Matching-based joint uplink and downlink user association for energy-efficient hetnets

We consider the energy efficiency-aware user association problem in heterogeneous networks (HetNets). Since there is a large disparity between the transmit power levels of the macro and small cell base stations, the traditional downlink received signal strength indicator (RSSI) based user association results in extreme load imbalance and high uplink interference, which degrades the spectrum efficiency and energy efficiency, especially in uplink. Focusing on this, we propose energy efficiency-aware user association taking into account both uplink and downlink. However, the formulated non-convex and combinatorial problem cannot be distributively solved by convex optimization techniques or common game-theoretic methods. Thus, we apply the matching theory to develop a distributed user association algorithm, which reaches a stable matching upon convergence. Lastly, by comparing with two downlink-centric user association methods, the proposed algorithm is shown to have a superior performance in terms of energy efficiency, uplink capacity and energy saving.