Improving Selection Methods for Evolutionary Algorithms by Clustering

This study applies clustering in population selection to improve the efficiency of evolutionary algorithms. The main contributions include: (a) Proposes a novel selection framework that uses the number of clusters for a population as the measurement the population diversity, (b) Proposes clustering-ranking selection, an instance of this framework, and discusses its mathematical principle by PD-SP equation, (c) Gives experiments over CLPSO (comprehensive learning particle swarm optimization). Experiment result shows that the proposed selection method outperforms canonical exponential ranking on all the sixteen-benchmark functions for both 10-D and 30-D problems except a function for 30-D problem.