A variant on randomly sampling distribution parameters at the point of creating the initial population. With Latin hypercube sampling (LHS), the parameter space is utilised fully, giving a wider initial range of individuals.
This initialisation method is default in SciPy's differential evolution implementation, as well as in other optimisation methods.
A variant on randomly sampling distribution parameters at the point of creating the initial population. With Latin hypercube sampling (LHS), the parameter space is utilised fully, giving a wider initial range of individuals.
This initialisation method is default in SciPy's differential evolution implementation, as well as in other optimisation methods.