HiGHS Incorrect declaration of infeasiblilty

Transferred from #1164 since incorrect infeasibility in the model that originally prompted the issue has been fixed by #1513

However, issue-1517.lp (aka model_198.lp) is still declared as being infeasible, incorrectly, when presolve is used

Nov 16 '23 21:11 jajhall

@jajhall, I am investigating this now. Do you happen to have the optimal objective value returned by GUROBI or CPLEX for this problem?

Apr 11 '25 08:04 fwesselm

It's very time-consuming to solve, and after 800s, Gurobi gives bounds of 1.2424e+07 1.2422e+07 0.01%

Apr 11 '25 09:04 jajhall

It's very time-consuming to solve, and after 800s, Gurobi gives bounds of 1.2424e+07 1.2422e+07 0.01%

Thank you, @jajhall. I have a potential fix that I need to qualify:

        Nodes      |    B&B Tree     |            Objective Bounds              |  Dynamic Constraints |       Work
Src  Proc. InQueue |  Leaves   Expl. | BestBound       BestSol              Gap |   Cuts   InLp Confl. | LpIters     Time
    .......
     14244    1332      4990  37.30%   12400496.83574  12435341.54001     0.28%     8571   1949   9989    13515k 15755.3s

Apr 11 '25 12:04 fwesselm

Issue seems to be fixed.

Nov 10 '25 15:11 fwesselm