Check if the weird performance observed for the inverted PBI is a side-effect of the estimation of the nadir point. While it makes sense to estimate the ideal point from the current population, the estimation of the nadir point may require a modification, so that it contains the "historic" worst function values instead of the "current" worst values.
Check if the weird performance observed for the inverted PBI is a side-effect of the estimation of the nadir point. While it makes sense to estimate the ideal point from the current population, the estimation of the nadir point may require a modification, so that it contains the "historic" worst function values instead of the "current" worst values.