Downlink cell association optimization for heterogeneous networks via dual coordinate descent

This paper considers the optimal association of remote user terminals to different cells in a heterogeneous network for load balancing. Assuming fixed transmit powers at the base-stations, we adopt a network utility maximization formulation with a proportional fairness objective and show that the downlink user association problem can be solved efficiently using a pricing approach where the prices are updated in the dual domain via coordinate descent. As compared to the previously proposed subgradient method, the proposed coordinate descent algorithm does not require the base-stations to synchronize in their price updates, while still guaranteeing convergence, which makes it particularly suitable for distributed implementation. Simulations show that the proposed method has fast convergence while achieving near-optimal solution.