Create evolutionary algorithms to search for useful parameterizations.

[SurrealVectors]

Create evolutionary algorithms to search for useful parameterization of neural types, then structures, then nervous systems.

There should be four measures used:

1. Effectiveness: How successful the individual is at the presented task(s).
2. Stability: The effectiveness of similar/related individuals in the population. This measure implies an improved adaptability and modularity. It can also improve the rate of finding useful parameterizations as it focuses on regions of effectiveness within the parameter space.
3. Simplicity: The size of the neural net, number of connections, etc. The individual should be no more complicated than necessary for the given task(s).
4. Variety: How dissimilar is the individual to others in the generation. There should be a variety of useful parameterizations searched for.