Near-optimal user-cell association schemes for real-world networks

The growing demand for wireless bandwidth makes WiFi deployments denser and pushes cellular networks to adopt a denser, small-cell architecture. In such dense environments, users have multiple options when it comes to selecting an access point (AP) to associate with the network, and the user-cell association scheme that is used has a large impact on overall network performance. Industry is currently using simplistic, sub-optimal approaches to select an AP for each user, while academia has produced optimal schemes under unrealistic assumptions, which prevents them from ever being used in practice. In this paper we design high-performance user-cell association schemes which can be deployed in real-world networks. The performance of the proposed schemes is shown to be near-optimal via both formal performance bounds under realistic assumptions, and simulation results under real-world setups.