Feature #1988

Updated by wuttke 4 months ago

Especially when using genetic minimizer, the last results are not necessarily the optimal values.
One can see the best parameter values in the fitlog window, but this is not convenient.

And quite generally, starting values are susceptible to being spoiled by nonconverging fits, or when a fit job is interrupted.

Therefore, after each fit step we should update a register that holds the parameters and the chi2 (or whatever) of the best fit attained so far. Upon stop, parameters should be set to that value.