[Cbc] How to get rid of the assertion failure?

[John J Forrest]

Given that you are using a pre-compiled binary, the simplest thing to do
would be to change your model.  I can't say for sure but I would guess that
somehow 0.0 was divided by 0.0 giving NAN.

Giving all variables different objective coefficient values should avoid
all problems AND make it run faster as the dual algorithm will not have to
deal with a degenerate objective.  If you think some variables are more
important than others give them larger costs, otherwise just do some thing
like coefficient values random multiple of 0.01 and between 100.00 and
200.00.

Then tell Cbc to stop after first solution.

John Forrest


                                                                                                                                             
  From:       "Yunjing XU" <yunjingxu at gmail.com>                                                                                             
                                                                                                                                             
  To:         cbc at list.coin-or.org                                                                                                           
                                                                                                                                             
  Date:       12/14/2007 02:28 AM                                                                                                            
                                                                                                                                             
  Subject:    [Cbc] How to get rid of the assertion failure?                                                                                 
                                                                                                                                             





Dear all:

I use Cbc with Clp to get the integer solutions for a set of linear
constraints. I do not care the objective value, so I set the objective
coefficients all zero. However, I still get the following assertion failure
massage in some cases when I run the program in Windows:

Assertion Failed!
File: OsiClpSolverInterface.cpp
Line: 497
Expression: modelPtr_->objectiveValue()<1.0e100

I use the pre-compiled binary of Cbc-1.1.0. Could you tell me how can I get
rid of it while keep the result correct?

Thank you.
Sincerely,
Yunjing XU

--
Yunjing XU
Software Institute, Nanjing University
Tel:+86-25-5874-5159, +86-1580-147-0200
Email: yunjingxu at gmail.com _______________________________________________
Cbc mailing list
Cbc at list.coin-or.org
http://list.coin-or.org/mailman/listinfo/cbc