In this paper, an energy efficiency maximization (EEM) optimization problem for the multiuser multicarrier energy-constrained amplify-and-forward (AF) multi-relay network is formulated under the total source transmit power budget and energy-causality constraints. We consider that each relay node is solely powered by the source nodes, employing energy harvesting time-switching (EHTS) protocol to harvest the energy through the ambient radio-frequency (RF) signal transmitted from the source nodes under the simultaneous wireless information and power transfer (SWIPT) paradigm. First, we propose a subcarrier and energy causality-based multi-relay selection policy. Second, we jointly optimize the parameters that control the energy efficiency (EE) of the system namely multi-relay selection, subcarrier pairing, user allocation, power allocation, and RF EHTS time block, that renders the problem to be a mixed integer non-linear programming problem (MINLP) which remains NP-hard to solve. Hence, we remodel the problem to a tractable quasi-concave form by applying a string of convex transformations. Later, we propose an iterative EEM algorithm to optimize the multi-parameter problem. Further, a suboptimal and best relay selection algorithm is studied by trading-off between complexity and performance. The effectiveness of the proposed algorithms is demonstrated by simulation results.