Multi-Dimensional Resource Allocation Aware User Clustering in User-Centric Overlapped Virtual Cells

It is highly possible that the fifth generation (5G) wireless communication system adopts a ultra-dense networks (UDNs) which can provide extremely high throughput and data rates. However, due to irregular deployment of numerous base stations (BSs), the interference environment becomes rather complex, thus causing severe inter-cell interference and spectrum deficiency. Virtual cell can be foreseen as an effective way to solve such challenges and to improve spectrum efficiency (SE) by cooperative and coordinated transmission of the BS in virtual cells. However, virtual cell with overlapped BSs has constraints set by the antenna number of transmitter and receiver, so we propose a new user clustering scheme based on a mathematic graph approach for such case. Considering a user can act as the role of both interfering and interfered, we design a novel edge weight with a balanced cooperative transmission strategy. We also incorporate into edge weight design the gain of space and power dimensional cooperation in the sub-sequent resource allocation. To form user clustering dynamically, we utilize the clustering algorithm entitled Greedy Tree Growing Algorithm (GTGA) which is based on the K-means clustering algorithm. The proposed algorithm adjusts the size of user-clusters by a threshold depending on the statistics of edge weight which is shown by simulation to follow a Gaussian distribution. The simulation results demonstrate that the proposed approach can effectively reduce the complexity, mitigate interference and improve SE of UDN.