Joint Optimization of Beam Selection and Power Allocation for mmWave Communication under Nonuniform Beam Cardinality

With the support of massive multiple input multiple output (MIMO), millimeter wave (mmWave) communication technology has emerged as a promising solution to meet the increasing demand on data traffic. This paper addresses the joint optimization of beam selection and power allocation to maximize the spectral efficiency of multiple users across multiple base stations (BSs), where the number of available beams has spatial differences (i.e., nonuniform cardinalities) for different BSs. A mixed-integer programming problem is formulated under the considerations of complex interference environments (i.e., including intra-BS interference and inter-BS interference) and nonuniform cardinality, which is transformed into the convex optimization problem by relaxing the binary indicator of beam selection and utilizing the logarithmic approximation of spectral efficiency. The Lagrangian dual method and the gradient descent method are then exploited to solve the transformed optimization problem. Corroborated by simulation results, the proposed algorithm can be converged iteratively and is superior to the state of the art in terms of spectral efficiency.