Self-optimization of uplink power and decoding order in heterogeneous networks

This work demonstrates the distributed joint selfoptimization of power and decoding order, to alleviate uplink intercell interference in a heterogeneous network, with signal to interference ratio constraints. This problem can be formulated as a potential game of coupled action space among the basestations. However, best response dynamics of this game require global channel knowledge at all cells. This is unlike when power is the only optimization variable, in which case the users can autonomously reach the unique sum power minimizing equilibrium with locally measurable information at its basestation. Thus in order to propose a distributed scheme, by exploiting the properties of small-cells, this paper designs a perturbation to the original action space and a messaging scheme among neighbourhoods of base-stations. The resulting perturbed game has best response dynamics that converge with much less information than the original game, under three proposed sufficient constraints. Monte Carlo simulations demonstrate that the perturbed game has performance close to that of the original game.