Decentralized weighted sum rate maximization in mimo-ofdma femtocell networks

Abstract — We propose a decentralized strategy for downlink interference management in overlaid MIMO femtocell-macrocell deployments operating in the same band under OFDMA access. The resource allocation, consisting of MIMO precoders at the macro base station (MBS) and femto access points (FAPs) and possibly the resource block (RB) assignment, is designed to maximize the Weighted Sum Rate (WSR) of the system. Being the overall problem non-convex, we propose to solve it in a decentralized way where each transmitter designs its resource allocation using interference prices exchanged with other transmitters (FAPs and MBS) by exploiting the wired backhaul connection. The problem includes a rate restriction per FAP that accounts for the limited bitrate provided by the Internet Service Provider (ISP) wired backhaul. Two efficient algorithms are provided for: 1) optimizing MIMO precoders and 2) optimizing both MIMO precoders and RB assignment. Numerical results compare decentralized strategies based on pricing exchange and on pure competition in dense FAP deployments. Keywords- MIMO; interference price; femtocell; limited-capacity backhaul; self-organisation; game theory I