Dynamic Partitional Clustering Using Evolution Strategies

Cin-Young Lee and Erik K. Antonsson

In The Third Asia-Pacific Conference on Simulated Evolution and Learning
(October 2000).


A novel evolution strategy implementing variable length genomes is developed to address the problem of dynamic partitional clustering. As opposed to static, dynamic clustering does not require a priori specification of the cluster number. Results of the algorithm are presented and discussed for 2-D touching and non-touching cluster test cases.