Traffic-aware link scheduling with interference alignment for multi-user MIMO networks

In this paper, we revisit the minimum-length link scheduling problem (MLSP) in wireless networks. Unlike previous scheduling solutions that adopt the interference avoidance approach, we embrace interference among neighboring links via a new technique of interference alignment (IA). Specifically, we use IA to cancel the interference among cooperative neighboring links, while controlling uncoordinated interference from non-neighboring links by traditional link scheduling. This novel approach of scheduling is modeled as an integer linear program based on an extended multiple-input and multiple-output (MIMO)-Pipe model, which not only captures the complex interference relationship between links in multi-user MIMO networks with IA, but also provides a realistic understanding of the interaction of achievable rate and suffered interference both for IA links and non-IA links in multi-user MIMO networks. By resolving the formulated programs, we further investigate the effect of node density and antenna number on the performance gain brought by IA cooperations. Simulation results show that the benefits of IA in MLSP problem generally increase with node density, but decrease with the number of antennas.