Latency-Aware Routing and Spectrum Assignment with Deep Reinforcement Learning

This paper evaluates different aspects of the performance of a solution for routing and spectrum allocation in elastic optical networks. The evaluation includes the experimentation under different traffic loads and the use of a previously trained deep reinforcement learning (DRL) agent. The obtained results show that the DRL agent further outperforms a traditional algorithm as network resources become scarce. Moreover, the results show the proper operation of the pre-Trained agent when provisioning lightpaths. © 2022 IEEE.The research leading to these results has received funding from the European Union's Horizon 2020 research and innovation programme, B5G-OPEN Project, grant agreement No. 101016663, Spanish Thematic Network Go2Edge (RED2018-102585-T), FPI scholarship PRE2019-091447, and AURORAS grant RTI2018-099178-B-I00 funded by MCIN/AEI/ 10.13039/501100011033 and by ERDF A way of making Europe