Decentralized Beam Pair Selection in Multi-Beam Millimeter-Wave Networks

• Published in: IEEE transactions on communications 2018-06, Vol.66 (6), p.2722-2737
• Main Authors: Yanping Liu, Xuming Fang, Ming Xiao, Mumtaz, Shahid
• Format: Article
• Language: English
• Subjects: Algorithms; beam pair selection; Complexity; Complexity theory; Computer simulation; Convergence; Exchanging; Game theory; Games; Heterogeneous multi-beam cloud radio access networks(HMBCRANs); Machine learning; Millimeter wave technology; Millimeter waves; millimeter-wave (mmWave) networks; multi-beam concurrent transmission; Optimization; Power consumption; Power line communications; Radio access networks; Stability analysis; State of the art
• ISSN: 0090-6778; 1558-0857; 1558-0857
• DOI: 10.1109/TCOMM.2018.2800756

Multi-beam concurrent transmission is one of promising solutions for a millimeter-wave (mmWave) network to provide seamless handover, robustness to blockage, and continuous connectivity. Nevertheless, one of the major obstacles in multi-beam concurrent transmissions is the optimization of beam pair selection, which is essential to improve the mmWave network performance. Therefore, in this paper, we propose a novel heterogeneous multi-beam cloud radio access network (HMBCRAN) architecture which provides seamless mobility and coverage for mmWave networks. We also design a novel acquirement method for candidate beam pair links (BPLs) in HMBCRANs architecture, which reduces user power consumption, signaling overhead, and overall time consumption. Based on HMBCRANs architecture and the resulted candidate BPLs for each user equipment, a beam pair selection optimization problem aiming at maximizing network sum rate is formulated. To find the solution efficiently, the considered problem is reformulated as a non-operative game with local interaction, which only needs local information exchanging among players. A decentralized algorithm based on HMBCRANs architecture and binary log-linear learning is proposed to obtain the optimal pure strategy Nash equilibrium of the proposed game, in which a concurrent multi-player selection scheme and an information exchanging protocol among players are developed to reduce the complexity and signal overheads. The stability, optimality, and complexity of the proposed algorithm are analyzed via theoretical and simulation method. The results prove that the proposed scheme has better convergence speed and sum rate against the state-of-the-art schemes.