TY - GEN
T1 - A linear estimation-of-distribution GP system
AU - Poli, Riccardo
AU - McPhee, Nicholas Freitag
PY - 2008
Y1 - 2008
N2 - We present N-gram GP, an estimation of distribution algorithm for the evolution of linear computer programs. The algorithm learns and samples a joint probability distribution of triplets of instructions (or 3-grams) at the same time as it is learning and sampling a program length distribution. We have tested N-gram GP on symbolic regressions problems where the target function is a polynomial of up to degree 12 and lawn-mower problems with lawn sizes of up to 12×12. Results show that the algorithm is effective and scales better on these problems than either linear GP or simple stochastic hill-climbing.
AB - We present N-gram GP, an estimation of distribution algorithm for the evolution of linear computer programs. The algorithm learns and samples a joint probability distribution of triplets of instructions (or 3-grams) at the same time as it is learning and sampling a program length distribution. We have tested N-gram GP on symbolic regressions problems where the target function is a polynomial of up to degree 12 and lawn-mower problems with lawn sizes of up to 12×12. Results show that the algorithm is effective and scales better on these problems than either linear GP or simple stochastic hill-climbing.
UR - http://www.scopus.com/inward/record.url?scp=47249095000&partnerID=8YFLogxK
UR - http://www.scopus.com/inward/citedby.url?scp=47249095000&partnerID=8YFLogxK
U2 - 10.1007/978-3-540-78671-9_18
DO - 10.1007/978-3-540-78671-9_18
M3 - Conference contribution
AN - SCOPUS:47249095000
SN - 3540786708
SN - 9783540786702
T3 - Lecture Notes in Computer Science (including subseries Lecture Notes in Artificial Intelligence and Lecture Notes in Bioinformatics)
SP - 206
EP - 217
BT - Genetic Programming - 11th European Conference, EuroGP 2008, Proceedings
T2 - 11th European Conference on Genetic Programming, EuroGP 2008
Y2 - 26 March 2008 through 28 March 2008
ER -