Accessibility navigation


Dynamic Niche Clustering: a fuzzy variable radius niching technique for multimodal optimisation in GAs

Gan, J. and Warwick, K. (2001) Dynamic Niche Clustering: a fuzzy variable radius niching technique for multimodal optimisation in GAs. In: Proceedings of the 2001 Congress on Evolutionary Computation. IEEE, pp. 215-222. ISBN 0780366573

Full text not archived in this repository.

To link to this article DOI: 10.1109/CEC.2001.934392

Abstract/Summary

This paper describes the recent developments and improvements made to the variable radius niching technique called Dynamic Niche Clustering (DNC). DNC is fitness sharing based technique that employs a separate population of overlapping fuzzy niches with independent radii which operate in the decoded parameter space, and are maintained alongside the normal GA population. We describe a speedup process that can be applied to the initial generation which greatly reduces the complexity of the initial stages. A split operator is also introduced that is designed to counteract the excessive growth of niches, and it is shown that this improves the overall robustness of the technique. Finally, the effect of local elitism is documented and compared to the performance of the basic DNC technique on a selection of 2D test functions. The paper is concluded with a view to future work to be undertaken on the technique.

Item Type:Book or Report Section
Refereed:Yes
Divisions:Faculty of Science > School of Systems Engineering
ID Code:21613
Publisher:IEEE

Centaur Editors: Update this record

Page navigation