Article
Automation & Control Systems
Lifeng Zheng, Liang Ran, Huaqing Li, Liping Feng, Zheng Wang, Qingguo Lu, Dawen Xia
Summary: This paper focuses on a class of decentralized convex optimization problems and introduces a synchronous full-decentralized primal-dual proximal splitting algorithm and its randomized version. The problems are solved through local information exchange without global coordination, and the convergence results are obtained with the help of asymmetric forward-backward-adjoint splitting technique. Numerical simulations demonstrate the effectiveness and practicability of the algorithms.
JOURNAL OF THE FRANKLIN INSTITUTE-ENGINEERING AND APPLIED MATHEMATICS
(2022)
Article
Automation & Control Systems
Huiyan Li, Jingyuan Zhan, Hai-Tao Zhang, Xiang Li
Summary: This article investigates the consensus of constrained linear heterogeneous multiagent systems under prediction and optimization. Two types of distributed analytical model predictive controllers are proposed by optimizing the consensus problems constrained to state equations and general linear constraints. Furthermore, stability conditions for the two types of controllers are derived, clarifying the relationship between the communication topology and dynamics of heterogeneous agents. Simulation examples of networked heterogeneous agents illustrate the convergence and validity of the proposed controllers.
IEEE TRANSACTIONS ON SYSTEMS MAN CYBERNETICS-SYSTEMS
(2023)
Article
Automation & Control Systems
Zhan Yu, Daniel W. C. Ho, Deming Yuan
Summary: This article introduces the multiagent optimization problem and proposes a distributed randomized gradient-free mirror descent method. The method employs the non-Euclidean Bregman divergence and generalizes the classical gradient descent method without using subgradient information. It achieves an approximate O(1/root T) convergence rate, recovering the best known optimal rate. Additionally, a decentralized reciprocal weighted averaging (RWA) approximating sequence is investigated, with convergence shown to hold over time-varying graphs. The article provides new insights for searching minimizers in distributed algorithms.
IEEE TRANSACTIONS ON AUTOMATIC CONTROL
(2022)
Article
Computer Science, Artificial Intelligence
Zhongguo Li, Bo Liu, Zhengtao Ding
Summary: This article proposes a distributed algorithm for training neural networks with private data sets, utilizing constant learning rates to enhance tracking ability and establishing convergence by exploring error dynamics of connected agents. Simulation results validate the effectiveness of the algorithm.
IEEE TRANSACTIONS ON NEURAL NETWORKS AND LEARNING SYSTEMS
(2022)
Article
Automation & Control Systems
Xuan Wang, Shaoshuai Mou, Brian D. O. Anderson
Summary: Inspired by integral feedback, this study proposes a distributed constant gain algorithm for multiagent networks to solve convex optimization problems with local linear constraints. Utilizing an undirected graph model for agent interactions, the algorithm achieves the optimum solution with an exponential convergence rate. The algorithm inherits beneficial requirements on communication bandwidth and robustness against disturbance from integral feedback. Analytical proof and numerical simulations validate the effectiveness of the proposed algorithm in solving constrained optimization problems.
IEEE TRANSACTIONS ON AUTOMATIC CONTROL
(2023)
Article
Automation & Control Systems
Chao Huang, Brian D. O. Anderson, Hao Zhang, Huaicheng Yan
Summary: For a group of networked agents, f-consensus means reaching a consensus on the value of a desired function, f, based on the initial state of the individual agents. This paper demonstrates how f-consensus problems can often be converted into distributed convex optimization (DCO) problems, which can be easily solved using existing DCO algorithms. This approach offers computational advantages and can solve specific classes of f-consensus problems, including weighted power mean consensus and kth smallest value or kth order statistic consensus.
Article
Computer Science, Artificial Intelligence
Zhe Wei, Wenwen Jia, Wei Bian, Sitian Qin
Summary: With the development of artificial intelligence and big data, distributed optimization has shown great potential in machine learning. This paper proposes a novel neural network for cooperatively solving nonsmooth distributed optimization problems, and its effectiveness and practicality are demonstrated through simulation results and a practical application.
NEURAL COMPUTING & APPLICATIONS
(2023)
Article
Automation & Control Systems
Hongzhe Liu, Wei Xing Zheng, Wenwu Yu
Summary: This article studies convex optimization problems with general constraints and proposes a distributed algorithm to solve the problem. The optimality condition of the optimization problem is developed using saddle point theory, and a continuous-time primal-dual algorithm is constructed accordingly.
IEEE TRANSACTIONS ON AUTOMATIC CONTROL
(2022)
Article
Automation & Control Systems
Zhan Yu, Daniel W. C. Ho, Deming Yuan, Jie Liu
Summary: This article discusses the distributed stochastic multiagent-constrained optimization problem over a time-varying network with a specific class of communication noise. A non-Euclidean method, based on the Bregman projection-based mirror descent scheme, is proposed and its convergence behavior is investigated. The method, known as the distributed stochastic composite mirror descent type method (DSCMD-N), provides a more general algorithm framework and new error bounds.
IEEE TRANSACTIONS ON CYBERNETICS
(2023)
Letter
Automation & Control Systems
Qian Xu, Zao Fu, Bo Zou, Hongzhe Liu, Lei Wang
Summary: This letter investigates the optimization problem using a combination of gradient descent method with push-sum algorithm framework to design distributed iterative formulas for time-varying and unbalanced graphs. The convergence property of the generated variable sequence under the proposed iterative formulas is analyzed.
IEEE-CAA JOURNAL OF AUTOMATICA SINICA
(2022)
Article
Automation & Control Systems
Xiaotian Wang, Housheng Su
Summary: A novel model of transmission-constrained consensus over random networks is proposed, considering the impact of information distortions and stochastic information flow caused by environmental conditions. The model utilizes heterogeneous functions and a directed random graph to represent transmission constraints and the characteristics of information flow. Using stochastic stability theory and the martingale convergence theorem, it is proved that agent states will converge to a consensus value with probability 1 despite information distortions and randomness.
IEEE TRANSACTIONS ON CYBERNETICS
(2023)
Article
Computer Science, Artificial Intelligence
Haris E. Psillakis, Konstantinos A. Oikonomidis
Summary: This paper proposes a new framework for solving the optimal consensus problem of agents with continuous dynamics. By introducing novel auxiliary continuous variables and using suitable smoothing functions, the continuity of the variables is guaranteed. It also designs adaptive fuzzy distributed controllers to approximate the unknown system nonlinearities and ensure optimal consensus.
IEEE TRANSACTIONS ON FUZZY SYSTEMS
(2023)
Article
Automation & Control Systems
Apostolos Rikos, Christoforos N. Hadjicostis
Summary: This article explores the distributed average consensus problem in multiagent systems and introduces a novel distributed algorithm based on quantized values, utilizing event-driven updates to achieve a common consensus value for all agents in finite time.
IEEE TRANSACTIONS ON AUTOMATIC CONTROL
(2021)
Article
Automation & Control Systems
Fei Chen, Jin Jin, Linying Xiang, Wei Ren
Summary: This article presents a scaling-function approach for distributed optimization of unbalanced multiagent networks under convex constraints. The algorithm does not require agents' out-degree information or estimation of the left eigenvector of the Laplacian matrix. Numerical examples validate the theoretical findings.
IEEE TRANSACTIONS ON AUTOMATIC CONTROL
(2022)
Article
Computer Science, Artificial Intelligence
Lifu Ding, Gangfeng Yan, Jianing Liu
Summary: This paper analyzes the application of Multiagent Reinforcement Learning (MARL) in engineering problems and discovers that strict global constraints can lead to sparse rewards. To address this issue, a fully distributed and convergent MARL algorithm based on Reward Recorder is proposed. Simulation examples demonstrate that the proposed algorithm has high stability and excellent decision-making ability.
INTERNATIONAL JOURNAL OF INTELLIGENT SYSTEMS
(2022)
Article
Optics
Honglin Yuan, Ke Lu, Qingshan Liu
Summary: This study investigates the issue of image blurring in star tracker under dynamic conditions and proposes a model-based approach to simulate and generate realistic blurred images. Experimental results demonstrate that the proposed method can effectively simulate various types of image blurs and plays a crucial role in the centroid extraction of stars and the dynamic performance of the star tracker.
JOURNAL OF THE OPTICAL SOCIETY OF AMERICA B-OPTICAL PHYSICS
(2022)
Article
Engineering, Multidisciplinary
Xiaoxuan Wang, Shaofu Yang, Zhenyuan Guo, Shiping Wen, Tingwen Huang
Summary: This paper addresses distributed nonsmooth optimization problems and proposes a distributed multi-agent network system based on consensus protocol and projected output feedback, which can converge to the optimal solution of the optimization problem.
IEEE TRANSACTIONS ON NETWORK SCIENCE AND ENGINEERING
(2022)
Article
Multidisciplinary Sciences
Kai Cao, Qiyu Gong, Yiguang Hong, Lin Wan
Summary: uniPort is a unified single-cell data integration framework that leverages different methods to handle heterogeneous data, constructs a shared latent space and reference atlas, and provides a label transfer framework for deconvolution of spatially resolved transcriptomic data without embedding latent space.
NATURE COMMUNICATIONS
(2022)
Article
Automation & Control Systems
Ziqin Chen, Peng Yi, Li Li, Yiguang Hong
Summary: In this work, a distributed algorithm is proposed for time-varying convex optimization over networks with quantized communications. The algorithm utilizes dynamic quantization scheme to reduce information loss, and it is capable of asymptotically tracking the optimal solution even with quantization information loss, as validated by theoretical analysis and numerical simulation.
IEEE TRANSACTIONS ON CYBERNETICS
(2023)
Article
Automation & Control Systems
Peng Yi, Jinlong Lei, Jie Chen, Yiguang Hong, Guodong Shi
Summary: In this article, the authors studied the convergence and convergence rate of distributed linear equation protocols over a *-mixing random network. They proposed a distributed projection consensus algorithm and proved its exponential convergence rate in the mean-squared sense for networks with exact solutions. Additionally, they proposed a distributed randomized projection consensus algorithm and established an exponential rate of convergence. Moreover, they proved that a distributed gradient-descent-like algorithm with diminishing step-sizes can drive all nodes' states to a least-squares solution at a sublinear rate for networks without exact solutions.
IEEE TRANSACTIONS ON AUTOMATIC CONTROL
(2023)
Article
Automation & Control Systems
Songsong Cheng, Shu Liang, Yuan Fan, Yiguang Hong
Summary: Tracking methods have gained popularity in distributed optimization, as they achieve linear convergence with a constant step-size for strongly convex optimization. This article presents a counterexample on constrained optimization, demonstrating that the direct extension of gradient tracking using projections cannot guarantee correctness. Instead, projected gradient tracking algorithms with diminishing step-sizes are proposed for distributed strongly convex optimization with different constraint sets and unbalanced graphs. The basic algorithm achieves an O(ln T/T) convergence rate, which is improved to O(1/T) with an epoch iteration scheme.
IEEE TRANSACTIONS ON AUTOMATIC CONTROL
(2023)
Article
Automation & Control Systems
Shijie Huang, Jinlong Lei, Yiguang Hong
Summary: This article discusses the distributed Nash equilibrium (NE) seeking of strongly monotone aggregative games over a multiagent network. A distributed algorithm is proposed, which involves multiple rounds of communication and achieves convergence to the NE with a linear convergence rate. Furthermore, a single-round communication version of the algorithm is studied, which also achieves linear convergence rate under certain conditions. Numerical simulations are provided to verify the results.
IEEE TRANSACTIONS ON AUTOMATIC CONTROL
(2023)
Article
Computer Science, Artificial Intelligence
Hui Shuai, Lele Wu, Qingshan Liu
Summary: This article proposes a unified framework called MTF-Transformer for 3D Human Pose Estimation, which can adaptively handle varying view numbers and video length without camera calibration. The framework consists of Feature Extractor, Multi-view Fusing Transformer (MFT), and Temporal Fusing Transformer (TFT). MTF-Transformer achieves competitive results compared to state-of-the-art methods and generalizes well to dynamic capture with an arbitrary number of unseen views.
IEEE TRANSACTIONS ON PATTERN ANALYSIS AND MACHINE INTELLIGENCE
(2023)
Article
Mathematics, Applied
Haomin Bai, Wenying Xu, Shaofu Yang, Jinde Cao
Summary: This paper investigates the problem of distributed generalized Nash equilibrium (GNE) tracking in a dynamic environment. It proposes a distributed inertial online game (D-IOG) algorithm to track Nash equilibrium (NE) in the absence of coupled constraints. Furthermore, it introduces a modified D-IOG algorithm based on primal-dual and mirror descent methods to handle time-varying coupled constraints. Simulation examples are provided to demonstrate the effectiveness of the proposed algorithms.
Article
Automation & Control Systems
Lingfei Wang, Carmela Bernardo, Yiguang Hong, Francesco Vasca, Guodong Shi, Claudio Altafini
Summary: This article investigates a two-timescale opinion dynamics model, named the concatenated Friedkin-Johnsen (FJ) model, which describes the evolution of the opinions of a group of agents over a sequence of discussion events. The topology of the underlying graph changes with the event, in the sense that the agents can participate or less to an event, and the agents are stubborn, with stubbornness that can vary from one event to the other. Concatenation refers to the fact that the final opinions of an event become initial conditions of the next event. We show that a concatenated FJ model can be represented as a time-varying product of stochastic transition matrices having a special form. Conditions are investigated under which a concatenated FJ model can achieve consensus in spite of the stubbornness. Four different sufficient conditions are obtained, mainly based on the special topological structure of our stochastic matrices.
IEEE TRANSACTIONS ON AUTOMATIC CONTROL
(2023)
Article
Computer Science, Artificial Intelligence
Zhen Zhang, Shaofu Yang, Wenying Xu
Summary: This paper proposes a decentralized second-order communication-efficient algorithm called CC-DQM, which combines event-triggered communication with compressed communication to solve the decentralized optimization problem. The theoretical analysis shows that the proposed algorithm can achieve exact linear convergence even with compression error and intermittent communication if the local objective functions are strongly convex and smooth. Numerical experiments demonstrate its satisfactory communication efficiency.
Article
Computer Science, Artificial Intelligence
Xiaoxuan Wang, Shaofu Yang, Zhenyuan Guo, Tingwen Huang
Summary: This article focuses on developing distributed optimization strategies for a class of machine learning problems over a directed network of computing agents. By introducing a second-order Nesterov accelerated dynamical system and the projected primal-dual method, the constraints in the problems are effectively dealt with, and the theoretical results are validated by practical problems.
IEEE TRANSACTIONS ON NEURAL NETWORKS AND LEARNING SYSTEMS
(2023)
Article
Automation & Control Systems
Guanpu Chen, Peng Yi, Yiguang Hong, Jie Chen
Summary: In this article, we focus on solving distributed constrained optimization problems. We propose a distributed projection-free dynamics based on the Frank-Wolfe method to avoid projection operations caused by constraints. The approach finds a feasible descent direction through solving an alternative linear suboptimization problem. We also analyze the convergence of the continuous-time dynamical systems and derive a discrete-time scheme with a proved convergence rate of $O(1/k)$. Moreover, we compare our proposed dynamics with existing distributed projection-based dynamics and other distributed Frank-Wolfe algorithms to demonstrate its advantage.
IEEE TRANSACTIONS ON CYBERNETICS
(2023)
Article
Automation & Control Systems
Xiaoxuan Wang, Shaofu Yang, Zhenyuan Guo, Quanbo Ge, Shiping Wen, Tingwen Huang
Summary: This article presents a solution to the k-winner-take-all (kWTA) problem with large-scale inputs in a distributed setting. A multiagent system is proposed, where each agent has a 1-D system and interacts with others through binary consensus protocols. The system convergence is proven using differential inclusion theory, and a novel comparison filter is introduced to eliminate the resolution ratio requirement on the input signal.
IEEE TRANSACTIONS ON CYBERNETICS
(2023)
Article
Computer Science, Theory & Methods
Zhaoyang Chengu, Guanpu Chen, Yiguang Hong
Summary: In this paper, a hypergame framework is used to analyze a single-leader-multiple-followers Stackelberg security game with misinformed situations. The study investigates the strategic stability and cognitive stability, and discovers mild stable conditions for the equilibria with misperception and deception to become hyper Nash equilibrium.
IEEE TRANSACTIONS ON INFORMATION FORENSICS AND SECURITY
(2022)