diplomsko delo
Abstract
Uporaba strojnega učenja v računalniških igrah postaja vse bolj pogosta za razvoj vedenja inteligentih agentov. Najpogostejši pristop k problemu je uporaba spodbujevanega učenja, ki se je že večkrat izkazalo za učinkovito pri iskanju robustnih rešitev. V diplomski nalogi smo, kot alternativno rešitev, uporabili genetske algoritme, ki so kljub njihovi enostavnosti le redko uporabljeni za razvoj vedenja inteligentnih agentov. Učinkovitost implementacije smo primerjali s splošno razširjeno rešitvijo ML-Agents, ki je osnovana na spodbujevanem učenju. Primerjava med pristopoma je bila izvedena na dveh popularnih igrah, pod primerljivimi pogoji. Naši rezultati nakazujejo, da je uporaba genetskih algoritmov smiselna za enostavnejše scenarije, medtem ko se v bolj kompleksnih primerih, ko je za reševanje danega problema zahtevano kompleksnejše vedenje, naša rešitev ni obnesla najbolje.
Keywords
genetski algoritmi;spodbujevano učenje;računalniške igre;računalništvo in informatika;univerzitetni študij;diplomske naloge;
Data
Language: |
Slovenian |
Year of publishing: |
2019 |
Typology: |
2.11 - Undergraduate Thesis |
Organization: |
UL FRI - Faculty of Computer and Information Science |
Publisher: |
[J. Bevc] |
UDC: |
004.8(043.2) |
COBISS: |
1538316483
|
Views: |
687 |
Downloads: |
183 |
Average score: |
0 (0 votes) |
Metadata: |
|
Other data
Secondary language: |
English |
Secondary title: |
Use of genetic algorithms for development of intelligent agents in games |
Secondary abstract: |
Machine learning techniques are already commonly applied for developing the behaviour of intelligent agents in video games. Most commonly the development of agents is executed via reinforced learning, a relatively simple approach, capable of producing robust solutions to various learning challenges. In the presented thesis we tested whether genetic algorithms could be a viable alternative to reinforced learning. Even though genetic algorithms are very simple and easy to implement they have not seen much use when it comes to development of intelligent agents. To compare the quality of our genetic algorithms based solution, we compared it with ML-Agents, a widespread framework for development of intelligent agents, based on reinforced learning. The comparison of both learning methods. was executed on two popular games under comparable conditions. Our results suggest that genetic algorithms could represent a viable alternative to reinforced learning, but only in simple scenarios. When applied to more complex scenarios, our implementation of genetic results did not fare extremely well. |
Secondary keywords: |
genetic algorithms;reinforcement learning;computer games;computer and information science;diploma; |
Type (COBISS): |
Bachelor thesis/paper |
Study programme: |
1000468 |
Embargo end date (OpenAIRE): |
1970-01-01 |
Thesis comment: |
Univ. v Ljubljani, Fak. za računalništvo in informatiko |
Pages: |
24 str. |
ID: |
11216349 |