Statistical QoS-driven power control and source adaptation for D2D communications.

Xiang Mi,Limin Xiao,Ming Zhao,Xibin Xu,Shidong Zhou,Jing Wang

DOI: https://doi.org/10.1109/MILCOM.2016.7795344

2016-01-01

Abstract:Device-to-device (D2D) communication has been considered as a key technique in the fifth generation cellular networks. Power control is critical to ensure satisfactory performance. However, most existing power control schemes focus on delay-unaware metrics, such as throughput and power consumption, thus only applying to delay-insensitive scenarios. To overcome this problem, we in this paper consider a network where multiple D2D pairs with delay quality-of-service (QoS) requirements share a certain spectrum band. We propose an effective power control and source adaptation policy that maximizes the system throughput while satisfying each D2D pair's delay QoS requirement. Specifically, we formulate a constraint optimization problem which aims at maximizing the QoS-guaranteed system throughput. By using convex approximation, we convert the original non-convex optimization problem into a convex one. The optimal solution is derived via the Lagrangian approach. Moreover, to solve for the dual function with given dual variables, we propose an iterative algorithm based on alternating optimization and prove its convergence to the global optimum. Simulation results show that our proposed policy can improve the QoS-guaranteed system throughput significantly compared with the baselines.