Quantifying the Severity of the Permutation Problem in Neuroevolution

S. Haflidason; Richard Neville

Quantifying the Severity of the Permutation Problem in Neuroevolution

S. Haflidason, Richard Neville

Research output: Chapter in Book/Conference proceeding › Conference contribution › peer-review

Abstract

Abstract. In this paper we investigate the likely severity of the PermutationProblem on a standard Genetic Algorithm used for the evolutionaryoptimisation of Neural Networks. We present a method forcalculating the expected number of permutations in an initial populationgiven a particular representation and show that typically this number isvery low. This low expectation coupled with the empirical evidence suggeststhat the severity of the Permutation Problem is low in general,and so not a common cause of poor performance in Neuroevolutionaryalgorithms.

Original language	English
Title of host publication	host publication
Publication status	Published - Sept 2009
Event	The Fourth International Workshop on Natural Computing (IWNC 2009), Sept. 23-25, 2009, Himeji International Exchange Center, Egret, 3rd floor 68-290 Honmachi, Himeji, URL HTTP://www.iwnc2009.org/ & HTTP://www.himeji-iec.or.jp/ , http://www.springerlink.com/content/m283x169461r6455/ - Hosted by: National Institute of Information and Communications Technology; Supported by: NICT, Osaka Electro-Communication University, University of Hyogo, Himeji-city Duration: 23 Sept 2009 → 25 Sept 2009 http://books.google.co.uk/books?id=P3POhHgIek0C&pg=PA149&lpg=PA149&dq=Quantifying+the+Severity+of+the+Permutation+Problem+in+Neuroevolution&source=bl&ots=Hq7nhTXRYi&sig=L-Jb26iExK1zL19yDacK6PPa7GQ&hl=en&sa=X&ei=l1WeUdiQKeqv4QTY64HgCQ&ved=0CC4Q6AEwAA#v=onepage&q=Quantifying%20the%20Severity%20of%20the%20Permutation%20Problem%20in%20Neuroevolution&f=falsehttp://link.springer.com/content/pdf/10.1007%2F978-4-431-53868-4_17.pdf

Conference

Conference	The Fourth International Workshop on Natural Computing (IWNC 2009), Sept. 23-25, 2009, Himeji International Exchange Center, Egret, 3rd floor 68-290 Honmachi, Himeji, URL HTTP://www.iwnc2009.org/ & HTTP://www.himeji-iec.or.jp/ , http://www.springerlink.com/content/m283x169461r6455/
City	Hosted by: National Institute of Information and Communications Technology; Supported by: NICT, Osaka Electro-Communication University, University of Hyogo, Himeji-city
Period	23/09/09 → 25/09/09
Internet address	http://books.google.co.uk/books?id=P3POhHgIek0C&pg=PA149&lpg=PA149&dq=Quantifying+the+Severity+of+the+Permutation+Problem+in+Neuroevolution&source=bl&ots=Hq7nhTXRYi&sig=L-Jb26iExK1zL19yDacK6PPa7GQ&hl=en&sa=X&ei=l1WeUdiQKeqv4QTY64HgCQ&ved=0CC4Q6AEwAA#v=onepage&q=Quantifying%20the%20Severity%20of%20the%20Permutation%20Problem%20in%20Neuroevolution&f=falsehttp://link.springer.com/content/pdf/10.1007%2F978-4-431-53868-4_17.pdf

Access to Document

Cite this

Haflidason, S., & Neville, R. (2009). Quantifying the Severity of the Permutation Problem in Neuroevolution. In host publication http://books.google.co.uk/books?id=P3POhHgIek0C&pg=PA149&lpg=PA149&dq=Quantifying+the+Severity+of+the+Permutation+Problem+in+Neuroevolution&source=bl&ots=Hq7nhTXRYi&sig=L-Jb26iExK1zL19yDacK6PPa7GQ&hl=en&sa=X&ei=l1WeUdiQKeqv4QTY64HgCQ&ved=0CC4Q6AEwAA#v=onepage&q=Quantifying%20the%20Severity%20of%20the%20Permutation%20Problem%20in%20Neuroevolution&f=false

@inproceedings{4b8816b5c0464839bfa331df08add9dd,

title = "Quantifying the Severity of the Permutation Problem in Neuroevolution",

abstract = "Abstract. In this paper we investigate the likely severity of the PermutationProblem on a standard Genetic Algorithm used for the evolutionaryoptimisation of Neural Networks. We present a method forcalculating the expected number of permutations in an initial populationgiven a particular representation and show that typically this number isvery low. This low expectation coupled with the empirical evidence suggeststhat the severity of the Permutation Problem is low in general,and so not a common cause of poor performance in Neuroevolutionaryalgorithms.",

author = "S. Haflidason and Richard Neville",

year = "2009",

month = sep,

language = "English",

booktitle = "host publication",

note = "The Fourth International Workshop on Natural Computing (IWNC 2009), Sept. 23-25, 2009, Himeji International Exchange Center, Egret, 3rd floor 68-290 Honmachi, Himeji, URL HTTP://www.iwnc2009.org/ & HTTP://www.himeji-iec.or.jp/ , http://www.springerlink.com/content/m283x169461r6455/ ; Conference date: 23-09-2009 Through 25-09-2009",

url = "http://books.google.co.uk/books?id=P3POhHgIek0C&pg=PA149&lpg=PA149&dq=Quantifying+the+Severity+of+the+Permutation+Problem+in+Neuroevolution&source=bl&ots=Hq7nhTXRYi&sig=L-Jb26iExK1zL19yDacK6PPa7GQ&hl=en&sa=X&ei=l1WeUdiQKeqv4QTY64HgCQ&ved=0CC4Q6AEwAA#v=onepage&q=Quantifying%20the%20Severity%20of%20the%20Permutation%20Problem%20in%20Neuroevolution&f=falsehttp://link.springer.com/content/pdf/10.1007%2F978-4-431-53868-4_17.pdf",

}

Haflidason, S & Neville, R 2009, Quantifying the Severity of the Permutation Problem in Neuroevolution. in host publication. The Fourth International Workshop on Natural Computing (IWNC 2009), Sept. 23-25, 2009, Himeji International Exchange Center, Egret, 3rd floor 68-290 Honmachi, Himeji, URL HTTP://www.iwnc2009.org/ & HTTP://www.himeji-iec.or.jp/ , http://www.springerlink.com/content/m283x169461r6455/, Hosted by: National Institute of Information and Communications Technology; Supported by: NICT, Osaka Electro-Communication University, University of Hyogo, Himeji-city, 23/09/09. <http://books.google.co.uk/books?id=P3POhHgIek0C&pg=PA149&lpg=PA149&dq=Quantifying+the+Severity+of+the+Permutation+Problem+in+Neuroevolution&source=bl&ots=Hq7nhTXRYi&sig=L-Jb26iExK1zL19yDacK6PPa7GQ&hl=en&sa=X&ei=l1WeUdiQKeqv4QTY64HgCQ&ved=0CC4Q6AEwAA#v=onepage&q=Quantifying%20the%20Severity%20of%20the%20Permutation%20Problem%20in%20Neuroevolution&f=false>

TY - GEN

T1 - Quantifying the Severity of the Permutation Problem in Neuroevolution

AU - Haflidason, S.

AU - Neville, Richard

PY - 2009/9

Y1 - 2009/9

N2 - Abstract. In this paper we investigate the likely severity of the PermutationProblem on a standard Genetic Algorithm used for the evolutionaryoptimisation of Neural Networks. We present a method forcalculating the expected number of permutations in an initial populationgiven a particular representation and show that typically this number isvery low. This low expectation coupled with the empirical evidence suggeststhat the severity of the Permutation Problem is low in general,and so not a common cause of poor performance in Neuroevolutionaryalgorithms.

AB - Abstract. In this paper we investigate the likely severity of the PermutationProblem on a standard Genetic Algorithm used for the evolutionaryoptimisation of Neural Networks. We present a method forcalculating the expected number of permutations in an initial populationgiven a particular representation and show that typically this number isvery low. This low expectation coupled with the empirical evidence suggeststhat the severity of the Permutation Problem is low in general,and so not a common cause of poor performance in Neuroevolutionaryalgorithms.

M3 - Conference contribution

BT - host publication

T2 - The Fourth International Workshop on Natural Computing (IWNC 2009), Sept. 23-25, 2009, Himeji International Exchange Center, Egret, 3rd floor 68-290 Honmachi, Himeji, URL HTTP://www.iwnc2009.org/ & HTTP://www.himeji-iec.or.jp/ , http://www.springerlink.com/content/m283x169461r6455/

Y2 - 23 September 2009 through 25 September 2009

ER -

Quantifying the Severity of the Permutation Problem in Neuroevolution

Abstract

Conference

Access to Document

Fingerprint

Cite this