Paper
15 October 2012 Benchmarking a derivative-free minimax optimizer for EM design
Author Affiliations +
Proceedings Volume 8454, Photonics Applications in Astronomy, Communications, Industry, and High-Energy Physics Experiments 2012; 84541H (2012) https://doi.org/10.1117/12.2000051
Event: Photonics Applications in Astronomy, Communications, Industry, and High-Energy Physics Experiments 2012, 2012, Wilga, Poland
Abstract
The paper presents a model-based derivative-free minimax optimizer and a benchmarking methodology. The optimizer is dedicated to such simulation-based design scenarios, when a single simulation is time-consuming and not very accurate, and so when fast improvement of design is expected rather then high accuracy optimization. A benchmarking methodology is formulated to compare efficiency of optimizers in such scenarios. A set of electromagnetic (EM) designs, for which a design simulation involves solution of very large partial differential equations, is used to exemplify the methodology. Presented results demonstrate how accuracy requirements and computational budget change ranking of optimizers. The proposed optimizer is shown to provide competetively rapid and reliable initial design improvement for modest 5-10% tolerance, but for higher accuracy demands and larger computational budget other solvers become more competitive.
© (2012) COPYRIGHT Society of Photo-Optical Instrumentation Engineers (SPIE). Downloading of the abstract is permitted for personal use only.
Leszek J. Opalski "Benchmarking a derivative-free minimax optimizer for EM design", Proc. SPIE 8454, Photonics Applications in Astronomy, Communications, Industry, and High-Energy Physics Experiments 2012, 84541H (15 October 2012); https://doi.org/10.1117/12.2000051
Lens.org Logo
CITATIONS
Cited by 1 scholarly publication.
Advertisement
Advertisement
RIGHTS & PERMISSIONS
Get copyright permission  Get copyright permission on Copyright Marketplace
KEYWORDS
Optimization (mathematics)

Algorithm development

Tolerancing

Computer simulations

Electromagnetism

Expectation maximization algorithms

Waveguides

Back to Top