Number of the records: 1
Doubly Trained Evolution Control for the Surrogate CMA-ES
- 1.
SYSNO ASEP 0466878 Document Type C - Proceedings Paper (int. conf.) R&D Document Type Conference Paper Title Doubly Trained Evolution Control for the Surrogate CMA-ES Author(s) Pitra, Zbyněk (UIVT-O) RID, ORCID, SAI
Bajer, L. (CZ)
Holeňa, Martin (UIVT-O) SAI, RIDSource Title Parallel Problem Solving from Nature - PPSN XIV. - Cham : Springer, 2016 / Handl J. ; Hart E. ; Lewis P.R. ; López-Ibáñez M. ; Ochoa G. ; Paechter B. - ISSN 0302-9743 - ISBN 978-3-319-45822-9 Pages s. 59-68 Number of pages 10 s. Publication form Print - P Action PPSN XIV. International Conference on Parallel Problem Solving from Nature /14./ Event date 17.09.2016 - 21.09.2016 VEvent location Edinburgh Country GB - United Kingdom Event type WRD Language eng - English Country CH - Switzerland Keywords black-box optimization ; surrogate model ; evolution control ; Gaussian process Subject RIV IN - Informatics, Computer Science R&D Projects NV15-33250A GA MZd - Ministry of Health (MZ) Institutional support UIVT-O - RVO:67985807 UT WOS 000387962100006 EID SCOPUS 84988531113 DOI 10.1007/978-3-319-45823-6_6 Annotation This paper presents a new variant of surrogate-model utilization in expensive continuous evolutionary black-box optimization. This algorithm is based on the surrogate version of the CMA-ES, the Surrogate Covariance Matrix Adaptation Evolution Strategy (S-CMA-ES). Similarly to the original S-CMA-ES, expensive function evaluations are saved through a surrogate model. However, the model is retrained after the points in which its prediction was most uncertain have been evaluated by the true fitness in each generation. We demonstrate that within small budget of evaluations, the new variant of S-CMA-ES improves the original algorithm and outperforms two state-of-the-art surrogate optimizers, except a few evaluations at the beginning of the optimization process. Workplace Institute of Computer Science Contact Tereza Šírová, sirova@cs.cas.cz, Tel.: 266 053 800 Year of Publishing 2017
Number of the records: 1