Primerjava evolucijskih algoritmov po vzorih fizikalnih fenomenov : magistrsko delo

Klakočer, Tadej

| | SLO | ENG | Cookies and privacy

Bigger font | Smaller font

First page > Show document

Show document

Title:	Primerjava evolucijskih algoritmov po vzorih fizikalnih fenomenov : magistrsko delo
Authors:	ID Klakočer, Tadej (Author) ID Črepinšek, Matej (Mentor) More about this mentor... ID Mernik, Marjan (Comentor)
Files:	MAG_Klakocer_Tadej_2019.pdf (2,69 MB) MD5: F1B09DEB6B95A589F1322BEE59815FD7 PID: 20.500.12556/dkum/b87e869b-22f5-4619-abba-b623f44d4c97
Language:	Slovenian
Work type:	Master's thesis/paper
Typology:	2.09 - Master's Thesis
Organization:	FERI - Faculty of Electrical Engineering and Computer Science
Abstract:	Dandanes se razvija vedno več različnih algoritmov, med katerimi so nekateri uspešnejši kot drugi. Pri primerjavi uspešnosti pa naletimo na oviro, saj ne obstaja poenoten postopek za primerjavo uspešnosti, zato je težko primerjati različne algoritme in ugotavljati, kateri je boljši. Na kratko smo povzeli klasifikacijo evolucijskih algoritmov. Vsak algoritem smo še podrobneje opisali, nekatere pojave prikazali s slikovnim materialom in zapisali psevdokode. Za primer smo vzeli algoritme fizikalnih fenomenov (EMO, GSA, CBO, LSA, CSS). Navedeni evolucijski algoritmi posnemajo fizikalne pojave v naravi, kot so električne strele, trkdelcev, gravitacija ipd. Najprej smo poustvarili rezultate, ki so navedeni v člankih, nato smo izvedli še tekmovanje med evolucijskimi algoritmi po sistemu šahovskega rangiranja v ogrodju EARS. V magistrskem delu smo tudi podrobneje opisali ogrodje in postopek, kako smo izvedli omenjeni turnir. Ocena uspešnosti temelji na tekmovanju vsakega algoritma z vsakim in za določen problem se turnir izvede tudi večkrat. Za eksperiment smo vzeli nabor matematičnih problemov iz eksperimentov CEC2014 in CEC2015. Uporabljene probleme smo na kratko opisali. Po izvedbi eksperimenta smo analizirali dobljene rezultate, npr. število porabljenih ovrednotenj, najboljša najdena rešitev idr., nato pa smo jih predstavili v obliki tabel rangov in intervala zaupanja. Ugotovili smo, da si nekateri avtorji prilagodijo testne primere, ki pa jih pri rezultatih ne omenjajo. Prav tako menimo, da bi bilo treba vzpostaviti enoten sistem ocenjevanja in primerjavo uspešnosti med algoritmi.
Keywords:	ogrodje EARS, evolucijski algoritmi, fizikalni zakoni, primerjava, šahovski sistem rangiranja
Place of publishing:	Maribor
Place of performance:	Maribor
Publisher:	[T. Klakočer]
Year of publishing:	2019
Number of pages:	VI, 79 str.
PID:	20.500.12556/DKUM-68566
UDC:	004.8.021(043.2)
COBISS.SI-ID:	22574102
NUK URN:	URN:SI:UM:DK:EHUWQX8I
Publication date in DKUM:	04.09.2019
Views:	1423
Downloads:	138
Metadata:
Categories:	KTFMB - FERI
:	Copy citation

Average score:	(0 votes)
Your score:	Voting is allowed only for logged in users.
Share:

Hover the mouse pointer over a document title to show the abstract or click on the title to get all document metadata.

Licences

License:	CC BY-SA 4.0, Creative Commons Attribution-ShareAlike 4.0 International

Link:	http://creativecommons.org/licenses/by-sa/4.0/
Description:	This Creative Commons license is very similar to the regular Attribution license, but requires the release of all derivative works under this same license.
Licensing start date:	26.09.2017

Secondary language

Language:	English
Title:	Comparison of EA inspired by physical phenomena
Abstract:	Nowadays, more and more different algorithms are being developed, some being more effective than others. When comparing the effectiveness, we are faced with an obstacle, as there is no unified procedure for the comparison of it. That is why it is difficult to compare different algorithms and determine which is better. We have shortly summarized the classification of algorithms. We explained each algorithm in more detail, presented some phenomena with photographic material, and written pseudocodes. We exemplified using algorithms of physical phenomena (EMO, GSA, CBO, LSA, CSS). These algorithms mimic physical phenomena in nature, such as lightning, particle collisions, gravity, etc. First, we recreated the results listed in the articles, and then we performed a competition between algorithms according to the chess rating system in the EARS framework. In the master's thesis, we also described in more detail the framework and the process of how we performed the aforementioned tournament. The evaluation of effectiveness is based on the competition of each algorithm with all others and in case of a specific problem, the tournament is carried out multiple times.For our experiment, we used a set of mathematical problems from the experiments CEC2014 and CEC2015. We briefly described the functions used. After the experiment was performed, the results obtained were analyzed, e.g. the number of evaluations used, the best solution found, etc., and then presented in the form of ranking tables and confidence intervals. We have found that some authors adapt the test cases, which they do not mention in the results. We also believe that a unified system of assessing and comparing algorithms should be established.
Keywords:	EARS framework, evolutionary algorithm, comparing, physical laws, chess rating system

Comments

Leave comment

You must log in to leave a comment.

Comments (0)

0 - 0 / 0

There are no comments!

Back