Julian Hall
Strange: I'll look at this. If I remember correctly, `computeExactDualObjectiveValue()` is only used to determine the dual objective value with unperturbed costs to determine whether to terminate via the dual...
OK I guessed something significant had triggered your interest in this. I need to do some sanity performance checks on latest - since I don't think I've done anything to...
Has this issue been resolved? @lkapelevich @galabovaa
If there's an error in presolve - and there are bugs to fix - then run with presolve=off
We've got a few instances where one or more bugs in presolve lead to the incorrect optimal objective. Small examples like yours are helpful. Just this one from you should...
[subset_select_highs_in.mps.txt](https://github.com/ERGO-Code/HiGHS/files/11337884/subset_select_highs_in.mps.txt) isn't the right file, as I get the same optimal objective value (of 39) with/without presolve, and with Gurobi. However, the log file make the (real) model look useful...
Thanks. I've verified that all these give the errors indicated in the name
I've made some progress: if the presolved MIP is saved and then solved with `--presolve=off`, the the correct objective is found for all your instances. It's not a fix, but...
It's not a matter of tolerances. I think I found a logical bug this evening. Thanks for your efforts though. I'll reply in more detail later. In walking the dog...
I've been working a lot on this for the past few days, and have learned more about presolve, particularly ours! If the presolved problem is saved, and then solves to...