The effects of crossover and mutation operators on variable length linear structures
Jonathan E. Rowe, Nicholas Freitag McPhee · OpenGrey (Institut de l'Information Scientifique et Technique) · 2001
In the search space of variable length strings, it is possible to define crossover and mutation operators that are equivalent to those used in genetic programming on tree structures. We study the effects of these operators on the lengths of strings within a population. It is shown that the distributions by which different string lengths are sampled are strongly biased. To investigate these biases, the effects of repeated application of the operators (without regard for fitness) is considered, and in some cases the fixed-point distributions are found.