Energy-Efficient Beamforming Optimization for MU-MISO Systems: A Penalty-Based Deep Learning Method

Energy-efficient communication is a key and increasingly important requirement in modern wireless systems, particularly in multi-user multiple-input single-output (MU-MISO) downlink scenarios that demand high performance and scalability. This paper proposes a deep learning (DL)-based beamforming design that directly and explicitly targets energy efficiency as the primary optimization objective in such systems. To address the mismatch between conventional DL architectures and truly energy-efficient operation, a penalty-based loss function is carefully introduced to guide the network away from unnecessary and inefficient full-power usage. Simulation results clearly show that the proposed method consistently achieves comparable or even superior energy efficiency compared to both conventional optimization and DL-based approaches.