Optimal distributed stochastic mirror descent for strongly convex optimization

doi:10.1016/j.automatica.2017.12.053

CSpace > 系统科学研究所

	Optimal distributed stochastic mirror descent for strongly convex optimization
	Yuan, Deming 1,2; Hong, Yiguang3 ; Ho, Daniel W. C.4; Jiang, Guoping 2
	2018-04-01
发表期刊	AUTOMATICA
ISSN	0005-1098
卷号	90 页码:196-203
摘要	In this paper we consider convergence rate problems for stochastic strongly-convex optimization in the non-Euclidean sense with a constraint set over a time-varying multi-agent network. We propose two efficient non-Euclidean stochastic subgradient descent algorithms based on the Bregman divergence as distance-measuring function rather than the Euclidean distances that were employed by the standard distributed stochastic projected subgradient algorithms. For distributed optimization of non-smooth and strongly convex functions whose only stochastic subgradients are available, the first algorithm recovers the best previous known rate of O(ln(T)/T) (where T is the total number of iterations). The second algorithm is an epoch variant of the first algorithm that attains the optimal convergence rate of O(1/T), matching that of the best previously known centralized stochastic subgradient algorithm. Finally, we report some simulation results to illustrate the proposed algorithms. (C) 2018 Elsevier Ltd. All rights reserved.
关键词	Distributed stochastic optimization Strong convexity Non-Euclidean divergence Mirror descent Epoch gradient descent Optimal convergence rate
DOI	10.1016/j.automatica.2017.12.053
语种	英语
资助项目	Natural Science Fund for Excellent Young Scholars of Jiangsu Province[BK20170099] ; National Natural Science Foundation of China[61573344] ; National Natural Science Foundation of China[61733018] ; National Natural Science Foundation of China[61374180] ; Research Grants Council of the Hong Kong Special Administrative Region, China[CityU 11300415]
WOS研究方向	Automation & Control Systems ; Engineering
WOS类目	Automation & Control Systems ; Engineering, Electrical & Electronic
WOS记录号	WOS:000427217600022
出版者	PERGAMON-ELSEVIER SCIENCE LTD
引用统计
文献类型	期刊论文
条目标识符	http://ir.amss.ac.cn/handle/2S8OKBNM/29756
专题	系统科学研究所
通讯作者	Yuan, Deming
作者单位	1.Nanjing Univ Sci & Technol, Sch Automat, Nanjing 210094, Jiangsu, Peoples R China 2.Nanjing Univ Posts & Telecommun, Coll Automat, Nanjing 210023, Jiangsu, Peoples R China 3.Chinese Acad Sci, Acad Math & Syst Sci, Key Lab Syst & Control, Beijing 100190, Peoples R China 4.City Univ Hong Kong, Dept Math, Kowloon, Hong Kong, Peoples R China
推荐引用方式 GB/T 7714	Yuan, Deming,Hong, Yiguang,Ho, Daniel W. C.,et al. Optimal distributed stochastic mirror descent for strongly convex optimization[J]. AUTOMATICA,2018,90:196-203.
APA	Yuan, Deming,Hong, Yiguang,Ho, Daniel W. C.,&Jiang, Guoping.(2018).Optimal distributed stochastic mirror descent for strongly convex optimization.AUTOMATICA,90,196-203.
MLA	Yuan, Deming,et al."Optimal distributed stochastic mirror descent for strongly convex optimization".AUTOMATICA 90(2018):196-203.