The growing demand for personalized communication capabilities is driving transport networks to evolve towards autonomous management focused on user needs. The article provides a thorough analysis of the concept of autonomous transport networks through the implementation of artificial intelligence, which allows automating processes at different levels of the network infrastructure. The scenarios for the use of AI in IP-over-DWDM transport systems are considered for tasks such as traffic forecasting, ensuring high data transmission quality, anomaly detection, optimization of network resources, and proactive failure management. Particular attention is paid to the role of artificial intelligence in improving key aspects of network operation. The central part of the study is the proposed management architecture built on open and standard SDN APIs. It allows for efficient distribution of the transport network for multi-tier systems and provides access to normalized data in real time, which becomes the basis for autonomous operation. The integration of artificial intelligence helps optimize resource use, improve service quality, reduce downtime and reduce operating costs, while ensuring high scalability of the network infrastructure. The use of machine and deep learning methods makes it possible to implement adaptive network management, taking into account the changing level of loads and uncertain events. This approach opens up the prospect of creating self-managed, self-healing and self-learning networks that can adapt to conditions without human intervention. In conclusion, the article emphasizes that the future of transport systems is closely related to the full integration of artificial intelligence, standardized platforms and open ecosystems. This ensures the sustainable development of telecommunications infrastructure in the face of growing demands for productivity and reliability. The use of AI allows you to reduce costs, improve the quality of service and adapt networks to the constantly increasing demands of consumers.
A. Kondrus (Wed,) studied this question.