Optimization of QoS in 5G optical networks for futuristic high-speed, low-latency applications
-
Piyush Kulshreshtha
Abstract
5G networks support emerging high-speed and low-latency applications, such as remote healthcare and connected cars. Such applications require a consistent quality of service (QoS) to be supported by the underlying optical networks. A combination of spatio-temporal graph neural networks (STGNNs) and twin-delayed deep deterministic policy gradient (TD3), with modifications, is proposed here to solve the problem of QoS optimization through metrics, such as delay, throughput, and jitter. The STGNNs extract the dynamic topology-related features from the network. Modified TD3 helps in learning an adaptive routing policy for optimization of QoS metrics. The proposed model, labelled as spatio-temporal optical network recurrent model with TD3 (STORM-TD3) is validated through simulation. Compared to DDPG, the existing approach, it reduces the end-to-end delay by 9 %, jitter by 1.65 %, maintaining the same level of throughput, and is recommended for deployment in 5G optical networks for QoS optimization.
Acknowledgments
OMNeT++ Simulation Tool (network tool).
-
Research ethics: Not applicable.
-
Informed consent: Not applicable.
-
Author contributions: The author(s) have accepted responsibility for the entire content of this manuscript and approved its submission.
-
Use of Large Language Models, AI and Machine Learning Tools: None declared.
-
Conflict of interest: No conflict of interest.
-
Research funding: None declared.
-
Data availability: Not applicable.
References
1. Cao, X, Yoshikane, N, Popescu, I, Tsuritani, T, Morita, I. Software-defined optical networks and network abstraction with functional service design. J Opt Commun Netw 2017;9:C65–75. https://doi.org/10.1364/jocn.9.000c65.Search in Google Scholar
2. Jain, A, Zamir, AR, Savarese, S, Saxena, A. Structural-rnn: deep learning on spatio-temporal graphs. In: Proceedings of the IEEE conference on computer vision and pattern recognition. NV, USA: Las Vegas; 2016:5308–17 pp.10.1109/CVPR.2016.573Search in Google Scholar
3. Fujimoto, S, Hoof, H, Meger, D. Addressing function approximation error in actor-critic methods. In: International conference on machine learning (PMLR). Sweden: Stockholm; 2018:1587–96 pp.Search in Google Scholar
4. Chakraborty, S, Turuk, AK, Sahoo, B. Federated learning enabled software-defined optical network with intelligent control plane architecture. Comput Electr Eng 2024;118:109329. https://doi.org/10.1016/j.compeleceng.2024.109329.Search in Google Scholar
5. Schwarzmann, S, Marquezan, CC, Trivisonno, R, Nakajima, S, Barriac, V, Zinner, T. Ml-based qoe estimation in 5g networks using different regression techniques. IEEE Trans Netw Serv Manag 2022;19:3516–32. https://doi.org/10.1109/tnsm.2022.3179924.Search in Google Scholar
6. Martín, I, Troia, S, Hernandez, JA, Rodriguez, A, Musumeci, F, Maier, G, et al.. Machine learning-based routing and wavelength assignment in software-defined optical networks. IEEE Trans Netw Serv Manag 2019;16:871–83. https://doi.org/10.1109/tnsm.2019.2927867.Search in Google Scholar
7. Sakthivel, S, Belgaum, MR, Sajak, AAB, Alam, MM, Su’ud, MM. Machine learning-driven QoT prediction for enhanced optical networks in DWDM system. IEEE Access 2025;13:80445–60. https://doi.org/10.1109/access.2025.3565502.Search in Google Scholar
8. Preciado-Velasco, JE, Gonzalez-Franco, JD, Anias-Calderon, CE, Nieto-Hipolito, JI, Rivera-Rodriguez, R. 5G/B5G service classification using supervised learning. Appl Sci 2021;11:4942. https://doi.org/10.3390/app11114942.Search in Google Scholar
9. Paudyal, P, Shen, S, Yan, S, Simeonidou, D. Toward deployments of ML applications in optical networks. IEEE Photon Technol Lett 2021;33:537–40. https://doi.org/10.1109/lpt.2021.3074586.Search in Google Scholar
10. Devi, SS, Kirubanandasarathy, N. An enhanced quality of transmission‐aware optical networking based on deep reinforcement learning. Int J Commun Syst 2025;38:e70204. https://doi.org/10.1002/dac.70204.Search in Google Scholar
© 2025 Walter de Gruyter GmbH, Berlin/Boston
