EvoRIC: Reinforcement Learning Fine-Tuned LLM-empowered RAN Intelligent Control Toward Autonomous O-RAN

2026-08-07

AI Summary

This work addresses the limitations of traditional machine learning methods—poor generalization in wireless access networks—and the high computational overhead and lack of domain knowledge in general-purpose large language models (LLMs). To bridge this gap, the paper proposes EvoRIC, a novel framework that, for the first time, integrates an LLM fine-tuned via proximal policy optimization (PPO)-based reinforcement learning into the O-RAN intelligent controller. Leveraging a hierarchical architecture comprising non-real-time and near-real-time RICs, EvoRIC dynamically aligns semantic reasoning with network performance objectives. The framework continuously evolves the LLM through environment interaction data, substantially enhancing its decision-making capability in wireless scenarios. Experimental evaluation in integrated access and backhaul (IAB) networks demonstrates EvoRIC’s strong generalization and effectiveness, laying a foundation for autonomous O-RAN systems.

Abstract

Despite recent advances in applying artificial intelligence (AI) techniques to radio access network (RAN), critical challenges remain: traditional machine learning (ML) algorithms suffer from limited generalization across varying network topologies, whereas general-purpose large language models (LLMs) face high computational demands and lack domain-specific knowledge. To address these gaps, this article introduces the evolving RAN intelligent controller (RIC) (EvoRIC) framework, a hierarchical architecture that enables continuous evolution by leveraging a non-real-time RIC (non-RT RIC) for global model updates and a near-real-time RIC (near-RT RIC) for local execution, dynamically empowering LLMs with domain-specific decision-making capabilities. Within this framework, we employ a reinforcement learning-based fine-tuning (RLFT) mechanism where an LLM operates as an actor within a proximal policy optimization (PPO) agent. By leveraging the interaction tuples collected from the wireless environment, the LLM's parameters are iteratively updated to align semantic reasoning with rigorous network performance objectives. We evaluate the generalization and efficacy of the proposed EvoRIC framework within integrated access and backhaul (IAB) networks, and finally, discuss the open challenges and future directions of the EvoRIC framework toward realizing autonomous O-RAN.

Problem

• Radio Access Network (RAN)
• Large Language Models (LLMs)
• Generalization
• Domain-specific Knowledge
• Autonomous O-RAN

Innovation

• Reinforcement Learning Fine-Tuning
• Large Language Models
• RAN Intelligent Controller
• O-RAN
• Proximal Policy Optimization