Optimizing Wireless Networks with Deep Unfolding: A Comparative Study on Two Deep Unfolding Mechanisms for Energy-Efficient Power Control

This work proposes two deep unfolding-based models for energy-efficient power control in multi-cell wireless networks. The first model (MASUM) combines domain knowledge and data-driven deep learning, while the second model (FUM) fully unfolds the closed-form solution of the power control problem. The proposed models achieve high accuracy and fast inference speed compared to traditional iterative solutions.