Deep RMCSA for Resource Allocation in Spectrally-Spatially Flexible Optical Networks

Abstract

A gradual transition from traditional fixed frequency networks towards Spectrally-Spatially Flexible Optical Networks (SS-FONs) will ensure that networks continue to meet increasing Internet bandwidth demands. The DeepRMCSA algorithm, proposed in this paper, uses deep reinforcement learning to determine the optimal policies for solving the Routing, Modulation, Core and Spectrum Assignment problem in SS-FONs. We evaluate the performance of our algorithm by comparing it with other approaches used in the literature.