Created by W.Langdon from gp-bibliography.bib Revision:1.8098
GEMS genetic evolution of machine language software Breeding system similar to crowding and Tackett's Softbrood selection (max litter size of 12). GA like crossover acts on code and contents of memory. Pool of 1500 member 0.20 mutation rate.
{"}indicates that the problem difficulty, over the range of the test and in terms of required spawns, while increasing rapidly, does not appear to be combinatorial or exponential{"} (suggests O(n**3) ).
Discussion of statistics of number of useful terminals in random and later populations.
Memory initialised to random values. {"}Cultural memory{"} cf \cite{spector:1996:ctiGP}.
Steady state GA. 2 types of Mutation (20 percent). While JP jump and subroutines are discussed the problem does not need iteration to solve it.
part of \cite{rosca:1995:ml}",
Genetic Programming entries for Ronald L Crepeau