A Routing and Spectrum Assignment Algorithm for Electric Power Elastic Optical Networks Based on Deep Reinforcement Learning

In order to improve the spectrum utilization of electric power elastic optical networks and reduce the training time of routing and spectrum allocation processes, the paper proposes a deep reinforcement learning based electric power elastic optical network routing and spectrum allocation algorithm. This algorithm establishes working and protection paths for high priority services such as power control during the routing stage, for reliable transmission of services, and uses comprehensive network information to represent the state during the spectrum allocation stage, and introduce heuristic strategies into the reward function. The simulation results show that the routing and spectrum allocation algorithm can get higher rewards, and after appropriate training, it can achieve better routing and spectrum allocation schemes among nodes, thereby improving the service carrying capacity of the electric power elastic optical network, reducing packet loss rate, and achieving reliable transmission of high priority services.