Power Allocation for Average Consensus in Wireless Networks with Link Failures

Abstract:
We study the problem of power allocation for average consensus in wireless networks with unidirectional, stochastic link failures. While several prior works address consensus protocols under time-varying or random topologies, they often assume symmetric (undirected) connectivity or do not explicitly consider power allocation. In contrast, we model the network as a weighted directed graph where link success probabilities depend asymmetrically on the nodes’ transmit power. We propose an optimization framework that allocates power to maximize the algebraic connectivity of the expected Laplacian, leading to faster convergence under a total power constraint.