System-Level Analysis of Heterogeneous Networks Under Imperfect Traffic Hotspot Localization

In this paper, we study the impact of imperfect small-cell positioning with respect to traffic hotspots in cellular networks. To derive the throughput distribution in macrocells and small cells, we first perform static-level analysis of the system considering a nonuniform distribution of user locations. Second, we introduce the dynamics of the system, which is characterized by random arrivals and departures of users after finite service duration, with the service rates and distribution of radio conditions outfitted from the first part of the work. When dealing with the dynamics of the system, macrocells and small cells are modeled by multiclass processor sharing (PS) queues. Macrocells and small cells are assumed to operate in the same bandwidth. Consequently, they are coupled due to the mutual interference generated by each cell to the other. We derive several performance metrics such as the mean flow throughput and the gain, if any, generated from deploying small cells to manage traffic hotspots. Our results show that in the case that the hotspot is near the macro base station (BS), even a perfect positioning of the small cell will not yield improved performance due to the high interference experienced at macro- and small-cell users. However, in case the hotspot is located far enough from the macro-BS, performing errors in small-cell positioning is tolerated (since related results show positive gains), and it is still beneficial in offloading traffic from the congested macrocell. The best performance metrics also depend on several other important factors such as the users' arrival intensity, the capacity of the cell, and the size of the traffic hotspot.