Since the objective of the project is to witness the evolution in traits of a species, defining a theoretical limit to the traits not only makes the decision biased, but also irrelevant since the range of the values in traits are already known.
Better approach
Modify the existing functions to be independent of species-constant values (eg. theoretical_maximum_base_height) and instead use these existing species-constant traits as trainable parameters which can be modified gradually via the yearly genetic learning or dynamic reinforcement learning.
Since the objective of the project is to witness the evolution in traits of a species, defining a theoretical limit to the traits not only makes the decision biased, but also irrelevant since the range of the values in traits are already known.
Better approach Modify the existing functions to be independent of species-constant values (eg.
theoretical_maximum_base_height
) and instead use these existing species-constant traits as trainable parameters which can be modified gradually via the yearly genetic learning or dynamic reinforcement learning.