Optimizacijski algoritem na osnovi iskanja hrane bakterij

diplomsko delo

Žiga Stupan (Author), Iztok Fister (Mentor), Iztok Fister (Co-mentor)

Abstract

V diplomskem delu smo raziskali področje optimizacije in optimizacijskih algoritmov po vzorih iz narave. Opisali smo optimizacijski algoritem na osnovi iskanja hrane bakterij, njegove biološke osnove, modifikacije in aplikacije. V empiričnem delu smo osnovno različico algoritma implementirali v programskem jeziku Python, kot razširitev ogrodja za razvoj in preizkušanje algoritmov po vzorih iz narave NiaPy. Implementiran algoritem smo preizkusili na desetih popularnih testnih funkcijah različnih dimenzij (10, 20 in 30). Rezultate smo nato primerjali z rezultati treh popularnih algoritmov po vzorih iz narave (DE, PSO in BA). Ugotovili smo, da osnovni algoritem BFO v večini primerov močno zaostaja za prej omenjenimi algoritmi v kakovosti najdenih rešitev.

Keywords

algoritmi po vzorih iz narave;inteligenca rojev;optimizacija na osnovi iskanja hrane bakterij;diplomske naloge;

Data

Language:	Slovenian
Year of publishing:	2021
Typology:	2.11 - Undergraduate Thesis
Organization:	UM FERI - Faculty of Electrical Engineering and Computer Science
Publisher:	[Ž. Stupan]
UDC:	004.421:004.8(043.2)
COBISS:	94976003
Views:	273
Downloads:	30
Average score:	0 (0 votes)
Metadata:

Other data

Secondary language:	English
Secondary title:	Bacterial foraging optimization algorithm
Secondary abstract:	In this thesis, we explored the area of optimization and nature-inspired algorithms. We described the bacterial foraging optimization algorithm, it's biological foundations, modifications and applications. In the empirical part of the thesis, we implemented the basic version of the algorithm in Python, as an extension of the NiaPy microframework for designing and testing nature-inspired algorithms. We then tested the algorithm on 10 popular benchmark test functions in different dimensions (10, 20, and 30) and compared the results with those obtained by three other popular nature-inspired algorithms (DE, PSO and BA). The results have shown that, in most cases, the classic BFO algorithm gets severely outclassed by all the aforementioned algorithms in terms of solution quality.
Secondary keywords:	nature-inspired algorithms;swarm intelligence;bacterial foraging optimization;
Type (COBISS):	Bachelor thesis/paper
Thesis comment:	Univ. v Mariboru, Fak. za elektrotehniko, računalništvo in informatiko, Računalništvo in informacijske tehnologije
Pages:	XVI, 43 str.
ID:	13284074