Scalable Average Consensus with Compressed Communications

TL;DR

This paper introduces a scalable decentralized average consensus algorithm that uses compressed communication to efficiently reach agreement across large networks, with proven convergence and practical validation.

Contribution

It presents a novel consensus method that accommodates biased compression operators and arbitrary network topologies, improving scalability and communication efficiency.

Findings

Algorithm converges to the true average

Scales linearly with network size n

Numerical experiments confirm theoretical results

Abstract

We propose a new decentralized average consensus algorithm with compressed communication that scales linearly with the network size n. We prove that the proposed method converges to the average of the initial values held locally by the agents of a network when agents are allowed to communicate with compressed messages. The proposed algorithm works for a broad class of compression operators (possibly biased), where agents interact over arbitrary static, undirected, and connected networks. We further present numerical experiments that confirm our theoretical results and illustrate the scalability and communication efficiency of our algorithm.