Distributed Push-Sum Algorithm for Multi-Agent optimization Via One-Point Gradient Estimator

This paper studies the distributed optimization problem over a multi-agent network that consists of multiple nodes, where the objective function of the problem is the sum of the local objective functions of nodes. The goal of the network is to minimize the collective objective function through local...

Ausführliche Beschreibung

Gespeichert in:
Bibliographische Detailangaben
Veröffentlicht in:Chinese Control Conference S. 6013 - 6018
Hauptverfasser: Wang, Cong, Yuan, Deming, Zhang, Baoyong
Format: Tagungsbericht
Sprache:Englisch
Veröffentlicht: Technical Committee on Control Theory, Chinese Association of Automation 01.07.2019
Schlagworte:
Automation
Convergence
Distributed optimization
Game theory
Linear programming
Multi-agent systems
One-Point Gradient Estimator
Optimization
Push-Sum
Smart grids
ISSN:1934-1768
Online-Zugang:Volltext
Tags: Tag hinzufügen
Keine Tags, Fügen Sie den ersten Tag hinzu!
Beschreibung
Zusammenfassung:This paper studies the distributed optimization problem over a multi-agent network that consists of multiple nodes, where the objective function of the problem is the sum of the local objective functions of nodes. The goal of the network is to minimize the collective objective function through local interactions among nodes. We propose an efficient distributed optimization algorithm that is based on push-sum algorithm and one-point gradient estimator, which removes the needs for doubly stochastic weight matrix and the information of subgradients of the objective functions. The convergence of the algorithm is established by deriving an O(1/T 1/6 ) (with T being the number of iterations) rate of convergence for the proposed algorithm.
ISSN:1934-1768
DOI:10.23919/ChiCC.2019.8865859