University of Twente Student Theses
Exploring Performance of Reinforcement Learning compared to other Heuristics in Low Predictability Environments : Applications in Digital Finance
Weisfelt, Mette (2025) Exploring Performance of Reinforcement Learning compared to other Heuristics in Low Predictability Environments : Applications in Digital Finance.
|
PDF
3MB |
| Abstract: | Reinforcement Learning has become increasingly popular in addressing issues in financial environments. A key strength of Reinforcement Learning lies in its sequential decision-making capability, where each choice influences future outcomes. This mirrors real-world financial markets, such as long-term investments. This research paper explores the performance of Reinforcement Learning by playing two games with different complexities. The games provide multiple controlled environments where agents must make decisions that are difficult to predict, simulating the challenges faced in real-world financial contexts. Results show that the complexity and predictability of generators significantly influence model performance. Simpler heuristic algorithms outperform Inferring Models in simple, predictable environments. The ability to ignore noise and focus on fundamental patterns often outweighs attempts to learn nonexistent patterns. Reward signal adjustments and hyperparameter tuning can increase performance of the Reinforcement Learning models. |
| Item Type: | Essay (Bachelor) |
| Faculty: | EEMCS: Electrical Engineering, Mathematics and Computer Science |
| Subject: | 54 computer science, 83 economics |
| Programme: | Computer Science BSc (56964) |
| Link to this item: | https://purl.utwente.nl/essays/105140 |
| 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