engineering,
University of Aerospace Technology
(Moscow State Aviation Institute),
Moscow, Russia

reply via email to

[Prev in Thread] Current Thread [Next in Thread]