University of Twente Student Theses
Decision Support for Traffic Management using Reinforcement Learning
Heijnen, Alex (2024) Decision Support for Traffic Management using Reinforcement Learning.
|
PDF
4MB |
| Abstract: | This research investigates the use of reinforcement learning for traffic optimization of a network of connected traffic lights which should be suitable as decision support for traffic engineers. A supervised learning approach is also done for evaluating the prediction accuracy of traffic done by a neural network. For the RL approach a two-layer traffic control system is advocated for, combining max pressure for local optimization and perimeter control for global optimization. Unfortunately, the results for the RL approach are not promising, having poor performance and taking a long time. In theory the solution approach should work well, and future research should try to improve upon it. Supervised learning is found to be a good alternative to give additional insights for traffic engineers. |
| Item Type: | Essay (Master) |
| Clients: | Technolution, Gouda, The Netherlands |
| Faculty: | BMS: Behavioural, Management and Social Sciences |
| Subject: | 54 computer science, 55 traffic technology, transport technology |
| Programme: | Industrial Engineering and Management MSc (60029) |
| Link to this item: | https://purl.utwente.nl/essays/105224 |
| 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