EvolutionStrategy ignores bounds on design variables

[FriesischScott]

Adding bounds to design variables has no effect on the result of the optimization with EvolutionStrategy.

For example: Finding the optimum of Himmelblau's function with x,y in [0, 5] returns the minimum at [3.58, -1.84].

The algorithm should either respect the bounds or give a warning that the optimization is performed unbounded.