Candidate oversampling prefers two to tango: Estimation of distribution algorithms

David Wallin; Conor Ryan; R. Muhammad Atif Azad

doi:10.1145/2001858.2001897

Candidate oversampling prefers two to tango: Estimation of distribution algorithms

David Wallin, Conor Ryan, R. Muhammad Atif Azad

Department of Computer Science & Information Systems

University of Limerick

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

Abstract

Recent work has enhanced the Evolutionary Bayesian Classifier-based Optimization Algorithm (EBCOA) by oversampling the next generation and identifying promising solutions without actually evaluating their fitness values. In order to model the existing generation, that work considered two classes of solutions, that is, high performing solutions (H-Group) and poorly performing solutions (L-Group). In this study, we test the utility of using two classes instead of using a single class, as is the norm in standard Estimation of Distribution Algorithms (EDAs). Our results show that a dual class model is preferable when oversampling is used.

Original language	English
Title of host publication	Genetic and Evolutionary Computation Conference, GECCO'11 - Companion Publication
Pages	65-66
Number of pages	2
DOIs	https://doi.org/10.1145/2001858.2001897
Publication status	Published - 2011
Event	13th Annual Genetic and Evolutionary Computation Conference, GECCO'11 - Dublin, Ireland Duration: 12 Jul 2011 → 16 Jul 2011

Publication series

Name	Genetic and Evolutionary Computation Conference, GECCO'11 - Companion Publication

Conference

Conference	13th Annual Genetic and Evolutionary Computation Conference, GECCO'11
Country/Territory	Ireland
City	Dublin
Period	12/07/11 → 16/07/11

Keywords

eda
estimation of distribution
evolutionary computation
probabilistic model
probabilistic model-building

Access to Document

10.1145/2001858.2001897

Cite this

@inproceedings{a138f25cb66043bc95a09443255b7943,

title = "Candidate oversampling prefers two to tango: Estimation of distribution algorithms",

abstract = "Recent work has enhanced the Evolutionary Bayesian Classifier-based Optimization Algorithm (EBCOA) by oversampling the next generation and identifying promising solutions without actually evaluating their fitness values. In order to model the existing generation, that work considered two classes of solutions, that is, high performing solutions (H-Group) and poorly performing solutions (L-Group). In this study, we test the utility of using two classes instead of using a single class, as is the norm in standard Estimation of Distribution Algorithms (EDAs). Our results show that a dual class model is preferable when oversampling is used.",

keywords = "eda, estimation of distribution, evolutionary computation, probabilistic model, probabilistic model-building",

author = "David Wallin and Conor Ryan and Azad, {R. Muhammad Atif}",

year = "2011",

doi = "10.1145/2001858.2001897",

language = "English",

isbn = "9781450306904",

series = "Genetic and Evolutionary Computation Conference, GECCO'11 - Companion Publication",

pages = "65--66",

booktitle = "Genetic and Evolutionary Computation Conference, GECCO'11 - Companion Publication",

note = "13th Annual Genetic and Evolutionary Computation Conference, GECCO'11 ; Conference date: 12-07-2011 Through 16-07-2011",

}

Wallin, D, Ryan, C & Azad, RMA 2011, Candidate oversampling prefers two to tango: Estimation of distribution algorithms. in Genetic and Evolutionary Computation Conference, GECCO'11 - Companion Publication. Genetic and Evolutionary Computation Conference, GECCO'11 - Companion Publication, pp. 65-66, 13th Annual Genetic and Evolutionary Computation Conference, GECCO'11, Dublin, Ireland, 12/07/11. https://doi.org/10.1145/2001858.2001897

Candidate oversampling prefers two to tango: Estimation of distribution algorithms. / Wallin, David; Ryan, Conor; Azad, R. Muhammad Atif.
Genetic and Evolutionary Computation Conference, GECCO'11 - Companion Publication. 2011. p. 65-66 (Genetic and Evolutionary Computation Conference, GECCO'11 - Companion Publication).

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

TY - GEN

T1 - Candidate oversampling prefers two to tango

T2 - 13th Annual Genetic and Evolutionary Computation Conference, GECCO'11

AU - Wallin, David

AU - Ryan, Conor

AU - Azad, R. Muhammad Atif

PY - 2011

Y1 - 2011

N2 - Recent work has enhanced the Evolutionary Bayesian Classifier-based Optimization Algorithm (EBCOA) by oversampling the next generation and identifying promising solutions without actually evaluating their fitness values. In order to model the existing generation, that work considered two classes of solutions, that is, high performing solutions (H-Group) and poorly performing solutions (L-Group). In this study, we test the utility of using two classes instead of using a single class, as is the norm in standard Estimation of Distribution Algorithms (EDAs). Our results show that a dual class model is preferable when oversampling is used.

AB - Recent work has enhanced the Evolutionary Bayesian Classifier-based Optimization Algorithm (EBCOA) by oversampling the next generation and identifying promising solutions without actually evaluating their fitness values. In order to model the existing generation, that work considered two classes of solutions, that is, high performing solutions (H-Group) and poorly performing solutions (L-Group). In this study, we test the utility of using two classes instead of using a single class, as is the norm in standard Estimation of Distribution Algorithms (EDAs). Our results show that a dual class model is preferable when oversampling is used.

KW - eda

KW - estimation of distribution

KW - evolutionary computation

KW - probabilistic model

KW - probabilistic model-building

UR - http://www.scopus.com/inward/record.url?scp=80051921715&partnerID=8YFLogxK

U2 - 10.1145/2001858.2001897

DO - 10.1145/2001858.2001897

M3 - Conference contribution

AN - SCOPUS:80051921715

SN - 9781450306904

T3 - Genetic and Evolutionary Computation Conference, GECCO'11 - Companion Publication

SP - 65

EP - 66

BT - Genetic and Evolutionary Computation Conference, GECCO'11 - Companion Publication

Y2 - 12 July 2011 through 16 July 2011

ER -

Candidate oversampling prefers two to tango: Estimation of distribution algorithms

Abstract

Publication series

Conference

Keywords

Access to Document

Other files and links

Fingerprint

Cite this