The Factorized Distribution Algorithm and the Minimum Relative Entropy Principle

Mühlenbein, Heinz; Höns, Robin

doi:10.4230/DagSemProc.06061.9

Document

The Factorized Distribution Algorithm and the Minimum Relative Entropy Principle

Authors Heinz Mühlenbein, Robin Höns

Part of: Volume: Dagstuhl Seminar Proceedings, Volume 6061
Part of: Series: Dagstuhl Seminar Proceedings (DagSemProc)
License: Creative Commons Attribution 4.0 International license
Publication Date: 2006-07-07

PDF

File

PDF

DagSemProc.06061.9.pdf

Filesize: 263 kB
27 pages

Document Identifiers

DOI: 10.4230/DagSemProc.06061.9
URN: urn:nbn:de:0030-drops-5973

Subject Classification

Keywords

Junction tree
minimum relative entropy
maximum likelihood
Bethe-Kikuchi approximation

Metrics

Access Statistics
Total Accesses (updated on a weekly basis)

0

PDF Downloads

0

Metadata Views

Abstract

We assume that the function to be optimized is additively decomposed (ADF). Then the interaction graph $G_{ADF}$ can be used to compute exact or approximate factorizations. For many practical problems only approximate factorizations lead to efficient optimization algorithms. The relation between the approximation used by the FDA algorithm and the minimum relative entropy principle is discussed. A new algorithm is presented, derived from the Bethe-Kikuchi approach in statistical physics. It minimizes the relative entropy to a Boltzmann distribution with fixed $eta$. We shortly compare different factorizations and algorithms within the FDA software. We use 2-d Ising spin glass problems and Kaufman's n-k function as examples.

Cite As Get BibTex

Heinz Mühlenbein and Robin Höns. The Factorized Distribution Algorithm and the Minimum Relative Entropy Principle. In Theory of Evolutionary Algorithms. Dagstuhl Seminar Proceedings, Volume 6061, pp. 1-27, Schloss Dagstuhl – Leibniz-Zentrum für Informatik (2006) https://doi.org/10.4230/DagSemProc.06061.9

Author Details

Heinz Mühlenbein

Robin Höns

Any Issues?

Feedback on the Current Page

Thanks for your feedback!

Feedback submitted to Dagstuhl Publishing

Could not send message

Please try again later or send an E-mail