-
作者：Zhang, Xu; Vasconcelos, Marcos M.
作者单位：Xidian University; Xidian University; State University System of Florida; Florida A&M University; Florida State University
摘要：Collecting the most informative data from a large dataset distributed over a network is a fundamental problem in many fields, including control, signal processing, and machine learning. In this article, we establish a connection between selecting the most informative data and finding the top-k elements of a multiset. The top-k selection in a network can be formulated as a distributed nonsmooth convex optimization problem known as quantile estimation. Unfortunately, the lack of smoothness in th...
-
作者：Zhao, Feiran; Sha, Xingyu; You, Keyou
作者单位：Tsinghua University; Tsinghua University
摘要：Learning policies in an asynchronous parallel way is essential to numerous successes of reinforcement learning for solving complex problems. However, their convergence has not been rigorously evaluated. To improve the theoretical understanding, we adopt the asynchronous parallel zero-order policy gradient (AZOPG) method to solve the continuous-time linear quadratic regulation problem. Specifically, multiple workers independently perform system rollouts to estimate zero-order policy gradients (...
-
作者：Faruqi, Abdul Hannan; Chatterjee, Anindya
作者单位：Indian Institute of Technology System (IIT System); Indian Institute of Technology (IIT) - Kanpur
摘要：Linear time-invariant single-input single-output (SISO) systems which satisfy a parity interlacing property (PIP) can be stabilized with a stable controller in a single feedback loop. We consider such stabilization of plants with rational transfer functions of relative degree 0, 1, or 2. Finding such controllers requires an interpolant U(s) with specific properties. Existing methods for finding U(s) use an iterative manual calculation or, when the plant's right half plane zeros are simple, a m...
-
作者：Zhang, Qi; Wang, Lin; Wang, Xiaofan; Chen, Guanrong
作者单位：Shanghai Jiao Tong University; Shanghai University; Shanghai Institute of Technology; City University of Hong Kong
摘要：Differing from existing research on intervention strategies such as leader selection and edge addition, we investigate the impact of intervention timing in opinion dynamics. We employ the leader-based DeGroot model to formulate the evolution of opinions in social networks, wherein leaders represent organizations or parties that influence public opinion. We propose an optimal timing selection problem, in which a leader maximizes public opinion at a specific time by strategically selecting inter...
-
作者：Gao, Rui; Yang, Guang-Hong
作者单位：Northeastern University - China; Northeastern University - China
摘要：In this article, we are concerned with the problem of distributed state estimation for discrete-time linear systems using a network of agents, where the measurement of each agent suffers from the lack of detectability with the system dynamics. The existing results on this topic require stringent condition either on the network connectivity or on the detectability. In contrast, a new form of distributed observer, which ensures that all agents asymptotically estimate the system state under the m...
-
作者：Sasahara, Hampei; Dan, Gyorgy; Amin, Saurabh; Sandberg, Henrik
作者单位：Institute of Science Tokyo; Royal Institute of Technology; Massachusetts Institute of Technology (MIT); Royal Institute of Technology
摘要：Eco-friendly freight operations are crucial for decarbonizing the transportation sector. Systematic analysis of policy measures requires a principled modeling approach. While the commonly used model referred to as a routing game considers the congestible nature of transportation facilities, existing models fail to account for environmental factors. This article aims at providing a mathematical framework to study strategic interaction between owners of mixed fleets comprising both internal comb...
-
作者：Zhang, Weihao; Zhao, Di; Liang, Shu
作者单位：Tongji University; Tongji University; Tongji University; Tongji University
摘要：This article introduces the concept of soft safety for nonlinear dynamical systems, examined from an input-output perspective, and establishes criteria for verifying soft safety in interconnected systems. Diverging from traditional safety approaches that enforce rigid/hard constraints on system states, our proposal involves integrating soft barrier functions with respect to system inputs and outputs to characterize soft safety. Particularly, when the soft barrier function takes a quadratic for...
-
作者：Li, Qiang; Wai, Hoi-To
作者单位：Chinese University of Hong Kong
摘要：This article studies the effect of data homogeneity on multiagent stochastic optimization. We consider the decentralized stochastic gradient (DSGD) algorithm and perform a refined convergence analysis. Our analysis is explicit on the similarity between Hessian matrices of local objective functions, which captures the degree of data homogeneity. We illustrate the impact of our analysis through studying the transient time, defined as the minimum number of iterations required for a distributed al...
-
作者：Craft, Kyle J.; DeMars, Kyle J.
作者单位：Texas A&M University System; Texas A&M University College Station
摘要：Bayes' rule, although a powerful framework for performing state estimation, is often intractable for real-world, nonlinear dynamic systems. As a result, estimation algorithms typically rely on a simplifying assumption, such as the linearity of the measurement model or Gaussianity of the likelihood function. For nonlinear, non-Gaussian systems, these approximations can introduce statistical inconsistencies into the underlying estimator. To mitigate approximation errors, a homotopic scheme is pr...
-
作者：Xu, Tao; Yi, Xiaojian; Duan, Zhisheng; Chen, Guanrong; Wen, Guanghui
作者单位：Beijing Institute of Technology; Nanyang Technological University; Beijing Institute of Technology; Peking University; City University of Hong Kong; Southeast University - China
摘要：As reported in the studies of dynamic average tracking, many existing solutions are not robust to initialization, in the sense that their design and implementation require specific and stringent initial conditions. All agents need to be reinitialized if the original initial conditions are violated due to network disruptions. This article aims to overcome this common issue existing in the control problem of distributed event-driven dynamic average tracking for networked multiple linear systems ...
