Evolution and Artificial Life

Genetic algorithms and Q learning

How is evolution a form of reinforcement learning?
How is evolutionary "learning" different from Q learning?
How could evolutionary "learning" by combined with Q learning?
How could evolution "look ahead"?

Genetic algorithms with "natural" selection

Who should get to mate?
Should mating be explicitly rewarded or punished or neither?
How can the population size be controlled?

Other issues

Representation: how are traits coded?

Exploration and exploitation, random vs. fixed decisions

Controlling initial decline

The effect of easy and tough environments

The effect of increasing toughness

Implementation: new classes

Coursework & grading

Other resources

Contact instructor