
Agentic AI Optimizes Cell-free O-RAN
This arXiv paper proposes an agentic AI framework using LLM-based agents for intent-driven optimization in cell-free O-RAN. A supervisor translates intents into objectives and rate requirements, while specialized agents handle weighting, O-RU management via DRL, and monitoring. PEFT enables LLM sharing, reducing active O-RUs by 41.93% and memory by 92%.






