Vylepšení metaheuristických algoritmů pomocí symbolické regrese

Abstract

Optimization methods are still under development, and researchers are working on the improvement of current methods. The purpose of this thesis is to find an improvement of three metaheuristic algorithms - SOMA, PSO, and DE. The analytic programming is used as a method of symbolic regression for this purpose. The beginning of this thesis consists of descriptions of SOMA, PSO, and DE, as well as of analytic programming. All algorithms were implemented in the C++ programming language and experiments were performed. The results are evaluated at the end of this thesis. Significant improvement was found for the SOMA algorithm. For PSO and DE, improvements were observed for some of the objective functions.