Constrained effcient global optimization with probabilistic support vector machines

Anirban Basudhar, Sylvain Lacaze, Samy Missoum

Research output: Chapter in Book/Report/Conference proceedingConference contribution

6 Citations (Scopus)

Abstract

This paper presents a methodology for constrained effcient global optimization (EGO) using support vector machines (SVMs). The proposed SVM-based method has several advantages. It is more general because it is applicable to a wider variety of problems compared to current techniques. These include problems with discontinuous and binary (pass/fail) states and multiple constraints. In this paper, the objective function is ap- proximated using Kriging while the constraint boundary is approximated using an SVM classifier. The probability of misclassification by the SVM is calculated using a probabilistic support vector machine (PSVM). The existing PSVM models have certain limitations that make them unsuitable for application in the proposed methodology. Therefore, a modified PSVM model is also proposed to overcome these limitations. Several constrained EGO for- mulations are implemented and compared in this paper. The results are also compared to EGO implementations with Kriging-based constraint approximations from the literature.

Original languageEnglish (US)
Title of host publication13th AIAA/ISSMO Multidisciplinary Analysis and Optimization Conference 2010
DOIs
StatePublished - 2010
Externally publishedYes
Event13th AIAA/ISSMO Multidisciplinary Analysis and Optimization Conference, MAO 2010 - Ft. Worth, TX, United States
Duration: Sep 13 2010Sep 15 2010

Other

Other13th AIAA/ISSMO Multidisciplinary Analysis and Optimization Conference, MAO 2010
CountryUnited States
CityFt. Worth, TX
Period9/13/109/15/10

Fingerprint

Constrained optimization
Global optimization
Support vector machines
Classifiers

ASJC Scopus subject areas

  • Aerospace Engineering
  • Mechanical Engineering

Cite this

Basudhar, A., Lacaze, S., & Missoum, S. (2010). Constrained effcient global optimization with probabilistic support vector machines. In 13th AIAA/ISSMO Multidisciplinary Analysis and Optimization Conference 2010 https://doi.org/10.2514/6.2010-9230

Constrained effcient global optimization with probabilistic support vector machines. / Basudhar, Anirban; Lacaze, Sylvain; Missoum, Samy.

13th AIAA/ISSMO Multidisciplinary Analysis and Optimization Conference 2010. 2010.

Research output: Chapter in Book/Report/Conference proceedingConference contribution

Basudhar, A, Lacaze, S & Missoum, S 2010, Constrained effcient global optimization with probabilistic support vector machines. in 13th AIAA/ISSMO Multidisciplinary Analysis and Optimization Conference 2010. 13th AIAA/ISSMO Multidisciplinary Analysis and Optimization Conference, MAO 2010, Ft. Worth, TX, United States, 9/13/10. https://doi.org/10.2514/6.2010-9230
Basudhar A, Lacaze S, Missoum S. Constrained effcient global optimization with probabilistic support vector machines. In 13th AIAA/ISSMO Multidisciplinary Analysis and Optimization Conference 2010. 2010 https://doi.org/10.2514/6.2010-9230
Basudhar, Anirban ; Lacaze, Sylvain ; Missoum, Samy. / Constrained effcient global optimization with probabilistic support vector machines. 13th AIAA/ISSMO Multidisciplinary Analysis and Optimization Conference 2010. 2010.
@inproceedings{454cc3c332474bb8ae72a6a53bd65847,
title = "Constrained effcient global optimization with probabilistic support vector machines",
abstract = "This paper presents a methodology for constrained effcient global optimization (EGO) using support vector machines (SVMs). The proposed SVM-based method has several advantages. It is more general because it is applicable to a wider variety of problems compared to current techniques. These include problems with discontinuous and binary (pass/fail) states and multiple constraints. In this paper, the objective function is ap- proximated using Kriging while the constraint boundary is approximated using an SVM classifier. The probability of misclassification by the SVM is calculated using a probabilistic support vector machine (PSVM). The existing PSVM models have certain limitations that make them unsuitable for application in the proposed methodology. Therefore, a modified PSVM model is also proposed to overcome these limitations. Several constrained EGO for- mulations are implemented and compared in this paper. The results are also compared to EGO implementations with Kriging-based constraint approximations from the literature.",
author = "Anirban Basudhar and Sylvain Lacaze and Samy Missoum",
year = "2010",
doi = "10.2514/6.2010-9230",
language = "English (US)",
isbn = "9781600869549",
booktitle = "13th AIAA/ISSMO Multidisciplinary Analysis and Optimization Conference 2010",

}

TY - GEN

T1 - Constrained effcient global optimization with probabilistic support vector machines

AU - Basudhar, Anirban

AU - Lacaze, Sylvain

AU - Missoum, Samy

PY - 2010

Y1 - 2010

N2 - This paper presents a methodology for constrained effcient global optimization (EGO) using support vector machines (SVMs). The proposed SVM-based method has several advantages. It is more general because it is applicable to a wider variety of problems compared to current techniques. These include problems with discontinuous and binary (pass/fail) states and multiple constraints. In this paper, the objective function is ap- proximated using Kriging while the constraint boundary is approximated using an SVM classifier. The probability of misclassification by the SVM is calculated using a probabilistic support vector machine (PSVM). The existing PSVM models have certain limitations that make them unsuitable for application in the proposed methodology. Therefore, a modified PSVM model is also proposed to overcome these limitations. Several constrained EGO for- mulations are implemented and compared in this paper. The results are also compared to EGO implementations with Kriging-based constraint approximations from the literature.

AB - This paper presents a methodology for constrained effcient global optimization (EGO) using support vector machines (SVMs). The proposed SVM-based method has several advantages. It is more general because it is applicable to a wider variety of problems compared to current techniques. These include problems with discontinuous and binary (pass/fail) states and multiple constraints. In this paper, the objective function is ap- proximated using Kriging while the constraint boundary is approximated using an SVM classifier. The probability of misclassification by the SVM is calculated using a probabilistic support vector machine (PSVM). The existing PSVM models have certain limitations that make them unsuitable for application in the proposed methodology. Therefore, a modified PSVM model is also proposed to overcome these limitations. Several constrained EGO for- mulations are implemented and compared in this paper. The results are also compared to EGO implementations with Kriging-based constraint approximations from the literature.

UR - http://www.scopus.com/inward/record.url?scp=84880831816&partnerID=8YFLogxK

UR - http://www.scopus.com/inward/citedby.url?scp=84880831816&partnerID=8YFLogxK

U2 - 10.2514/6.2010-9230

DO - 10.2514/6.2010-9230

M3 - Conference contribution

AN - SCOPUS:84880831816

SN - 9781600869549

BT - 13th AIAA/ISSMO Multidisciplinary Analysis and Optimization Conference 2010

ER -