Article
Computer Science, Artificial Intelligence
Anbiao Wu, Ye Yuan, Yuliang Ma, Guoren Wang
Summary: In this study, a method called ATDGEB is proposed for encoding nodes in time-dependent graphs into vectors based on nodes' local structure and special properties, outperforming existing embedding methods in terms of node clustering, reachability prediction, and link prediction.
KNOWLEDGE-BASED SYSTEMS
(2022)
Article
Chemistry, Multidisciplinary
Xian Tang, Junfeng Zhou, Yunyu Shi, Xiang Liu, Keng Lin
Summary: Given a directed graph, the k-hop reachability query is used to check for the existence of a directed path from u to v with a length of at most k. Existing algorithms for addressing k-hop reachability queries can be inefficient due to costly graph traversal operations. To improve query performance, we propose an approach based on a vertex cover, which constructs an index covering all reachability information using a small set of vertices. Our experimental results demonstrate that our approach significantly outperforms existing approaches in terms of query response time.
APPLIED SCIENCES-BASEL
(2023)
Article
Computer Science, Information Systems
Imane Hocine, Said Yahiaoui, Ahcene Bendjoudi, Nadia Nouali-Taboudjemat
Summary: This paper presents an Overlay Graph-based Distributed Reachability Indexing approach (ODRI) to handle big graphs by processing smaller subgraphs in parallel through overlay graphs, significantly reducing index construction and query processing time, ensuring scalability, and preserving reachability properties. Experimentally, the approach outperforms state-of-the-art methods and is scalable in terms of partition numbers, regardless of graph distribution.
INFORMATION SCIENCES
(2021)
Article
Computer Science, Information Systems
Chenming Yang, Jingjing Li, Ke Lu, Bryan Hooi, Liang Zhou
Summary: This paper proposes a new learning objective called Continuous-time Graph Directed Information Maximization (CGDIM) to learn informative node presentations for temporal networks. By maximizing the directed information, the proposed CGDIM captures the time causal relations among edges with continuous time and improves the performance of backbone models.
INFORMATION SCIENCES
(2023)
Article
Automation & Control Systems
Corrado Possieri, Mattia Frasca, Alessandro Rizzo
Summary: We use reinforcement learning methods to model the reachability probabilities in stochastic directed graphs and show that the transition probabilities can be modeled via a difference inclusion, which can be interpreted as a Markov decision process. By using this framework, we propose a methodology to design reward functions that provide upper and lower bounds on the reachability probabilities of a set of nodes in stochastic digraphs. The effectiveness of this technique is demonstrated through its application to the diffusion of epidemic diseases over time-varying contact networks generated by mobile agents' proximity patterns.
IEEE TRANSACTIONS ON AUTOMATIC CONTROL
(2023)
Article
Operations Research & Management Science
Z. Bartosiewicz
Summary: Positive reachability of a positive system on a time scale means that system trajectories starting from the origin can reach all the points of the nonnegative cone. Positive stabilization of such a system means feedback stabilization that preserves positivity of the system. It is shown that if a positive (linear or nonlinear) continuous-time system is positively reachable from the origin, then it is positively feedback stabilizable. On the other hand, such implication does not hold in general for systems on discrete time scales, even linear ones, though it holds for one-dimensional systems. These differences between continuous-time and discrete-time systems follow from different characterizations of positive reachability for these two classes of systems.
Article
Automation & Control Systems
Ning Xu, Yun Chen, Anke Xue, Guangdeng Zong
Summary: This paper addresses the problems of finite-time stability and stabilization for continuous-time switched positive linear time-delay systems. By constructing a multiple piecewise copositive Lyapunov-Krasovskii functional, finite-time stability conditions are established, and finite-time stabilization is achieved by designing a state-feedback controller in the form of linear programming. The proposed controller construction approach reduces conservatism.
JOURNAL OF THE FRANKLIN INSTITUTE-ENGINEERING AND APPLIED MATHEMATICS
(2022)
Article
Automation & Control Systems
Yanan Zhu, Wei Ren, Wenwu Yu, Guanghui Wen
Summary: This paper investigates resource allocation problem for a group of agents communicating over a strongly connected directed graph. Two continuous-time algorithms are proposed for weight-balanced and weight-unbalanced graphs, with convergence analysis and performance evaluation through numerical simulations.
IEEE TRANSACTIONS ON SYSTEMS MAN CYBERNETICS-SYSTEMS
(2021)
Article
Computer Science, Artificial Intelligence
Shih-Gu Huang, Jing Xia, Liyuan Xu, Anqi Qiu
Summary: We developed a deep learning framework for predicting cognition and disease using fMRI. The framework consists of two neural networks for learning spatial and temporal information of functional time series and functional connectivity features. It also includes an attention component for generating a spatial attention map. Experimental results demonstrate that the framework is generalizable and outperforms other machine learning techniques in cognition and age prediction.
MEDICAL IMAGE ANALYSIS
(2022)
Article
Automation & Control Systems
Vladimir Sinyakov, Antoine Girard
Summary: This article discusses the computation of efficient symbolic abstractions for continuous-time control systems. The new abstraction algorithm generates symbolic models with the same number of states but fewer transitions compared to the standard algorithm, while maintaining at least the same controllability. The proposed algorithm solves a region-to-region control synthesis problem using the theory of viscosity solutions and differential equations with discontinuous righthand side. Symbolic controls in the new algorithm are essentially feedback controllers that solve the control synthesis problem. The approach can produce deterministic abstract systems or systems with a singleton input alphabet for certain control systems with suitable discretization parameters. Examples comparing the two abstraction algorithms are provided.
IEEE TRANSACTIONS ON AUTOMATIC CONTROL
(2023)
Article
Multidisciplinary Sciences
Niall Rodgers, Peter Tino, Samuel Johnson
Summary: In many real, directed networks, the strongly connected component is very small and not following the current theory based on random graphs. This has significant implications for other properties of real networks and the behavior of complex systems. Strong connectivity depends on the network's overall direction or hierarchical ordering, measured by trophic coherence. The targeted attack on edges running counter to the overall direction can disrupt the connectivity structure, indicating that dynamical processes on networks can be influenced by a small fraction of edges.
PROCEEDINGS OF THE NATIONAL ACADEMY OF SCIENCES OF THE UNITED STATES OF AMERICA
(2023)
Article
Computer Science, Artificial Intelligence
Junfeng Zhou, Jeffrey Xu Yu, Yaxian Qiu, Xian Tang, Ziyang Chen, Ming Du
Summary: Answering reachability queries is a fundamental graph operation. This paper proposes a novel approach called RCN reduction to reduce the input graph and accelerate query performance. Efficient algorithms are also proposed to improve the reduction ratio. Based on the result of RCN reduction, a labeling scheme is further proposed for query acceleration. Extensive experimental results confirm the effectiveness of the approach.
IEEE TRANSACTIONS ON KNOWLEDGE AND DATA ENGINEERING
(2023)
Article
Engineering, Multidisciplinary
Made Widhi Surya Atman, Azwirman Gusrialdi
Summary: This paper fills a gap in the literature by proposing a fully distributed algorithm to verify and ensure the strong connectivity of a directed network. Inspired by the maximum consensus algorithm, the proposed algorithm does not require information of the overall network topology, and can verify and ensure the strong connectivity of a directed graph in a finite number of steps. It is scalable and preserves the privacy of the overall network's topology.
IEEE TRANSACTIONS ON NETWORK SCIENCE AND ENGINEERING
(2022)
Article
Computer Science, Information Systems
Jason J. R. Liu, Maoqi Zhang, James Lam, Baozhu Du, Ka-Wai Kwok
Summary: This article aims to design proportional-derivative (PD) controllers for interval positive linear systems in the continuous-time domain, ensuring both closed-loop system stability and positivity. The work proposes a systematic framework and presents methodology and algorithm, which are validated by numerical examples.
INFORMATION SCIENCES
(2021)
Article
Quantum Science & Technology
Rebekah Herrman, Thomas G. Wong
Summary: This paper investigates the simplification methods of quantum walks on dynamic graphs, proposes six scenarios for graph simplification, and provides examples of how to simplify dynamic graphs to achieve parallel single-qubit gates.
QUANTUM INFORMATION PROCESSING
(2022)
Article
Automation & Control Systems
Giulia De Pasquale, Maria Elena Valcher
Summary: This paper addresses the consensus problem in networked agents with a communication graph that splits into clusters. By introducing a modified version of DeGroot's law and constraining the agents' conservatism about their own opinions, a distributed algorithm is proposed to achieve consensus.
Article
Automation & Control Systems
Giulia De Pasquale, Maria Elena Valcher
Summary: A model for the interplay between homophily-based appraisal dynamics and influence-based opinion dynamics has been proposed, and a simplified version of the model is found to be equally accurate and effective. The equilibria reached by the model correspond to structurally balanced agent networks.
EUROPEAN JOURNAL OF CONTROL
(2022)
Article
Automation & Control Systems
Giulia De Pasquale, Maria Elena Valcher
Summary: This paper investigates the herdability property of linear time-invariant state space models, focusing on the system's capability to be driven towards the positive orthant. The study explores the herdability of matrix pairs (A, B), where A represents the adjacency matrix of a multi-agent network and B is a selection matrix that identifies a subset of network leaders. It examines the cases when the associated graph G(A) is directed and clustering balanced or has a tree topology with a single leader.
Article
Automation & Control Systems
Ettore Fornasini, Maria Elena Valcher
Summary: This article aims to investigate the conditions under which a Boolean control network can accept a state feedback control law and make the resulting Boolean network reconstructible. A result is proposed to reduce the problem size significantly and mitigate the curse of dimensionality commonly encountered in dealing with logical systems of large sizes. A necessary and sufficient condition for problem solvability is provided, based on the algebra of noncommutative polynomials. Finally, a procedure is presented for designing a possible state feedback controller that achieves the desired result when this condition holds.
IEEE TRANSACTIONS ON AUTOMATIC CONTROL
(2023)
Proceedings Paper
Automation & Control Systems
Giulia De Pasquale, Maria Elena Valcher
Summary: In this paper, two multi-dimensional Hagselmann-Krause (HK) models for opinion dynamics are considered, which describe how individuals adjust their opinions on multiple topics based on the influence of their peers. The models differ in the criterion for individuals to decide whom they want to be influenced from. It is shown that in the average-based model, individuals' opinions reach consensus only when their average opinions do so. In the uniform affinity model, individuals influence each other on each single topic only if their opinions do not differ more than a given tolerance, and the range of opinions on each single topic is non-increasing.
2022 IEEE 61ST CONFERENCE ON DECISION AND CONTROL (CDC)
(2022)
Proceedings Paper
Automation & Control Systems
Giulia De Pasquale, Maria Elena Valcher
Summary: This paper proposes a mathematical model for a Transactive Memory System (TMS) in the context of cooperative learning. The model examines the intertwined dynamics between individual expertise and the interaction network among cooperators. The findings suggest that if all agents are non-stubborn, they are able to acquire the competence of the most expert members in the group. However, when dealing with all stubborn agents, the ability to pass on tasks depends on the connectedness properties of the interaction graph.
2022 EUROPEAN CONTROL CONFERENCE (ECC)
(2022)
Proceedings Paper
Automation & Control Systems
Giulia De Pasquale, Maria Elena Valcher
Summary: The paper introduces a discrete time binary model based on the homophily social mechanism that dynamically reduces cognitive dissonance among agents in a social network, driving the network from initial structural imbalance towards social balance. The (V, Sigma)-factorization characterizes nonstructurally balanced equilibrium points and provides an interpretation in terms of structurally balanced classes, suitable for analyzing convergence to equilibrium in small-size networks.
2021 EUROPEAN CONTROL CONFERENCE (ECC)
(2021)
Proceedings Paper
Automation & Control Systems
Giulia De Pasquale, Maria Elena Valcher
Summary: This paper addresses two forms of consensus for multi-agent systems with specific decision classes, showing that simple modifications of DeGroot's algorithm can achieve tripartite consensus or sign consensus under certain assumptions about cooperative/antagonistic relationships between agents.
2021 AMERICAN CONTROL CONFERENCE (ACC)
(2021)
Proceedings Paper
Automation & Control Systems
Giulia De Pasquale, Maria Elena Valcher
2020 59TH IEEE CONFERENCE ON DECISION AND CONTROL (CDC)
(2020)
Proceedings Paper
Automation & Control Systems
Giulia De Pasquale, Yvonne R. Sturz, Maria Elena Valcher, Roy S. Smith
2020 59TH IEEE CONFERENCE ON DECISION AND CONTROL (CDC)
(2020)
Article
Automation & Control Systems
Ettore Fornasini, Maria Elena Valcher
IEEE CONTROL SYSTEMS LETTERS
(2020)
Proceedings Paper
Automation & Control Systems
Maria Elena Valcher, Gianfranco Parlangeli
2019 23RD INTERNATIONAL CONFERENCE ON SYSTEM THEORY, CONTROL AND COMPUTING (ICSTCC)
(2019)
Article
Computer Science, Theory & Methods
Fabio A. Schreiber, Maria Elena Valcher
INTERNATIONAL JOURNAL OF NEXT-GENERATION COMPUTING
(2019)
Proceedings Paper
Automation & Control Systems
Maria Elena Valcher
2019 18TH EUROPEAN CONTROL CONFERENCE (ECC)
(2019)
Proceedings Paper
Automation & Control Systems
Gianfranco Parlangeli, Maria Elena Valcher
2018 EUROPEAN CONTROL CONFERENCE (ECC)
(2018)