MAGC-RSA: Multi-Agent Graph Convolutional Reinforcement Learning for Distributed Routing and Spectrum Assignment in Elastic Optical Networks

This paper proposes MAGC-RSA, a Multi-Agent Graph Convolutional Reinforcement Learning approach, to solve the Routing and Spectrum Assignment (RSA) problem in a distributed manner. A blocking probability reduction of 80% can be achieved compared to the Shortest Path First-Fit approach.