Low-Rank Matrix Recovery for Topological Interference Management

• Published in: International Conference on Wireless Communications and Signal Processing pp. 860 - 864
• Main Authors: Jiang, Xue, Zheng, Baoyu, Wang, Lei, Hou, Xiaoyun
• Format: Conference Proceeding
• Language: English
• Published: IEEE 21.10.2020
• Subjects: Interference; Interference alignment; Interference channels; low-rank matrix recovery; Mathematical model; Matrix decomposition; Minimization; Receivers; sparse and low-rank matrix decomposition; Sparse matrices; topological interference management
• ISSN: 2472-7628
• DOI: 10.1109/WCSP49889.2020.9299683

Low-rank matrix completion plays an important role in modeling and computational methods for topological interference management (TIM), but in many applications affected by noise, these networks topological information cannot be fully directly observed, and one encounters the problem of recovering the topology information matrix given only incomplete observations. To reduce the noise effect and make topological interference management scalable in multi-user wireless networks, we present an algorithmic approach to investigating the achievable degrees-of-freedom (DoF) by recasting the topological interference management problem as a low-rank matrix recovery (LRMR) problem. Furthermore, we propose two TIM algorithms to solve the low-rank matrix recovery problem for partially connected interference channels. One is nuclear norm and l 1- norm minimization TIM algorithm and the other is nuclear norm and Frobenius norm minimization TIM algorithm. Finally, the simulation results show the proposed TIM algorithms to solve the low-rank matrix recovery problem is efficient.