Hierarchical Optimization: Reordering of scaling options

[yannikschaelte]

Concerning the files logLikelihoodHierarchical etc.:

Since all observables in one scalingsoptions.obs_groups group have the same mode anyway (multiple/single/absolut), there seems to be no need for a field scalingsoptions.obj(j), but the mode can be given at the group level directly. This might be possible also for other situations.

Unless I miss some desired flexibility.

[yannikschaelte]

Update: It is currently discussed anyway whether to have this in PESTO or AMICI.