Optimal client association, airtime sharing and contention resolution in throughput fair multi-cell WLANs with hidden APs

In this paper we study the problem of client association, airtime sharing and contention resolution in a multi-cell WLAN to achieve fair download bandwidth sharing. We first model the problem as a utility maximization problem, then, because of the integral nature of association, the problem being NP-hard, we decompose it into two sub-problems: the client association problem on one hand, and the joint airtime sharing and contention resolution problem on the other. We show that, for each fixed association pattern, the corresponding joint airtime sharing and contention resolution problem can be solved in polynomial time. To solve the client association problem, we propose the so-called Swap localized algorithm. The Swap procedure iteratively swaps clients between pairs of neighbour access points (APs) if the swap leads to an improvement in the global utility. Simulation results show that, the proposed algorithm is significantly better than RSSI-based heuristic, and quickly leads to a near-optimal solution.