Institutional Repository of Vilnius University
Evoliuciniai skaičiavimai ir jų taikymai /
Abstract
dc:descriptionThis Bachelor’s thesis investigates evolutionary and metaheuristic algorithms and their appli- cation in solving complex, high-dimensional optimization problems. The primary focus is on over- coming the ”curse of dimensionality” and the zero-gradient (information void) challenges. During the research, Genetic Algorithm (GA), Memetic Algorithm (MA), and Cuckoo Search (CS) algori- thms were implemented, tested, and compared by optimizing multi-component Gaussian functions across various dimensions. The results demonstrated that while the Memetic Algorithm ensures faster convergence in terms of iterations, it demands significantly more computational resources per generation than the stan- dard GA. To prevent premature convergence in high dimensions, a ”Shotgun” search strategy, uti- lizing quasi-random Sobol sequences, was successfully applied, enabling the optimization of 20D spaces. Additionally, it was found that the optimal values for CS control parameters (population size and nest abandonment probability) are directly dependent on the dimensionality of the problem.
Degree
thesis:*- Grantor dc:publisher
- Institutional Repository of Vilnius University
- Year dc:date
- 2026
Author and committee
dc:creator, dc:contributor.*- Author dc:creator
-
- Nasvytis, Justas,
Rights
dc:rights- Statement dc:rights
-
- info:eu-repo/semantics/openAccess
- Language dc:language
- lit
Identifiers
dc:identifier.*- Repository record dc:identifier
- https://repository.vu.lt/VU:ELABAETD308337662&prefLang=en_US