[Cbc] Possible bug

[Gleb Belov]

Hi,

here is an instance erroneously reported feasible with the current
Cbc-stable:
https://www.dropbox.com/s/4s7c5dszer647ee/budget-bug.osicbc.mps?dl=0
Output of cbc standalone (Kubuntu 16.04.4):
Coin0505I Presolved problem not optimal, resolve after postsolve
Cgl0013I Postprocessed model is infeasible - possible tolerance issue - try
without preprocessing
2 relaxed row infeasibilities - summing to 185.283
Accuracy problem on post-processing - maybe try without pre-processing
2 relaxed row infeasibilities - summing to 185.283
2 relaxed row infeasibilities - summing to 185.283

Result - Optimal solution found

Could someone have a look please?

The original MiniZinc model, describing some simple constraints which can
be switched off leading to correct result, and discussion:
http://www.minizinc.org/forum.html?place=msg%2Fminizinc%2F1adai3q65ts%2Fiw3Q4JykAwAJ

Thank you,
Gleb
-- 
Dr Gleb Belov       Monash University      +61 3 9903 1622


-- 
Dr Gleb Belov       Monash University      +61 3 9903 1622
-------------- next part --------------
An HTML attachment was scrubbed...
URL: <http://list.coin-or.org/pipermail/cbc/attachments/20180615/8287834e/attachment.html>