Complex Network Model Research Articles

Semantic Web-Based Identification of Key Causative Factors and Chains for Team Error in NPPs' ACR

Team errors are identified as a major contributor to incidents in safety-critical systems like nuclear power plants (NPPs). Safety in NPPs is more dependent on team performance than individual, and reducing team errors requires identifying key causative factors and chains. Current studies primarily use qualitative analysis whereas lacking quantitative methods. This study addressed this by constructing a complex network model of team error based on semantic features and coding of the Semantic Web and determining criteria for selecting key causative factors by three node failure strategies. The top 5 team error key PSFs and top 5 high-risk causative chains were identified in the results. Finally, a set of control strategies was proposed further. This research can guide identifying key causative factors which could be a good starting point to scrutinize the team performance and establish control strategies for team error in NPPs.

Situational Awareness and Intelligence Mining in National Security Research Based on Complex Network Models

Situational Awareness and Intelligence Mining in National Security Research Based on Complex Network Models

Deep learning resilience inference for complex networked systems

Resilience, the ability to maintain fundamental functionality amidst failures and errors, is crucial for complex networked systems. Most analytical approaches rely on predefined equations for node activity dynamics and simplifying assumptions on network topology, limiting their applicability to real-world systems. Here, we propose ResInf, a deep learning framework integrating transformers and graph neural networks to infer resilience directly from observational data. ResInf learns representations of node activity dynamics and network topology without simplifying assumptions, enabling accurate resilience inference and low-dimensional visualization. Experimental results show that ResInf significantly outperforms analytical methods, with an F1-score improvement of up to 41.59% over Gao-Barzel-Barabási framework and 14.32% over spectral dimension reduction. It also generalizes to unseen topologies and dynamics and maintains robust performance despite observational disturbances. Our findings suggest that ResInf addresses an important gap in resilience inference for real-world systems, offering a fresh perspective on incorporating data-driven approaches to complex network modeling.

Executive myopia, industry partners, and financial performance: a simulation study using complex network and propagation model

Executive myopia, industry partners, and financial performance: a simulation study using complex network and propagation model

Exploring an adaptive management model for “status-optimization-regulation” of mining city in transition: A case study of Huangshi, China

Mining cities face the challenges of ecological transformation and sustainable development after mineral depletion. Thus, ecological space optimization and adaptive management are pivotal after ecological restoration project, but easily neglected. Taking Huangshi in Hubei Province as an example, a top-down adaptive management model for regional mining ecological space integrating status-optimization-regulation (SOR) was established based on land system resilience (LSR) evaluation, circuit theory, complex network model and community discovery algorithm. The results showed that i) As the LSR increased, the land use structure shifted from non-ecological zones to forest and cropland, indicating that balancing agriculture and forest protection was crucial for increasing LSR in Huangshi. i) The ecological networks (ENs) of Huangshi had an irregular fishnet pattern, with densely intricate corridors in the south and broader, sparser ones in the north, and numerous ecological barriers near the mining area. Furthermore, the distribution of ecological sources and corridors displayed complementarity and hierarchy. iii) The optimized ENs exhibited higher connectivity and efficiency under disturbance, reducing corridor redundancy and migration costs. iv) The ENs clusters were classified into five types based on ecological connectivity: ecological buffer zone, priority restoration zone, moderate restoration zone, natural restoration zone, and moderate development zone, and tailored ecological regulatory strategies were proposed. The findings provide practice-oriented guidance for the sustainable development of mining cities, and offer a direct approach to support conscious, clear, and coherent adaptive management of ecological restoration in rapidly changing environments.

Assessment of Ventilation Potential and Construction of Wind Corridors in Chengdu City Based on Multi-Source Data and Multi-Model Analysis

The establishment of urban ventilation corridors (UVCs) aims to mitigate the urban heat island effect. While most studies focus on the construction and assessment of the environmental benefit of UVCs, they often overlook the analysis of UVCs’ topological features. This research integrates multi-source data including 3D urban buildings, historical meteorological observations, high-resolution remote sensing, and land use planning, combined with multiple models, including geographic information system spatial analysis, circuit theory, and complex networks. Based on an assessment of urban ventilation potential, the circuit model was applied to extract UVCs aligned with the prevailing wind direction for both summer and winter seasons. Complex network modeling was employed to analyze the topological features of the ventilation network. From the analytical results, a multi-level wind corridor system for Chengdu was quantitatively developed. The results indicate that the city’s overall ventilation resistance is high, with notable spatial clustering, and the southeastern region faces substantial ventilation obstructions. A total of 143 critical ventilation nodes were identified, with the number of air inlets and outlets in summer being significantly fewer than in winter. However, the cooling effect of ventilation corridors in the prevailing summer wind direction is superior to that in winter. The ventilation network comprises 16 communities with distinct ventilation characteristics, exhibiting moderate connectivity, lacking small-world properties, and showing congestion and instability.

AI in the Courtroom: Enhancing Legal Decision-Making through Predictive Modelling

Legal outcome prediction or legal judgement prediction is the process of making predictions of the possible outcome of a legal proceeding based on its case file. This work focuses on the importance of allowing less privileged or marginalised groups to have better access to law services. This work discusses the implementation of 13 different deep learning models and the results obtained for each of them. The range of models employed encompasses a spectrum from basic perceptron models to more complex convolutional network models and encoder-based models. The models like multi-layer perceptron (MLP), convolutional neural network (CNN) model, BERT, LegalBERT, DistilBERT, RoBERTa and the different combinations of these encoder-based models are implemented. The input to these models is the case files and their corresponding judgement status. Four different classes of petitions and appeals, submitted to the court, were considered to train all the models. The efficiency of these models was tested in terms of various performance metrics like training accuracy, testing accuracy, training loss, validation loss and [Formula: see text]1 score.

A clustered-LPSEIRS malware propagation model in complex networks

In this paper, we present a new malware propagation model that integrates epidemic spread, clustering, and link prediction techniques, tailored for complex network networks. Our model is based on the clustered-link prediction-susceptible-exposed-infected-recovered (clustered-LPSEIRS) epidemic model, which simulates malware dissemination within the network. Our findings reveal a significant decrease in the rate of malware spread compared to the traditional SEIR model, with this enhancement in containment attributed to the integration of clustering and link prediction methods. We also compute the basic reproduction ratio (R0\\documentclass[12pt]{minimal} \\usepackage{amsmath} \\usepackage{wasysym} \\usepackage{amsfonts} \\usepackage{amssymb} \\usepackage{amsbsy} \\usepackage{mathrsfs} \\usepackage{upgreek} \\setlength{\\oddsidemargin}{-69pt} \\begin{document}$$R_{0}$$\\end{document}) for our model, providing insights into the potential ramifications of malware within the network. By examining parameter variations, we enhance our understanding of the model's behavior under diverse scenarios. Additionally, we assess the influence of clustering and link prediction on mitigating malware spread, emphasizing its effectiveness in diminishing the overall impact.

A complex neural network model by Hilbert Transform

The phase information of the optical wave plays a vital role in processing wave-related signals. In deep learning fields, complex-valued neural networks are put forward on the concept of complex amplitudes for full utilization of phase information. To build a complex-valued neural network, the common way is to exploit Fourier Transform of the observed signal to extract amplitude and phase information. However, this will lead to spectrum waste for a real-valued signal by introducing negative frequencies that have no physical meaning. To this end, we attempt to use Hilbert Transform as an alternative to yield a single sideband spectrum and avoid negative frequencies from interacting with positive ones. On the other hand, Fourier transform is a global analysis thus it tells nothing about the time domain. As our key insight, we further explore the usage of instantaneous frequency calculated by Hilbert Transform and propose a new method of constructing complex input from a time–frequency angle. Simple pixel-wise classification experiments are carried out on two hyperspectral datasets and MNIST dataset. Experimental results have demonstrated that Hilbert Transform with instantaneous frequency performs better by a large margin than Fourier Transform owing to the additional time information.

Application of artificial intelligence technologies at substantiation of effective control algorithms for the electrical engineering complex of urban electric transport traction electrical equipment

Background. Studies on the current state and development trends of urban ground public electric transportation in Russia highlight the urgent need for innovative technologies. These innovations should focus on designing and operating new types of rolling stock, traction electrical equipment, and promising types of electric traction. At the same time, special attention should be paid to building autonomous control systems for electric transportation using artificial intelligence technologies. Aim. This study aims to explore the application of neural networks to develop algorithms for effectively controlling the electrical engineering complex of traction electrical equipment in urban ground rail electric transport. Materials and Methods. The research utilized data from studies on the traction electrical equipment modes of rolling stock. These studies were conducted through both computer simulations and field experiments under real operating conditions of urban electric transport using different control algorithms. By applying probability theory and mathematical statistics, the analysis identified the correlations between operational and energy parameters of rolling stock movement and the operational mode of its traction electrical equipment. Results. The identified correlation dependencies informed the design of an effective network architecture, including its size and complexity, as well as the composition of their training samples. This led to the development of an original, simplified algorithm for determining effective control parameters for the electrical complex of traction electrical equipment during the movement of a vehicle on a given section of track. Conclusion. The research concluded that using “simple” neural networks for calculating the parameters of effective control of traction electrical equipment operation modes in urban electric transport provides higher speed and sufficient accuracy compared to complex neural network models. These results are valuable for developers of intelligent control systems for streetcar transportation.

Computing Zagreb Connection Indices for the Cartesian Product of Path and Cycle Graphs

Topological indices ( ) are a class of graph-based descriptors widely used in chemiformatics and Quantitative Structure-Activity Relationship (QSAR) studies. TIs capture key structural features of molecules by encoding graph-theoretic information, offering a quantitative representation of molecular topology. Each mathematical network is associated with a specific numerical value determined by a function. Among the Zagreb connection indices have been extensively researched. This study delves into the Cartesian product of cycle graphs with path graphs, elucidating the comprehensive implications of . These indices encompass the first , second , and third . Furthermore, comprehensive results of the modified first , modified second , and modified third are presented, along with the first multiplicative ZCI second multiplicative ZCI (SMZCI), and third multiplicative ZCI Moreover, modified first multiplicative ZCI ), modified second multiplicative ZCI and modified third multiplicative ZCI ( are also calculated. To provide precision, both the graphical and numerical analyses of the computed findings are aligned for the two Cartesian products. The foundation for mathematically modeling complex networks and chemical structures lies in graph theory. Topological indices ( ) are a class of graph-based descriptors widely used in chemiformatics and quantitative structure-activity relationship (QSAR) studies. TIs capture key structural features of molecules by encoding graph-theoretic information, offering a quantitative representation of molecular topology. Each mathematical network is associated with a specific numerical value determined by a topological index function. Among these the Zagreb connection indices are extensively researched TIs. This article delves into the Cartesian product of cycle graphs with path graphs, elucidating the comprehensive implications of . These indices encompass the first, Second and third. Furthermore, we present the comprehensive results of the modified , modified and modified , and also present the first multiplicative Zagreb connection index , and , along with their modified counterparts like modified first multiplicative Zagreb connection index , and . These analysis encompass two distinct types of graphs: cycle graphs and path graphs. To provide further precision, we align both the graphical and numerical analysis of our computed findings for these two Cartesian products.

Risk analysis of accident-causing evolution in chemical laboratory based on complex network

Chemical laboratories usually have many high risks from various dangerous chemicals, dangerous devices, and experimental operators involved in the laboratory. Once a severe accident occurs, the accident consequence will significantly impact scientific research innovation and social progress. Aiming at laboratory safety problems, an accident evolution model is established to reveal the critical accident-causing factors and the most likely evolutionary path of accidents in the process of accident evolution. Firstly, the 24 model is used to identify the cause factors of laboratory accidents and the causal relationship between factors. Then, the complex network model of laboratory accident evolution is constructed with accident-causing factors as nodes and causal relationships between factors as edges. Secondly, using the TOPSIS method, four typical complex network characteristic indexes, including degree centrality, betweenness centrality, closeness centrality, and eigenvector centrality, are constructed into a comprehensive evaluation index to reveal the key influencing factors in the accident evolution process. Finally, the fuzzy number is introduced to quantify the uncertainty of the accident evolution path. The most likely evolutionary path is obtained through the Dijkstra and risk entropy theory. Based on the data of 151 laboratory accidents from 2001 to 2022, the accident evolution analysis of chemical laboratories was carried out. The evolution results show that the maximum probability value is the evolution path of fire accidents, and the minimum probability value is the evolution path of electric shock accidents, which accords with the general statistical law of accidents. When applied to the actual laboratory, the safety management level of the laboratory can be effectively improved by controlling the key influencing factors and cutting off the accident evolution path.

Complex network analysis for accident causes modelling to enhance process safety in chemical enterprises

AbstractHazardous chemicals often cause catastrophic accidents, and accidents often result from intricate interactions among various causes. Due to the varying risk factors in different areas of chemical enterprises, to achieve more precise prevention, a more detailed study of the accident risk factors in each area is necessary. Therefore, this study focuses on analyzing critical accident causes and their interrelationships in different functional areas of chemical enterprises to enhance process safety by using a complex network model. Based on 90 accident information, complex network models are constructed for hazardous chemical warehouse areas (HCWAs), tank farm areas (TFAs), and production areas (PAs). Subsequently, a topological analysis of the complex network models is conducted. Based on the PageRank algorithm, 13 critical nodes are identified for HCWAs, while 14 for TFAs and 13 for PAs. Node degree analysis with confidence quantifies mutual influences, forming critical accident causal links for each area. The research results offer decision support for precise accident risk control, aiding in reducing future accidents and improving process system safety in chemical enterprises.

A Giza Pyramids Construction metaheuristic approach based on upper bound calculation for solving the network reliability problem

Network reliability optimization is an optimization problem that focuses on finding an optimal solution for a reliable network design. In network reliability optimization, the goal is to maximize network reliability so that the overall cost of the network is reduced at the same time. The discussion of the reliability of various systems in the field of industry and engineering is of great importance, that's why reliability optimization has received a lot of attention in recent decades. Since this problem is included in the category of NP-Hard problems, the use of soft computing methods will be highly effective in solving it. In this paper, an approach based on the Giza Pyramids Construction (GPC) metaheuristic algorithm is proposed to solve the network reliability problem. For this purpose, two independent single objective functions with constraints are defined, and then the reliability of the network is calculated through optimistic estimation using the upper bound method. In order to compare the performance, 12 types of diverse and complex network models have been generated and the proposed algorithm has been compared with 10 popular and state-of-the-art algorithms. Statistical analysis has been used to find significant differences in the performance of algorithms. Also, a real model of the university network has been generated, investigated, and solved as a case study. The results of experiments, statistical analysis, and observations show that the proposed algorithm has a better performance than other metaheuristic algorithms and the proposed approach in solving reliability is an effective and low-cost approach.

Research on the Influencing Factors of Carbon Emissions in Low-carbon Communities based on Complex Network Theory

Achieving carbon peak before 2030 and carbon neutrality before 2060 is a solemn commitment made by China to address global climate change, and is also one of the main goals for economic and social development in the 14th Five Year Plan and the 2035 vision period. The changes in carbon emissions are directly related to the progress of China's "carbon peak" and "carbon neutrality" goals. Therefore, in-depth research on the influencing factors of carbon emissions has become a key link in promoting the achievement of this goal. In the existing research on carbon emission influencing factors, countries mainly focus on macro scale low-carbon urban carbon emission influencing factors and micro scale low-carbon building full life cycle carbon emission influencing factors. However, there is relatively little research on the influencing factors of carbon emissions in low-carbon communities of urban micro units, and there is still considerable research space. This study conducted an in-depth analysis of the influencing factors of carbon emissions in low-carbon communities using complex network methods. By constructing a complex network model of factors affecting carbon emissions, we identified key nodes and pathways, and explored their interrelationships. The results indicate that factors such as energy structure, resident behavior, building design, and policy implementation play an important role in carbon emissions in low-carbon communities.

Multiple-ResNet GAN: An enhanced high-resolution image generation method for translation from fundus structure image to fluorescein angiography.

Fluorescein angiography (FA) is a diagnostic method for observing the vascular circulation in the eye. However, it poses a risk to patients. Therefore, generative adversarial networks have been used to convert retinal fundus structure images into FA images. Existing high-resolution image generation methods employ complex deep network models that are challenging to optimize, which leads to issues such as blurred lesion boundaries and poor capture of microleakage and microvessels. In this study, we propose a multiple-ResNet generative adversarial network (GAN) to improve model training, thereby enhancing the ability to generate high-resolution FA images. First, the structure of the multiple-ResNet generator is designed to enhance detail generation in high-resolution images. Second, the Gaussian error linear unit (GELU) activation function is used to help the model converge rapidly. The effectiveness of the multiple-ResNet is verified using the publicly available Isfahan MISP dataset. Experimental results show that our method outperforms other methods, achieving better quantitative results with a mean structural similarity of 0.641, peak signal-to-noise ratio of 18.25, and learned perceptual image patch similarity of 0.272. Compared with state-of-the-art methods, the results showed that using the multiple-ResNet framework and GELU activation function can improve the generation of detailed regions in high-resolution FA images.

Cascading failure model and resilience-based sequential recovery strategy for complex networks

Complex networks, which exhibit high connectivity, self-organization, small-world properties, and heterogeneity, are susceptible to the rapid spread of local failures, often resulting in cascading effects throughout the entire system. The paper introduces a cascading failure model based on biased random walks that incorporate betweenness centrality and the power-law distribution of node degrees. This model is used to investigate cascade failures triggered by extreme fluctuations in load that follow a Poisson distribution. Furthermore, we propose a resilience-based sequential recovery strategy that accounts for varying node recovery time and resource limitations, setting an upper limit on the number of nodes that can be in recovery simultaneously. The network’s robustness improves, and the variation in the power-law exponent during cascading failures and recovery decreases when the betweenness bias parameter is set to 1 instead of -1. The capacity parameter has the most significant and direct effect on improving the network’s robustness. Reducing node recovery time can improve the network’s initial invulnerability; however, its impact on final residual resilience remains limited. The power-law exponent of the initial network significantly affects residual resilience during the recovery process, with higher exponents leading to improved network performance. An appropriate increase in the number of nodes that can be in recovery simultaneously can enhance the overall recovery performance of the network. Extensive comparative simulations reveal substantial advantages of our proposed recovery strategy in enhancing network recovery.

Research on coal mine safety risk evolution and key hidden dangers under the perspective of complex network

In order to find out the main causes of coal mine safety accidents and improve the pertinence of coal mine safety risk management and control, the identification and analysis of coal mine safety risks and hidden dangers are carried out based on the analysis of coal mine accident reports. Combing the complex network theory, a complex network model for the evolution of coal mine safety risks is constructed. The key elements that affect coal mine safety risk accidents are obtained through quantitative research on the characteristic indicators of the complex network model of coal mine safety risks. And the key nodes of coal mine safety risk spread network are obtained through network interference to the overall efficiency. The research results show that the complex network of coal mine safety risks illustrate the characteristics of a small-world network, and the spread of a certain risk is likely to cause coal mine safety accidents. Strengthening the risk management and control of hidden dangers with higher intermediate centrality can isolate the spread of coal mine safety risks and reduce the possibility of coal mine accidents.

AutoPeptideML: a study on how to build more trustworthy peptide bioactivity predictors.

Automated machine learning (AutoML) solutions can bridge the gap between new computational advances and their real-world applications by enabling experimental scientists to build their own custom models. We examine different steps in the development life-cycle of peptide bioactivity binary predictors and identify key steps where automation cannot only result in a more accessible method, but also more robust and interpretable evaluation leading to more trustworthy models. We present a new automated method for drawing negative peptides that achieves better balance between specificity and generalization than current alternatives. We study the effect of homology-based partitioning for generating the training and testing data subsets and demonstrate that model performance is overestimated when no such homology correction is used, which indicates that prior studies may have overestimated their performance when applied to new peptide sequences. We also conduct a systematic analysis of different protein language models as peptide representation methods and find that they can serve as better descriptors than a naive alternative, but that there is no significant difference across models with different sizes or algorithms. Finally, we demonstrate that an ensemble of optimized traditional machine learning algorithms can compete with more complex neural network models, while being more computationally efficient. We integrate these findings into AutoPeptideML, an easy-to-use AutoML tool to allow researchers without a computational background to build new predictive models for peptide bioactivity in a matter of minutes. Source code, documentation, and data are available at https://github.com/IBM/AutoPeptideML and a dedicated web-server at http://peptide.ucd.ie/AutoPeptideML. A static version of the software to ensure the reproduction of the results is available at https://zenodo.org/records/13363975.

Analysis of risk influential factors of marine pilots during the embarkation and disembarkation

Due to the complex geographical conditions within the port waters, it is necessary to take pilotage operations. The embarkation and disembarkation (E&D) of marine pilots is the riskiest phase of pilotage work, with accidents occurring frequently. In order to examine the key Risk Influential Factors (RIFs) of E&D accident of marine pilots, in this study, a novel complex network model of RIFs of marine pilots during the E&D process is proposed. Firstly, based on the E&D accident investigation report of marine pilots, the causal chain of E&D accidents is extracted, and the RIFs leading to E&D accidents are identified. Secondly, a complex network model of RIFs of marine pilots during the E&D is constructed based on the complex networks theory, and the network reliability is verified. Finally, by comprehensive use of topology characteristics analysis, robustness analysis and module analysis, the key RIFs of pilot E&D accidents are identified. The research results show that pilot and management factors have important impact on pilot safety, ‘insufficient pilot ladder strength' is the most critical RIF. It should be focused on RIFs with high connectivity between modules and within modules, and the risk evolution can be blocked through the control of those RIFs. Abbreviations: BNs: Bayesian Networks; CDM: Critical Decision Method; DEMATEL: Decision-making Trial and Evaluation Laboratory; E&D: Embarkation and Disembarkation; FSA: Formal Safety Assessment; RIFs: Risk Influential Factors; RPD: Recognition-Primed Decision; SHERPA: Systematic Human Error Reduction and Prediction Approach; SOLAS: International Convention for the Safety of Life at Sea; VTS: Vessel Traffic Services; TECHR: Technique for Early Consideration of Human Reliability.