Congestion aware dynamic user association in Heterogeneous cellular network: A stochastic decision approach

In this paper, we propose a novel distributed optimization method for dynamic user association in a downlink Heterogeneous cellular network (Hetnet). We aim at maximizing the utilization of the base stations (BSs) in such a network by jointly considering the effect of the channel gains and load balancing of different BSs. Specifically, we first formulate the user association as a combinatorial optimization problem, whose global optimal solution is difficult to obtain. This is due to the high complexity caused by the large network scale and prohibitive signaling overhead. To address this issue, we then consider the optimization problem under the stochastic decision framework, and propose a distributed heuristic algorithm to independently and dynamically associate each user with the best BS. By posing a price factor to the BS evaluation update, the convergence of the heuristics is guaranteed. Numerical results indicate that the proposed heuristics can perform better than the conventional best-SNR (signal-to-noise ratio) method in the presence of large number of users, and achieves the nearly optimal solution.