Competitive design of multiuser MIMO interference systems based on game theory: A unified framework

In this paper we focus on the maximization of the information rates subject to transmit power constraints for noncooperative multiple-input multiple-output (MIMO) systems, using the same physical resources, i.e., time, bandwidth and space. To derive decentralized solutions that do not require any cooperation among the systems, the optimization problem is formulated as a static noncooperative game. The analysis of the game for arbitrary MIMO interference channels is quite involved, since it requires the study of a set of nonlinear nondifferentiable matrix-valued equations, based on the MIMO waterfilling solution. To overcome this difficulty, we provide a new interpretation of the waterfilling operator, for the general MIMO multiuser case, as a matrix projection. This key result allows us to simplify the study of the game and to obtain sufficient conditions for both uniqueness of the Nash equilibrium (NE) and convergence of the proposed totally asynchronous distributed algorithms. The proposed approach provides a general framework that encompasses all previous works, mostly concerned with the particular case of SISO Gaussian frequency-selective interference channel.