Generative AI-Driven Hierarchical Multi-Agent Framework for Zero-Touch Optical Networks

TL;DR

The paper tackles the challenge of achieving zero-touch optical networks by moving beyond single-agent GenAI solutions to a hierarchical multi-agent framework. It introduces a four-layer architecture (optical, digital twin, control, support) led by a Network Director, with division agents and AI Experts operating through a Shared Pool for task data and workflow orchestration. Field tests across planning, operation, and upgrade stages demonstrate autonomous DT building, QoT estimation, dynamic channel changes, and capacity upgrades, achieving fast task completion and high accuracy while maintaining security and explainability. The work highlights scalable, explainable autonomous network management with potential multimodal and edge-enabled enhancements, offering a practical pathway toward intelligent, adaptive optical networks.

Abstract

The rapid development of Generative Artificial Intelligence (GenAI) has catalyzed a transformative technological revolution across all walks of life. As the backbone of wideband communication, optical networks are expecting high-level autonomous operation and zero-touch management to accommodate their expanding network scales and escalating transmission bandwidth. The integration of GenAI is deemed as the pivotal solution for realizing zero-touch optical networks. However, the lifecycle management of optical networks involves a multitude of tasks and necessitates seamless collaboration across multiple layers, which poses significant challenges to the existing single-agent GenAI systems. In this paper, we propose a GenAI-driven hierarchical multi-agent framework designed to streamline multi-task autonomous execution for zero-touch optical networks. We present the architecture, implementation, and applications of this framework. A field-deployed mesh network is utilized to demonstrate three typical scenarios throughout the lifecycle of optical network: quality of transmission estimation in the planning stage, dynamic channel adding/dropping in the operation stage, and system capacity increase in the upgrade stage. The case studies, illustrate the capabilities of multi-agent framework in multi-task allocation, coordination, execution, evaluation, and summarization. This work provides a promising approach for the future development of intelligent, efficient, and collaborative network management solutions, paving the way for more specialized and adaptive zero-touch optical networks.

Figures (5)

• Figure 1: GenAI-enabled hierarchical multi-agent framework, including the Network Director, Shared Pool, Division Agents and AI Experts cross four network layers: optical layer, digital twin layer, control layer and support layer.
• Figure 2: The interaction and workflow among multi-agents for most daily maintenance tasks in the zero-touch optical network.
• Figure 3: Case 1: demonstration of a network planning scenario, using the multi-agent framework to assist in DT building and QoT estimation for service deployment.
• Figure 4: Case 2: demonstration of a regular network operation scenario, using the multi-agent framework to assist in analyzing network performance and executing dynamic channel dropping.
• Figure 5: Case 3: demonstration of a network upgrading scenario, using the multi-agent framework to assist in adding an 800Gb/s signal at an appropriate channel, while two 100Gb/s and five 400Gb/s signals have already existed on the same path.