Convergence of Stochastic PDMM
Sebastian O. Jordan (Student TU Delft)
T.W. Sherson (TU Delft - Signal Processing Systems)
R. Heusdens (TU Delft - Signal Processing Systems, Netherlands Defence Academy)
More Info
expand_more
Other than for strictly personal use, it is not permitted to download, forward or distribute the text or part of it, without the consent of the author(s) and/or copyright holder(s), unless the work is under an open content license such as Creative Commons.
Abstract
In recent years, the large increase in connected devices and the data that are collected by these devices have caused a heightened interest in distributed processing. Many practical distributed networks are of heterogeneous nature, because different devices in the network can have different specifications. Because of this, it is highly desirable that algorithms operating within these networks can operate asynchronously, since in that case there is no need for clock synchronisation between the nodes, and the algorithm is not slowed down by the slowest device in the network. In this paper, we focus on the primal-dual method of multipliers (PDMM), which is a promising distributed optimisation algorithm that is suitable for distributed optimisation in heterogeneous networks. Most theoretical work that can be found in existing literature focuses on synchronous versions of PDMM. In this work, we prove the convergence of stochastic PDMM, which is a general framework that can model variations such as asynchronous PDMM and PDMM with transmission losses.