Reinforcement-Learning based routing for packet-optical networks with hybrid telemetry
Networking and Internet Architecture
2024-06-24 v2 Machine Learning
Abstract
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.
Cite
@article{arxiv.2406.12602,
title = {Reinforcement-Learning based routing for packet-optical networks with hybrid telemetry},
author = {A. L. García Navarro and Nataliia Koneva and Alfonso Sánchez-Macián and José Alberto Hernández and Óscar González de Dios and J. M. Rivas-Moscoso},
journal= {arXiv preprint arXiv:2406.12602},
year = {2024}
}