Pareto optimization for MIMO interference channel

Interference alignment (IA) is a transmission technique for exploiting all available degrees of freedom (DoF) in the frequency- or time-selective interference channel with an arbitrary number of users, which can achieve linear capacity scaling with number of users. However, standard IA approach has limit performance in the low-to-moderate signal-to-noise ratio (SNR) region. Recently, linear combination designs have been proposed to improve performance of sum rate at finite SNR which jointly considers the interference and the signal optimization simultaneously, but is sensitive to the weight selection. Therefore, it motivates us to propose a normal-boundary intersection (NBI)-based design that jointly optimizes the rank of interference and its power with evenly spread parameters. The underlying problem is transferred into limited subproblem that can be efficiently solved. Simulation results validate the effectiveness of the proposed algorithm that achieves the Pareto optimality and provides evenly-distributed Pareto frontiers.