Graph Sequence Attention Network-Enabled Reinforcement Learning for Time-Aware Robust Routing in OSU-Based OTN (Th2A.18)
Presenter: Huangxu Ma, Beijing University of Posts & Telecommun
We propose a time-aware robust routing in OSU-based OTN through the newly designed graph sequence attention network-enabled reinforcement learning. Simulation results show > 28% OSU frame loss reduction compared to the baselines.
Authors:Huangxu Ma, Beijing University of Posts & Telecommun / Jiawei Zhang, Beijing University of Posts & Telecommun / Yuefeng Ji, Beijing University of Posts & Telecommun