[Numpy-discussion] Linear assignment problem: Jonker-Volgenant algorithm

Gael Varoquaux gael.varoquaux at normalesup.org
Mon May 16 11:55:17 EDT 2011


On Mon, May 16, 2011 at 08:49:58AM -0700, Hoyt Koepke wrote:
> On Mon, May 16, 2011 at 12:18 AM, Gael Varoquaux
> <gael.varoquaux at normalesup.org> wrote:
> > Jonker-Volgenant algorithm for best linear assignment in Python, using
> > numpy. Unsuprisingly, it is proving time-costly. I cannot afford to spend
> > too much time on this, as it not to solve a problem of mine, but for the
> > scikits.learn. Thus I was wondering if someone had a BSD-licensed Python
> > version of the algorithm that he would be willing to share.

> There are a number available.  The lemon graph library
> (http://lemon.cs.elte.hu/trac/lemon) has solvers for this problem and
> has python bindings.  It's under the boost license (is that ok?).   It
> might be a bit heavyweight for this, though, but it's great software.

Thanks for the suggestion. Unfortunately, I do not want to add compiled
code (or an external dependency) to the scikit for this solver, as it is
a fairly minor step for us. I particular chances are that it will never
be a computational bottleneck on our poblems.



More information about the NumPy-Discussion mailing list