Convergence Rates for the Distribution of Program
Outputs
W. Langdon. GECCO 2002: Proceedings of the Genetic and
Evolutionary Computation Conference, page 812--819. New York, Morgan Kaufmann Publishers, (9-13 July 2002)
Abstract
Fitness distributions (landscapes) of programs tend to
a limit as they get bigger. Markov chain convergence
theorems give general upper bounds on the linear
program sizes needed for convergence. Tight bounds
(exponential in N, N log N, and smaller) are given for
five computer models (any, average, cyclic, bit flip
and Boolean). Mutation randomizes a genetic algorithm
population in 0.25 (l+1)(log(l)+4) generations. Results
for a genetic programming (GP) like model are confirmed
by experiment.
GECCO 2002: Proceedings of the Genetic and
Evolutionary Computation Conference
year
2002
month
9-13 July
pages
812--819
publisher
Morgan Kaufmann Publishers
publisher_address
San Francisco, CA 94104, USA
size
8 pages
isbn
1-55860-878-8
notes
GECCO-2002 A joint meeting of the eleventh
International Conference on Genetic Algorithms
(ICGA-2002) and the seventh Annual Genetic Programming
Conference (GP-2002) Part of
langdon:2002:GECCO
%0 Conference Paper
%1 langdon:2002:crlp
%A Langdon, W. B.
%B GECCO 2002: Proceedings of the Genetic and
Evolutionary Computation Conference
%C New York
%D 2002
%E Langdon, W. B.
%E Cantú-Paz, E.
%E Mathias, K.
%E Roy, R.
%E Davis, D.
%E Poli, R.
%E Balakrishnan, K.
%E Honavar, V.
%E Rudolph, G.
%E Wegener, J.
%E Bull, L.
%E Potter, M. A.
%E Schultz, A. C.
%E Miller, J. F.
%E Burke, E.
%E Jonoska, N.
%I Morgan Kaufmann Publishers
%K Average Distance, Fitness Landscapes, Markov Minorization, Mutation Random Total Variation algorithms, analysis, computer convergence eigenvalues, genetic programming, time, walk
%P 812--819
%T Convergence Rates for the Distribution of Program
Outputs
%U http://www.cs.ucl.ac.uk/staff/W.Langdon/ftp/papers/gecco2002/gecco-2002-14.pdf
%X Fitness distributions (landscapes) of programs tend to
a limit as they get bigger. Markov chain convergence
theorems give general upper bounds on the linear
program sizes needed for convergence. Tight bounds
(exponential in N, N log N, and smaller) are given for
five computer models (any, average, cyclic, bit flip
and Boolean). Mutation randomizes a genetic algorithm
population in 0.25 (l+1)(log(l)+4) generations. Results
for a genetic programming (GP) like model are confirmed
by experiment.
%@ 1-55860-878-8
@inproceedings{langdon:2002:crlp,
abstract = {Fitness distributions (landscapes) of programs tend to
a limit as they get bigger. Markov chain convergence
theorems give general upper bounds on the linear
program sizes needed for convergence. Tight bounds
(exponential in N, N log N, and smaller) are given for
five computer models (any, average, cyclic, bit flip
and Boolean). Mutation randomizes a genetic algorithm
population in 0.25 (l+1)(log(l)+4) generations. Results
for a genetic programming (GP) like model are confirmed
by experiment.},
added-at = {2008-06-19T17:35:00.000+0200},
address = {New York},
author = {Langdon, W. B.},
biburl = {https://www.bibsonomy.org/bibtex/29b6a2c94c309d859d596b90d8822a02c/brazovayeye},
booktitle = {GECCO 2002: Proceedings of the Genetic and
Evolutionary Computation Conference},
editor = {Langdon, W. B. and Cant{\'u}-Paz, E. and Mathias, K. and Roy, R. and Davis, D. and Poli, R. and Balakrishnan, K. and Honavar, V. and Rudolph, G. and Wegener, J. and Bull, L. and Potter, M. A. and Schultz, A. C. and Miller, J. F. and Burke, E. and Jonoska, N.},
interhash = {f87ef5c7a003f03281df7e6d9c5fdca3},
intrahash = {9b6a2c94c309d859d596b90d8822a02c},
isbn = {1-55860-878-8},
keywords = {Average Distance, Fitness Landscapes, Markov Minorization, Mutation Random Total Variation algorithms, analysis, computer convergence eigenvalues, genetic programming, time, walk},
month = {9-13 July},
notes = {GECCO-2002 A joint meeting of the eleventh
International Conference on Genetic Algorithms
(ICGA-2002) and the seventh Annual Genetic Programming
Conference (GP-2002) Part of
\cite{langdon:2002:GECCO}},
pages = {812--819},
publisher = {Morgan Kaufmann Publishers},
publisher_address = {San Francisco, CA 94104, USA},
size = {8 pages},
timestamp = {2008-06-19T17:45:00.000+0200},
title = {Convergence Rates for the Distribution of Program
Outputs},
url = {http://www.cs.ucl.ac.uk/staff/W.Langdon/ftp/papers/gecco2002/gecco-2002-14.pdf},
year = 2002
}