<html>
<head>
<meta content="text/html; charset=windows-1252"
http-equiv="Content-Type">
</head>
<body text="#000000" bgcolor="#FFFFFF">
Dear Marko,<br>
<br>
yes you are right, the behaviour is kind of strange. In this
situation SCIP does not even know that the problem is feasible.
After analysing your instance I saw that during presolving SCIP
recognize that a linear constraint is parallel to the objective.
Thus the right hand side will be used for the objective cutoff but
unfortunately also for checking the (absolute and relative) gap
limit. I fixed this by checking the gap limit only if SCIP has found
at least one solution.<br>
<br>
Should I send you a patch for this? Or do you use the windows
binaries of our homepage? I can compile and send them to you if you
need them.<br>
<br>
<br>
Best regards,<br>
<br>
Benny<br>
<br>
<br>
<div class="moz-cite-prefix">On 09/03/2015 09:56 AM, Horváth Markó
wrote:<br>
</div>
<blockquote cite="mid:55E7FD16.9060603@gmail.com" type="cite">
<meta http-equiv="content-type" content="text/html;
charset=windows-1252">
Dear all,<br>
<br>
I'm using SCIP-3.1.2 with SoPlex 2.2.0 on windows (but this
problem also arises with SCIP-3.1.1 or SCIP-3.1.0).<br>
<br>
The problem arises when I want to solve certain MIPs with absolute
gap limit.<br>
I have attached a small example file. In this MIP the optimal
solution is equal to 0, and of course SCIP can solve easily. But
when I set absolute gap to 11 (that is, set <i>limits/absgap</i>=11)
SCIP terminates according to gap limit, but without solution:<br>
<br>
<i>SCIP Status : solving was interrupted [gap limit
reached]</i><i><br>
</i><i>Solving Time (sec) : 0.01</i><i><br>
</i><i>Solving Nodes : 0</i><i><br>
</i><i>Primal Bound : +7.12610000000000e+000 (0 solutions)</i><i><br>
</i><i>Dual Bound : +0.00000000000000e+000</i><i><br>
</i><i>Gap : infinite<br>
<br>
</i>So, what I have got is trivially not correct, I guess.<br>
How can we reach gap limit without a feasible solution? And why do
SCIP set finite primal bound without any feasible solution? (Okay,
it is indeed an upper bound for the optimum (according to bounds 0
<= x417 <= 7), and it is lower than the absolute gap, but
SCIP should not say that the gap limit reached since we have not
got any solution.)<br>
<br>
Best regards,<br>
Marko<br>
<br>
<fieldset class="mimeAttachmentHeader"></fieldset>
<br>
<pre wrap="">_______________________________________________
Scip mailing list
<a class="moz-txt-link-abbreviated" href="mailto:Scip@zib.de">Scip@zib.de</a>
<a class="moz-txt-link-freetext" href="http://listserv.zib.de/mailman/listinfo/scip">http://listserv.zib.de/mailman/listinfo/scip</a>
</pre>
</blockquote>
<br>
<pre class="moz-signature" cols="72">--
______________________________
Benjamin Müller
Zuse Institute Berlin
Takustr. 7, 14195 Berlin
<a class="moz-txt-link-abbreviated" href="mailto:benjamin.mueller@zib.de">benjamin.mueller@zib.de</a>
+49 30 841 85-195</pre>
</body>
</html>