Reinforcement-Learning based routing for packet-optical networks with hybrid telemetry
Fuente:
arXiv
Gespeichert in:
| Hauptverfasser: | , , , , , |
|---|---|
| Format: | Preprint |
| Veröffentlicht: |
2024
|
| Schlagworte: | |
| Online-Zugang: | |
| Tags: |
Tag hinzufügen
Keine Tags, Fügen Sie den ersten Tag hinzu!
|
| _version_ | 1866910496792248320 |
|---|---|
| author | Navarro, A. L. García Koneva, Nataliia Sánchez-Macián, Alfonso Hernández, José Alberto de Dios, Óscar González Rivas-Moscoso, J. M. |
| author_facet | Navarro, A. L. García Koneva, Nataliia Sánchez-Macián, Alfonso Hernández, José Alberto de Dios, Óscar González Rivas-Moscoso, J. M. |
| contents | This article provides a methodology and open-source implementation of Reinforcement Learning algorithms for finding optimal routes in a packet-optical network scenario. The algorithm uses measurements provided by the physical layer (pre-FEC bit error rate and propagation delay) and the link layer (link load) to configure a set of latency-based rewards and penalties based on such measurements. Then, the algorithm executes Q-learning based on this set of rewards for finding the optimal routing strategies. It is further shown that the algorithm dynamically adapts to changing network conditions by re-calculating optimal policies upon either link load changes or link degradation as measured by pre-FEC BER. |
| format | Preprint |
| id |
arxiv_https___arxiv_org_abs_2406_12602 |
| institution | arXiv |
| publishDate | 2024 |
| record_format | arxiv |
| spellingShingle | Reinforcement-Learning based routing for packet-optical networks with hybrid telemetry Navarro, A. L. García Koneva, Nataliia Sánchez-Macián, Alfonso Hernández, José Alberto de Dios, Óscar González Rivas-Moscoso, J. M. Networking and Internet Architecture Machine Learning This article provides a methodology and open-source implementation of Reinforcement Learning algorithms for finding optimal routes in a packet-optical network scenario. The algorithm uses measurements provided by the physical layer (pre-FEC bit error rate and propagation delay) and the link layer (link load) to configure a set of latency-based rewards and penalties based on such measurements. Then, the algorithm executes Q-learning based on this set of rewards for finding the optimal routing strategies. It is further shown that the algorithm dynamically adapts to changing network conditions by re-calculating optimal policies upon either link load changes or link degradation as measured by pre-FEC BER. |
| title | Reinforcement-Learning based routing for packet-optical networks with hybrid telemetry |
| topic | Networking and Internet Architecture Machine Learning |
| url | https://arxiv.org/abs/2406.12602 |