Inter-Core Crosstalk Aware Deep Reinforcement Learning Based Resource Allocation in Multicore Elastic Optical Networks

We present a deep reinforcement learning (DRL) framework for the routing, modulation, core selection, and spectrum allocation (RMCSA) of the multicore elastic optical networks. Specifically, the DRL agent senses the impact of the inter-core crosstalk and makes the RMCSA decisions accordingly. Simulation results show that our method outperforms previous methods.