Traffic-Aware Resource Allocation for Multi-User Beamforming

Yu-Hsuan Liu,Kate Ching-Ju Lin

DOI: https://doi.org/10.1109/tmc.2022.3141787

IF: 6.075

2022-01-01

IEEE Transactions on Mobile Computing

Abstract:5G New Radio (NR), a new radio access technology, has been proposed to enhance the flexibility, scalability, and efficiency of 5G networks. By leveraging phased-array antennas, a base station (BS) can adaptively form multiple directional beams to serve geo-distributed user equipments (UEs) concurrently. However, the imperfect beam pattern of a phased-array antenna may create side lobes, leading to inter-beam interference. While recent efforts have focused on beam selection that mitigates inter-beam interference and maximizes the sum-rate, we notice that the selected beams may not be fully utilized in an OFDMA-based system. The root cause is that only a fixed set of beams can be configured at a time to serve a wide frequency band but some resource blocks (i.e., subcarriers) may not be able to be allocated to any UEs due to limited traffic demands. To address the above resource underutilization problem, this paper presents traffic-aware joint beam configuration and resource allocation, which explicitly considers user traffic demands and configures beams that can optimally utilize the spectrum resources. We derive an approximation model that produces a suboptimal solution solvable by open-source solvers and further develop a low-complexity greedy algorithm for large-scale networks. Our simulation results show that the proposed traffic-aware allocation and beam configuration scheme achieves better utilization, especially when users have heterogeneous traffic demands. The overall effective throughput can be significantly enhanced as compared to conventional sum-rate maximization allocation.

computer science, information systems,telecommunications