The application of reinforcement learning to NATM tunnel design
The New Austrian Tunnelling Method (NATM) tunnel design is performed by testing support classes against the geological profile. We propose to replace this manual process with reinforcement learning, a generic framework within the realm of artificial intelligence that solves control tasks. Previous s...
| Main Authors: | , , |
|---|---|
| Format: | Article |
| Language: | English |
| Published: |
KeAi Communications Co., Ltd.
2022-12-01
|
| Series: | Underground Space |
| Subjects: | |
| Online Access: | http://www.sciencedirect.com/science/article/pii/S2467967422000356 |
| _version_ | 1797756972817186816 |
|---|---|
| author | Enrico Soranzo Carlotta Guardiani Wei Wu |
| author_facet | Enrico Soranzo Carlotta Guardiani Wei Wu |
| author_sort | Enrico Soranzo |
| collection | DOAJ |
| description | The New Austrian Tunnelling Method (NATM) tunnel design is performed by testing support classes against the geological profile. We propose to replace this manual process with reinforcement learning, a generic framework within the realm of artificial intelligence that solves control tasks. Previous studies have demonstrated this possibility, albeit with methodological simplifications. We coupled the Finite Difference Method with a Python script, used the output of the first to train the machine learning model implemented in the latter and improved the choice of the support classes. Through benchmark tests, we demonstrated that our method was capable of choosing the optimal support classes for various geological sets and showed the relation between its performance and the number of training episodes. |
| first_indexed | 2024-03-12T18:08:17Z |
| format | Article |
| id | doaj.art-4fea234ee65f4abcb42d036136b08348 |
| institution | Directory Open Access Journal |
| issn | 2467-9674 |
| language | English |
| last_indexed | 2024-03-12T18:08:17Z |
| publishDate | 2022-12-01 |
| publisher | KeAi Communications Co., Ltd. |
| record_format | Article |
| series | Underground Space |
| spelling | doaj.art-4fea234ee65f4abcb42d036136b083482023-08-02T09:18:27ZengKeAi Communications Co., Ltd.Underground Space2467-96742022-12-01769901002The application of reinforcement learning to NATM tunnel designEnrico Soranzo0Carlotta Guardiani1Wei Wu2Corresponding author.; University of Natural Resources and Life Sciences, Feistmantelstraße 4, Vienna 1180, AustriaUniversity of Natural Resources and Life Sciences, Feistmantelstraße 4, Vienna 1180, AustriaUniversity of Natural Resources and Life Sciences, Feistmantelstraße 4, Vienna 1180, AustriaThe New Austrian Tunnelling Method (NATM) tunnel design is performed by testing support classes against the geological profile. We propose to replace this manual process with reinforcement learning, a generic framework within the realm of artificial intelligence that solves control tasks. Previous studies have demonstrated this possibility, albeit with methodological simplifications. We coupled the Finite Difference Method with a Python script, used the output of the first to train the machine learning model implemented in the latter and improved the choice of the support classes. Through benchmark tests, we demonstrated that our method was capable of choosing the optimal support classes for various geological sets and showed the relation between its performance and the number of training episodes.http://www.sciencedirect.com/science/article/pii/S2467967422000356Deep Q-NetworkNATMReinforcement learningSupport classesTunnelling |
| spellingShingle | Enrico Soranzo Carlotta Guardiani Wei Wu The application of reinforcement learning to NATM tunnel design Underground Space Deep Q-Network NATM Reinforcement learning Support classes Tunnelling |
| title | The application of reinforcement learning to NATM tunnel design |
| title_full | The application of reinforcement learning to NATM tunnel design |
| title_fullStr | The application of reinforcement learning to NATM tunnel design |
| title_full_unstemmed | The application of reinforcement learning to NATM tunnel design |
| title_short | The application of reinforcement learning to NATM tunnel design |
| title_sort | application of reinforcement learning to natm tunnel design |
| topic | Deep Q-Network NATM Reinforcement learning Support classes Tunnelling |
| url | http://www.sciencedirect.com/science/article/pii/S2467967422000356 |
| work_keys_str_mv | AT enricosoranzo theapplicationofreinforcementlearningtonatmtunneldesign AT carlottaguardiani theapplicationofreinforcementlearningtonatmtunneldesign AT weiwu theapplicationofreinforcementlearningtonatmtunneldesign AT enricosoranzo applicationofreinforcementlearningtonatmtunneldesign AT carlottaguardiani applicationofreinforcementlearningtonatmtunneldesign AT weiwu applicationofreinforcementlearningtonatmtunneldesign |