Distributed user association for delay-load tradeoffs in energy harvesting enabled HetNets

Energy harvesting is an emerging technology to save on-grid power and reduce carbon emissions in wireless communications. Terminals can utilize renewable energy obtained from environment by energy harvesting facilities. This paper investigates user association problem under a flow-level traffic model in energy harvesting enabled heterogeneous networks. In this paper, all the base stations are equipped with energy harvesting facilities and assumed to be solely powered by renewable energy. The average traffic delay is denoted using queuing theory, and a fairness index is redefined to evaluate the level of load balancing. Then a convex optimization problem is formulated in order to reduce average traffic delay and improve load balancing level. To this end, an iterative algorithm completed by user and base station respectively is proposed. Simulation results show that the user association algorithm improves the load balancing level with limited sacrifice on average traffic delay compared with the existing delay-optimal algorithm, and enables a flexible tradeoff between the average traffic delay and load balancing.