Open s-c-e opened 5 years ago
I am facing a similar problem with an LP file I have. It gets solved on NEOS based CBC (version 2.9.9 Revision # 2366), but on my installation (CBC 2.9.0, 2.9.8 and 2.9.9 (Revision # 2403)), unless I turn preprocess
to off
, it does not solve.
I think also in your problem, if you only turn preprocess
to off
(and not presolve
), it should be enough to make it work, since it is the cut generators that proclaim the problem infeasible!
What I understand is that this happens in problems where the coefficient matrix has huge range of values.
Attachments: https://github.com/s-c-e/cbc-trac-migration-attachments/blob/master/trac-ticket-156.zip
Hello all,
I’m having issues solving my model with Cbc. Attached you can find the mps. GLPK and Cplex find feasible solutions. Cplex finds the optimal solution within a minute, while GLPK needs several hours to find a decent solution.
Cbc results:
Continuous objective value is 122787 - 0.61 seconds Cgl0003I 7 fixed, 0 tightened bounds, 321 strengthened rows, 0 substitutions Cgl0000I Cut generators found to be infeasible! (or unbounded) Pre-processing says infeasible or unbounded
When setting the parameters presolve and preprocess to off, Cbc (sometimes) finds an optimal solution with objective value 480069. However, the solution provided by Cplex is 300274.
Do you know why cbc fails to solve this problem to optimality?