Julian Hall
This is a bug when updating the calculation of the prima-dual integral, due to a case I'd missed.
You don't give all the HiGHS logging... I can't reproduce this behaviour with `v1.7.2` using `gaswellrig.mps`, but with `gaswellrig.lp`, the MIP solver seems to have stalled after the logging line...
Sorry, I see the HiGHS logging in the original comment now - it's obscured by your logging. Very odd. Unfortunately the `gaswellrig.lp` run I left finishes with the optimal solution,...
Closer, but still not there... The start of the presolve logging is the same as you observe, but there's a diversion towards the end, so the presolved model is different,...
Bingo! I've reproduced the bug! Thanks
No problem, I understand that it's hard to chase bugs when you're busy.
Perhaps I'm stating the obvious, but HiGHS terminates with a MIP gap of 0. The bug is that the dual bound has reached 931850, coinciding with the best objective value...
Thanks. It looks as if this was closed due to a comment typo in #1942 I've run the model again - in debug - and it triggers an assert suggesting...