Institutional Repository of Vilnius University
Edukacinių žaidimų įrankio kūrimas naudojant DKM /
Abstract
dc:descriptionThis paper analyzes the principles of educational game development by integrating psycholo- gical engagement theories (e.g., the flow state) with modern artificial intelligence technologies. To address the limitations of static content and the issue of rote memorization, this research explores the application of procedural content generation using Large Language Models (LLMs). Following a comparative analysis of technologies, the Google Gemini 3.0 Flash model was selected for the project. The study evaluated the model’s accuracy in generating mathematical content for grades 5–12 and applied prompt engineering methods (few-shot prompting and persona adoption) to en- sure factual and technical validity. Additionally, vector graphics (SVG) research was conducted: the models’ spatial perception deficiencies in zero-shot prompts were assessed, the most effective combined prompt method for direct SVG code generation was identified, and the challenges various LLMs face when writing programmatic algorithms to create drawings were analyzed. In the prac- tical component of the study, a fully functional version of the ”Memory” card game was developed using the Godot engine and GDScript within a Model-View-Controller (MVC) architectural pattern, capable of generating task sets in real-time across six different topics.
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, Augustas,
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:ELABAETD308337629&prefLang=en_US