Energy-Spectral-Efficiency Analysis and Optimization of Heterogeneous Cellular Networks: A Large-Scale User-Behavior Perspective

Heterogeneous cellular networks (HCNs) are capable of meeting the explosive mobile-traffic demands. However, the conventional base station (BS) deployment strategy is unsuitable for supporting the often unpredictable nonuniform mobile-traffic demands, as governed by the large-scale user behavior (LUB). This results in the inefficient exploitation of the system's resources. In this paper, we develop an analytical framework for characterizing the achievable energy-spectral-efficiency (ESE) of HCNs, which explicitly quantifies the relationship between the network's ESE and the randomly time-varying LUBs as well as other network deployment parameters. Specifically, we model the quantitative impact of the geographical mobile-traffic intensity, the load migration factor, the users' required service rate and the per-tier BS densities on the achievable ESE of the network, while considering the area-spectral-efficiency requirements. Importantly, a closed-form ESE expression is derived, which enables us to explicitly analyze the properties of the network's ESE. Furthermore, the optimal LUB-aware BS deployment strategy is proposed for maximizing the ESE under specific outage constraints. Using numerical simulations, we verify the accuracy of the analytical ESE expression and quantify the impact of several relevant system parameters on the achievable ESE. Furthermore, we evaluate the achievable ESE performance of the network under diverse time-varying LUB scenarios. Our work, therefore, provides valuable insights for designing future ultradense HCNs.