B5G-OPEN @ OFC2022 #2
“Evaluation of Deep Reinforcement Learning for Restoration in Optical Networks”
By C. Hern´andez-Chulde, R. Casellas, R. Mart´ınez, R. Vilalta, R. Munoz – CTTC.
Abstract: A deep reinforcement learning-based agent is presented to perform autonomous lightpath restoration upon a link failure event. The agent is evaluated against other heuristic algorithms under different traffic load and failure duration scenarios.
