We study the problem of decentralized optimization over time-varying networks with strongly convex smooth cost functions. In our approach, nodes run a multi-step gossip procedure after making each gradient update, thus ensuring approximate consensus at each iteration, while the outer loop is based on accelerated Nesterov scheme. The algorithm achieves precision ε>0 in O(κgχlog2(1/ε)) communication steps and O(κglog(1/ε)) gradient computations at each node, where κg is the global function number and χ characterizes connectivity of the communication network. In the case of a static network, χ=1/γ where γ denotes the normalized spectral gap of communication matrix W. The complexity bound includes κg, which can be significantly better than the worst-case condition number among the nodes.
@article{arxiv.2009.11069,
title = {Towards accelerated rates for distributed optimization over time-varying networks},
author = {Alexander Rogozin and Vladislav Lukoshkin and Alexander Gasnikov and Dmitry Kovalev and Egor Shulgin},
journal= {arXiv preprint arXiv:2009.11069},
year = {2022}
}