University of Twente Student Theses
As of Friday, 8 August 2025, the current Student Theses repository is no longer available for thesis uploads. A new Student Theses repository will be available starting Friday, 15 August 2025.
Genetic Algorithms for Controller Generation of Understandable Bomberman Controllers
Wijma, Tim (2025) Genetic Algorithms for Controller Generation of Understandable Bomberman Controllers.
|
PDF
4MB |
| Abstract: | This thesis explores the use of genetic algorithms (GAs) to automatically generate understandable agent controllers for the game Bomberman using guarded command programs (GCPs). This approach evolves agents represented by a fixed-size set of human-readable, condition-action rules within the Pommerman environment. Experiments analyzing various mutation and crossover strategies confirmed that the GCP structure is effective at producing interpretable agent logic. The results demonstrate that the GA successfully evolved agents capable of basic survival tactics, primarily by learning to avoid self-destructive behaviors. However, the key finding of this research is that agent performance was fundamentally constrained by the fitness function, which failed to adequately reward complex strategic and tactical maneuvers. This work concludes that while GAs combined with GCPs provide a viable framework for generating understandable AI, achieving high-level strategic competence is critically dependent on the design of a more sophisticated fitness function that can guide evolution beyond simple behaviors. |
| Item Type: | Essay (Bachelor) |
| Faculty: | EEMCS: Electrical Engineering, Mathematics and Computer Science |
| Subject: | 54 computer science |
| Programme: | Computer Science BSc (56964) |
| Link to this item: | https://purl.utwente.nl/essays/107351 |
| Export this item as: | BibTeX EndNote HTML Citation Reference Manager |
Show download statistics for this publication
Show download statistics for this publicationRepository Staff Only: item control page