The concept of information gain has been adopted as tool to study the effectiveness of population-based optimizers; using this approach, it is possible to infer convergence properties for population-based optimizers. The experimental results have shown characteristic phase transition between exploration and exploitation phase during the evolutionary process and, moreover, the evidence that gain maximization offers a robust theoretical framework to study the convergence of stochastic optimizers.
|Titolo:||Entropic Divergence for Population Based Optimization Algorithms|
|Data di pubblicazione:||2010|
|Appare nelle tipologie:||4.1 Contributo in Atti di convegno|