shrinkParSet: Shrink param set towards a point.
In berndbischl/focussearch: Global optimization using focussed random search

Shrinks a paramset towards the best found point. Numeric values are shrinked to an interval around the best obtained value of half of the previous length in each iteration, while for discrete variables, a random, not best-performing level is dropped.

1	shrinkParSet(par.set, x.df, check.feasible = FALSE)

`par.set`	[`ParamSet`] Param set.
`x.df`	[('data.frame')] 'data.frame' containing the x values to shrink around.
`check.feasible`	['logical(1)'] Should feasibility of the parameters be checked? If feasibility is not checked, and invalid params are present, no shrinking will be done.