Distributed weighted sum throughput maximization in multi-cell wireless networks

This paper addresses the weighted sum throughput maximization problem on the interference channel. It proposes a sub-optimal algorithm that enables to perform a distributed power allocation, which performs well at any SINR regime. The sub-optimal method consists of two phases: first, identify the links that should be set to zero, and then, operate in the high SINR regime for the remaining links. In the single channel case, this simple, fully distributed algorithm is shown to perform almost as well as the centralized optimal algorithm. In the multi-channel case, the algorithm is more efficient than iterative water-filling in order to decrease the buffer size.