Accelerated consensus via Min-Sum Splitting
نویسندگان
چکیده
•Common choice is Metropolis-Hastings: W M H i j = { 1/(2dmax) if {i , j } ∈ E 1−di /(2dmax) if i = j 0 otherwise •Rate of convergence is controlled by ρ(W −11T /n). • min{ρ(W −11T /n) : W symmetrical} is a convex problem (SDP). •Optimal matrix yields slow rate O(D2), achieved by W M H . •Lower-bound: Ω(D), where D is graph diameter. •To get fast rates, two approaches have been developed independently: Lifted Markov chains5 and multi-step gradient methods1. Q. Can we get fast rates with Min-Sum?
منابع مشابه
Additivity of Heegaard genera of bounded surface sums
Let M be a surface sum of 3-manifolds M1 and M2 along a bounded connected surface F and ∂i be the component of ∂Mi containing F . If Mi has a high distance Heegaard splitting, then any minimal Heegaard splitting of M is the amalgamation of those of M,M and M∗, where M i = Mi \ ∂i × I, and M∗ = ∂1 × I ∪F ∂2 × I. Furthermore, once both ∂i \ F are connected, then g(M) = Min {
متن کاملHigher order operator splitting methods via Zassenhaus product formula: Theory and applications
In this paper, we contribute higher order operator-splitting method improved by Zassenhaus product. We apply the contribution to classical and iterative splitting methods. The underlying analysis to obtain higher order operator-splitting methods is presented. While applying the methods to partial differential equations, the benefits of balancing time and spatial scales are discussed to accelera...
متن کاملA Douglas-Rachford Type Primal-Dual Method for Solving Inclusions with Mixtures of Composite and Parallel-Sum Type Monotone Operators
In this paper we propose two different primal-dual splitting algorithms for solving inclusions involving mixtures of composite and parallel-sum type monotone operators which rely on an inexact Douglas-Rachford splitting method, however applied in different underlying Hilbert spaces. Most importantly, the algorithms allow to process the bounded linear operators and the set-valued operators occur...
متن کاملA multi-block alternating direction method with parallel splitting for decentralized consensus optimization
Decentralized optimization has attracted much research interest for resource-limited networked multi-agent systems in recent years. Decentralized consensus optimization, which is one of the decentralized optimization problems of great practical importance, minimizes an objective function that is the sum of the terms from individual agents over a set of variables on which all the agents should r...
متن کاملSolving Systems of Monotone Inclusions via Primal-dual Splitting Techniques
In this paper we propose an algorithm for solving systems of coupled monotone inclusions in Hilbert spaces. The operators arising in each of the inclusions of the system are processed in each iteration separately, namely, the single-valued are evaluated explicitly (forward steps), while the set-valued ones via their resolvents (backward steps). In addition, most of the steps in the iterative sc...
متن کامل