Couenne - Another bug. For an infeasible problem, status shows locally solved.

[guanghuidatafuelx]

using AmplNLWriter, Couenne_jll model = Model(() -> AmplNLWriter.Optimizer(Couenne_jll.amplexe)) ... optimize!(model) println("primal_status=", primal_status(model)) println("termination_status=", termination_status(model)) print("raw_status(model)=", raw_status(model))

The couenne.opt file is couenne.opt1.txt

The .lp file is Couenne_model1.lp.txt

The .nl file is Couenne_problem1.nl.txt

But this problem is actually infeasible because Spots_x[i]<=5, i=1,..,5. 4315 Spots_x[1] + 4315 Spots_x[2] + 4315 Spots_x[3] + 4315 Spots_x[4] + 4315 Spots_x[5] <=(4315)(5)(5)=107865.

So the constraint:

4315 Spots_x[1] + 4315 Spots_x[2] + 4315 Spots_x[3] + 4315 Spots_x[4] + 4315 Spots_x[5] ≥ 107876.0

will never be satisfied.

However, the status coming from the solver are:

primal_status(model)=FEASIBLE_POINT. termination_status(model)=LOCALLY_SOLVED

[guanghuidatafuelx]

A note: both SCIP and Juniper identified this problem as infeasible.