A distributed interference alignment algorithm using min‐maxing strategy

AbstractInterference alignment is a joint‐transmission strategy that significantly increases the sum rate of interference channels at high signal‐to‐noise ratios (SNRs). The recent iterative interference alignment approaches are incapable of guaranteeing the best sum‐rate performance with the increase of the SNR amongst different K‐user interference channels, especially at high SNR regime. In this paper, a new interference alignment algorithm is developed to improve the sum‐rate performance of K‐user multi‐input multi‐output (MIMO) and multi‐carrier interference channels by minimising the interference leakage and maximising the desired power concurrently, which is called by min‐maxing strategy. For a K‐user MIMO interference channel, we design transmit precoding matrices and receive decoding matrices through an efficient iterative algorithm based on min‐maxing strategy in a distributed way, in which each receiver maximises the desired signal power, whereas it preserves the minimum leakage interference as a constraint. This optimisation problem is reformulated and relaxed into a standard semidefinite programming form. The convergence of the proposed algorithm is proved as well. Furthermore, a simplified min‐maxing algorithm is proposed for rank‐deficient interference channels to achieve the targeted performance with less complexity. The numerical simulations show that the proposed algorithm proffers significant sum‐rate improvement in K‐user MIMO interference channels compared with the existing algorithms at high SNR regime. Moreover, the simplified algorithm matches the optimal performance in the systems of rank‐deficient channels. Finally, the developed min‐maxing algorithm has been extended to K‐user multi‐carrier interference channels, which outperform the previous approaches in terms of sum rate in several scenarios. Copyright © 2014 John Wiley & Sons, Ltd.