Intelligent Agent-Based Dynamic Reliability Evaluation for Optical Networks: A Comprehensive Framework and Case Study

Reliability holds a central position in optical network management research. Continuous advancements in optical transmission and signal processing technologies have significantly increased the transmission capacity and coverage of optical networks. Numerous factors, including component defects, aging, environmental conditions, human factors, and power supply, influence the reliable operation of optical networks. Consequently, accurately evaluating the reliability of optical networks poses a significant challenge. However, current reliability evaluation methods inadequately consider optical networks' dynamic, time-dependent, and function-dependent characteristics. In this article, a reliability evaluation framework is proposed based on a five-level agent hierarchy. This frame-work describes the operational, state transition, and data feedback mechanisms. The framework is capable of evaluating the reliability of optical networks under multi-factor and multi-state conditions. The framework is applied to optical and power interconnection networks with different topologies and three types of interconnections. Reliability and availability metrics validate the framework's effectiveness and scalability. This article offers a valuable reference for optical network management and provides helpful guidance for designing more robust optical network systems in the future.