Throughput Maximization for a Buffer-Aided Successive Relaying Network Employing Energy Harvesting

Energy harvesting (EH)-assisted nodes are capable of significantly prolonging the lifetime of future wireless networks, provided that they rely on appropriate transmission policies, which accommodate the associated stochastic energy arrival. In this paper, a successive-relaying-based network using rechargeable source and relay nodes having limited buffers for both their energy and data storage is considered. The maximization of the network throughput with noncausal knowledge of energy arrivals by the deadline $T$ is formulated as a nonconvex optimization problem, and it is solved using the interior-point optimization (IPOPT) method. The performance of the low-complexity suboptimal scheme was found to reach its maximum when the two phases of the successive relaying protocol have equal duration. The optimal and suboptimal schemes are capable of achieving up to 92% and 88% of the throughput performance of the benchmark scheme. The suboptimal scheme's throughput performance is consistently about 90% of that of the optimal scheme. For asymmetric data (or energy) buffer sizes, it was found that the throughput performance depends on the total (i.e., collective) data (or energy) buffer capacity available in the network and not just on the smallest data buffer.