The Weighted Sum Rate Maximization in MIMO Interference Networks: The Minimax Lagrangian Duality and Algorithm

Abstract

We take a new perspective on the weighted sum-rate maximization in multiple-input multiple-output (MIMO) interference networks, by formulating an equivalent max-min problem. This seemingly trivial reformulation has significant implications: the Lagrangian duality of the equivalent max-min problem provides an elegant way to establish the sum-rate duality between an interference network and its reciprocal when such a duality exists, and more importantly, suggests a novel iterative minimax algorithm for the weighted sum-rate maximization. Moreover, the design and convergence proof of the algorithm use only general convex analysis. They apply and extend to any max-min problems with similar structure, and thus provide a general class of algorithms for such optimization problems. This paper presents a promising step and lends hope for establishing a general framework based on the minimax Lagrangian duality for characterizing the weighted sum-rate and developing efficient algorithms for general MIMO interference networks.