Additional (lower level) API

[SimonBlanke]

This would add another API "level", which enables more control over the optimization run and will be closer to the original mathematical idea of gradient-free-optimization. This API could improve the way how GFO can be used as a dependency in other packages, which might require more control.

It could look somewhat like this:

n_iter = 100

opt = RandomSearch()

for _ in range(n_iter):
  new_position = opt.iterate()
  new_score = objective_function(new_position)
  opt.evaluate(new_score)