<html>
<head>
<meta content="text/html; charset=ISO-8859-1"
http-equiv="Content-Type">
</head>
<body bgcolor="#FFFFFF" text="#000000">
Hi John, Matteo, list,<br>
Thanks for your replies.<br>
I'd be very interested to see if they give some more mileage, if
you're kind enough to put them in.<br>
I had, indeed, found "fragments" of evidence in the code that
someone had started at least.<br>
From the direct formulation it's not at all obvious how these
zero-half cuts emerge, and cplex is not very<br>
forthcoming, naturally: I just get<i> (in its log file)<br>
</i><i><br>
</i><i>Clique cuts applied: 17</i><i><br>
</i><i>Zero-half cuts applied: 61</i><i><br>
</i><i>Gomory fractional cuts applied: 6</i><br>
<br>
I'd be delighted if they turned out to work well.<br>
<br>
Best wishes all,<br>
James.<br>
<br>
<br>
<div class="moz-cite-prefix">John Forrest wrote:<br>
</div>
<blockquote cite="mid:50ED8491.8040504@fastercoin.com" type="cite">
<pre wrap="">James,
Implementation of zero-half cuts?
Some years ago, while I was still at IBM, there was a submission of code
- but I have no record of who submitted them! I have just found the
code on a disk and would be happy to test and put them in with due
appreciation if the author would just e-mail me saying they were her/his
work. IBM had very strict rules on accepting code, which might have
caused problems. Do names such as Cgl012cut or Cgldikbd or
CglZeroHalf mean anything to anyone?
John Forrest
On 09/01/13 10:43, <a class="moz-txt-link-abbreviated" href="mailto:James.Gunning@csiro.au">James.Gunning@csiro.au</a> wrote:
</pre>
<blockquote type="cite">
<pre wrap="">Hello All,
I'm new to the list, and have been learning a bit about cbc. First, a big thanks to the
developers for making such a big tool available to all, & for all the effort that's gone in etc.
I'm trying to use it to solve some binary QP problems for a binary vector x, of form
min x^T.Q.x. I've done the standard tricks to unroll this out to a binary linear IP problem
(new variables p_ij for products, inequalities to force these to match x_i, x_j)
Test runs on cplex show that it uses zero-half cuts a good deal on the problem.
Has anyone tried implementing zero-half cuts in cbc? Any success? Anything
in the pipeline planned by some clever contributors?
Best wishes all,
James.
_______________________________________________
Cbc mailing list
<a class="moz-txt-link-abbreviated" href="mailto:Cbc@list.coin-or.org">Cbc@list.coin-or.org</a>
<a class="moz-txt-link-freetext" href="http://list.coin-or.org/mailman/listinfo/cbc">http://list.coin-or.org/mailman/listinfo/cbc</a>
</pre>
</blockquote>
<pre wrap="">
_______________________________________________
Cbc mailing list
<a class="moz-txt-link-abbreviated" href="mailto:Cbc@list.coin-or.org">Cbc@list.coin-or.org</a>
<a class="moz-txt-link-freetext" href="http://list.coin-or.org/mailman/listinfo/cbc">http://list.coin-or.org/mailman/listinfo/cbc</a>
</pre>
</blockquote>
<br>
</body>
</html>