文献类型：期刊文献

英文题名：Differentially Private Distributed Optimization With an Event-Triggered Mechanism

作者：Mao, Shuai[1];Yang, Minglei[2];Yang, Wen[2];Tang, Yang[2];Zheng, Wei Xing[3];Gu, Juping[1];Werner, Herbert[4]

机构：[1]Nantong Univ, Dept Elect Engn, Nantong 226019, Peoples R China;[2]East China Univ Sci & Technol, Key Lab Smart Mfg Energy Chem Proc, Minist Educ, Shanghai 200237, Peoples R China;[3]Western Sydney Univ, Sch Comp Data & Math Sci, Sydney, NSW 2751, Australia;[4]Hamburg Univ Technol, Inst Control Syst, D-21073 Hamburg, Germany

年份：2023

卷号：70

期号：7

起止页码：2943

外文期刊名：IEEE TRANSACTIONS ON CIRCUITS AND SYSTEMS I-REGULAR PAPERS

收录：;EI(收录号：20231814033253);WOS:【SCI-EXPANDED(收录号:WOS:000976065300001)】；

基金：This work was supported in part by the Natural Science Foundation of China under Grant 62233005, Grant 62122026, and Grant 62293502, in part by the Smart Grid Joint Fund of State Key Program of National Natural Science Foundation of China under Grant U2066203, in part by the Program of Shanghai Academic Research Leader, China, under Grant 20XD1401300, in part by the Sino-German Center for Research Promotion under Grant M-0066, in part by the Project of Key Research and Development Plan of Jiangsu Province under Grant BE2021063, and in part by the Programme of Introducing Talents of Discipline to Universities (the 111 Project) under Grant B17017.

语种：英文

外文关键词：Distributed optimization; event-triggered mechanism; differential privacy

摘要：This study concentrates on the differential private distributed optimization problem with an event-triggered mechanism, whose goals include preserving the privacy of agents' initial states and local cost functions and improving communication efficiency. A distributed event-triggered mechanism is integrated into the differentially private subgradient-push distributed optimization algorithm and then a new algorithm named as DP-ETSP is designed, where the real-time information propagation among agents is avoided. Additionally, under the proposed event-triggered mechanism, an analysis of mean-square consensus and optimality over time-varying directed networks is made when the added Laplace noises meet some specific decaying conditions. Convergence rate results are further established under a specific stepsize, which are equal to the rate of stochastic gradient-push algorithm without event-triggered communication. Moreover, the differential privacy preservation performance is analyzed and the rule for selecting privacy level is discussed. Finally, the feasibility and effectiveness of DP-ETSP are verified in two simulation cases.