Energy Efficient Competitive Resource Allocation in MIMO networks

TL;DR

This paper studies energy-efficient resource allocation in MIMO networks, establishing conditions for stable equilibria and proposing decentralized algorithms that outperform existing methods in terms of simplicity and effectiveness.

Contribution

It introduces new conditions for the existence and uniqueness of Nash equilibria in energy-efficient MIMO games and develops a decentralized algorithm with improved performance.

Findings

Conditions outperform state-of-the-art in OFDM networks

Decentralized algorithm converges reliably

Tighter theoretical bounds via contraction mapping

Abstract

This paper considers the competitive resource allocation problem in Multiple-Input Multiple-Output (MIMO) interfering channels, when users maximize their energy efficiency. Considering each transmitter-receiver pair as a selfish player, conditions on the existence and uniqueness of the Nash equilibrium of the underlying noncooperative game are obtained. A decentralized asynchronous algorithm is proven to converge towards this equilibrium under the same conditions. Two frameworks are considered for the analysis of this game. On the one hand, the game is rephrased as a Quasi-Variational Inequality (QVI). On the other hand, the best response of the players is analyzed in light of the contraction mappings. For this problem, the contraction approach is shown to lead to tighter results than the QVI one. When specializing the obtained results to OFDM networks, the obtained conditions appear to significantly outperform state-of-the-art works, and to lead to much simpler decentralized algorithms. Numerical results finally assess the obtained conditions in different settings.