Joint Beamwidth And Energy Optimization For Multi-User Millimeter Wave Communications

This paper aims to optimize beamwidths of the base station (BS) and the user equipments (UEs), and optimize energy allocations for beam training and data transmission of each UE for the downlink of an energy-constrained multi-user millimeter (mmWave) communication system. To this end, we formulate a multi-objective optimization problem, in which we target to minimize the beam training time (maximize beamwidth), minimize the total energy usage and maximize the sum-rate simultaneously while maintaining the minimum QoS for each UE. After energy is allocated to each UE, the allocated energy is efficiently split for beam training and data transmission. To resolve the formulated problem, we propose a new approach in-line with the epsilon-constraint method where the lower and upper bounds for the beamwidth and energy usage are derived, transforming the problem to a single objective optimization of the downlink sum rate. This single objective optimization is found out to be non-convex in nature and in general, is hard to solve. We, thus, develop a numerical algorithm to achieve the Pareto optimal solutions within the derived bounds. Finally, we carry out simulations to evaluate the performance of the proposed algorithm. Also, we analyze the effect of channel coherence time on the Pareto solutions obtained using the proposed algorithm.