Logo
ScienceStack
Table of Contents
Fetching ...
Sign In

A hybrid solution for 2-UAV RAN slicing

Nathan Boyer

TL;DR

The paper tackles 2-UAV RAN slicing by formulating a joint UAV placement and bandwidth allocation optimization, while demonstrating that a hybrid solution combining classical optimization with AI can outperform a purely data-driven approach. It defines a binary placement variable and bandwidth split, with a QoS-based objective to maximize the number of satisfied users, and evaluates multiple ML strategies, including reinforcement learning and supervised learning from optimal solutions. Key contributions include an efficient data-preprocessing pipeline, a hybrid agent that places UAVs using AI but allocates bandwidth via optimization, and comparative analyses showing favorable trade-offs between solution quality and computation time, with generalization to varying user distributions. The results indicate practical viability for UAV-based RAN slicing and point toward scalability to more drones and richer channel models. The work advances the state of UAV-aided network slicing by marrying optimization rigor with data-driven acceleration, enabling responsive, SLA-driven deployments.

Abstract

It's possible to distribute the Internet to users via drones. However it is then necessary to place the drones according to the positions of the users. Moreover, the 5th Generation (5G) New Radio (NR) technology is designed to accommodate a wide range of applications and industries. The NGNM 5G White Paper \cite{5gwhitepaper} groups these vertical use cases into three categories: - enhanced Mobile Broadband (eMBB) - massive Machine Type Communication (mMTC) - Ultra-Reliable Low-latency Communication (URLLC). Partitioning the physical network into multiple virtual networks appears to be the best way to provide a customised service for each application and limit operational costs. This design is well known as \textit{network slicing}. Each drone must thus slice its bandwidth between each of the 3 user classes. This whole problem (placement + bandwidth) can be defined as an optimization problem, but since it is very hard to solve efficiently, it is almost always addressed by AI in the litterature. In my internship, I wanted to prove that viewing the problem as an optimization problem can still be useful, by building an hybrid solution involving on one hand AI and on the other optimization. I use it to achieve better results than approaches that use only AI, although at the cost of slightly larger (but still reasonable) computation times.

A hybrid solution for 2-UAV RAN slicing

TL;DR

The paper tackles 2-UAV RAN slicing by formulating a joint UAV placement and bandwidth allocation optimization, while demonstrating that a hybrid solution combining classical optimization with AI can outperform a purely data-driven approach. It defines a binary placement variable and bandwidth split, with a QoS-based objective to maximize the number of satisfied users, and evaluates multiple ML strategies, including reinforcement learning and supervised learning from optimal solutions. Key contributions include an efficient data-preprocessing pipeline, a hybrid agent that places UAVs using AI but allocates bandwidth via optimization, and comparative analyses showing favorable trade-offs between solution quality and computation time, with generalization to varying user distributions. The results indicate practical viability for UAV-based RAN slicing and point toward scalability to more drones and richer channel models. The work advances the state of UAV-aided network slicing by marrying optimization rigor with data-driven acceleration, enabling responsive, SLA-driven deployments.

Abstract

It's possible to distribute the Internet to users via drones. However it is then necessary to place the drones according to the positions of the users. Moreover, the 5th Generation (5G) New Radio (NR) technology is designed to accommodate a wide range of applications and industries. The NGNM 5G White Paper \cite{5gwhitepaper} groups these vertical use cases into three categories: - enhanced Mobile Broadband (eMBB) - massive Machine Type Communication (mMTC) - Ultra-Reliable Low-latency Communication (URLLC). Partitioning the physical network into multiple virtual networks appears to be the best way to provide a customised service for each application and limit operational costs. This design is well known as \textit{network slicing}. Each drone must thus slice its bandwidth between each of the 3 user classes. This whole problem (placement + bandwidth) can be defined as an optimization problem, but since it is very hard to solve efficiently, it is almost always addressed by AI in the litterature. In my internship, I wanted to prove that viewing the problem as an optimization problem can still be useful, by building an hybrid solution involving on one hand AI and on the other optimization. I use it to achieve better results than approaches that use only AI, although at the cost of slightly larger (but still reasonable) computation times.
Paper Structure (14 sections, 1 equation, 11 figures, 2 tables)

This paper contains 14 sections, 1 equation, 11 figures, 2 tables.

Table of Contents

  1. Probleme presentation and literature analysis
  2. Problem definition and solution via mathematical optimization
  3. Optimization problem
  4. Instances generation
  5. Improvement of the optimization algorithm
  6. Machine learning to solve the problem
  7. Reinforcement learning
  8. Supervised learning from the optimal solution
  9. Learn only what's necessary
  10. Generalizing to a variable number of users
  11. Comparison of the different methods
  12. Areas for improvement
  13. Conclusion
  14. Appendix detailling how the neural networks are built

Figures (11)

  • Figure 1: Representation of the convex enveloppe of the 4 clusters
  • Figure 2: Performance when users are random
  • Figure 3: Performance when users are in 2 clusters
  • Figure 4: Position error
  • Figure 5: Bandwidth error
  • ...and 6 more figures