ARM: Anonymous rating mechanism for discrete power control

Abstract

Wireless interference management through continuous power control has been extensively studied in the literature. However, practical systems often adopt discrete power control with a limited number of power levels and MCSs (Modulation Coding Schemes). In general, discrete power control is NP-hard due to its combinatorial nature. To tackle this challenge, we propose an innovative approach of interference management: ARM (Anonymous Rating Mechanism). Inspired by the successes of the simple Anonymous Rating Mechanism in Internet and E-commerce, we develop ARM as distributed near-optimal algorithm for solving the discrete power control problem (i.e., the joint scheduling, power allocation, and modulation coding adaption problem) under the physical interference model. We show that ARM achieves a close-to-optimal network throughput with a very low control overhead. We also characterize the performance gap of ARM due to the loss of rating information, and study the trade-off between such gap and the convergence time of ARM. Through comprehensive simulations under various network scenarios, we find that the optimality gap of ARM is small and such a small gap can be achievable with only a small number of power levels. Furthermore, the performance degradation is marginal if only limited local network information is available.