Matching with externalities for decoupled uplink-downlink user association in full-duplex small cell networks

In multi-tier cellular/small cell networks, decoupled uplink-downlink association (DUDe) enables a user to be associated with different base stations (BSs) for uplink (UL) and downlink (DL) transmissions. In this paper, we consider the overall UL and DL rate maximization problem of the users in such a network with a provisioning for decoupled association. In particular, we formulate the UL and DL user association problem as a matching game where users and BSs rank one another using well-defined preference metrics such that their total UL and DL throughput is maximized. The preference profiles of the users are dynamic subject to the unique interference conditions in DUDe networks resulting due to dynamic user associations. As such, considering the impact of externalities becomes crucial in a matching game. To this end, we propose an algorithm based on swap matching that enables players to find a stable association such that no user tries to change her association. Simulations are conducted to show the efficacy of the proposed swap matching algorithm over traditional user association in DUDe networks in which simple user association criteria are used for both the UL and DL transmissions (i.e., path-loss in the UL and received signal power in the DL).