Energy-Efficient Parallel Two-Way Relaying Networks with Energy Harvesting

In this paper, the energy efficient parallel two-way relay (P-TWR) transmission networks with multiple energy harvesting nodes are studied. Relay selection (RS), power allocation (PA) and rate policies across multiple time intervals are jointly designed. The problem of energy efficiency (EE) maximization under the constraints of the energy causality and a pre-defined rate threshold is formulated. Though the optimization problem is NP-hard, we are able to convert it into a convex one. Specifically, by applying the Lagrange dual decomposition method, we separate PA from RS and rate policy such that PA can be solved first. Further, with the obtained optimum PA solution, RS is mapped into a maximum weighted bipartite graph matching problem and solved with the Kuhn-Munkres algorithm. Finally, the optimum rate policy is obtained with the efficient interior point method. The performance comparisons with "Depleted Energy" and "Uniform Power" policies demonstrate that our policies can provide notable EE gains. It is also shown that a high pre-defined rate threshold costs a significant EE penalty. Finally, it is illustrated that the proposed algorithm converges speedily.