Dynamics And Performance Of A Linear Genetic Programming System

Genetic Programming (“GP”) is a machine learning algorithm that uses a series of selection and transformation steps, modeled after biological evolution, to evolve computer programs. This thesis explores key research elements in the design of a widely-used GP system and compares its performance to other machine learning algorithms. The issues addressed include the emergence of “introns” or “code bloat” in GP runs and the role of mutation in biologic.