KirlikSayin with unbounded objectives about multiobjectivealgorithms.jl HOT 3 CLOSED

It might be confusing since the nadir point is bounded if the problem is. However in KirlikSayin it is the worst point that is computed:

y U is an over-estimator of the nadir point y N

after Equation (2).

In the current bi-objective case it seems that we should compute the nadir point as it is really simple from the lexicographic points. That makes this warning unexpected.

It might be a good idea to have dedicated functions to compute the nadir point and the worst case point.

edit: https://doi.org/10.1016/S0377-2217(02)00595-7

from multiobjectivealgorithms.jl.

Does this new warning help? https://github.com/jump-dev/MultiObjectiveAlgorithms.jl/pull/64/files

We could provide a dedicated function for the nadir point in the biobjective case, but it seems like KirlikSaying etc are most useful for N > 2.

from multiobjectivealgorithms.jl.

The new warning is more explicit and will probably avoid some confusion 👍🏼

I've just suggested changes to the warning to call it anti-ideal point instead of nadir point.

from multiobjectivealgorithms.jl.