The new version of the spsann package includes some bug fixes and a few modifications. Users now can
choose how optimCLHS computes objective function values: as in the original paper or as in the FORTRAN
implementation. Users now also must inform the weights passed to optimCLHS as to guarantee that s/he is
aware of what s/he is doing. The same apples to other functions that deal with multi-objective optimization
problems: optimACDC and optimSPAM. Another important modification in the current version of spsann is
the possibility to use a finite set of candidate locations by setting cellsize = 0. This is useful when
optimizing sample points only in the feature space and should reduce the computation time needed to find the
solution.