Integrating Queuing Theory and GBM for Dynamic Resource Allocation in 6G Networks

6G networks promise seamless connectivity and unprecedented advancements compared to previous generations. Small cells are at the core of 6G's success, which improves network coverage, reduces delays, and increases capacity by placing base stations close to users. However, the dense deployment of small cells in 6G networks introduces significant challenges, particularly in resource allocation due to huge state and action spaces. More specifically, resource allocation decisions must be computed for thousands of cells. Also, frequent resource reallocation due to high user mobility increases the algorithmic overhead. These challenges negatively impact critical performance metrics such as latency, throughput, and packet loss. To address these challenges, this paper proposes a dynamic resource allocation model that integrates queuing-theory-based small cell state determination with Gradient Boosting Machine (GBM) predictions. This integrated approach empowers dynamic resource adjustment by ensuring the timely allocation of bandwidth, computing power, and energy usage. The simulation results demonstrate the effectiveness of the proposed approach in terms of latency, throughput, and packet loss compared to conventional resource allocation methods.