[Top][All Lists]
[Date Prev][Date Next][Thread Prev][Thread Next][Date Index][Thread Index]
Re: [Help-glpk] GLPK vs. SCIP for solving MIP
|
From: |
David Bremner |
|
Subject: |
Re: [Help-glpk] GLPK vs. SCIP for solving MIP |
|
Date: |
Fri, 02 Oct 2009 07:39:43 -0300 |
|
User-agent: |
EMIKO/1.14.1 (Choanoflagellata) FLIM/1.14.9 (Goj ō) APEL/10.7 EasyPG/0.0.16 Emacs/23.1 (i486-pc-linux-gnu) MULE/6.0 (HANACHIRUSATO) |
Yaron Kretchmer wrote:
>Hi All.
>I have a large (10^5 variables) MIP problem which I'm struggling to
>get GLPSOL to solve.
>I saw on http://scip.zib.de/ some benchmarks which claim that GLPK is
>16X slower in solving MIP.
If I recall correctly SCIP has some techniques from constraint
programming that are very effective for some classes of problems, but
it does depend on problem structure, and probably how the problem is
modelled. So at a guess, some modifications would be required to take
advantage of the speedups. The only reliable way to find out if
benchmarks are relevant to try (small instances of) your problems on
the solver in question.
Perhaps someone with more personal experience SCIP can tell you more.
d