Multi-Agentic AI for Conflict-Aware rApp Policy Orchestration in Open RAN

ABSTRACT

This research introduces a multi-agent framework designed to automate the generation and orchestration of rApp policies in Open Radio Access Networks (Open RAN). The system utilizes three specialized LLM agents (Perception, Reasoning, and Refinement) to resolve conflicts and synthesize intent-aligned control pipelines. The methodology achieves a 70% improvement in deployment accuracy and a significant reduction in reasoning costs.