Modelling GA Dynamics: 1st
Modelling GA Dynamics: 1st
This paper is an introduction to the mathematical modelling of the dynamics of genetic algorithms. The distinguishing feature of this approach is that we consider macroscopic properties of the system. After some brief introductory remarks, we look at a generational GA, with tournament selection, tackling the ones-counting problem. Initially we ignore recombination. We start with a two parameter model of the evolution. This is sufficient to explain the qualitative features of the dynamics, although it does not give a good quantitative agreement with simulations. We show how the agreement can be improved by using more parameters to describe the population and by introducing finite population corrections. Finally, we come back to recombination and show how this can be modelled.
3-540-67396-2
59-86
Prügel-Bennett, A.
b107a151-1751-4d8b-b8db-2c395ac4e14e
Rogers, A.
f9130bc6-da32-474e-9fab-6c6cb8077fdc
2001
Prügel-Bennett, A.
b107a151-1751-4d8b-b8db-2c395ac4e14e
Rogers, A.
f9130bc6-da32-474e-9fab-6c6cb8077fdc
Prügel-Bennett, A. and Rogers, A.
(2001)
Modelling GA Dynamics: 1st.
In,
Kallel, L., Naudts, B. and Rogers, A.
(eds.)
Theoretical Aspects of Evolutionary Computing.
(Natural Computing)
Theoretical Aspects of Evolutionary Computing (01/01/01)
Springer, .
Record type:
Book Section
Abstract
This paper is an introduction to the mathematical modelling of the dynamics of genetic algorithms. The distinguishing feature of this approach is that we consider macroscopic properties of the system. After some brief introductory remarks, we look at a generational GA, with tournament selection, tackling the ones-counting problem. Initially we ignore recombination. We start with a two parameter model of the evolution. This is sufficient to explain the qualitative features of the dynamics, although it does not give a good quantitative agreement with simulations. We show how the agreement can be improved by using more parameters to describe the population and by introducing finite population corrections. Finally, we come back to recombination and show how this can be modelled.
More information
Published date: 2001
Additional Information:
Address: Berlin
Venue - Dates:
Theoretical Aspects of Evolutionary Computing, 2001-01-01
Organisations:
Agents, Interactions & Complexity, Southampton Wireless Group
Identifiers
Local EPrints ID: 263205
URI: http://eprints.soton.ac.uk/id/eprint/263205
ISBN: 3-540-67396-2
PURE UUID: f999d124-7777-40a2-952f-57a164aeadb5
Catalogue record
Date deposited: 23 Nov 2006
Last modified: 14 Mar 2024 07:26
Export record
Contributors
Author:
A. Prügel-Bennett
Author:
A. Rogers
Editor:
L. Kallel
Editor:
B. Naudts
Editor:
A. Rogers
Download statistics
Downloads from ePrints over the past year. Other digital versions may also be available to download e.g. from the publisher's website.
View more statistics