In aerospace "skin & stringer" structures, the lightest mass structures may want to allow the skins to buckle... but not the stringers. The current optimization capabilities only recognize a single, system provided "buckling load factor" which applies to the structure as a whole.
Provide some kind of ability to allow (ignore) buckling of some items but not others. I'm not sure of the best way to implement that. Perhaps allow "BLF measures" to be assigned to specific components or simulation features in exactly the same way as one can measure/monitor stress. In effect, allow a criteria like "optimize the stringer dimensions for the lowest mass that keeps the BLF of the stringers >1.5. Such an analysis would effectively "not care" if the skins were (or were not) buckling.