ADORN: Adaptive Drift handling for Open RAN using Reinforcement Learning

2026-07-09

AI Summary

This work addresses the challenge of data drift in Open RAN caused by dynamic traffic patterns, which degrades the performance of AI/ML models while conventional retraining approaches incur high computational overhead and risk violating service-level agreements (SLAs). To overcome these limitations, the authors propose a reinforcement learning–based adaptive retraining mechanism that formulates retraining decisions as a Markov decision process. A Q-learning agent learns an optimal policy balancing prediction accuracy against retraining cost. Furthermore, a multi-expert LSTM ensemble architecture enhances model robustness and mitigates catastrophic forgetting. Experimental results demonstrate that, compared to greedy and random baselines, the proposed method significantly reduces retraining overhead while consistently maintaining system performance within SLA constraints.

Abstract

Dynamic traffic variations in Open Radio Access Networks (O-RAN) lead to drift, which degrades the performance of Artificial Intelligence/Machine Learning (AI/ML) models. Traditional retraining approaches maintain forecasting accuracy but incur high computational cost and may lead to violations of Service Level Agreements (SLAs). This work proposes a Q-learning-based adaptive retraining approach that formulates the retraining decision as a Markov Decision Process (MDP), where a Reinforcement Learning (RL) agent learns a policy that balances forecasting accuracy and retraining cost. The proposed approach incorporates a multi-expert Long Short-Term Memory (LSTM) ensemble to mitigate catastrophic forgetting and improve robustness across diverse traffic conditions. Experimental results show that the proposed approach effectively reduces retraining overhead compared to greedy and random baselines, while maintaining system performance within predefined limits.

Problem

• Open RAN
• concept drift
• AI/ML model degradation
• retraining cost
• Service Level Agreements

Innovation

• Reinforcement Learning
• Adaptive Retraining
• Open RAN
• LSTM Ensemble
• Markov Decision Process