Toggle Main Menu Toggle Search

Open Access padlockePrints

Weasel Words, Genetic Algorithms and Coarse Acknowledgements

Lookup NU author(s): Dr Victor Khomenko

Downloads

Full text for this publication is not currently held within this repository. Alternative links are provided below where available.


Abstract

The ‘weasel words’ example by Richard Dawkins, allegedly proving the possibility of evolution, is examined. It is considered in the framework of genetic algorithms, and a new model, which better reflects the process of natural selection, is proposed. Computer simulations of these two models are performed and compared with each other, and it is shown that in contrast to the original ‘weasel words’ example, which converges within a few hundred generations, the proposed new model does not converge even within a million generations. It is argued that the primary reason for the quick convergence of the former model is fine acknowledgements, which are not very realistic in the real world. In the latter model, the acknowledgments are ‘coarser’, which makes it much harder to converge. Moreover, an additional problem of multiextremity, which is not present in the former model due to its simplicity, clearly manifests itself in the new, less artificial, setting.


Publication metadata

Author(s): Khomenko V

Publication type: Article

Publication status: Published

Journal: Technical Journal

Year: 2004

Volume: 18

Issue: 1

Pages: 78-85

Print publication date: 01/01/2004

ISSN (print): 1833-6213

Publisher: Answers in Genesis


Share