Clustering for Interference Alignment in a Multiuser Interference Channel

Interference alignment (IA) has been shown as a promising technique to achieve the optimal capacity scaling of a multiuser interference channel at asymptotically high signal-to-noise-ratio (SNR). However, in practical communication systems, mitigating all interference via IA is not necessary since some of them only have negligible effect due to path loss. Moreover, IA feasibility constraint hinders its application to all users. Clustered IA provides a mechanism to mitigate the IA feasibility constraint and maximize the achievable rate. However, how to form IA clusters properly has not been well addressed. In this paper, we consider the problem of IA clustering in a multiuser interference channel at finite SNR. By exploiting the statistics of the channel state information (CSI), two novel clustering algorithms based on graph partitioning and heuristic are proposed. It is shown that the proposed schemes lead to significant gain in achievable rate when compared with non-cooperative transmission scheme.