Distributed Gradient Methods with Variable Number of Working Nodes

TL;DR

This paper introduces a distributed gradient method where nodes probabilistically participate in updates, reducing communication and computation costs while maintaining convergence to the optimal solution under certain conditions.

Contribution

It proposes a novel probabilistic activation scheme for distributed gradient methods that ensures convergence and reduces computational and communication costs.

Findings

Converges in mean square sense when activation probability approaches one.

Achieves linear convergence rate with linearly increasing activation probability.

Significantly reduces communication and gradient evaluations compared to standard methods.

Abstract

We consider distributed optimization where $N$ nodes in a connected network minimize the sum of their local costs subject to a common constraint set. We propose a distributed projected gradient method where each node, at each iteration $k$, performs an update (is active) with probability $p_k$, and stays idle (is inactive) with probability $1-p_k$. Whenever active, each node performs an update by weight-averaging its solution estimate with the estimates of its active neighbors, taking a negative gradient step with respect to its local cost, and performing a projection onto the constraint set; inactive nodes perform no updates. Assuming that nodes' local costs are strongly convex, with Lipschitz continuous gradients, we show that, as long as activation probability $p_k$ grows to one asymptotically, our algorithm converges in the mean square sense (MSS) to the same solution as the standard distributed gradient method, i.e., as if all the nodes were active at all iterations. Moreover, when $p_k$ grows to one linearly, with an appropriately set convergence factor, the algorithm has a linear MSS convergence, with practically the same factor as the standard distributed gradient method. Simulations on both synthetic and real world data sets demonstrate that, when compared with the standard distributed gradient method, the proposed algorithm significantly reduces the overall number of per-node communications and per-node gradient evaluations (computational cost) for the same required accuracy.