Secure 5G Network Slicing with Reinforcement Learning-Based Intrusion Detection
Keywords:
5G network slicing, intrusion detection, reinforcement learning, security architecture, policy governance, adaptive security, network softwarizationAbstract
The emergence of fifth-generation (5G) mobile networks has introduced network slicing as a fundamental architectural enabler, allowing the creation of multiple isolated logical networks over a shared physical infrastructure. While slicing enables tailored service delivery for diverse use cases such as enhanced mobile broadband, ultra-reliable low-latency communications, and massive machine-type communications, it simultaneously expands the attack surface and introduces novel security vulnerabilities. Traditional intrusion detection systems, which rely on static signature-based or shallow learning methods, struggle to adapt to the dynamic and distributed nature of slice environments. This paper presents a comprehensive systems-level analysis of reinforcement learning (RL)-based intrusion detection mechanisms as a security solution for 5G network slicing. We examine the architectural trade-offs between detection accuracy, computational overhead, and slice isolation integrity. The discussion emphasizes structural considerations including policy governance, resource allocation fairness, and the sustainability of RL models in production deployments. We further explore the implications of continuous learning, adversarial robustness, and cross-slice data sharing for regulatory compliance and operational resilience. Through comparative analysis with conventional approaches and illustrative deployment scenarios, we demonstrate that RL-based intrusion detection offers a promising pathway toward adaptive, self-optimizing security, provided that challenges related to model interpretability, training convergence, and inter-slice privacy are systematically addressed. The paper concludes with forward-looking perspectives on the integration of RL agents with zero-trust architectures and the role of human-in-the-loop oversight in mission-critical slice operations.
References
1. Afaq, M., Liu, Y., & Hu, J. (2020). Network slicing for 5G: A survey. IEEE Communications Surveys & Tutorials, 22(4), 3525–3545.
2. Alavian, A., & Movaghar, A. (2019). Security challenges and solutions in 5G network slicing: A survey. Journal of Network and Computer Applications, 145, 102421.
3. Alnoman, A., & Anpalagan, A. (2018). Network slicing with NFV and SDN: A survey of concepts, architectures, and challenges. IEEE Access, 6, 73363–73385.
4. Bhardwaj, S., Kaur, R., & Singh, M. (2021). Intrusion detection in 5G mobile networks using deep learning: A review. Computers & Security, 108, 102378.
5. Casas, P., D’Alconzo, A., & Callegari, C. (2020). Machine learning for network security: Challenges and opportunities. Computer Networks, 171, 107131.
6. Chen, J., & Chen, Y. (2019). A survey of intrusion detection in 5G network slicing. Journal of Communications and Networks, 21(4), 345–358.
7. D’Angelo, G., & Palmieri, F. (2021). Deep reinforcement learning for adaptive intrusion detection in IoT networks. IEEE Internet of Things Journal, 8(18), 14056–14066.
8. Deng, S., Wang, J., & Zhang, Y. (2022). Reinforcement learning for network security: A comprehensive survey. ACM Computing Surveys, 55(3), 1–37.
9. Fischer, A., & Weiler, M. (2020). Explainable reinforcement learning: A review. Artificial Intelligence Review, 54(3), 2025–2052.
10. Fragkiadakis, A., & Siris, V. (2019). Anomaly-based intrusion detection in 5G networks using deep autoencoders. Security and Communication Networks, 2019, 1–12.
11. Ghafouri, M., & Kavian, Y. (2021). Safe reinforcement learning for cybersecurity applications. IEEE Transactions on Information Forensics and Security, 16, 4120–4133.
12. Ghavimi, F., & Chen, H. (2020). Resource allocation in 5G network slicing: A game-theoretic approach. IEEE Transactions on Wireless Communications, 19(6), 3915–3928.
13. Li, Q. (2026). QoS Assurance Mechanism for 5G Network Slicing Based on the Deep Reinforcement Learning PPO Algorithm. arXiv preprint arXiv:2605.03345.
14. Lin, S., & Thiran, P. (2021). Federated reinforcement learning for multi-tenant network security. In Proceedings of the IEEE Conference on Computer Communications (INFOCOM) (pp. 1–9). IEEE.
15. Liu, W., & Wu, J. (2022). Adversarial robustness of deep reinforcement learning for intrusion detection. Neurocomputing, 472, 124–135.
16. Mao, Y., & Zhang, C. (2019). A survey of mobile edge computing for 5G network slicing. IEEE Communications Magazine, 57(9), 72–78.
17. Marotta, M., & Cerqueira, E. (2020). Human-in-the-loop approaches for autonomous network security. IEEE Security & Privacy, 18(5), 48–56.
18. Mijumbi, R., & Serrat, J. (2016). Network function virtualization: State-of-the-art and research challenges. IEEE Communications Surveys & Tutorials, 18(1), 236–262.
19. Nencioni, G., & Helvik, B. (2021). Security isolation in network slicing: Threats and countermeasures. IEEE Transactions on Network and Service Management, 18(3), 3425–3439.
20. Ordonez, A., & Zhang, L. (2020). Deep Q-network for intrusion detection in 5G core networks. In IEEE International Conference on Communications (ICC) (pp. 1–6). IEEE.
21. Raza, M., & Shafiq, M. (2022). Catastrophic forgetting in online learning for network anomaly detection. Pattern Recognition Letters, 158, 87–93.
22. Santos, J., & Solano, P. (2021). Policy-based security orchestration for 5G network slices. IEEE Transactions on Network and Service Management, 18(4), 4558–4572.
23. Shen, X., & Gao, L. (2020). Zero-trust architecture for 5G network slicing. IEEE Network, 34(6), 186–192.
24. Taleb, T., & Afrin, H. (2019). Network slicing security: A survey of threat landscape and defense mechanisms. IEEE Communications Surveys & Tutorials, 21(4), 3462–3492.
25. Wang, L., & Liu, K. (2022). Reinforcement learning for network slicing: A survey of resource allocation and security. Journal of Network and Systems Management, 30(4), 1–35.
Downloads
Published
Issue
Section
License
Copyright (c) 2026 Journal of Data Intelligence and AI Systems
This work is licensed under a Creative Commons Attribution 4.0 International License.
This article is published under the Creative Commons Attribution 4.0 International License (CC BY 4.0), which permits unrestricted use, distribution, and reproduction in any medium, provided the original author and source are credited.
