Latin Hypercube Sampling (LHS) is a method for generating a set of points in a multi-dimensional space such that each point is equally spaced in each dimension. This is useful for sampling a function over a multi-dimensional space in a way that is more efficient than a grid-based approach.
The standard LHS algorithm does not take into account sampled points prior to the current iteration and also does not sample from a pool but rather generates a new set of points based on the maximum and minimum values of each dimension.
The sample method in this package, however is an adjusted method that takes into consideration both reference_condition that are already known samples, and conditions which is the condition pool to sample from.
The pool method works like a standard LHS algorithm.
Attention: the pool method ignores the allowed_values parameter and instead uses the value_range parameter to generate the pool.