Targeting Solutions in Bayesian Multi-Objective Optimization: Sequential and Batch Versions

David Gaudrie 1, 2, 3, 4, 5 Rodolphe Le Riche 5, 1, 6, 3, 4 Victor Picheny 7 Benoit Enaux 2 Vincent Herbert 2
1 Mines Saint-Étienne MSE - École des Mines de Saint-Étienne
2 PSA - PSA Peugeot - Citroën
3 FAYOL-ENSMSE - Institut Henri Fayol
4 FAYOL-ENSMSE - Département Génie mathématique et industriel
5 LIMOS - Laboratoire d'Informatique, de Modélisation et d'Optimisation des Systèmes
6 LCG-ENSMSE - UMR 5146 - Laboratoire Claude Goux
7 INRA Toulouse - Institut national de la recherche agronomique [Toulouse]
Abstract : Multi-objective optimization aims at finding trade-off solutions to conflicting objectives. These constitute the Pareto optimal set. In the context of expensive-to-evaluate functions, it is impossible and often non-informative to look for the entire set. As an end-user would typically prefer a certain part of the objective space, we modify the Bayesian multi-objective optimization algorithm which uses Gaussian Processes to maximize the Expected Hypervolume Improvement, to focus the search in the preferred region. The cumulated effects of the Gaussian Processes and the targeting strategy lead to a particularly efficient convergence to the desired part of the Pareto set. To take advantage of parallel computing, a multi-point extension of the targeting criterion is proposed and analyzed.
Keywords : Gaussian processes Bayesian optimization Computer experiments Preference-based optimization Parallel optimization
Document type :
Journal articles
Complete list of metadatas
https://hal-emse.ccsd.cnrs.fr/emse-01957614
Contributor : Florent Breuil <>
Submitted on : Monday, December 17, 2018 - 2:34:22 PM
Last modification on : Tuesday, September 24, 2019 - 9:26:27 AM

Links full text

arxiv

Identifiers

Collections

EMSE | DEMO-ENSMSE | FAYOL-ENSMSE | INRA | INSMI | INSTITUT-TELECOM | LIMOS | TDS-MACS | PRES_CLERMONT

Citation

David Gaudrie, Rodolphe Le Riche, Victor Picheny, Benoit Enaux, Vincent Herbert. Targeting Solutions in Bayesian Multi-Objective Optimization: Sequential and Batch Versions. Annals of Mathematics and Artificial Intelligence, Springer Verlag, 2019, pp 1-26. ⟨10.1007/s10472-019-09644-8⟩. ⟨emse-01957614⟩

Export

BibTeX TEI DC DCterms EndNote

Share

Metrics

Record views

83

Contact

support.ccsd.cnrs.fr
hal.support@ccsd.cnrs.fr
Logo CCSD