Logo
ScienceStack
Table of Contents
Fetching ...
Sign In

Agentic AI-RAN: Enabling Intent-Driven, Explainable and Self-Evolving Open RAN Intelligence

Zhizhou He, Yang Luo, Xinkai Liu, Mahdi Boloursaz Mashhadi, Mohammad Shojafar, Merouane Debbah, Rahim Tafazolli

TL;DR

This article surveys how agentic AI systems with explicit planning, tool use, memory, and self-management offer a natural way to structure long-lived control loops in Open RAN, and introduces a small set of agentic primitives that improve slice life-cycle and RRM performance compared to conventional baselines and ablations that remove individual primitives.

Abstract

Open RAN (O-RAN) exposes rich control and telemetry interfaces across the Non-RT RIC, Near-RT RIC, and distributed units, but also makes it harder to operate multi-tenant, multi-objective RANs in a safe and auditable manner. In parallel, agentic AI systems with explicit planning, tool use, memory, and self-management offer a natural way to structure long-lived control loops. This article surveys how such agentic controllers can be brought into O-RAN: we review the O-RAN architecture, contrast agentic controllers with conventional ML/RL xApps, and organise the task landscape around three clusters: network slice life-cycle, radio resource management (RRM) closed loops, and cross-cutting security, privacy, and compliance. We then introduce a small set of agentic primitives (Plan-Act-Observe-Reflect, skills as tool use, memory and evidence, and self-management gates) and show, in a multi-cell O-RAN simulation, how they improve slice life-cycle and RRM performance compared to conventional baselines and ablations that remove individual primitives. Security, privacy, and compliance are discussed as architectural constraints and open challenges for standards-aligned deployments. This framework achieves an average 8.83\% reduction in resource usage across three classic network slices.

Agentic AI-RAN: Enabling Intent-Driven, Explainable and Self-Evolving Open RAN Intelligence

TL;DR

This article surveys how agentic AI systems with explicit planning, tool use, memory, and self-management offer a natural way to structure long-lived control loops in Open RAN, and introduces a small set of agentic primitives that improve slice life-cycle and RRM performance compared to conventional baselines and ablations that remove individual primitives.

Abstract

Open RAN (O-RAN) exposes rich control and telemetry interfaces across the Non-RT RIC, Near-RT RIC, and distributed units, but also makes it harder to operate multi-tenant, multi-objective RANs in a safe and auditable manner. In parallel, agentic AI systems with explicit planning, tool use, memory, and self-management offer a natural way to structure long-lived control loops. This article surveys how such agentic controllers can be brought into O-RAN: we review the O-RAN architecture, contrast agentic controllers with conventional ML/RL xApps, and organise the task landscape around three clusters: network slice life-cycle, radio resource management (RRM) closed loops, and cross-cutting security, privacy, and compliance. We then introduce a small set of agentic primitives (Plan-Act-Observe-Reflect, skills as tool use, memory and evidence, and self-management gates) and show, in a multi-cell O-RAN simulation, how they improve slice life-cycle and RRM performance compared to conventional baselines and ablations that remove individual primitives. Security, privacy, and compliance are discussed as architectural constraints and open challenges for standards-aligned deployments. This framework achieves an average 8.83\% reduction in resource usage across three classic network slices.
Paper Structure (29 sections, 2 equations, 4 figures, 2 tables)

This paper contains 29 sections, 2 equations, 4 figures, 2 tables.

Table of Contents

  1. Introduction
  2. Motivation and Background
  3. Contribution
  4. Fundamentals
  5. O-RAN Architecture
  6. Agentic Components
  7. Agentic vs Traditional ML/RL vs MARL
  8. Task landscape
  9. Network Slice Life-cycle
  10. RRM and Closed-Loop Control
  11. Security, Privacy, and Compliance
  12. Agentic Primitives
  13. Plan-Act-Observe-Reflect
  14. Skills as Tool-Use
  15. Memory & Evidence
  16. ...and 14 more sections

Figures (4)

  • Figure 1: Architecture of O-RAN
  • Figure 2: O-RAN task landscape with three clusters (network slice life-cycle, radio resource management closed loops, and security/privacy/compliance) and their relation to KPM, KQI, and SLA metrics.
  • Figure 3: Agentic primitives for O-RAN, aligned with control-loop cadences. The loop plans over skills, sequences and gates incremental commits, observes summarized KPM/KQI, and records evidence. Memory spans short-term caches at Near-RT, episodic logs per decision, and long-term knowledge curated at Non-RT.
  • Figure 4: Stacked, min–max normalized KPIs for the full agentic controller, ablations, and the conventional baseline, illustrating the trade-off between QoS and operational cost.