[Coin-discuss] Cbc Changeset #131 "out assert" (05/12/05)

Takuma Kawai takuma-kawai at exa-corp.co.jp
Mon Jan 28 19:13:40 EST 2008


Hi,

I've experienced the following message while solving a MIP problem with
Cbc 04122005(12 Apr 2005)

> checkSolution infeas! Retrying wihout scaling.
> Assertion failed: solver_->isProvenOptimal(), file  CbcModel.cpp, line
3874

This error first appeared when I modified the problem from pure integer
programming to mixed integer programming.
(I took away the integer constraint from a slack variable.)

It looks like 'Retrying without scaling' function has been disabled
because of the infeasible assertion.
It is obvious that this problem never becomes infeasible.

I looked up in the Changeset list and found that the assertion code
which caused the above message has been commented out by Changeset 131
(05/12/05 09:44:34).
https://projects.coin-or.org/Cbc/changeset/131

//  assert(solver_->isProvenOptimal());

I suppose this change has been made for fixing the problem.
(The error did not show up when I commented it out.)

I know it's from a long ago, but I appreciate if anyone tells me how
this change has been made.
Does anyone see any adverse effect if I just comment out the code from
Cbc 04122005?

For your reference, I've listed out the changes made to Cbc from
04/12/05 to 05/12/05.

#,Timestamp,Author,Mesasge,CbcModel?
Changeset107,04/13/05 09:48:23,forrest,OsiCbc? conditional,N
Changeset108,04/18/05 15:17:00,forrest,get around bad coding in CbcMain? ,Y
Changeset109,04/20/05 10:47:17,forrest,in case vol in clp ,N
Changeset110,04/20/05 11:00:49,forrest,changing default to cuts in search ,N
Changeset111,04/20/05 16:40:45,forrest,make rounding default ,N
Changeset112,04/21/05 15:46:12,forrest,out solver3 ,N
Changeset113,05/02/05 09:18:08,rlh,*** empty log message *** ,N
Changeset114,05/02/05 11:51:05,rlh,removed area column and updated text.,N
Changeset115,05/02/05 11:52:10 ,rlh,*** empty log message *** ,N
Changeset116,05/02/05 12:45:36,forrest,decided odd hole too slow,N
Changeset117,05/02/05 12:47:38,forrest,decided oddd hole too slow,N
Changeset118,05/02/05 16:01:57,forrest,always back to correct level ,N
Changeset119,05/02/05 16:37:10,rlh,Revisions for book chapter,N
Changeset120,05/03/05 09:03:29,rlh,Rennovated Cbc user guide for INFORMS
Tutorial book chapter,N
Changeset121,05/05/05 09:46:09,forrest,new example,N
Changeset122,05/06/05 10:23:53,forrest,major changes,Y
Changeset123,05/06/05 11:24:14,forrest,more samples ,N
Changeset124,05/06/05 11:28:07,forrest,more examples ,N
Changeset125,05/07/05 18:57:18,forrest,sample for Jon,N
Changeset126,05/08/05 09:06:12,forrest,bit faster,N
Changeset127,05/08/05 09:31:15,forrest,scaling off,N
Changeset128,05/09/05 09:49:53,forrest,add reference ,N
Changeset129,05/12/05 09:33:59,forrest,make createBranch non const,N
Changeset130,05/12/05 09:37:34,forrest,fix const object,N
Changeset131,05/12/05 09:44:34,forrest,out assert,Y
Changeset132,05/12/05 12:01:01,forrest,correct external iteration count,Y
Changeset133,05/16/05 10:04:13,forrest,out assert ,N

Takuma Kawai




More information about the Coin-discuss mailing list