Genetic Programming with One-Point Crossover

Abstract

In recent theoretical and experimental work onschemata in genetic programming we have proposed a newsimpler form of crossover in which the same crossoverpoint is selected in both parent programs. We call thisoperator one-point crossover because of its similaritywith the corresponding operator in genetic algorithms.One-point crossover presents very interestingproperties from the theory point of view. In this paperwe describe this form of crossover as well as a newvariant called strict one-point crossover highlightingtheir useful theoretical and practical features. Wealso present experimental evidence which shows thatone-point crossover compares favourably with standardcrossover.