Throughput Maximization in Cellular Networks With Wireless Backhaul and Energy Harvesting

This article investigates the throughput maximization problem in cellular networks where a macro cell is overlaid with multiple small cells, and each small-cell base station (sBS) provides data links for wireless backhaul and energy links for energy harvesting (EH) nodes. The data links enable communication between sBSs, while the energy links facilitate radio-frequency-based energy harvesting by EH nodes. Efficient scheduling of these links is crucial for enhancing network throughput while fulfilling the energy requirements of the EH nodes. To tackle this challenge, a mixed integer non-linear program (MINLP) and a minimizing power problem are formulated to derive a link schedule for both types of links with the minimum power consumption. However, obtaining the optimal schedule is challenging due to the complexity of the initial formulations. To address this, a practical problem decomposition method is proposed, along with a single-stage technique that simultaneously maximizes the number of scheduled data links and minimizes transmit power. The optimization process is further improved using benders decomposition (BD), complemented by the machine learning-aided multi-cut method for enhanced convergence. Simulation results demonstrate that the proposed approach achieves near-optimal link scheduling with reduced iterations. Numerical analysis validates the effectiveness of the problem decomposition method, particularly when the EH nodes have an energy requirement of 0.1 Joules.