No idea. I would expect that to be part of the research... especially if you were trying (for instance) to set up three well running habitats initially and come back 100 years later and find seven habitats that were running with additional optimizations that the initial three lacked. You want persistence, growth, and innovation instead of just grim survival.

If you seeded them with "astronaut quality" people then you'll probably get regression to the mean in subsequent generations, and if the initial group was barely surviving then their descen... (read more)