Article
Engineering, Multidisciplinary
I-Cheng Lin, Osman Yagan, Carlee Joe-Wong
Summary: In networked systems, initial failures at a small part may trigger cascading failures, which can lead to the breakdown of the entire system. This vulnerability is more severe in modern systems with interdependent networks. The paper studies the robustness of such systems and proposes a dynamic coupling coefficient strategy to improve the system's resilience.
IEEE TRANSACTIONS ON NETWORK SCIENCE AND ENGINEERING
(2023)
Article
Physics, Multidisciplinary
Ning Wang, Zi-Yang Jin, Jiao Zhao
Summary: This study explores the robustness of interdependent networks considering different attack strategies, connection strategies, and load distribution mechanisms. Findings suggest that cascade failures are more evident in networks with assorted connections where load nodes are more concentrated. Additionally, networks with a DC strategy (connect nodes with high load in network A and nodes with low load in network B) exhibit the worst robust level against cascading failures under intentional attacks.
PHYSICA A-STATISTICAL MECHANICS AND ITS APPLICATIONS
(2021)
Article
Physics, Multidisciplinary
Mengyu Lv, Linqiang Pan, Xueming Liu
Summary: Previous research on the robustness of interdependent networks focused on undirected networks, while directed networks were only limited to random or targeted attacks. However, some failure scenarios such as earthquakes, floods, and epidemics cannot be described by these attacks as they are localized. This study introduces a theoretical framework to analyze the robustness of interdependent directed networks under localized attacks. It is found that for degree homogeneous networks, network robustness under localized attacks is similar to that under random attacks. For degree heterogeneous networks, localized attacks are more likely to lead to collapse than random attacks. The findings provide insights into network robustness and can contribute to the design of robust interdependent systems.
PHYSICA A-STATISTICAL MECHANICS AND ITS APPLICATIONS
(2023)
Article
Mathematics, Interdisciplinary Applications
Mattia Frasca, Lucia Valentina Gambuzza
Summary: This paper introduces a distributed control strategy to prevent dynamically-induced cascading failures in power grids. By modeling power grids using complex networks and nonlinear dynamics, the approach successfully prevents cascading failures in various power grid models.
CHAOS SOLITONS & FRACTALS
(2021)
Article
Engineering, Electrical & Electronic
Xinkai Fan, Ekaterina Dudkina, Lucia Valentina Gambuzza, Mattia Frasca, Emanuele Crisostomi
Summary: In this work, a novel model based on a structure-preserving approach is developed to describe a power grid. The model takes into consideration classic voltage and frequency protection mechanisms. By studying the Italian power grid, it is found that more realistic models are crucial in determining the size of cascading failures and the sequence of involved links.
ELECTRIC POWER SYSTEMS RESEARCH
(2022)
Article
Engineering, Industrial
Anna Varbella, Blazhe Gjorgiev, Giovanni Sansavini
Summary: Past events have shown that widespread blackouts are often caused by cascading failures in the power grid. Understanding the underlying mechanisms of these failures is important in developing strategies to minimize their risks. Real-time detection of precursors to cascading failures can help operators take preventive measures. This study proposes a data-driven methodology for online estimation of cascading failure risks by using deep learning techniques. The results show that the developed Graph Neural Network (GNN) model has an accuracy and balanced accuracy above 96% on selected test datasets, outperforming the Feed-forward Neural Network (FNN) model.
RELIABILITY ENGINEERING & SYSTEM SAFETY
(2023)
Article
Engineering, Multidisciplinary
Abdorasoul Ghasemi, Hermann de Meer
Summary: This article discusses the cascading failure process in interdependent power-communication networks, proposes weak and strong interdependency models, and presents a congestion-aware load balancing scheme. The study examines the cascade process in the system and captures the potential positive and negative effects of interdependency between the networks.
IEEE TRANSACTIONS ON NETWORK SCIENCE AND ENGINEERING
(2023)
Article
Engineering, Electrical & Electronic
Lin Ding, Lunxiao Xie, Xiao-Ke Xu
Summary: The study introduces a fresh framework for systematically analyzing cascading failures in interdependent networks composed of power and communication layers with different types of rich-clubs. Results indicate that networks with a weaker rich-club in the communication layer or a stronger inter-layer rich-club demonstrate better robustness, and optimal network performance can be achieved by considering both factors. These findings highlight the importance of rich-club connectivity in designing optimization strategies against cascading failures.
IEEE JOURNAL ON EMERGING AND SELECTED TOPICS IN CIRCUITS AND SYSTEMS
(2022)
Article
Computer Science, Information Systems
Rachad Atat, Muhammad Ismail, Shady S. Refaat, Erchin Serpedin, Thomas Overbye
Summary: Future smart grids involve multiple interactions between power grid and communication network, requiring joint vulnerability analysis to reduce the risk of large-scale cascading failures. A proposed attack strategy targets the most influential nodes in both subsystems, showing comparable or larger damage impact compared to traditional exhaustive search attack with lower complexity.
IEEE SYSTEMS JOURNAL
(2022)
Article
Multidisciplinary Sciences
Malgorzata Turalska, Ananthram Swami
Summary: This paper explores optimal conditions for controlling cascading failures in interdependent networks. The study finds that coupling a layer operating in a supercritical regime with a layer in a subcritical zone can reduce the probability of cascades. Additionally, there are parasitic and mutualistic control strategies identified in the research.
SCIENTIFIC REPORTS
(2021)
Article
Computer Science, Information Systems
Vetrivel Subramaniam Rajkumar, Alexandru Stefanov, Alfan Presekal, Peter Palensky, Jose Luis Rueda Torres
Summary: Cascading effects in the power grid can be caused by cyber attacks, which have become more prevalent with the increasing digitalisation of power systems. By studying historic blackouts and identifying key cyber-physical factors, we demonstrate the connection between cyber attacks and cascading failures. Our synthetic case-study and simulation results provide evidence for the impact of cyber attacks on power grids.
Article
Computer Science, Information Systems
Aliasghar Salehpour, Irfan Al-Anbagi, Kin-Choong Yow, Xiaolin Cheng
Summary: In this paper, a novel cyber-attack failure propagation model in smart grids is proposed, which accurately models cascading failures based on power and communication characteristics and interdependencies. The model identifies the most-vulnerable nodes to improve system robustness.
Article
Computer Science, Information Systems
Xiaoliang Wang, Fei Xue, Qigang Wu, Shaofeng Lu, Lin Jiang, Yue Hu
Summary: This article summarizes power grids as temporal weighted networks (TWNs) which have fixed topological structure but time varying weight distribution. It proposes an inverse-community (IC) structure to intuitively identify the risk of cascading failures and upgrades the conventional modularity to inverse-modularity (IM) to quantify the characteristic of IC structure in power networks. Additionally, a security/economic dispatch (SED) method is designed to handle the optimal power dispatch issues considering the risk of cascading failures represented by IM and the cost of power network operation. Simulation results and tests on the IEEE 118-bus system demonstrate the effectiveness of the proposed SED method in mitigating cascading failure risks.
IEEE INTERNET OF THINGS JOURNAL
(2023)
Article
Computer Science, Artificial Intelligence
Wei Huang, Yuxin Gao, Tianyi Zhang, Hua Gao
Summary: In this paper, a novel coupled model is proposed to address the robustness issue in cyber-physical systems under cascading failures. The Q-learning algorithm is adopted to find the optimal recovery sequence with minimal recovery times. Comparative analysis shows that branch recovery strategy requires less recovery cost and the Q-learning-based algorithm effectively finds the optimal sequence.
NEURAL COMPUTING & APPLICATIONS
(2023)
Article
Mathematics, Interdisciplinary Applications
Alessandra Corso, Lucia Valentina Gambuzza, Federico Malizia, Giovanni Russo, Vito Latora, Mattia Frasca
Summary: In this article, a method is proposed to reconstruct the active links of a power network using a second-order Kuramoto model and dynamically induced cascading failures. The method reconstructs the active links based on the evolution of the node phases and angular velocities. It is found that a small number of samples is crucial to accurately reconstruct the temporal sequence of faults, contradicting the need for larger data sets in the presence of noise. The number of samples used in the algorithm must be selected to balance prediction error and temporal resolution of the active link reconstruction.
JOURNAL OF COMPLEX NETWORKS
(2022)
Article
Computer Science, Hardware & Architecture
Novella Bartolini, Ting He, Viviana Arrigoni, Annalisa Massini, Federico Trombetti, Hana Khamfroush
IEEE-ACM TRANSACTIONS ON NETWORKING
(2020)
Article
Computer Science, Hardware & Architecture
Michael Lin, Novella Bartolini, Michael Giallorenzo, Thomas F. La Porta
IEEE-ACM TRANSACTIONS ON NETWORKING
(2020)
Article
Computer Science, Information Systems
Ala Khalifeh, Khalid A. Darabkh, Ahmad M. Khasawneh, Issa Alqaisieh, Mohammad Salameh, Ahmed AlAbdala, Shams Alrubaye, Anwar Alassaf, Samer Al-HajAli, Radi Al-Wardat, Novella Bartolini, Giancarlo Bongiovannim, Kishore Rajendiran
Summary: This paper proposes a new framework using WSNs for remote sensing and monitoring in smart city applications and suggests utilizing unmanned aerial vehicles as data mules. The experimental evaluation revealed inconsistencies between the performance metrics advertised and actual measurements in hardware datasheets.
Article
Computer Science, Information Systems
Novella Bartolini, Andrea Coletta, Gaia Maselli, Alar Khalifeh
Summary: The paper introduces a new concept - Weighted Progressive Coverage, aiming to address the problem of drone trip scheduling. The authors proposed an efficient algorithm and validated its superiority through simulations and prototype experiments.
IEEE TRANSACTIONS ON MOBILE COMPUTING
(2021)
Article
Computer Science, Information Systems
Andrea Coletta, Novella Bartolini, Gaia Maselli, Annalyse Kehs, Peter McCloskey, David P. Hughes
Summary: This article introduces a crowdsensing framework based on a mobile application for early detection of plant diseases and sustainable food production. By analyzing plant images and utilizing an AI engine, the application can recognize health issues in plants. The study demonstrates that this framework outperforms current solutions in terms of monitoring accuracy and completeness, while being more cost-effective.
IEEE INTERNET OF THINGS JOURNAL
(2022)
Article
Computer Science, Information Systems
Xiaojie Liu, Xingwei Wang, Jie Jia, Jianhui Lv, Novella Bartolini
Summary: This article introduces an effective UAV network system that provides long-lasting communication services while minimizing the number of UAVs through UAV deletion and insertion. A distributed algorithm based on virtual Coulomb force and Voronoi diagram is proposed to optimize UAV deployment, improve communication coverage, and turn off redundant UAVs. The proposed moving and sleeping schemes help improve communication coverage and save energy, respectively, resulting in a minimal number of UAVs required for sufficient coverage.
IEEE SYSTEMS JOURNAL
(2022)
Article
Computer Science, Information Systems
Novella Bartolini, Andrea Coletta, Gaia Maselli
Summary: Aerial drones are increasingly being used in monitoring tasks, but current trajectory planning solutions rely on perfect knowledge of ongoing events. We propose a solution for a squad of drones to autonomously inspect an area of interest under uncertainty of time and location of target events. Through extensive simulations and real-field experiments, we show that our proposal outperforms existing algorithms in terms of speed of discovering new events, percentage of visited events, and inspection delay.
IEEE TRANSACTIONS ON MOBILE COMPUTING
(2023)
Article
Computer Science, Hardware & Architecture
Viviana Arrigoni, Novella Bartolini, Annalisa Massini, Federico Trombetti
Summary: This paper proposes a Bayesian approach to address the issue of a large number of monitoring paths in Boolean Network Tomography (BNT) for failure identification. A polynomial-time greedy strategy is also introduced to approximate the posterior probabilities of node failures. Additionally, a monitoring technique considering information aging in dynamic failure scenarios is proposed.
IEEE-ACM TRANSACTIONS ON NETWORKING
(2023)
Proceedings Paper
Computer Science, Information Systems
Novella Bartolini, Andrea Coletta, Flavio Giorgi, Gaia Maselli, Matteo Prata, Domenicomichele Silvestri
Summary: Swarms of UAVs are a key technology for communication in harsh environments. This paper introduces a novel data offloading approach called Stop & Route, which takes advantage of controllable mobility to facilitate network routing.
2023 18TH WIRELESS ON-DEMAND NETWORK SYSTEMS AND SERVICES CONFERENCE, WONS
(2023)
Article
Computer Science, Information Systems
Novella Bartolini, Andrea Coletta, Flavio Giorgi, Gaia Maselli, Matteo Prata, Domenicomichele Silvestri
Summary: Swarms of UAVs are an important technology for supporting communication in disrupted or unavailable environments. This paper introduces a novel data offloading approach called Stop & Offload, which utilizes controllable device mobility to facilitate network routing. Extensive simulations demonstrate the superiority of this approach.
COMPUTER COMMUNICATIONS
(2023)
Proceedings Paper
Automation & Control Systems
Novella Bartolini, Andrea Coletta, Matteo Prata, Camilla Serino
Summary: Flying Ad-hoc Networks (FANETs) are crucial for safety critical scenarios, but face challenges in wide-area deployment. The connected deployment problem aims to ensure multi-hop low-latency communications for monitoring tasks.
2021 IEEE/RSJ INTERNATIONAL CONFERENCE ON INTELLIGENT ROBOTS AND SYSTEMS (IROS)
(2021)
Proceedings Paper
Computer Science, Hardware & Architecture
Novella Bartolini, Andrea Coletta, Andrea Gennaro, Gaia Maselli, Matteo Prata
Summary: MAD is a packet routing protocol specifically tailored for networks of aerial vehicles, utilizing device controllable mobility to facilitate network routing and supported by a reinforcement learning approach for movement-assisted delivery. Through extensive simulations, MAD is shown to outperform previous solutions in performance metrics including average packet delay, delivery ratio, and communication overhead, with a slight loss in average device availability.
2021 IEEE 41ST INTERNATIONAL CONFERENCE ON DISTRIBUTED COMPUTING SYSTEMS (ICDCS 2021)
(2021)
Article
Computer Science, Information Systems
Viviana Arrigoni, Novella Bartolini, Annalisa Massini
Summary: In this article, the author investigates the problem of node identifiability in Boolean Network Topography, provides theoretical bounds on the minimum number of necessary measurement paths, and offers an algorithmic approach to designing network topologies and path deployment. Through extensive simulations on synthetic and real network topologies, the author evaluates the tightness of the proposed lower bounds compared to state-of-the-art heuristic methods.
Proceedings Paper
Computer Science, Theory & Methods
Novella Bartolini, Andrea Coletta, Gaia Maselli, Mauro Piva, Domenicomichele Silvestri
Summary: In recent years, there has been a significant increase in the number of Unmanned Aerial Vehicles (UAVs) used in various scenarios, promising high coverage of monitored areas but facing constraints such as limited power. Gen-Path, a genetic algorithm developed for efficient scheduling of multi-round UAV missions, has shown to improve existing solutions in terms of coverage and energy cost through simulations.
IEEE CONFERENCE ON COMPUTER COMMUNICATIONS WORKSHOPS (IEEE INFOCOM WKSHPS 2021)
(2021)
Proceedings Paper
Computer Science, Hardware & Architecture
Viviana Arrigoni, Novella Bartolini, Annalisa Massini, Federico Trombetti
Summary: Boolean Network Tomography (BNT) allows for network failure localization through end-to-end monitoring paths, but efficiency is compromised in real scenarios when multiple failures occur simultaneously due to the large size of the solution space. To address this, a stochastic optimization-based progressive failure localization approach is proposed to maximize identification capabilities with a limited number of monitoring probes, employing greedy strategies for complexity management. Experimentation on real network topologies demonstrates the practical effectiveness of the approach, showcasing its superiority over existing solutions.
IEEE CONFERENCE ON COMPUTER COMMUNICATIONS (IEEE INFOCOM 2021)
(2021)
