K. Imamura, R. Heckendorn, T. Soule, and J. Foster. Genetic Programming, Proceedings of the 5th European
Conference, EuroGP 2002, volume 2278 of LNCS, page 172--181. Kinsale, Ireland, Springer-Verlag, (3-5 April 2002)
Abstract
We introduce a new method, N-Version Genetic
Programming (NVGP), for building fault tolerant
software by building an ensemble of automatically
generated modules in such a way as to maximize their
collective fault masking ability. The ensemble itself
is an example of n-version modular redundancy for fault
tolerance, where the output of the ensemble is the most
frequent output of n independent modules. By maximising
collective fault masking, NVGP approaches the fault
tolerance expected from n version modular redundancy
with independent faults in component modules. The
ensemble comprises individual modules from a large pool
generated with genetic programming, using operators
that increase the diversity of the population. Our
experimental test problem classified promoter regions
in Escherichia coli DNA sequences. For this problem,
NVGP reduced the number and variance of errors over
single modules produced by GP, with statistical
significance.
Genetic Programming, Proceedings of the 5th European
Conference, EuroGP 2002
year
2002
month
3-5 April
pages
172--181
publisher
Springer-Verlag
series
LNCS
volume
2278
publisher_address
Berlin
isbn
3-540-43378-3
notes
EuroGP'2002, part of lutton:2002:GP, UCI ML
e.coli benchmark (balanced training 35 positives, 35
negatives). beowulf. 2-gram (16 possible). linear gp
(MIPS like). max length 80. 4 read/write registers
(memory). 5 crossover types. Inversion (!). 2 mutation
operators, tournament fitness=correlation coefficient.
40 isolated islands (demes) each 100 individuals.
ensemble = composition from (randomly chosen)
island.
ensemble is qualified if number of its errors <= number
of errors expected if its components were _independent_
14% to 58% improvement in error rate for ensemble (of
30) compared to single GP (pop 100).
%0 Conference Paper
%1 imamura:2002:EuroGP
%A Imamura, Kosuke
%A Heckendorn, Robert B.
%A Soule, Terence
%A Foster, James A.
%B Genetic Programming, Proceedings of the 5th European
Conference, EuroGP 2002
%C Kinsale, Ireland
%D 2002
%E Foster, James A.
%E Lutton, Evelyne
%E Miller, Julian
%E Ryan, Conor
%E Tettamanzi, Andrea G. B.
%I Springer-Verlag
%K algorithms, genetic programming
%P 172--181
%T $N$-version Genetic Programming via Fault Masking
%U http://link.springer-ny.com/link/service/series/0558/papers/2278/22780172.pdf
%V 2278
%X We introduce a new method, N-Version Genetic
Programming (NVGP), for building fault tolerant
software by building an ensemble of automatically
generated modules in such a way as to maximize their
collective fault masking ability. The ensemble itself
is an example of n-version modular redundancy for fault
tolerance, where the output of the ensemble is the most
frequent output of n independent modules. By maximising
collective fault masking, NVGP approaches the fault
tolerance expected from n version modular redundancy
with independent faults in component modules. The
ensemble comprises individual modules from a large pool
generated with genetic programming, using operators
that increase the diversity of the population. Our
experimental test problem classified promoter regions
in Escherichia coli DNA sequences. For this problem,
NVGP reduced the number and variance of errors over
single modules produced by GP, with statistical
significance.
%@ 3-540-43378-3
@inproceedings{imamura:2002:EuroGP,
abstract = {We introduce a new method, N-Version Genetic
Programming (NVGP), for building fault tolerant
software by building an ensemble of automatically
generated modules in such a way as to maximize their
collective fault masking ability. The ensemble itself
is an example of n-version modular redundancy for fault
tolerance, where the output of the ensemble is the most
frequent output of n independent modules. By maximising
collective fault masking, NVGP approaches the fault
tolerance expected from n version modular redundancy
with independent faults in component modules. The
ensemble comprises individual modules from a large pool
generated with genetic programming, using operators
that increase the diversity of the population. Our
experimental test problem classified promoter regions
in Escherichia coli DNA sequences. For this problem,
NVGP reduced the number and variance of errors over
single modules produced by GP, with statistical
significance.},
added-at = {2008-06-19T17:35:00.000+0200},
address = {Kinsale, Ireland},
author = {Imamura, Kosuke and Heckendorn, Robert B. and Soule, Terence and Foster, James A.},
biburl = {https://www.bibsonomy.org/bibtex/26345346a17f7ce79b50dfe3023fd5dec/brazovayeye},
booktitle = {Genetic Programming, Proceedings of the 5th European
Conference, EuroGP 2002},
editor = {Foster, James A. and Lutton, Evelyne and Miller, Julian and Ryan, Conor and Tettamanzi, Andrea G. B.},
interhash = {0ccdb16f14d65c33cb2cd11b70d65028},
intrahash = {6345346a17f7ce79b50dfe3023fd5dec},
isbn = {3-540-43378-3},
keywords = {algorithms, genetic programming},
month = {3-5 April},
notes = {EuroGP'2002, part of \cite{lutton:2002:GP}, UCI ML
e.coli benchmark (balanced training 35 positives, 35
negatives). beowulf. 2-gram (16 possible). linear gp
(MIPS like). max length 80. 4 read/write registers
(memory). 5 crossover types. Inversion (!). 2 mutation
operators, tournament fitness=correlation coefficient.
40 isolated islands (demes) each 100 individuals.
ensemble = composition from (randomly chosen)
island.
ensemble is qualified if number of its errors <= number
of errors expected if its components were _independent_
14% to 58% improvement in error rate for ensemble (of
30) compared to single GP (pop 100).},
pages = {172--181},
publisher = {Springer-Verlag},
publisher_address = {Berlin},
series = {LNCS},
timestamp = {2008-06-19T17:42:11.000+0200},
title = {{$N$}-version Genetic Programming via Fault Masking},
url = {http://link.springer-ny.com/link/service/series/0558/papers/2278/22780172.pdf},
volume = 2278,
year = 2002
}