基于机器学习的光网络路由与频谱分配机制研究

对于光网络而言,路由与频谱分配是制约光层资源利用率和光网络容量的重要问题之一。机器学习迅速发展,为光网络管理与控制的智能化、自动化提供了新的发展方向。本文回顾了近年来基于机器学习的路由频谱分配相关研究,介绍了光网络控制中常见的机器学习算法,描述了基于神经网络和强化学习的路由频谱分配机制,最后分析了当前研究在泛化性和可靠性等方面的潜在挑战。

Research on RSA Mechanism of Optical Networks based on Machine Learning

For optical networks, routing and spectrum allocation are one of the most important problems that restrict the utilization of optical layer resources and the capacity of optical networks. The rapid development of Machine Learning(ML) provides a new direction for the intelligent and automatic management and control of optical network. This paper reviewed the recent research on ML-based routing spectrum allocation, introduced the common ML algorithms in optical network control, described the routing spectrum allocation mechanism based on neural network and Reinforcement Learning, and finally analyzed the potential challenges in generalization and reliability of current research.