Coalition Game-Based Beamwidth Selection for D2D Users Underlying Ultra Dense mmWave Networks

With the advent of 5G mobile communication era, ultra-dense millimeter wave (mmWave) networks have become an important architecture to improve network performance. However, the susceptibility to attenuation becomes a bottleneck for long-distance transmission of mmWave links. D2D communication can alleviate the bottleneck utilizing its short-range direct communication and its ability to offload traffic for base stations. The adoption of high- directivity signal beams at cellular users (CUs) and D2D users (DUs) reusing the full mmWave frequency band can greatly improve the spectrum efficiency, but at the same time, it will also cause unexpected co-channel interference. This paper proposes a beamwidth selection scheme for D2D pairs underlying mmWave network, which automatically selects the optimal beamwidth for DUs via non-cooperative coalition game, thereby reducing the interference between DUs and CUs and increasing network capacity. In the proposed algorithm, D2D users continuously update their beamwidth strategies to explore and evaluate the potential coalition structure in the direction of improving system capacity. Finally, a stable partition structure with the optimal network utility is reached. Simulation results show that the proposed algorithm can achieve the close performance compared with particle swarm optimization (PSO) algorithm, but the complexity is sharply reduced.