Header

UZH-Logo

Maintenance Infos

Gaussian Adaptation Revisited – An Entropic View on Covariance Matrix Adaptation


Müller, Christian L; Sbalzarini, Ivo F (2010). Gaussian Adaptation Revisited – An Entropic View on Covariance Matrix Adaptation. Lecture Notes in Computer Science, 6024:432-441.

Abstract

We revisit Gaussian Adaptation (GaA), a black-box optimizer for discrete and continuous problems that has been developed in the late 1960’s. This largely neglected search heuristic shares several interesting features with the well-known Covariance Matrix Adaptation Evolution Strategy (CMA-ES) and with Simulated Annealing (SA). GaA samples single candidate solutions from a multivariate normal distribution and continuously adapts its first and second moments (mean and covariance) such as to maximize the entropy of the search distribution. Sample-point selection is controlled by a monotonically decreasing acceptance threshold, reminiscent of the cooling schedule in SA. We describe the theoretical foundations of GaA and analyze some key features of this algorithm. We empirically show that GaA converges log-linearly on the sphere function and analyze its behavior on selected non-convex test functions.

Abstract

We revisit Gaussian Adaptation (GaA), a black-box optimizer for discrete and continuous problems that has been developed in the late 1960’s. This largely neglected search heuristic shares several interesting features with the well-known Covariance Matrix Adaptation Evolution Strategy (CMA-ES) and with Simulated Annealing (SA). GaA samples single candidate solutions from a multivariate normal distribution and continuously adapts its first and second moments (mean and covariance) such as to maximize the entropy of the search distribution. Sample-point selection is controlled by a monotonically decreasing acceptance threshold, reminiscent of the cooling schedule in SA. We describe the theoretical foundations of GaA and analyze some key features of this algorithm. We empirically show that GaA converges log-linearly on the sphere function and analyze its behavior on selected non-convex test functions.

Statistics

Citations

7 citations in Web of Science®
6 citations in Scopus®
Google Scholar™

Altmetrics

Downloads

87 downloads since deposited on 05 Jul 2013
14 downloads since 12 months
Detailed statistics

Additional indexing

Item Type:Journal Article, refereed, original work
Communities & Collections:Special Collections > SystemsX.ch
Special Collections > SystemsX.ch > Research, Technology and Development Projects > WingX
Dewey Decimal Classification:570 Life sciences; biology
Language:English
Date:2010
Deposited On:05 Jul 2013 11:59
Last Modified:07 Dec 2017 21:38
Publisher:Springer
ISSN:0302-9743
Publisher DOI:https://doi.org/10.1007/978-3-642-12239-2_45

Download

Download PDF  'Gaussian Adaptation Revisited – An Entropic View on Covariance Matrix Adaptation'.
Preview
Filetype: PDF
Size: 250kB
View at publisher