Evolving Presentations of Genetic Information: Motivation, Methods, and Analysis

We present a hybrid genetic meta-algorithm that while applying a standard näıve genetic algorithm attempts to concurrently evolve a presentation of the genetic information that makes simple single-point crossovers increasingly effective. The idea is to attach positional information to genes and evolve them also by crossover and mutation, resulting in a presentation in which successful schema have shorter defining length and more likely to survive crossover. Initial results are promising and act as a proof-of-concept; improvements, future directions and an ideal solution are also discussed.