Spectrum Allocation for Hyper-Dense Small Cell Networks: A Partially-Distributed Approach

In this paper, we investigate the problem of spectrum allocation for small cell networks with dense deployment of access points in large quantities. Since the formulated problem is intractable in the context of large-scale networks, finding the optimal solution is very complex. To tackle this issue, here, we develop a partially-distributed approach to achieve desirable performance. Specifically, the proposed scheme mainly consists of three phases: dividing network into disjoint clusters, sub-channel assignment within each cluster and interference elimination among clusters. Finally, numerical results show that the partially-distributed solution outperforms some existing approaches. Meanwhile, the proposed scheme can achieve desirable performance, which is close to that of the global optimization method, with a faster convergence speed.