Fast and Optimal Power Control Games in Multiuser MIMO Networks

TL;DR

This paper introduces a distributed game-theoretic power control algorithm for multiuser MIMO networks that improves equilibrium uniqueness and convergence speed, with modifications to enhance sum-rate performance.

Contribution

The paper presents a novel, modified power control algorithm that guides the game to an optimal equilibrium, improving sum-rate and convergence speed in multiuser MIMO networks.

Findings

Noticeable sum-rate improvement with the modified algorithm

Enhanced convergence speed using the inexact method

Greater uniqueness probability of Nash equilibria

Abstract

In this paper, we analyze the problem of power control in a multiuser MIMO network, where the optimal linear precoder is employed in each user to achieve maximum point- to-point information rate. We design a distributed power control algorithm based on the concept of game theory and contractive functions that has a couple of advantages over the previous designs (e.g. more uniqueness probability of Nash equilibria and asynchronous implementation). Despite these improvements, the sum-rate of the users does not increase because the proposed algorithm can not lead the power control game to an efficient equilibrium point. We solve this issue by modifying our algorithm such that the game is led to the equilibrium that satisfies a particular criterion. This criterion can be chosen by the designer to achieve a certain optimality among the equilibria. Furthermore, we propose the inexact method that helps us to boost the convergence speed of our modified algorithms. Lastly, we show that pricing algorithms can also be a special case of our modified algorithms. Simulations show a noticeable improvement in the sum-rate when we modify our proposed algorithm.