Parallel Programs are More Evolvable than Sequential
Programs
K. Leung, K. Lee, and S. Cheang. Genetic Programming, Proceedings of EuroGP'2003, volume 2610 of LNCS, page 107--118. Essex, Springer-Verlag, (14-16 April 2003)
Abstract
This paper presents a novel phenomenon of the Genetic
Parallel Programming (GPP) paradigm - the GPP
accelerating phenomenon. GPP is a novel Linear Genetic
Programming representation for evolving parallel
programs running on a Multi-ALU Processor (MAP). We
carried out a series of experiments on GPP with
different number of ALUs. We observed that parallel
programs are more evolvable than sequential programs.
For example, in the Fibonacci sequence regression
experiment, evolving a 1-ALU sequential program
requires 51 times on average of the computational
effort of an 8-ALU parallel program. This paper
presents three benchmark problems to show that the GPP
can accelerate evolution of parallel programs. Due to
the accelerating evolution phenomenon of GPP over
sequential program evolution, we could increase the
normal GP's evolution efficiency by evolving a parallel
program by GPP and if there is a need, the evolved
parallel program can be translated into a sequential
program so that it can run on conventional hardware.
%0 Conference Paper
%1 leung03
%A Leung, Kwong Sak
%A Lee, Kin Hong
%A Cheang, Sin Man
%B Genetic Programming, Proceedings of EuroGP'2003
%C Essex
%D 2003
%E Ryan, Conor
%E Soule, Terence
%E Keijzer, Maarten
%E Tsang, Edward
%E Poli, Riccardo
%E Costa, Ernesto
%I Springer-Verlag
%K algorithms, genetic programming
%P 107--118
%T Parallel Programs are More Evolvable than Sequential
Programs
%U http://www.springerlink.com/openurl.asp?genre=article&issn=0302-9743&volume=2610&spage=107
%V 2610
%X This paper presents a novel phenomenon of the Genetic
Parallel Programming (GPP) paradigm - the GPP
accelerating phenomenon. GPP is a novel Linear Genetic
Programming representation for evolving parallel
programs running on a Multi-ALU Processor (MAP). We
carried out a series of experiments on GPP with
different number of ALUs. We observed that parallel
programs are more evolvable than sequential programs.
For example, in the Fibonacci sequence regression
experiment, evolving a 1-ALU sequential program
requires 51 times on average of the computational
effort of an 8-ALU parallel program. This paper
presents three benchmark problems to show that the GPP
can accelerate evolution of parallel programs. Due to
the accelerating evolution phenomenon of GPP over
sequential program evolution, we could increase the
normal GP's evolution efficiency by evolving a parallel
program by GPP and if there is a need, the evolved
parallel program can be translated into a sequential
program so that it can run on conventional hardware.
%@ 3-540-00971-X
@inproceedings{leung03,
abstract = {This paper presents a novel phenomenon of the Genetic
Parallel Programming (GPP) paradigm - the GPP
accelerating phenomenon. GPP is a novel Linear Genetic
Programming representation for evolving parallel
programs running on a Multi-ALU Processor (MAP). We
carried out a series of experiments on GPP with
different number of ALUs. We observed that parallel
programs are more evolvable than sequential programs.
For example, in the Fibonacci sequence regression
experiment, evolving a 1-ALU sequential program
requires 51 times on average of the computational
effort of an 8-ALU parallel program. This paper
presents three benchmark problems to show that the GPP
can accelerate evolution of parallel programs. Due to
the accelerating evolution phenomenon of GPP over
sequential program evolution, we could increase the
normal GP's evolution efficiency by evolving a parallel
program by GPP and if there is a need, the evolved
parallel program can be translated into a sequential
program so that it can run on conventional hardware.},
added-at = {2008-06-19T17:35:00.000+0200},
address = {Essex},
author = {Leung, Kwong Sak and Lee, Kin Hong and Cheang, Sin Man},
biburl = {https://www.bibsonomy.org/bibtex/2121dc19eb01ac0d5f4d945212d0f0161/brazovayeye},
booktitle = {Genetic Programming, Proceedings of EuroGP'2003},
editor = {Ryan, Conor and Soule, Terence and Keijzer, Maarten and Tsang, Edward and Poli, Riccardo and Costa, Ernesto},
interhash = {bca309fa177485116aeb45b66a9297bf},
intrahash = {121dc19eb01ac0d5f4d945212d0f0161},
isbn = {3-540-00971-X},
keywords = {algorithms, genetic programming},
month = {14-16 April},
notes = {EuroGP'2003 held in conjunction with EvoWorkshops
2003},
organisation = {EvoNet},
pages = {107--118},
publisher = {Springer-Verlag},
publisher_address = {Berlin},
series = {LNCS},
timestamp = {2008-06-19T17:45:27.000+0200},
title = {Parallel Programs are More Evolvable than Sequential
Programs},
url = {http://www.springerlink.com/openurl.asp?genre=article&issn=0302-9743&volume=2610&spage=107},
volume = 2610,
year = 2003
}