Work

First Demonstration of Multi-Agent LLM System for Million-Scale Optical Link Management in Global Production AIDCs

Year: 2026

Type: preprint

Abstract: We present the first LLM-powered multi-agent system for autonomous fault management across millions of optical links in production AIDCs. Refined via SFT and continuous memory evolution, it achieves 97.7% F1 and over 60% fault-incident reduction, outperforming SOTA LLMs on a ten-week field data evaluation.

Source: arXiv (Cornell University)

DOI: 10.48550/arxiv.2608.23145

Authors: Su, Jingyi, Zhang, Yihao, Fu, Dianxuan, Fei, Leiyan, Wang, Juan, Dai, Mengfan, Liu, Qing, Wu, Xiong, Jiang, Yufeng, Chen, Cheng, Zhang, Bowen, Wang, Peilong, Chen, Xi, He, Zonglong, Yu, Hongchen, Ye, Zhicheng, Hu, Weisheng, Zhuge, Qunbi

Institutions: Shanghai Jiao Tong University, Baidu (China), Huawei Technologies (China)

Language: English

Topic: Mobile Agent-Based Network Management

Subfield: Computer Networks and Communications

Field: Computer Science

Domain: Physical Sciences

Open Access status: green

Funders: National Natural Science Foundation of China, Shanghai Jiao Tong University

Awards: G726549461, G8649715131

OpenAlex ID: W7204195666

Cites:

Cited by:

Related to: 10