Usporedba rada evolucijskog algoritma i algoritma roja na standardnim testnim funkcijama

Čehulić, Lovro (2017) Usporedba rada evolucijskog algoritma i algoritma roja na standardnim testnim funkcijama. = Comparison of evolutionary and swarm algorithm on standard benchmark test functions. Undergraduate thesis , Sveučilište u Zagrebu, Fakultet strojarstva i brodogradnje, UNSPECIFIED. Mentor: Ćurković, Petar.

[img]
Preview
Text
Cehulic_2017_zavrsni_preddiplomski.pdf - Published Version Jezik dokumenta:Croatian

Download (3MB) | Preview
[img] Text
Čehulić_Lovro_autorska_izjava_zavrsni_2017.pdf - Published Version
Restricted to Repository staff only Jezik dokumenta:Croatian

Download (479kB)

Abstract (Croatian)

Genetski algoritam i optimizacija rojem čestica su metaheuristički optimizacijski alati. Genetski algoritam pripada području evolucijskog računarstva, a inspiriran je evolucijom na genetskoj razini. Optimizacija rojem čestica pripada području inteligencije roja i inspirirana je rojevima kukaca i jatima ptica čija je specijalizacija problem traženja. U ovome radu, ova dva optimizacijska algoritma testirana su na raznim testnim funkcijama te su dobivena dobra rješenja uz određena ograničenja. Odabrane testne funkcije su uglavnom multimodalni, višedimenzijski minimizacijski problemi. Optimizacijom rojem čestica postignuti su rezultati visoke točnosti za probleme koji su uspješno optimizirani. Genetski algoritam dao je rezultate koji nisu tako visoke točnosti, ali su njime pronađena rješenja onih funkcija za koje Optimizacija rojem čestica nije dala rezultate.

Abstract

Genetic Algorithm and Particle Swarm Optimization are Metaheuristic optimization tools. Genetic Algorithm belongs to the field of evolutionary computing, and it is inspired by evolution on genetic level. Particle Swarm Optimization belong to the field of swarm inteligence inspired by swarms of insects and flocks of birds whose specialty is search problem. In this thesis, these two optimization algorithms were tested on various (benchmark) test functions and have obtained good solutions with certain restrictions. The selected test functions are mainly multimodal, multidimensional minimization problems. Particle Swarm Optimization results have been achieved with high precision on the successfully optimize d problems. Genetic Algorithm has provided results with not as high accuracy, but it found solutions for those functions that Particle Swarm Optimization did not.

Item Type: Thesis (Undergraduate thesis)
Uncontrolled Keywords: Genetski algoritam; optimizacija rojem čestica; standardne testne funkcije.
Keywords (Croatian): Genetic Algorithm; Particle Swarm Optimization; standard (benchmark) test functions.
Subjects: TECHNICAL SCIENCE > Mechanical Engineering
Divisions: 900 Department of Robotics and Production System Automation > 920 Chair of Manufacturing and Assembly System Planning
Date Deposited: 24 Feb 2017 09:40
Last Modified: 13 Mar 2017 10:31
URI: http://repozitorij.fsb.hr/id/eprint/7435

Actions (login required)

View Item View Item

Downloads

Downloads per month over past year

Nema podataka za dohvacanje citata