CISUC

A Hybrid Evolutionary Algorithm for Cluster Geometry Optimization: the importance of structural elitism

Authors

Abstract

We report a detailed analysis on the effect of structural elitism in the global minimum search of atomic clusters by a hybrid evolutionary algorithm. Structural elitism allows to preserve a certain diversity in the population and is based on the classification of clusters as spheric, oblate, prolate or asymmetric according to their shape.
The search algorithm has been applied to perform the global optimization of Morse clusters between 30 and 50 atoms. Results show that the structural elitism enhances the ability of the algorithm to discover the putative global minima.

Keywords

cluster geometry optimization, elitism, selection

Subject

Evolutionary Optimization

Conference

Eight International Conference on Hybrid Intelligent Systems (HIS-2008), September 2008


Cited by

No citations found