Nástroj pro analýzu a vizualizaci hejnových algoritmů

Abstract

Swarm algorithms, or Swarm intelligences, are popular and increasingly used tools for optimization problems, which in the scope of this thesis means finding the global optimum - in our case, the minimum. With their help it is possible, by using a swarm or a flock, inspired by animal behaviour, to solve problems, which an individual would not be capable solving without help, or it would take him a disproportionately long time to solve. The topic of this thesis was to study the above mentioned algorithms and to program the selected algorithms, for this purpose the algorithms known by the abbreviations PSO, SOMA and WHALE were chosen and are explained below. In the next objectives of the work, it was necessary to study the possibilities of visualization in a multidimensional space and the properties of the population so that they could be programmed and practically used for the last point, which is the comparison of the populations of the three flocking algorithms.