Distributed Lightpath Establishment based on Physical Impairments in Large-Scale WDM Networks

点击翻译按钮获取中文摘要

In this research, in large-scale WDM networks, we proposed a dynamic lightpath establishment method based on physical layer impairments. This method utilizes reinforcement learning technique. We first defined a state set, an action set, and reward function for the reinforcement learning in order to improve the physical impairments. Then, we proposed a method that can improve fairness and provide the service differentiation based on physical layer impairments. Moreover, we reduced the number of states and actions that are used for the reinforcement learning for the easy implementation. Finally, we considered the interaction between the proposed method and GMPLS protocol.