Mühlenbein, Heinz ;
Höns, Robin
The Factorized Distribution Algorithm and the Minimum Relative Entropy Principle
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 BetheKikuchi 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 2d Ising spin glass problems and Kaufman's nk function as examples.
BibTeX  Entry
@InProceedings{mhlenbein_et_al:DSP:2006:597,
author = {Heinz M{\"u}hlenbein and Robin H{\"o}ns},
title = {The Factorized Distribution Algorithm and the Minimum Relative Entropy Principle},
booktitle = {Theory of Evolutionary Algorithms},
year = {2006},
editor = {Dirk V. Arnold and Thomas Jansen and Michael D. Vose and Jonathan E. Rowe},
number = {06061},
series = {Dagstuhl Seminar Proceedings},
ISSN = {18624405},
publisher = {Internationales Begegnungs und Forschungszentrum f{\"u}r Informatik (IBFI), Schloss Dagstuhl, Germany},
address = {Dagstuhl, Germany},
URL = {http://drops.dagstuhl.de/opus/volltexte/2006/597},
annote = {Keywords: Junction tree, minimum relative entropy, maximum likelihood, BetheKikuchi approximation}
}
2006
Keywords: 

Junction tree, minimum relative entropy, maximum likelihood, BetheKikuchi approximation 
Seminar: 

06061  Theory of Evolutionary Algorithms

Related Scholarly Article: 


Issue date: 

2006 
Date of publication: 

2006 