<font face="arial,helvetica,sans-serif">Hi Forrest,<br><br>Tested again and the bug is gone.<br><br>Thanks,<br><br>Haroldo<br></font><br><div class="gmail_quote">On Wed, Sep 7, 2011 at 11:07 AM, John Forrest <span dir="ltr"><<a href="mailto:john.forrest@fastercoin.com">john.forrest@fastercoin.com</a>></span> wrote:<br>
<blockquote class="gmail_quote" style="margin:0 0 0 .8ex;border-left:1px #ccc solid;padding-left:1ex;">Should be fixed in 2.7 and trunk<br>
<div class="im">On Fri, 2011-09-02 at 20:11 -0300, Haroldo Santos wrote:<br>
</div><div><div></div><div class="h5">> Did more experiments to see when this bug occurs.<br>
><br>
> To turn off all heuristics and cuts it makes it happens faster.<br>
><br>
> This is an instance<br>
> <a href="http://www.decom.ufop.br/haroldo/files/sprint_hidden06.lp.gz" target="_blank">http://www.decom.ufop.br/haroldo/files/sprint_hidden06.lp.gz</a><br>
> which is quickly solved by CPLEX and has optimal objective function<br>
> value 130.<br>
><br>
> CBC crashes at this point:<br>
><br>
> Cbc0010I After 1700 nodes, 656 on tree, 249 best solution, best<br>
> possible 99.390873 (139.08 seconds)<br>
> cbc: CbcBranchDynamic.cpp:551: virtual int<br>
> CbcBranchDynamicDecision::betterBranch(CbcBranchingObject*,<br>
> CbcBranchingObject*, double, int, double, int): Assertion `distance >=<br>
> 0.0' failed<br>
><br>
><br>
><br>
><br>
><br>
><br>
><br>
><br>
> On Wed, Aug 24, 2011 at 5:43 PM, Haroldo Santos<br>
> <<a href="mailto:haroldo.santos@gmail.com">haroldo.santos@gmail.com</a>> wrote:<br>
> Dear Forrest/CBC Developers,<br>
><br>
> There is this assertion failing in :<br>
> CbcBranchDynamic.cpp:551: virtual int<br>
> CbcBranchDynamicDecision::betterBranch(CbcBranchingObject*,<br>
> CbcBranchingObject*, double, int, double, int): Assertion<br>
> `distance >= 0.0' failed.<br>
><br>
> when I'm solving this MIP:<br>
> <a href="http://www.decom.ufop.br/haroldo/files/sprint_hidden04.lp.gz" target="_blank">http://www.decom.ufop.br/haroldo/files/sprint_hidden04.lp.gz</a><br>
><br>
><br>
> I did some debug and discover that this fails because at a<br>
> given point of the search the distance is -0.415667 .<br>
><br>
> Cbc0010I After 10000 nodes, 4845 on tree, 180 best solution,<br>
> best possible 142.25 (1909.96 seconds)<br>
> Cbc0010I After 10100 nodes, 4849 on tree, 180 best solution,<br>
> best possible 142.25 (1919.14 seconds)<br>
><br>
> Tested in Linux with CBC 2.7 stable, default compilation<br>
> options and no ThirdParty software included.<br>
><br>
> The stack trace at this point is:<br>
> #0 0x00007ffff779c5cb in<br>
> CbcBranchDynamicDecision::betterBranch (this=0xb3eee10,<br>
> thisOne=<br>
> 0x12d60060, changeUp=0, numInfUp=205, changeDown=0,<br>
> numInfDown=211) at CbcBranchDynamic.cpp:555<br>
> #1 0x00007ffff784bf40 in CbcNode::chooseDynamicBranch<br>
> (this=0x12d9a290, model=0x60dcd0,<br>
> lastNode=0x12aa7510, numberPassesLeft=4) at<br>
> CbcNode.cpp:3178<br>
> #2 0x00007ffff783521c in CbcModel::chooseBranch<br>
> (this=0x60dcd0, newNode=@0x7fffffff53a0,<br>
> numberPassesLeft=5, oldNode=0x12aa7510, cuts=...,<br>
> resolved=@0x7fffffff4d18, lastws=0x12e21f40,<br>
> lowerBefore=0x12ac54b0, upperBefore=0x1396c2e0,<br>
> branches=@0x7fffffff4b08) at CbcModel.cpp:12642<br>
> #3 0x00007ffff783b052 in CbcModel::doOneNode (this=0x60dcd0,<br>
> baseModel=0x60dcd0,<br>
> node=@0x7fffffff5298, newNode=@0x7fffffff53a0) at<br>
> CbcModel.cpp:14299<br>
> #4 0x00007ffff78108f5 in CbcModel::branchAndBound<br>
> (this=0x60dcd0, doStatistics=0)<br>
> at CbcModel.cpp:3797<br>
> #5 0x00007ffff7b7be8b in CbcMain1 (argc=11,<br>
> argv=0x7fffffffdef8, model=...,<br>
> callBack=0x7ffff7b67569 <dummyCallBack(CbcModel*, int)>)<br>
> at CbcSolver.cpp:5589<br>
> #6 0x00007ffff7b67bb5 in CbcMain1 (argc=11,<br>
> argv=0x7fffffffdef8, model=...) at CbcSolver.cpp:1111<br>
> #7 0x0000000000400f47 in main (argc=11, argv=0x7fffffffdef8)<br>
> at CoinSolve.cpp:336<br>
><br>
> Cheers,<br>
><br>
> Haroldo<br>
><br>
> --<br>
> =============================================================<br>
> Haroldo Gambini Santos<br>
> Computing Department - Universidade Federal de Ouro Preto -<br>
> UFOP<br>
> email: haroldo [at ] <a href="http://iceb.ufop.br" target="_blank">iceb.ufop.br</a><br>
> home/research page: <a href="http://www.decom.ufop.br/haroldo/" target="_blank">www.decom.ufop.br/haroldo/</a><br>
><br>
> "Computer science is no more about computers than astronomy<br>
> is about telescopes." Edsger Dijkstra<br>
><br>
><br>
><br>
><br>
> --<br>
> =============================================================<br>
> Haroldo Gambini Santos<br>
> Computing Department - Universidade Federal de Ouro Preto - UFOP<br>
> email: haroldo [at ] <a href="http://iceb.ufop.br" target="_blank">iceb.ufop.br</a><br>
> home/research page: <a href="http://www.decom.ufop.br/haroldo/" target="_blank">www.decom.ufop.br/haroldo/</a><br>
><br>
> "Computer science is no more about computers than astronomy<br>
> is about telescopes." Edsger Dijkstra<br>
><br>
</div></div><div><div></div><div class="h5">> _______________________________________________<br>
> Cbc mailing list<br>
> <a href="mailto:Cbc@list.coin-or.org">Cbc@list.coin-or.org</a><br>
> <a href="http://list.coin-or.org/mailman/listinfo/cbc" target="_blank">http://list.coin-or.org/mailman/listinfo/cbc</a><br>
<br>
<br>
_______________________________________________<br>
Cbc mailing list<br>
<a href="mailto:Cbc@list.coin-or.org">Cbc@list.coin-or.org</a><br>
<a href="http://list.coin-or.org/mailman/listinfo/cbc" target="_blank">http://list.coin-or.org/mailman/listinfo/cbc</a><br>
</div></div></blockquote></div><br><br clear="all"><br>-- <br>=============================================================<br>Haroldo Gambini Santos<br>Computing Department - Universidade Federal de Ouro Preto - UFOP<br>
email: haroldo [at ] <a href="http://iceb.ufop.br" target="_blank">iceb.ufop.br</a><br>home/research page: <a href="http://www.decom.ufop.br/haroldo/" target="_blank">www.decom.ufop.br/haroldo/</a><br> <br>"Computer science is no more about computers than astronomy <br>
is about telescopes." Edsger Dijkstra<br> <br>