[Coin-discuss] Row bounds in OsiGLPK and MIP hot starts

[Ted Ralphs]

Eduardo,

As it turns out, I've recently been investigating some of the issues you 
are referring to with one of my Ph.D students. It probably won't be 
officially released for another month or so, but we have completed an 
alpha version of a new SYMPHONY callable library for MIP that has an OSI 
interface on top of it, and uses an OSI interface underneath to link to 
the user's choice of LP solver. We have tried to make this new version 
of SYMPHONY look and feel like an LP solver, in the sense that there is 
an associated warm start class and the MIP solver can be warm started, 
even after problem modifications. One of the issues we will be 
considering in future work is how to hot start the solution process from 
a given heuristic solution. This effort is in the early stages, but it 
seems promising. A beta version should be released in a month or so. If 
you are interested in trying it out before then, I would be happy to 
send it to you, but it is still somewhat buggy.

Cheers,

Ted
-- 
Dr. Ted Ralphs
Assistant Professor
Industrial and Systems Engineering
Lehigh University
(610)758-4784
tkralphs at lehigh.edu
www.lehigh.edu/~tkr2