A detailed analysis of kernel parameters in Gaussian process-based optimization

Abstract : The global optimization of expensive-to-evaluate functions frequently occurs in many real-world applications. Among the methods developed for solving such problems, Efficient Global Optimization (EGO) is regarded as one of the state-of-the-art unconstrained continuous optimization algorithms. The most important control on the efficiency of EGO is the Gaussian process covariance function which must be chosen together with the objective function. Traditionally, a param-eterized family of covariance functions is considered whose parameters are learned by maximum likelihood or cross-validation. In this paper, we theoretically and empirically analyze the effect of length-scale covariance parameters and nugget on the design of experiments generated by EGO and the associated optimization performance.
Document type :
Reports
Complete list of metadatas

Cited literature [9 references]  Display  Hide  Download

https://hal.archives-ouvertes.fr/hal-01246677
Contributor : Hossein Mohammadi <>
Submitted on : Tuesday, February 9, 2016 - 7:28:06 PM
Last modification on : Tuesday, October 23, 2018 - 2:36:09 PM
Long-term archiving on : Saturday, November 12, 2016 - 3:49:50 PM

File

A detailed analysis of kernel ...
Files produced by the author(s)

Identifiers

  • HAL Id : hal-01246677, version 2

Citation

Hossein Mohammadi, Rodolphe Le Riche, Eric Touboul. A detailed analysis of kernel parameters in Gaussian process-based optimization. [Technical Report] Ecole Nationale Supérieure des Mines; LIMOS. 2015. ⟨hal-01246677v2⟩

Share

Metrics

Record views

362

Files downloads

195