检索结果-内蒙古大学图书馆

asynchronous distributed optimization via dual decomposition for delay-constrained flying ad hoc networks

COMPUTER COMMUNICATIONS 2019年 137卷 70-80页

作者： Wen, Shaojie Huang, Chuanhe Wuhan Univ Sch Comp Sci Wuhan Hubei Peoples R China Collaborat Innovat Ctr Geospatial Technol Wuhan Hubei Peoples R China

The rapidly changing network topologies bring some challenges to optimize congestion control, routing and power control in delay-constrained flying ad hoc networks (FANETs). Traditional methods designed for mobile ad hoc networks (MANETs) underperform due to their poor adaptability to the changing network topologies. In this work, we propose an optimization framework with end-to-end (E2E) delay constraint, and formulate the problems as a utility maximization problem with respect to link capacity, E2E delay and flow conservation by coupling data flows and decomposing the E2E delay constraint. Further, to optimize such a non-linear maximization problem, we employ primal-dual decomposition and the delay scale factors to cast the maximization problem into several sub-problems with lower complexity. Subsequently, we propose a delay-constrained path selection algorithm and an asynchronous distributed optimization method which requires only the local channel information and outdated messages to optimize different parameters. Finally, to highlight the advantages of the proposed method, we further give the related analysis from the viewpoint of the different solutions of the maximization problem. Simulation results demonstrate that the proposed framework obtains significant network performance in terms of energy efficiency and packet timeout rate, compared with traditional methods.

关键词： Flying ad hoc networks E2E delay constraint Dual decomposition asynchronous distributed optimization

来源：评论

学校读者我要写书评

暂无评论

distributed asynchronous double accelerated optimization for ethylene plant considering delays

引用

Chinese Journal of Chemical Engineering 2025年第2期78卷 245-250页

作者： Ting Wang Zhongmei Li Wenli Du State Key Laboratory of Industrial Control Technology East China University of Science and Technology Key Laboratory of Smart Manufacturing in Energy Chemical Process Ministry of Education East China University of Science and Technology

Considering the complexity of plant-wide optimization for large-scale industries, a distributed optimization framework to solve the profit optimization problem in ethylene whole process is proposed. To tackle the delays arising from the residence time for materials passing through production units during the process with guaranteed constraint satisfaction, an asynchronous distributed parameter projection algorithm with gradient tracking method is introduced. Besides, the heavy ball momentum and Nesterov momentum are incorporated into the proposed algorithm in order to achieve double acceleration properties. The experimental results show that the proposed asynchronous algorithm can achieve a faster convergence compared with the synchronous algorithm.

关键词： asynchronous distributed optimization Heavy ball Inequality constraints Nesterov Plant-wide optimization

来源：评论

学校读者我要写书评

暂无评论

asynchronous Message-Passing and Zeroth-Order optimization Based distributed Learning With a Use-Case in Resource Allocation in Communication Networks

引用

IEEE TRANSACTIONS ON SIGNAL AND INFORMATION PROCESSING OVER NETWORKS 2024年 10卷 916-931页

作者： Behmandpoor, Pourya Moonen, Marc Patrinos, Panagiotis KU Leuven Univ STADIUS Ctr Dynam Syst Signal Proc & Data Analyt Dept Elect Engn ESAT B-3000 Leuven Belgium

distributed learning and adaptation have received significant interest and found wide-ranging applications in machine learning and signal processing. While various approaches, such as shared-memory optimization, multi-task learning, and consensus-based learning (e.g., federated learning and learning over graphs), focus on optimizing either local costs or a global cost, there remains a need for further exploration of their interconnections. This paper specifically focuses on a scenario where agents collaborate towards a common task (i.e., optimizing a global cost equal to aggregated local costs) while effectively having distinct individual tasks (i.e., optimizing individual local parameters in a local cost). Each agent's actions can potentially impact other agents' performance through interactions. Notably, each agent has access to only its local zeroth-order oracle (i.e., cost function value) and shares scalar values, rather than gradient vectors, with other agents, leading to communication bandwidth efficiency and agent privacy. Agents employ zeroth-order optimization to update their parameters, and the asynchronous message-passing between them is subject to bounded but possibly random communication delays. This paper presents theoretical convergence analyses and establishes a convergence rate for nonconvex problems. Furthermore, it addresses the relevant use-case of deep learning-based resource allocation in communication networks and conducts numerical experiments in which agents, acting as transmitters, collaboratively train their individual policies to maximize a global reward, e.g., a sum of data rates.

关键词： Training optimization Transmitters Receivers Distance learning Computer aided instruction Convergence Communication networks Resource management Delays distributed learning and adaptation asynchronous distributed optimization zeroth-order optimization bounded delay deep learning-based resource allocation

来源：评论

学校读者我要写书评

暂无评论

distributed Stochastic Gradient Descent With Staleness: A Stochastic Delay Differential Equation Based Framework

引用

IEEE TRANSACTIONS ON SIGNAL PROCESSING 2025年 73卷 1708-1726页

作者： Yu, Siyuan Chen, Wei Poor, H. Vincent Tsinghua Univ Dept Elect Engn State Key Lab Space Network & Commun Beijing 100084 Peoples R China Tsinghua Univ Beijing Natl Res Ctr Informat Sci & Technol BNRist Beijing 100084 Peoples R China Princeton Univ Dept Elect & Comp Engn Princeton NJ 08544 USA

distributed stochastic gradient descent (SGD) has attracted considerable recent attention due to its potential for scaling computational resources, reducing training time, and helping protect user privacy in machine learning. However, stragglers and limited bandwidth may induce random computational/ communication delays, thereby severely hindering the learning process. Therefore, how to accelerate asynchronous SGD (ASGD) by efficiently scheduling multiple workers is an important issue. In this paper, a unified framework is presented to analyze and optimize the convergence of ASGD based on stochastic delay differential equations (SDDEs) and the Poisson approximation of aggregated gradient arrivals. In particular, we present the run time and staleness of distributed SGD without a memorylessness assumption on the computation times. Given the learning rate, we reveal the relevant SDDE's damping coefficient and its delay statistics, as functions of the number of activated clients, staleness threshold, the eigenvalues of the Hessian matrix of the objective function, and the overall computational/communication delay. The formulated SDDE allows us to present both the distributed SGD's convergence condition and speed by calculating its characteristic roots, thereby optimizing the scheduling policies for asynchronous/event-triggered SGD. It is interestingly shown that increasing the number of activated workers does not necessarily accelerate distributed SGD due to staleness. Moreover, a small degree of staleness does not necessarily slow down the convergence, while a large degree of staleness will result in the divergence of distributed SGD. Numerical results demonstrate the potential of our SDDE framework, even in complex learning tasks with non-convex objective functions.

关键词： distributed optimization machine learning stochastic gradient descent gradient staleness stochastic delay differential equations asynchronous distributed optimization straggler problem distributed optimization machine learning stochastic gradient descent gradient staleness stochastic delay differential equations asynchronous distributed optimization straggler problem

来源：评论

学校读者我要写书评

暂无评论

Hybrid LMMSE Transceiver optimization for distributed IoT Sensing Networks With Different Levels of Synchronization

引用

IEEE INTERNET OF THINGS JOURNAL 2021年第19期8卷 14458-14470页

作者： Liu, Heng Wang, Shuai Gong, Shiqi Zhao, Nan An, Jianping Quek, Tony Q. S. Beijing Inst Technol Sch Informat & Elect Beijing 100081 Peoples R China Univ Macau Dept Elect & Comp Engn Macau Peoples R China Peng Cheng Lab Shenzhen 518066 Peoples R China Dalian Univ Technol Dalian 116024 Peoples R China Singapore Univ Technol & Design Singapore 487372 Singapore

In this article, we investigate the analog-digital hybrid transceiver optimization for distributed Internet-of-Things (IoT) sensing networks consisting of a multiantenna fusion center (FC) and several multiantenna sensor nodes. Analog-digital hybrid transceiver is an economic way to realize tradeoffs between hardware cost and performance for multiantenna communications. Under the nonconvex unit modulus constraints and transmit power constraint at each sensor, two synchronization schemes are considered for the hybrid linear minimum mean-square error (LMMSE) transceiver optimization. First, a centralized algorithm is proposed, in which the hybrid transceivers are computed at the FC. Based on the framework of alternating direction method of multipliers (ADMMs), the unit modulus constraints can be satisfied by projecting the elements of analog transceivers onto the unit modulus circle. However, the centralized algorithm usually suffers from strict synchronous requirements and high communication overhead. In order to accommodate the inevitable computing and communication delays in distributed IoT sensing networks, an asynchronous distributed ADMM (AD-ADMM) algorithm is proposed. By using the aged information, the hybrid transceivers are computed at the sensors without the coordination of the FC. Thus, the AD-ADMM algorithm can greatly reduce the computation overhead of the FC and improve the scalability of IoT sensing networks. Simulation results are presented to show that both the centralized ADMM and AD-ADMM algorithms perform closely to the fully digital counterpart.

关键词： Transceivers optimization Sensors Internet of Things Radio frequency Convex functions Antenna arrays Alternating direction method of multipliers (ADMMs) asynchronous distributed optimization hybrid transceiver optimization Internet of Things (IoT)

来源：评论

学校读者我要写书评

暂无评论

Branch-Wise Parallel Successive Algorithm for Online Voltage Regulation in Distribution Networks

引用

IEEE TRANSACTIONS ON SMART GRID 2019年第6期10卷 6678-6689页

作者： Hu, Xiong Liu, Zhi-Wei Wen, Guanghui Yu, Xinghuo Li, Chaojie Huazhong Univ Sci & Technol Sch Artificial Intelligence & Automat Wuhan 430074 Peoples R China Huazhong Univ Sci & Technol Key Lab Image Proc & Intelligent Control Minist Educ Wuhan 430074 Peoples R China Southeast Univ Sch Math Nanjing 210096 Peoples R China RMIT Univ Sch Engn Melbourne Vic 3001 Australia Alibaba Grp Hangzhou 311100 Peoples R China

Due to high penetration of distributed energy resources, voltage violation has become one of the most critical issues in control of modern power systems with integration of renewable distributed generations (DGs). In this paper, an online distributed approach is proposed to regulate voltage in radial distribution networks via DGs. By fully considering the topological properties of the radial distribution network, the proposed approach successfully decomposes the global voltage regulation problem into a sequence of local reactive power optimization sub-problems with low dimensions. It has been further shown that the optimal solution of the global voltage regulation problem can be achieved by cyclically solving these local subproblems in a branch-wise parallel successive order. Compared with the existing literature, the proposed online distributed voltage regulation approach does not require synchronous updates of all the participants and can improve the performance of the distribution network during the process of its execution. The effectiveness, robustness to fast voltage fluctuations, convergence speed, and insensitivity to parameters of the proposed approach are validated by performing case studies.

关键词： Distribution network reactive power support voltage regulation parallel computing asynchronous distributed optimization

来源：评论

学校读者我要写书评

暂无评论

An asynchronous Fully distributed Traffic Assignment Algorithm for Optimizing Traffic Allocation Problem 38

An Asynchronous Fully Distributed Traffic Assignment Algorit...

引用

38th Chinese Control Conference (CCC)

作者： Zheng, Huaqing Wang, Jingyao Zeng, Jianping Xiamen Univ Dept Automat Xiamen 361005 Fujian Peoples R China Xiamen Univ Shenzhen Res Inst Shenzhen 518000 Peoples R China

ISBN: (纸本)9789881563972

With the development of Internet technology, IP video services have become more and more popular, and video traffic plays a more important role in Internet traffic. Future Internet technology needs to improve its ability to cope with the ever growing resource demands for video services, while being able to provide users with satisfied Quality-of-Experience (QoE). In the work of this paper, based on the distributed traffic allocation algorithm FDTAA which is proposed in the literature to maximize the utility function of the connectionless network,this paper extends FDTTA to fit into the asynchronous framework, and simultaneously considers the influence of communication delay on the asynchronous algorithm. In the synchronous algorithm,all regions need to wait for the slowest region due to that these regions are usually heterogeneous and require different amounts of calculation time, which results in the waste of computation resources. The results of the numerical experiment show that asynchronous FDTAA can improve this situation. In addition, the communication delay will deepen the severity of asynchronism of the algorithm. The greater the communication delay, the more severe asynchronism and the worse the convergence performance of the algorithm. In addition, The convergence performance of the algorithm can be improved by changing the step-size parameters of the algorithm, at the price of the increase of iterations and calculation time.

关键词： Traffic allocation problem FDTAA asynchronous distributed optimization Communication delay

来源：评论

学校读者我要写书评

暂无评论

Impact of Communication Delay on asynchronous distributed Optimal Power Flow Using ADMM

Impact of Communication Delay on Asynchronous Distributed Op...

引用

IEEE International Conference on Smart Grid Communications (SmartGridComm)

作者： Guo, Junyao Hug, Gabriela Tonguz, Ozan Carnegie Mellon Univ Dept Elect & Comp Engn Pittsburgh PA 15213 USA ETH Power Syst Lab Zurich Switzerland

ISBN: (纸本)9781538609439

distributed optimization has attracted lots of attention in the operation of power systems in recent years, where a large area is decomposed into smaller control regions each solving a local optimization problem with periodic information exchange with neighboring regions. However, most distributed optimization methods are iterative and require synchronization of all regions at each iteration, which is hard to achieve without a centralized coordinator and might lead to under-utilization of computation resources due to the heterogeneity of the regions. To address such limitations of synchronous schemes, this paper investigates the applicability of asynchronous distributed optimization methods to power system optimization. Particularly, we focus on solving the AC Optimal Power Flow problem and propose an algorithmic framework based on the Alternating Direction Method of Multipliers (ADMM) method that allows the regions to perform local updates with information received from a subset of but not all neighbors. Through experimental studies, we demonstrate that the convergence performance of the proposed asynchronous scheme is dependent on the communication delay of passing messages among the regions. Under mild communication delays, the proposed scheme can achieve comparable or even faster convergence compared with its synchronous counterpart, which can be used as a good alternative to centralized or synchronous distributed optimization approaches.

关键词： asynchronous distributed optimization ADMM communication delay optimal power flow power system communications

来源：评论

学校读者我要写书评

暂无评论

SYNTHESIS: A Semi-asynchronous Path-Integrated Stochastic Gradient Method for distributed Learning in Computing Clusters 22

SYNTHESIS: A Semi-Asynchronous Path-Integrated Stochastic Gr...

引用

23rd International Symposium on Theory, Algorithmic Foundations, and Protocol Design for Mobile Networks and Mobile Computing (MobiHoc)

作者： Liu, Zhuqing Zhang, Xin Liu, Jia Ohio State Univ Dept Elect & Comp Engn Columbus OH 43210 USA Iowa State Univ Dept Stat Ames IA USA

ISBN: (纸本)9781450391658

To increase the training speed of distributed learning, recent years have witnessed a significant amount of interest in developing both synchronous and asynchronous distributed stochastic variancereduced optimization methods. However, all existing synchronous and asynchronous distributed training algorithms suffer from various limitations in either convergence speed or implementation complexity. This motivates us to propose an algorithm called SYNTHESIS (semi-asynchronous path-integrated stochastic gradient search), which leverages the special structure of the variance-reduction framework to overcome the limitations of both synchronous and asynchronous distributed learning algorithms, while retaining their salient features. We consider two implementations of SYNTHESIS under distributed and shared memory architectures. We show that our SYNTHESIS algorithms have O(root N epsilon(-2) (Delta+ 1) +N) and O(root N epsilon(- 2) (Delta + 1)d +N) computational complexities for achieving an epsilon-stationary point in non-convex learning under distributed and shared memory architectures, respectively, where N denotes the total number of training samples and Delta represents the maximum delay of the workers. Moreover, we investigate the generalization performance of SYNTHESIS by establishing algorithmic stability bounds for quadratic strongly convex and non-convex optimization. We further conduct extensive numerical experiments to verify our theoretical findings.

关键词： Machine learning asynchronous distributed optimization

来源：评论

学校读者我要写书评

暂无评论

asynchronous distributed Matrix Factorization with Similar User and Item Based Regularization 10

Asynchronous Distributed Matrix Factorization with Similar U...

引用

10th ACM Conference on Recommender Systems (RecSys)

作者： Joshi, Bikash Iutzeler, Franck Amini, Massih-Reza Univ Grenoble Alpes CNRS Comp Sci Lab LIG Grenoble France Univ Grenoble Alpes CNRS Appl Math Lab LJK Grenoble France

ISBN: (纸本)9781450340359

We introduce an asynchronous distributed stochastic gradient algorithm for matrix factorization based collaborative filtering. The main idea of this approach is to distribute the user-rating matrix across different machines, each having access only to a part of the information, and to asynchronously propagate the updates of the stochastic gradient optimization across the network. Each time a machine receives a parameter vector, it averages its current parameter vector with the received one, and continues its iterations from this new point. Additionally, we introduce a similarity based regularization that constrains the user and item factors to be close to the average factors of their similar users and items found on subparts of the distributed user-rating matrix. We analyze the impact of the regularization terms on Movie-Lens (100K, 1M, 10M) and NetFlix datasets and show that it leads to a more efficient matrix factorization in terms of Root Mean Square Error (RMSE) and Mean Absolute Error (MAE), and that the asynchronous distributed approach significantly improves in convergence time as compared to an equivalent synchronous distributed approach.

关键词： Recommender Systems asynchronous distributed optimization Similarity-based Regularization

来源：评论

学校读者我要写书评

暂无评论

建议与咨询留下您的常用邮箱和电话号码，以便我们向您反馈解决方案和替代方法

时间限定

文献类型

馆藏选择

核心期刊

语言

文献类型

帮助

文字说明：

检索规则说明：

检索范例：

分类表

所选分类

限定检索结果

文献类型

馆藏范围

日期分布

学科分类号

主题

机构

作者

语言

请选择保存的检索档案：

请选择收藏分类：

通借通还

建议与咨询 留下您的常用邮箱和电话号码，以便我们向您反馈解决方案和替代方法

时间限定

文献类型

馆藏选择

核心期刊

语言

文献类型

帮助

文字说明：

检索规则说明：

检索范例：

分类表

所选分类

限定检索结果

文献类型

馆藏范围

日期分布

学科分类号

主题

机构

作者

语言

请选择保存的检索档案： 新增检索档案 确定 取消

请选择收藏分类： 新增自定义分类 确定 取消

通借通还

建议与咨询留下您的常用邮箱和电话号码，以便我们向您反馈解决方案和替代方法

请选择保存的检索档案：

请选择收藏分类：