Power and Beam Optimization for Uplink Millimeter-Wave Hotspot Communication Systems

Abstract

We propose an effective interference management and beamforming mechanism for uplink communication systems that yields fair allocation of rates. In particular, we consider a hotspot area of a millimeter-wave (mmWave) access network consisting of multiple user equipment (UE) in the uplink and multiple access points (APs) with directional antennas and adjustable beam widths and directions (beam configurations). This network suffers tremendously from multi-beam multi-user interference, and, to improve the uplink transmission performance, we propose a centralized scheme that optimizes the power, the beam width, the beam direction of the APs, and the UE - AP assignments. This problem involves both continuous and discrete variables, and it has the following structure. If we fix all discrete variables, except for those related to the UE-AP assignment, the resulting optimization problem can be solved optimally. This property enables us to propose a heuristic based on simulated annealing (SA) to address the intractable joint optimization problem with all discrete variables. In more detail, for a fixed configuration of beams, we formulate a weighted rate allocation problem where each user gets the same portion of its maximum achievable rate that it would have under non-interfered conditions. We solve this problem with an iterative fixed point algorithm that optimizes the power of UEs and the UE - AP assignment in the uplink. This fixed point algorithm is combined with SA to improve the beam configurations. Theoretical and numerical results show that the proposed method improves both the UE rates in the lower percentiles and the overall fairness in the network.