Article
Engineering, Multidisciplinary
Mingfei Chen, Dong Wang, Xiaodong Wang, Zheng-Guang Wu, Wei Wang
Summary: This paper studies the distributed aggregative optimization problem with local constraint sets over an undirected graph. A continuous-time algorithm with nonuniform gradient gains is proposed to find the optimal decision variable, which only requires the sign of relative state information between agents' neighbours and has an advantage in reducing communication cost. The effectiveness of the theoretical results is demonstrated through numerical examples.
IEEE TRANSACTIONS ON NETWORK SCIENCE AND ENGINEERING
(2023)
Article
Automation & Control Systems
Cong Wang, Shengyuan Xu, Deming Yuan, Baoyong Zhang, Zhengqiang Zhang
Summary: This article introduces a distributed online convex optimization algorithm for multiagent systems, which updates decisions through communication between nodes to handle Lipschitz continuous and strongly convex cost functions.
IEEE TRANSACTIONS ON CYBERNETICS
(2022)
Article
Automation & Control Systems
Haishan Ye, Luo Luo, Ziang Zhou, Tong Zhang
Summary: This paper addresses the decentralized convex optimization problem and proposes novel algorithms that achieve optimal or near-optimal computation and communication complexity. The theoretical results provide affirmative answers to an open problem and demonstrate that the algorithms can achieve a communication complexity matching the lower bound based on the global condition number. Additionally, the convergence of the algorithms only relies on the strong convexity of the global objective, without requiring convexity of the local functions.
JOURNAL OF MACHINE LEARNING RESEARCH
(2023)
Article
Computer Science, Artificial Intelligence
Xiwang Meng, Qingshan Liu
Summary: In this paper, a distributed algorithm is proposed for solving distributed optimization problems with closed convex set constraints in a multi-agent system under a weight-balanced graph. The algorithm considers uncertainties such as state noise and gradient disturbance and improves convergence rate by introducing a projection error term and auxiliary parameter through gradient tracking and projection methods.
Article
Automation & Control Systems
Tongyu Wang, Peng Yi
Summary: In this paper, a distributed Frank-Wolfe algorithm based on gradient tracking is proposed to solve a distributed convex aggregative optimization problem in a network. The algorithm minimizes the sum of cost functions by maintaining two estimates at each node. The proposed algorithm is projection-free and only involves solving a linear optimization problem at each step. Convergence and computational efficiency of the algorithm are demonstrated through numerical simulations.
INTERNATIONAL JOURNAL OF ROBUST AND NONLINEAR CONTROL
(2023)
Article
Automation & Control Systems
Cong Wang, Shengyuan Xu, Deming Yuan, Baoyong Zhang, Zhengqiang Zhang
Summary: This article discusses the distributed convex optimization problem over a multiagent system and investigates the impact of communication delays on the convergence results of the PS-DDA algorithm. The performance of the PS-DDA algorithm is demonstrated through numerical simulations.
IEEE TRANSACTIONS ON SYSTEMS MAN CYBERNETICS-SYSTEMS
(2023)
Article
Automation & Control Systems
Zhongyuan Zhao
Summary: The convex optimization problem in multi-agent systems is studied and a sample-based event-triggered optimization algorithm is proposed. A novel dynamic event-triggered condition is designed to deduce the communication burden and simplify the implementation of the algorithm. The global optimal solution can be obtained exponentially and the event-triggered time intervals are significantly prolonged. Additionally, the sampling control mechanism avoids Zeno behavior.
INTERNATIONAL JOURNAL OF CONTROL AUTOMATION AND SYSTEMS
(2022)
Article
Automation & Control Systems
Tao Dong, Huiyun Zhu, Wenjie Hu
Summary: This research proposes an event-triggered differentially privacy optimal consensus algorithm to address the privacy leakage issue by protecting the privacy of cost function of agents during the process of consensus computation. The study analyzes the accuracy and consensus of the algorithm, as well as the privacy preservation effects, demonstrating that the privacy of all agents' states can be preserved.
INTERNATIONAL JOURNAL OF CONTROL AUTOMATION AND SYSTEMS
(2021)
Article
Mathematics, Applied
Yaxin Huang
Summary: This paper investigates asymptotic consensus for multi-agent systems under the influence of stochastic noise in the measurement of neighbor information. The presence of intractable parametric uncertainties in the agent dynamics, combined with inherent nonlinearities, necessitates distributed adaptive compensation. The distributed protocol design focuses on reference consensus based on relative information and agent tracking based on individual estimates to attenuate the effects of measurement noises and ensure consensus in the almost sure sense. The proposed distributed adaptive protocol facilitates state tracking and prevents finite-time explosion caused by inherent nonlinearities, ultimately leading to average consensus.
COMMUNICATIONS IN NONLINEAR SCIENCE AND NUMERICAL SIMULATION
(2023)
Article
Engineering, Electrical & Electronic
Kexin Lv, Fan He, Xiaolin Huang, Jie Yang
Summary: This paper proposes a consensus-based distributed algorithm for GEP in multi-agent systems, which can effectively deal with problems with quadratic inseparable constraints.
Article
Automation & Control Systems
Yipeng Pang, Guoqiang Hu
Summary: This article presents a special type of distributed optimization problems, where the global cost function is convex, but each individual can be nonconvex. A Gaussian-smoothing technique is introduced and a gradient-free method is proposed to solve the problem. The algorithm's convergence is proven.
INTERNATIONAL JOURNAL OF ROBUST AND NONLINEAR CONTROL
(2022)
Article
Computer Science, Information Systems
Hao Yu, Hui Chen, Shengjie Zhao, Qingjiang Shi
Summary: This article introduces a distributed soft clustering algorithm for data analysis in IoT networks, which addresses challenges such as data volume and information security. Through experiments, the algorithm is shown to perform as well as centralized methods, offering stable clustering quality and practical applications. The algorithm uses distributed deterministic initialization and finite-time average-consensus algorithm for efficient computation and stability improvement.
IEEE INTERNET OF THINGS JOURNAL
(2021)
Article
Telecommunications
Mounssif Krouka, Anis Elgabli, Chaouki Ben Issaid, Mehdi Bennis
Summary: This paper proposes a novel distributed reinforcement learning algorithm, A-RLADMM, that tackles the challenge of wireless channel interference using the alternating direction method of multipliers (ADMM) and analog transmission scheme.
IEEE TRANSACTIONS ON COGNITIVE COMMUNICATIONS AND NETWORKING
(2022)
Article
Automation & Control Systems
Shi Pu, Wei Shi, Jinming Xu, Angelia Nedic
Summary: In this article, a distributed convex optimization approach is introduced, which achieves minimal cost functions through the push-pull gradient method for information exchange between nodes in a network. Experimental results demonstrate that this algorithm exhibits linear convergence in various network architectures, especially showing significant performance in random-gossip settings.
IEEE TRANSACTIONS ON AUTOMATIC CONTROL
(2021)
Article
Automation & Control Systems
Cong Wang, Shengyuan Xu, Deming Yuan, Yuming Chu, Zhengqiang Zhang
Summary: This paper addresses distributed convex optimization problems for multi-agent systems using the push-sum distributed dual averaging algorithm, while considering subgradient delays. The main result proves that the algorithm converges with sublinear growth of error, and a numerical example is provided to demonstrate its performance.
JOURNAL OF THE FRANKLIN INSTITUTE-ENGINEERING AND APPLIED MATHEMATICS
(2021)