Sequential Design of Computer Experiments for the Assessment of Fetus Exposure to Electromagnetic Fields. Issue 1 (2nd January 2016)

• Record Type: Journal Article
• Authors: Jala, Marjorie; Levy-Leduc, Céline; Moulines, Éric; Conil, Emmanuelle; Wiart, Joe
• Is Part Of: Technometrics. Volume 58:Issue 1(2016)
• Page Start: 30
• Page End: 42
• Publication Date: 2016-01-02
• Subjects: Design of experiments -- Gaussian process -- Quantile estimation -- Sequential approach; Statistical physics -- Periodicals; Chemistry -- Statistical methods -- Periodicals; Engineering -- Statistical methods -- Periodicals; 519.5
• DOI: 10.1080/00401706.2014.979951
• Languages: English
• ISSNs: 0040-1706

Abstract

Abstract : In this article, we describe four sequential sampling strategies for estimating the quantile of Y = f (X ), whereX has a known distribution in R d and f is a deterministic unknown, expensive-to-evaluate real-valued function. These approaches all consist in modeling f as a sample of a well-chosen Gaussian process and aim at estimating the quantile by using as few evaluations of f as possible. The different methodologies are first compared through various numerical experiments. Then, in the framework of the ANR-JST FETUS project, we apply our strategies to a real example corresponding to the exposure of a Japanese pregnant-woman model and her 26-week-old fetus to a plane wave. Finally, we compare our methodologies on a simplified geometric model designed for modeling the fetus exposure to plane waves. Supplementary materials for this article are available online.