An Integrated Online and Offline Multi-Agent AI Assistant for Network Management System Diagnosis

In modern network management, efficient network data monitoring, such as connectivity and traffic, is essential. Traditional diagnostic methods often involve manual inspection of complex configurations, leading to time-consuming and error-prone processes. This study introduces an advanced AI-based network management assistant designed to overcome these challenges. The system employs Large Language Models (LLMs), Retrieval-Augmented Generation (RAG) technology, and a Multi-Agent framework to effectively manage user queries and deliver precise diagnostics and solutions. By integrating user manuals, FAQs, and Rule-Based AI (R-BAI) technologies, the system automatically detects issues and generates detailed reports. Tasks are distributed among specialized agents to optimize efficiency and accuracy. Furthermore, the system features both online and offline feedback mechanisms, allowing real-time user suggestions and expert reviews to enhance learning capabilities. This comprehensive approach aims to simplify network management tasks and significantly improve operational efficiency. Preliminary results demonstrate enhanced diagnostic accuracy and reduced error rates, underscoring the system’s potential to transform network management practices.