Primal-dual $\varepsilon$-Subgradient Method for Distributed Optimization

TL;DR

This paper introduces a primal-dual $\varepsilon$-subgradient method for distributed optimization that handles nondifferentiable objectives and inexact subgradients, providing convergence guarantees and improved transient behavior.

Contribution

It proposes a novel projected primal-dual dynamics using approximate subgradients, with theoretical analysis of error bounds and convergence conditions for distributed optimization.

Findings

The method converges with an error depending on subgradient approximation accuracy.

Exact solutions are achievable when approximation errors are sufficiently small.

A componentwise normalized variant improves transient convergence behavior.

Abstract

This paper studies the distributed optimization problem when the objective functions might be nondifferentiable and subject to heterogeneous set constraints. Unlike existing subgradient methods, we focus on the case when the exact subgradients of the local objective functions can not be accessed by the agents. To solve this problem, we propose a projected primal-dual dynamics using only the objective function's approximate subgradients. We first prove that the formulated optimization problem can generally be solved with an error depending upon the accuracy of the available subgradients. Then, we show the exact solvability of this distributed optimization problem when the accumulated approximation error of inexact subgradients is not too large. After that, we also give a novel componentwise normalized variant to improve the transient behavior of the convergent sequence. The effectiveness of our algorithms is verified by a numerical example.