Page 351 - AI for Good Innovate for Impact

AI for Good Innovate for Impact

Additionally, the use case supports the development of smart city applications by ensuring
reliable and efficient network services. This reliability is essential for enabling real-time urban
services, improving the sustainability and livability of modern cities.

2�3 Future Work 4.3 - 5G
• Data Collection: Gather extensive network traffic and performance data to enhance the
accuracy of machine learning models (Synthetic data will be used for the ML training).
• Model Development: Develop advanced machine learning algorithms capable of real-
time decision-making for dynamic slice instance management by time series analysis of
the traffic load.
• Infrastructure Enhancement: Upgrade network monitoring and management systems to
support dynamic slice instance switching by the load balancer entity.

3 Use Case Requirements

• REQ-01: It is critical that the system dynamically monitors load conditions on network
slices and adjusts instance configurations in real-time to maintain optimal performance.
• REQ-02: It is critical that the solution employs machine learning techniques to accurately
predict traffic patterns and proactively adjust slice instances accordingly.
• REQ-03: It is critical that the platform ensures dynamic instance switching does not
degrade the Quality of Service (QoS) experienced by end-users.
• REQ-04: It is critical that the system integrates seamlessly with existing network
management frameworks to enable coordinated control and comprehensive monitoring.
• REQ-05: It is critical that the solution be inherently scalable to support a large number of
network slices with diverse and evolving service requirements.

4 Sequence Diagram

5 References

[1] Zhao, G., Wen, M., Hao, J., & Hai, T. (2021). Application of Dynamic Management of 5G
Network Slice Resource Based on Reinforcement Learning in Smart Grid. In The 10th

315