The Role of Transmitter Cooperation in Linear Interference Networks with Block Erasures

Abstract: In this work, we explore the potential and optimal use of transmitter cooperation in wireless interference networks with deep fading conditions. We consider a linear interference network with K transmitter-receiver pairs, where each transmitter can be connected to two neighboring receivers. Long-term fluctuations (shadow fading) in the wireless channel can lead to any link being erased with probability p. Each receiver is interested in one unique message that can be available at two transmitters. The considered rate criterion is the average per user degrees of freedom (puDoF) as K goes to infinity. Prior to this work, the optimal assignment of messages to transmitters were identified in the two limits as p goes to 0 and as p goes to 1. We identify new schemes that achieve average puDoF values that are higher than the state of the art for a significant part of the range 0 < p < 1. The key idea to our results is to understand that the role of cooperation shifts from increasing the probability of delivering a message to its intended destination at high values of p, to interference cancellation at low values of p. Our schemes are based on an algorithm that achieves the optimal DoF value in any network realization, when restricted to a given message assignment as well as the use of zero-forcing schemes.