Míry diverzity populace v evolučních algoritmech

Abstract

The thesis focuses on a basic analysis of diversity measures used in genetic algorithms and differential evolution. Its main goal is to analyze and describe the behavior of selected diversity measures on various discrete and continuous optimization problems and under different configurations of these evolutionary algorithms. For discrete optimization, the 3-SAT, GCP, and QAP problems were used to analyze the diversity measures of genetic algorithms. Rosenbrock and Ackley functions were selected for continuous optimization and testing of diversity measures in differential evolution. The results demonstrated that individual diversity measures respond very differently to both various problems and the specific properties of their instances.

Description

Delayed publication

Available after

Subject(s)

diversity, genetic algorithms, differential evolution, evolutionary algorithms, continuous optimization problems, discrete optimization problems

Citation