1. Uporaba genetskega programiranja za napoved cen kriptovalut : magistrsko deloTilen Heric, 2025, magistrsko delo Opis: Odkar so se pojavile kriptovalute in hitro za tem občutek hitrega zaslužka, se je rodilo zanimanje za napoved njihovih cen. Zaradi nepredvidljivosti in visoke volatilnosti gibanja cen tradicionalne metode ne zadoščajo pri ustvarjanju zanesljivih napovedi. Namen magistrskega dela je razvoj napovednega modela, ki bo znal samostojno napovedati ceno kriptovalute naslednjega dne. Cilj napovednega modela je čim natančnejša napoved za več kriptovalut. Model temelji na genetskem programiranju z uporabo simbolične regresije. Vhodni podatki algoritma so zgodovinski podatki in tehnični kazalniki. Z metodo iskanja po mreži smo optimizirali parametre genetskega algoritma. Rezultati napovedovanja kažejo, da je genetsko programiranje učinkovito pri napovedovanju cen kriptovalut, zato smo preizkusili, kako se obnese pri energentih, saj so ti nekoliko manj volatilni. Pri energentih se je model obnesel še bolj učinkovito. Zaključimo lahko, da je naša metoda primerno orodje za obravnavo problemov z visoko volatilnostjo in kompleksnostjo. Naloga je primer praktične uporabe naprednih evolucijskih algoritmov za reševanje realnih problemov in ponuja osnovo za nadaljnje raziskave na področju napovedovanja časovnih vrst.
Ključne besede: genetsko programiranje, napoved cen, kriptovalute, energenti
Objavljeno v DKUM: 15.10.2025; Ogledov: 0; Prenosov: 19
 Celotno besedilo (4,55 MB) |
2. Razvoj preproste večigralske bojne arene in umetne inteligence za pametne nasprotnike : diplomsko deloNino Franci, 2025, diplomsko delo Opis: V diplomskem delu sta predstavljena zasnova in razvoj preproste večigralske bojne arene z vgrajeno umetno inteligenco za nadzor nasprotnikov. Igra, razvita v igralnem pogonu Unity, vključuje agente, ki uporabljajo vedenjska drevesa, razvita z uporabo evolucijskih algoritmov. Sistem omogoča zavzemanje planetov, zbiranje virov in medsebojne boje. Rezultati testiranja so pokazali, da agenti uspešno razvijejo kompleksne taktike in učinkovito sodelujejo v različnih simulacijskih scenarijih. Delo predstavlja osnovo za nadaljnje raziskave umetne inteligence v simulacijskih okoljih in možnosti za nadaljnji razvoj in razširitev igre.
Ključne besede: Unity, umetna inteligenca, vedenjska drevesa, evolucijski algoritmi
Objavljeno v DKUM: 15.10.2025; Ogledov: 0; Prenosov: 9
Celotno besedilo (1,98 MB) |
3. Performance comparison of single-objective evolutionary algorithms implemented in different frameworksMiha Ravber, Marko Šmid, Matej Moravec, Marjan Mernik, Matej Črepinšek, 2025, izvirni znanstveni članek Opis: Fair comparison with state-of-the-art evolutionary algorithms is crucial, but is obstructed by differences in problems, parameters, and stopping criteria across studies. Metaheuristic frameworks can help, but often lack clarity on algorithm versions, improvements, or deviations. Some also restrict parameter configuration. We analysed source codes and identified inconsistencies between implementations. Performance comparisons across frameworks, even with identical settings, revealed significant differences, sometimes even with the authors’ own code. This questions the validity of comparisons using such frameworks. We provide guidelines to improve open-source metaheuristics, aiming to support more credible and reliable comparative studies.
Ključne besede: metaheuristics, evolutionary algorithm, metaheuristic optimization framework, algorithm comparison, benchmarking
Objavljeno v DKUM: 02.10.2025; Ogledov: 0; Prenosov: 4
Celotno besedilo (3,89 MB) |
4. Razvoj multiplatformne aplikacije za izbiro video vsebin s pomočjo ogrodja React Expo : diplomsko deloDejan Rojko, 2025, diplomsko delo Opis: V današnjem času, ko so uporabniki vsakodnevno izpostavljeni množici filmov in serij na različnih platformah, je iskanje primernega videa pogosto naporno in dolgotrajno. Namen diplomskega dela je bil razviti multiplatformno mobilno aplikacijo, ki omogoča enostavno in hitro izbiro filmov in serij s pomočjo ogrodja React Expo. Aplikacija je zasnovana na principu interakcije "swipe", kar omogoča hitro izbiro filmov na podlagi preteklih odločitev.
Razvoj aplikacije je temeljil na ogrodju React, ki omogoča hitro in učinkovito gradnjo uporabniških vmesnikov, in platformi Expo, ki poenostavi razvoj in distribucijo aplikacij za več platform hkrati (iOS, Android). Kot zaledna rešitev je bila uporabljena platforma Firebase za avtentikacijo uporabnikov in shranjevanje podatkov, medtem ko je bila končna točka TMDB uporabljena za pridobivanje filmov in serij ter njihovo filtriranje.
Aplikacija omogoča personalizirane predloge filmov na podlagi uporabniških preferenc in omogoča shranjevanje izbranih vsebin za kasnejši ogled. Poleg tega vključuje funkcionalnosti, kot so registracija, prijava, filtriranje filmov, ogled podrobnosti o filmih in spreminjanje nastavitev uporabniškega računa. Ta projekt se je soočal z različnimi tehnološkimi izzivi, kot sta integracija Firebase v multiplatformnem okolju in implementacija uporabniškega vmesnika, ki bi zagotavljal enostavno in prijetno uporabniško izkušnjo.
Ključne besede: multiplatformna mobilna aplikacija, React Expo, filtriranje filmov, uporabniška izkušnja, TMDB API.
Objavljeno v DKUM: 23.09.2025; Ogledov: 0; Prenosov: 29
Celotno besedilo (2,61 MB) |
5. Overcoming stagnation in metaheuristic algorithms with MsMA’s adaptive meta-level partitioningMatej Črepinšek, Marjan Mernik, Miloš Beković, Matej Pintarič, Matej Moravec, Miha Ravber, 2025, izvirni znanstveni članek Opis: Stagnation remains a persistent challenge in optimization with metaheuristic algorithms (MAs), often leading to premature convergence and inefficient use of the remaining evaluation budget. This study introduces , a novel meta-level strategy that externally monitors MAs to detect stagnation and adaptively partitions computational resources. When stagnation occurs, divides the optimization run into partitions, restarting the MA for each partition with function evaluations guided by solution history, enhancing efficiency without modifying the MA’s internal logic, unlike algorithm-specific stagnation controls. The experimental results on the CEC’24 benchmark suite, which includes 29 diverse test functions, and on a real-world Load Flow Analysis (LFA) optimization problem demonstrate that MsMA consistently enhances the performance of all tested algorithms. In particular, Self-Adapting Differential Evolution (jDE), Manta Ray Foraging Optimization (MRFO), and the Coral Reefs Optimization Algorithm (CRO) showed significant improvements when paired with MsMA. Although MRFO originally performed poorly on the CEC’24 suite, it achieved the best performance on the LFA problem when used with MsMA. Additionally, the combination of MsMA with Long-Term Memory Assistance (LTMA), a lookup-based approach that eliminates redundant evaluations, resulted in further performance gains and highlighted the potential of layered meta-strategies. This meta-level strategy pairing provides a versatile foundation for the development of stagnation-aware optimization techniques.
Ključne besede: optimization, metaheuristics, stagnation, meta-level strategy, algorithmic performance, duplicate solutions
Objavljeno v DKUM: 30.05.2025; Ogledov: 0; Prenosov: 8
Celotno besedilo (1,47 MB) |
6. Tackling blind spot challenges in metaheuristics algorithms through exploration and exploitationMatej Črepinšek, Miha Ravber, Luka Mernik, Marjan Mernik, 2025, izvirni znanstveni članek Opis: This paper defines blind spots in continuous optimization problems as global optima that are inherently difficult to locate due to deceptive, misleading, or barren regions in the fitness landscape. Such regions can mislead the search process, trap metaheuristic algorithms (MAs) in local optima, or hide global optima in isolated regions, making effective exploration particularly challenging. To address the issue of premature convergence caused by blind spots, we propose LTMA+ (Long-Term Memory Assistance Plus), a novel meta-approach that enhances the search capabilities of MAs. LTMA+ extends the original Long-Term Memory Assistance (LTMA) by introducing strategies for handling duplicate evaluations, shifting the search away from over-exploited regions and dynamically toward unexplored areas and thereby improving global search efficiency and robustness. We introduce the Blind Spot benchmark, a specialized test suite designed to expose weaknesses in exploration by embedding global optima within deceptive fitness landscapes. To validate LTMA+, we benchmark it against a diverse set of MAs selected from the EARS framework, chosen for their different exploration mechanisms and relevance to continuous optimization problems. The tested MAs include ABC, LSHADE, jDElscop, and the more recent GAOA and MRFO. The experimental results show that LTMA+ improves the success rates for all the tested MAs on the Blind Spot benchmark statistically significantly, enhances solution accuracy, and accelerates convergence to the global optima compared to standard MAs with and without LTMA. Furthermore, evaluations on standard benchmarks without blind spots, such as CEC’15 and the soil model problem, confirm that LTMA+ maintains strong optimization performance without introducing significant computational overhead.
Ključne besede: optimization, metaheuristics algorithm, algorithmic performance, duplicate solutions, nonrevisited solutions, blind spots, LTMA
Objavljeno v DKUM: 19.05.2025; Ogledov: 0; Prenosov: 4
Celotno besedilo (4,16 MB) |
7. Leveraging grammarware for active video game developmentMatej Črepinšek, Tomaž Kosar, Matej Moravec, Miha Ravber, Marjan Mernik, 2025, izvirni znanstveni članek Opis: This paper presents a grammarware-based approach to developing active video games (AVGs) for sensor-driven training systems. The GCGame domain-specific language (DSL) is introduced to define game logic, sensor interactions, and timing behavior formally. This approach ensures cross-platform consistency, supports real-time configurability, and simplifies the integration of optimization and visualization tools. The presented system, called GCBLE, serves as a case study, demonstrating how grammarware enhances modularity, maintainability, and adaptability in real-world physical interaction applications. The results highlight the potential of a DSL-driven design to bridge the gap between developers and domain experts in embedded interactive systems
Ključne besede: active video games, grammarware, internet of things, DSL, procedural level generation, evolutionary computation, game controllers
Objavljeno v DKUM: 23.04.2025; Ogledov: 0; Prenosov: 7
Celotno besedilo (4,32 MB) |
8. Maximum number of generations as a stopping criterion considered harmfulMiha Ravber, Shih-Hsi Liu, Marjan Mernik, Matej Črepinšek, 2022, izvirni znanstveni članek Opis: Evolutionary algorithms have been shown to be very effective in solving complex optimization problems. This has driven the research community in the development of novel, even more efficient evolutionary algorithms. The newly proposed algorithms need to be evaluated and compared with existing state-of-the-art algorithms, usually by employing benchmarks. However, comparing evolutionary algorithms is a complicated task, which involves many factors that must be considered to ensure a fair and unbiased comparison. In this paper, we focus on the impact of stopping criteria in the comparison process. Their job is to stop the algorithms in such a way that each algorithm has a fair opportunity to solve the problem. Although they are not given much attention, they play a vital role in the comparison process. In the paper, we compared different stopping criteria with different settings, to show their impact on the comparison results. The results show that stopping criteria play a vital role in the comparison, as they can produce statistically significant differences in the rankings of evolutionary algorithms. The experiments have shown that in one case an algorithm consumed 50 times more evaluations in a single generation, giving it a considerable advantage when max gen was used as the stopping criterion, which puts the validity of most published work in question.
Ključne besede: evolutionary algorithms, stopping criteria, benchmarking, algorithm termination, algorithm comparison
Objavljeno v DKUM: 28.03.2025; Ogledov: 0; Prenosov: 10
Celotno besedilo (1,40 MB)
Gradivo ima več datotek! Več... |
9. Razvoj aplikacije ios za igrifikacijo treninga z boksarsko vrečo : zaključno deloFilip Božić, 2024, diplomsko delo Opis: V diplomski nalogi smo predstavili razvoj mobilne aplikacije za igrifikacijo treninga z boksarsko vrečo na operacijskem sistemu iOS. Raziskali smo pomen in načela koncepta igrifikacije ter kako jih aplikacija uporablja za spodbujanje uporabnikov. Iz tehničnega vidika smo preučili in opisali nativni razvoj iOS aplikacij, Swift programski jezik kot sedanjost in prihodnost razvoja iOS aplikacij ter SwiftUI razvojno ogrodje za razvoj uporabniškega vmesnika. Rezultat diplomske naloge je mobilna aplikacija, ki spodbuja uporabnike in omogoča spremljanje uspešnosti treniranja z boksarsko vrečo.
Ključne besede: iOS, Swift, SwiftUI, mobilna aplikacija, igrifikacija, trening boksa
Objavljeno v DKUM: 14.10.2024; Ogledov: 0; Prenosov: 30
Celotno besedilo (4,16 MB) |
10. |
