Distributed Resource Allocation for Femtocell Networks: Regret Learning with Proportional Self-belief

摘要

Femtocell networks promise improvement in network quality and performance for dense wireless networks, but will suffer from inter-cell interference if resource management is not properly employed. This paper presents distributed joint resource allocation (sub-channel and power) to address co- and cross-tier interference issues in two-tier heterogeneous femtocell networks. Due to uncoordinated nature of femtocell base stations (HeNB) deployment, the interactions among self-interested HeNBs are formulated using game-theoretical tools. Then, we designed individual utility function for every HeNB in order to enforce cooperative behaviour among HeNBs as well as to avoid cross-tier interference towards macrocell user equipments within HeNB coverage. Based on the designed utility function, we propose a fully distributed adaptive learning algorithm with a proportional self-belief concept that can lead to correlated equilibrium with fast and decisive convergence. Finally, performance analysis on the proposed algorithm done in simulated environment showed positive results indicating improvements in terms of co- and cross-tier interference mitigation as compared to generic regret-based learning scheme and utility functions.