Interference alignment using quadratic programming

Interference alignment (IA) in the multi-user MIMO interference channel with constant channel coefficients was considered. An iterative algorithm for finding the interference alignment solutions was proposed for the case where all interference alignment multiplexing gains are one. The perfect alignment constraint was relaxed in order to obtain better sum rate performance. The problem was first converted to a convex quadratic program problem based on the small step size assumption and then all the precoding and decoding matrices were updated iteratively. Simulation results showed that the proposed algorithm achieves better sum rate performance compared with minimum interference leakage (MIL) algorithm at lower SNR, and almost the same at higher SNR.