Direct and Indirect Evolutionary Designs of Artificial Neural Networks

O. Alba-Cisneros; A. Espinal; G. López-Vázquez; M. A. Sotelo-Figueroa; O. J. Purata-Sifuentes; V. Calzada-Ledesma; R. A. Vázquez; H. Rostro-González

doi:10.1007/978-3-030-35445-9_31

Direct and Indirect Evolutionary Designs of Artificial Neural Networks

O. Alba-Cisneros, A. Espinal^*, G. López-Vázquez, M. A. Sotelo-Figueroa, O. J. Purata-Sifuentes, V. Calzada-Ledesma, R. A. Vázquez, H. Rostro-González

^*Corresponding author for this work

Research output: Book chapter › Chapter › peer-review

Abstract

Evolutionary Algorithms (EAs) and other kind of metaheuristics are utilized to either design or optimize the architecture of Artificial Neural Networks (ANNs) in order to adapt them for solving a specific problem; these generated ANNs are known as Evolutionary Artificial Neural Networks (EANNs). Their architecture components, including number of neurons or their kind of transfer functions, connectivity pattern, etc., can be defined through direct or indirect encoding schemes; the former, directly codifies ANN architecture components into the genotype of solutions, meanwhile the last one, requires to transform the solution’s genotype through mapping processes to generate an ANN architecture. In this work, architectures of three-layered feed-forward ANNs are designed by using both encoding schemes to solve several well-known benchmark datasets of supervised classification problems. The results of designed ANNs by using direct and indirect encoding schemes are compared.

Original language	English
Title of host publication	Studies in Computational Intelligence
Publisher	Springer
Pages	431-443
Number of pages	13
DOIs	https://doi.org/10.1007/978-3-030-35445-9_31
Publication status	Published - 2020
Externally published	Yes

Publication series

Name	Studies in Computational Intelligence
Volume	862
ISSN (Print)	1860-949X
ISSN (Electronic)	1860-9503

Keywords

Artificial neural networks
Differential evolution
Direct encoding scheme
Indirect encoding scheme
Pattern classification

Access to Document

10.1007/978-3-030-35445-9_31

Cite this

Alba-Cisneros, O., Espinal, A., López-Vázquez, G., Sotelo-Figueroa, M. A., Purata-Sifuentes, O. J., Calzada-Ledesma, V., Vázquez, R. A., & Rostro-González, H. (2020). Direct and Indirect Evolutionary Designs of Artificial Neural Networks. In Studies in Computational Intelligence (pp. 431-443). (Studies in Computational Intelligence; Vol. 862). Springer. https://doi.org/10.1007/978-3-030-35445-9_31

@inbook{caf30a64de3c464fa24688f0bd85fc44,

title = "Direct and Indirect Evolutionary Designs of Artificial Neural Networks",

abstract = "Evolutionary Algorithms (EAs) and other kind of metaheuristics are utilized to either design or optimize the architecture of Artificial Neural Networks (ANNs) in order to adapt them for solving a specific problem; these generated ANNs are known as Evolutionary Artificial Neural Networks (EANNs). Their architecture components, including number of neurons or their kind of transfer functions, connectivity pattern, etc., can be defined through direct or indirect encoding schemes; the former, directly codifies ANN architecture components into the genotype of solutions, meanwhile the last one, requires to transform the solution{\textquoteright}s genotype through mapping processes to generate an ANN architecture. In this work, architectures of three-layered feed-forward ANNs are designed by using both encoding schemes to solve several well-known benchmark datasets of supervised classification problems. The results of designed ANNs by using direct and indirect encoding schemes are compared.",

keywords = "Artificial neural networks, Differential evolution, Direct encoding scheme, Indirect encoding scheme, Pattern classification",

author = "O. Alba-Cisneros and A. Espinal and G. L{\'o}pez-V{\'a}zquez and Sotelo-Figueroa, {M. A.} and Purata-Sifuentes, {O. J.} and V. Calzada-Ledesma and V{\'a}zquez, {R. A.} and H. Rostro-Gonz{\'a}lez",

note = "Publisher Copyright: {\textcopyright} Springer Nature Switzerland AG 2020.",

year = "2020",

doi = "10.1007/978-3-030-35445-9_31",

language = "English",

series = "Studies in Computational Intelligence",

publisher = "Springer",

pages = "431--443",

booktitle = "Studies in Computational Intelligence",

address = "United States",

}

Alba-Cisneros, O, Espinal, A, López-Vázquez, G, Sotelo-Figueroa, MA, Purata-Sifuentes, OJ, Calzada-Ledesma, V, Vázquez, RA & Rostro-González, H 2020, Direct and Indirect Evolutionary Designs of Artificial Neural Networks. in Studies in Computational Intelligence. Studies in Computational Intelligence, vol. 862, Springer, pp. 431-443. https://doi.org/10.1007/978-3-030-35445-9_31

TY - CHAP

T1 - Direct and Indirect Evolutionary Designs of Artificial Neural Networks

AU - Alba-Cisneros, O.

AU - Espinal, A.

AU - López-Vázquez, G.

AU - Sotelo-Figueroa, M. A.

AU - Purata-Sifuentes, O. J.

AU - Calzada-Ledesma, V.

AU - Vázquez, R. A.

AU - Rostro-González, H.

PY - 2020

Y1 - 2020

N2 - Evolutionary Algorithms (EAs) and other kind of metaheuristics are utilized to either design or optimize the architecture of Artificial Neural Networks (ANNs) in order to adapt them for solving a specific problem; these generated ANNs are known as Evolutionary Artificial Neural Networks (EANNs). Their architecture components, including number of neurons or their kind of transfer functions, connectivity pattern, etc., can be defined through direct or indirect encoding schemes; the former, directly codifies ANN architecture components into the genotype of solutions, meanwhile the last one, requires to transform the solution’s genotype through mapping processes to generate an ANN architecture. In this work, architectures of three-layered feed-forward ANNs are designed by using both encoding schemes to solve several well-known benchmark datasets of supervised classification problems. The results of designed ANNs by using direct and indirect encoding schemes are compared.

AB - Evolutionary Algorithms (EAs) and other kind of metaheuristics are utilized to either design or optimize the architecture of Artificial Neural Networks (ANNs) in order to adapt them for solving a specific problem; these generated ANNs are known as Evolutionary Artificial Neural Networks (EANNs). Their architecture components, including number of neurons or their kind of transfer functions, connectivity pattern, etc., can be defined through direct or indirect encoding schemes; the former, directly codifies ANN architecture components into the genotype of solutions, meanwhile the last one, requires to transform the solution’s genotype through mapping processes to generate an ANN architecture. In this work, architectures of three-layered feed-forward ANNs are designed by using both encoding schemes to solve several well-known benchmark datasets of supervised classification problems. The results of designed ANNs by using direct and indirect encoding schemes are compared.

KW - Artificial neural networks

KW - Differential evolution

KW - Direct encoding scheme

KW - Indirect encoding scheme

KW - Pattern classification

UR - http://www.scopus.com/inward/record.url?scp=85080941799&partnerID=8YFLogxK

U2 - 10.1007/978-3-030-35445-9_31

DO - 10.1007/978-3-030-35445-9_31

M3 - Chapter

AN - SCOPUS:85080941799

T3 - Studies in Computational Intelligence

SP - 431

EP - 443

BT - Studies in Computational Intelligence

PB - Springer

ER -

Direct and Indirect Evolutionary Designs of Artificial Neural Networks

Abstract

Publication series

Keywords

Access to Document

Other files and links

Fingerprint

Cite this