Fast Algorithms for Distributed Optimization and Hypothesis Testing: A Tutorial

TL;DR

This tutorial presents fast algorithms for distributed optimization and hypothesis testing, demonstrating linear convergence times relative to network size by leveraging recent advances in average consensus algorithms.

Contribution

It introduces a unified approach to achieve linear-time convergence in distributed problems using a novel application of linear-time average consensus algorithms.

Findings

Convergence times scale linearly with network size.

Algorithms outperform previous methods in speed.

Applicable to various distributed optimization tasks.

Abstract

We consider several problems in the field of distributed optimization and hypothesis testing. We show how to obtain convergence times for these problems that scale linearly with the total number of nodes in the network by using a recent linear-time algorithm for the average consensus problem.