Dagstuhl Seminar Proceedings, Volume 6061
Dagstuhl Seminar Proceedings
DagSemProc
https://www.dagstuhl.de/dagpub/1862-4405
https://dblp.org/db/series/dagstuhl
1862-4405
Schloss Dagstuhl – Leibniz-Zentrum für Informatik
6061
2006
https://drops.dagstuhl.de/entities/volume/DagSemProc-volume-6061
06061 Abstracts Collection – Theory of Evolutionary Algorithms
From 05.02.06 to 10.02.06, the Dagstuhl Seminar 06061 ``Theory of Evolutionary Algorithms'' was held in the International Conference and Research Center (IBFI),
Schloss Dagstuhl.
During the seminar, several participants presented their current
research, and ongoing work and open problems were discussed. Abstracts of
the presentations given during the seminar as well as abstracts of
seminar results and ideas are put together in this paper. The first section
describes the seminar topics and goals in general.
Links to extended abstracts or full papers are provided, if available.
Evolutionary algorithms
evolutionary computation
theory
1-16
Regular Paper
Dirk V.
Arnold
Dirk V. Arnold
Thomas
Jansen
Thomas Jansen
Jonathan E.
Rowe
Jonathan E. Rowe
Michael D.
Vose
Michael D. Vose
10.4230/DagSemProc.06061.1
Creative Commons Attribution 4.0 International license
https://creativecommons.org/licenses/by/4.0/legalcode
06061 Executive Summary – Theory of Evolutionary Algoritms
The 2006 Dagstuhl Seminar ``Theory of Evolutionary Algorithms'' carried forward a series of Dagstuhl seminars that started in 2000 and has become an established event in the community. In the week from from 05.02.2006 to 10.02.2006, 56 researchers from 12 countries discussed their recent work and recent trends in evolutionary computation.
Evolutionary algorithms
evolutionary computation
theory
1-2
Regular Paper
Dirk V.
Arnold
Dirk V. Arnold
Thomas
Jansen
Thomas Jansen
Jonathan E.
Rowe
Jonathan E. Rowe
Michael D.
Vose
Michael D. Vose
10.4230/DagSemProc.06061.2
Creative Commons Attribution 4.0 International license
https://creativecommons.org/licenses/by/4.0/legalcode
A Mathematical Modelling Technique for the Analysis of the Dynamics of a Simple Continuous EDA
We describe a mathematical model for the infinite-population dynamics of a simple continuous EDA: UMDAc. Using this model, it is possible to numerically generate the dynamics of the algorithm on a fitness function of known form. The technique is compared with existing analysis and illustrated on a number of simple test problems. The model is also used to examine the effect of adding an amplification constant to the variance parameter of the UMDAc model.
Estimation of Distribution Algorithms
1-7
Regular Paper
Marcus
Gallagher
Marcus Gallagher
Bo
Yuan
Bo Yuan
10.4230/DagSemProc.06061.3
Creative Commons Attribution 4.0 International license
https://creativecommons.org/licenses/by/4.0/legalcode
A New Quartet Tree Heuristic for Hierarchical Clustering
We present a new quartet heuristic for
hierarchical clustering
from a given distance matrix.
We determine a dendrogram (ternary tree)
by a new quartet
method and a fast heuristic to implement it.
We do not assume that there is a true ternary tree that generated the
distances and which we with to recover as closeley as possible.
Our aim is to model the distance matrix as faithfully as possible
by the dendrogram. Our algorithm is essentially
randomized hill-climbing, using
parallellized Genetic Programming, where
undirected trees evolve in a random walk
driven by a prescribed fitness function.
Our method is capable of handling up to 60--80
objects in a matter of hours, while no existing quartet heuristic
can directly compute a quartet tree of more than about 20--30 objects
without running for years.
The method is implemented and available as public software
at www.complearn.org. We present applications in many areas
like music, literature, bird-flu (H5N1) virus clustering, and automatic
meaning discovery using Google.
Genetic programming
hierarchical clustering
quartet tree method
1-13
Regular Paper
Rudi
Cilibrasi
Rudi Cilibrasi
Paul M. B.
Vitany
Paul M. B. Vitany
10.4230/DagSemProc.06061.4
Creative Commons Attribution 4.0 International license
https://creativecommons.org/licenses/by/4.0/legalcode
How fast does the stationary distribution of the Markov chain modelling EAs concentrate on the homogeneous populations for small mutation rate?
The state space of the Markov chain modelling an evolutionary algorithm
is quite large especially if the population space and the search space are
large. I shell introduce an appropriate notion of "coarse graining" for
such Markov chains. Indeed, from the mathematical point of view, this can
be called a quotient of a Markov chain by an equivalence relation over the
state space. The newly obtained Markov chain has a significantly smaller
state space and it's stationary distribution is "coherent" with the
initial large chain. Although the transition probabilities of the
coarse-grained Markov chain are defined in terms of the stationary
distribution of the original big chain, in some cases it is possible to
deduce interesting information about the stationary distribution of the
original chain in terms of the quatient chain. I will demonstrate how
this method works. I shell also present some simple results and open
questions.
Markov chains
Evolutionary algorithms
coarse graining quotients of irreducible Markov chains
concentration on the uniform populations
1-9
Regular Paper
Boris S.
Mitavskiy
Boris S. Mitavskiy
Jonathan E.
Rowe
Jonathan E. Rowe
10.4230/DagSemProc.06061.5
Creative Commons Attribution 4.0 International license
https://creativecommons.org/licenses/by/4.0/legalcode
On Complexity of Optimized Crossover for Binary Representations
We consider the computational complexity of producing the best
possible offspring in a crossover, given two solutions of the
parents. The crossover operators are studied on the class of
Boolean linear programming problems, where the Boolean vector of
variables is used as the solution representation. By means of
efficient reductions of the optimized gene transmitting crossover
problems (OGTC) we show the polynomial solvability of the OGTC for
the maximum weight set packing problem, the minimum weight set
partition problem and for one of the versions of the simple plant
location problem. We study a connection between the OGTC for
linear Boolean programming problem and the maximum weight
independent set problem on 2-colorable hypergraph and prove the
NP-hardness of several special cases of the OGTC problem in
Boolean linear programming.
Genetic Algorithm
Optimized Crossover
Complexity
1-13
Regular Paper
Anton
Eremeev
Anton Eremeev
10.4230/DagSemProc.06061.6
Creative Commons Attribution 4.0 International license
https://creativecommons.org/licenses/by/4.0/legalcode
On Turing complete T7 and MISC F--4 program fitnes landscapes
We use the minimal instruction set F-4 computer to define a
minimal Turing complete T7 computer suitable for genetic
programming (GP) and amenable to theoretical analysis.
Experimental runs
and mathematical analysis of the T7,
show
the fraction of halting programs is drops to zero as bigger programs
are run.
Genetic programming
1-28
Regular Paper
William B.
Langdon
William B. Langdon
Riccardo
Poli
Riccardo Poli
10.4230/DagSemProc.06061.7
Creative Commons Attribution 4.0 International license
https://creativecommons.org/licenses/by/4.0/legalcode
Runtime Analysis of a Simple Ant Colony Optimization Algorithm
Ant Colony Optimization (ACO) has become quite popular in recent
years. In contrast to many successful applications, the theoretical foundation of
this randomized search heuristic is rather weak. Building up such a
theory is demanded to understand how these heuristics work as well as
to come up with better algorithms for certain problems. Up to now,
only convergence results have been achieved showing that optimal
solutions can be obtained in a finite amount of time. We present the
first runtime analysis of a simple ACO algorithm that
transfers many rigorous results with respect to the expected runtime
of a simple evolutionary algorithm to our algorithm. In addition,
we examine the choice of the evaporation factor, which is a crucial
parameter in such an algorithm, in greater detail and analyze its effect
with respect to the runtime.
Randomized Search Heuristics
Ant Colony Optimization
Runtime Analysis
1-17
Regular Paper
Frank
Neumann
Frank Neumann
Carsten
Witt
Carsten Witt
10.4230/DagSemProc.06061.8
Creative Commons Attribution 4.0 International license
https://creativecommons.org/licenses/by/4.0/legalcode
The Factorized Distribution Algorithm and the Minimum Relative Entropy Principle
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.
Junction tree
minimum relative entropy
maximum likelihood
Bethe-Kikuchi approximation
1-27
Regular Paper
Heinz
Mühlenbein
Heinz Mühlenbein
Robin
Höns
Robin Höns
10.4230/DagSemProc.06061.9
Creative Commons Attribution 4.0 International license
https://creativecommons.org/licenses/by/4.0/legalcode