Network Slice Admission Control for Enhanced Resource Utilization in 6G Multi-UAV Networks

In the context of 6G networks, Unmanned Aerial Vehicle (UAV) based edge nodes play a crucial role in delivering accessible, scalable, and low-latency services. Optimizing resource utilization in multi-UAV frameworks require further investigation into user association. Additionally, network slicing allows for the creation of multiple independent, coexisting logical networks, enabling different ground users to share computing, networking, and storage resources. However, due to the finite nature of these resources, service providers cannot meet all user requests. Users often overestimate their resource needs to avoid potential SLA violations. To address these challenges, we introduce a network slice admission control algorithm designed to optimize resource utilization and enhance user association in a UAV environment. We formulated the resource utilization and user association problem as a linear programming problem and solved it using standard optimization tools. To further improve resource utilization and user association, our algorithm employs a Bidirectional Long Short-Term Memory (BiLSTM) to dynamically pool resources. Our implementation and evaluation of this algorithm in a 6G-UAV environment demonstrates its superior performance compared to conventional approaches, as confirmed by our numerical analysis.