Joint optimization of stream allocation and beamforming and combining weights for the MIMO interference channel

We propose an algorithm to allocate streams and find the corresponding beamforming and combining weights that maximize the sum rate of a set of interfering multiple-input multiple-output (MIMO) links. Our algorithm iteratively computes the beamforming and combining weights of each link and determines how many streams, if any, are allocated to each link. Assigning zero streams to a link is desirable whenever the interference introduced by activating that link outweighs the throughput contributed by the link. We present numerical results to illustrate that our algorithm achieves a sum rate higher than previously reported algorithms at high interference, and that it achieves similar performance to the top-performing algorithms at medium and low interference. In one high-interference example with many links, our algorithm achieves a 65% higher sum rate than the best-known alternative.