Load-Aware Dynamic Access for Ultra-Dense Small Cell Networks: A Hypergraph Game Theoretic Solution

In this paper we research the load-aware channel allocation in ultra-dense small cell networks based on the hypergraph interference model. Cumulative interference is a hard nut to crack in ultra-dense networks because of the intensive distribution of low-powered and small-coverage small cells. The traditional binary graph interference model, which mainly focused on the pair-wise strong interference relation, can not capture the cumulative interference. Therefore, we use the hypergraph model to accurately describe the complex interference relation among small cells. The applications of hypergraph in wireless networks is in its infant stage. Considering the practical traffic demands of small cells, they can access multiple channels. To cope with this problem, we formulate the multi-channel access problem as a local altruistic hypergraph game and prove that it is an exact potential game, which admits at least one pure strategy Nash Equilibrium. To overcome the complexity of the centralized method and the constraint on the direct information exchange among small cells in hyperedges, a cloud-based centralized-distributed model is utilized. With the information shared in the cloud, a centralized-distributed learning algorithm can quickly search the Nash Equilibrium. The simulation results show that the proposed algorithm is superior to the existing binary graph-based schemes and significantly improves the communication efficiency.