Evolutionary Computation Based Student Scheduling

U svakodnevnom životu problemi raspoređivanja su vrlo česti i obično su visoke računske složenosti. Stoga su načini njihovog rješavanja od posebnog interesa. U okviru ovog rada implementiran je prototip programske implementacije sustava za raspoređivanja korisnika po radionicama koristeći genetski algoritam. Analiziran je utjecaj ponašanja algoritma u odnosu na različite konfiguracije parametara i heuristika. Pokazano je da razvijeni sustav uspješno raspoređuje korisnike po radionicama uz mogućnost konfiguracije raznih kriterija dobrote rasporeda.Scheduling problems are very common in everyday life and are of high computational complexity. As such, the methods for solving such problems are of particular interest. This thesis presents a prototype implementation of a system for scheduling user attendance for workshops based on the genetic algorithm. Analysis is conducted to determine the influence of various configuration parameters and heuristics on the algorithm's behaviour. The developed system is then demonstrated to successfully schedule users, with the ability to configure the various scheduling criteria