The Survey for a Hierarchical Resource Management Framework Enabled with LLM-based Agents
Derik Chan, Yibo Wu, DaChuan Bai, Tom Mak, Tony Chan, XinXu, Tao Qiu · 2024
A lot of researchers use NLP to change the world and society. Obviously LLM-based Agents are breakthrough in NLP world, they can think and then act to deliver the required services.We would propose to use LLM-based Agents to close the gap within the fast-growing industry and handle all end-to-end operation task not only within a network domain but across different network domain where a model talk to each other to make the work done. It is efficiency to perform this model-to-model communications for an end-to-end solution in the future network.In this paper we propose a Hierarchical Resource Management Framework where we should deploy LLM-based Agents to enable the intention capability of all the provider network assets and turn them into digital chains and carry daily network operations such as tasks like resources allocations, network route calculation and fault handling management automatically to improve the operational efficiency in a multi-resource multi-domain environment.LLM-based Agents enable all the Digital network assets with "Intelligent" and "Thinking", and access to digital chains.