Prior to the next stable release, MIDAS needs to have the functionality for user to be able provide initial chromosomes for the first generation of the optimizer. For population-based algorithms, there should be logic for the normal random sampling to fill in the rest of the population if the user does not provide enough initial chromosomes.