Online Algorithms for Efficient Beamforming in Next-Generation Wireless Networks

This thesis tackles key challenges in next-generation wireless networks, focusing on efficient beam alignment and energy harvesting. A major contribution is integrating fixed-budget best-arm identification for unimodal bandits into the multi-armed bandit (MAB) framework, which had not been previously explored in the MAB literature. This approach is particularly effective in solving the beam alignment problem in millimeter wave massive MIMO and Holographic Metasurface Transceiver-assisted systems. Additionally, the thesis enhances fairness and energy efficiency in Radio Frequency Energy Harvesting networks by optimizing power levels and ensuring fair energy allocation. These contributions offer theoretical guarantees and practical solutions for sustainable, high-performance wireless networks, advancing robust communication and energy management in future systems.