Genetic programming convergence

We study both genotypic and phenotypic convergence in GP floating point continuous domain symbolic regression over thousands of generations. Subtree fitness variation across the population is measured shown many cases to fall. In an expanding region about root node, genetic opcodes function evaluation values are identical or nearly identical. Bottom up (leaf root) analysis shows syntactic semantic (including entropy) similarity expand from outermost node. Despite large regions zero variation, continues evolve near crossover disruption suggests improved systems within existing memory use.