A Fully Distributed and Clustered Learning of Power Control in User-Centric Ultra-Dense HetNets

In this paper, we investigate a power control of uplink connection in the user-centric ultra-dense heterogeneous networks (HetNets), which are studied as different types of access points (APs). The main objective of this investigation is to engage users in a cooperative game to confront with the problem of per-user power control and coordinate multi-user interferences. Thus, we formulate the optimization problem as a sum of the users' cooperative energy efficiency (EE) functions. Firstly, a realistic and new model of the user device's power consumption is proposed. This model includes the power used for operating modes and signal processing of mobile devices during the uplink data transmission. Secondly, the EE optimization problem is formulated by maximizing the sum of the users' cooperative EE function subject to each user's quality-of-service (QoS) and a power constraint. Then, we propose a fully distributed and clustered learning scheme for solving the optimization problem, where neighboring users are clustered to engage in the cooperative game of power control in order to coordinate the multi-user interferences. It is theoretically proved that the size of the clusters has an impact on the sum of the users' cooperative EE. Additionally, our scheme can achieve convergence with imperfect channel feedback and the local knowledge of the user's constraints. Finally, simulation results confirm the theoretical analysis and demonstrate the robust performance of the proposed scheme compared with two benchmark methods.