IntelliFiber: Optical Network Optimization through Machine Learning Algorithms

The rapid evolution of the Internet and other communication systems has led to an explosion of available services and technology, making communication networks increasingly complex. The continued success of the communication network relies heavily on integrating Artificial Intelligence (AI) and Machine Learning (ML) in various aspects. While optical networks play a vital role in both the core and access networks of communication systems, they also confront significant difficulties due to the complexity of their systems and the need for human intervention in their daily operations. Deploying additional intelligence capabilities to enable autonomous and flexible network operations is crucial to overcoming current restrictions and addressing the difficulties of future optical networks [1]. Many modern optical network applications use ML approaches because of their demonstrated superiority in addressing complicated challenges. The current applications of Machine Learning (ML) in optical networks are thoroughly discussed in this paper. ML's usages are broken down into two groups, [3] [4] [5]: network management (control and resource management) and resilience (monitoring and detection) in optical networks. Different ML methods are evaluated and contrasted across use scenarios. In addition, a primer on ML applications is offered, covering such topics as the fundamentals of popular ML algorithms, ML paradigms, and the reasons for using ML. Finally, the challenges and potential remedies of Machine Learning use in optical networks are examined to stimulate further research and development.