Complexity最新文献

This study constructs a bilevel network model based on heterogeneous financial data to explore the complex network characteristics and risk transmission mechanisms in the stock market. Using the trading data and textual sentiment data of Shanghai Stock Exchange (SSE) 50 constituent stocks over the past 5 years, a daily return network model and a textual sentiment analysis network model are constructed, which are then combined to form a bilevel network. The study finds that the bilevel network model can more comprehensively capture the multidimensional relationships and risk transmission behaviors in the market, revealing the close connection between sentiment factors and returns. By analyzing the interlayer coupling characteristics of the bilevel network, we found that information and risks are efficiently transmitted between different network layers. This method not only provides a new perspective for financial market analysis but also offers a valuable theoretical basis and practical tools for risk management and market regulation. The results show that the bilevel network model has significant implications for understanding and preventing financial risks.

This study investigates a discrete-time predator–prey model that includes both prey refuge and memory effects. The research identifies the conditions under which fixed points exist and remain stable. A key focus is placed on analyzing different types of bifurcation—such as period doubling (PD), Neimark–Sacker (NS), and strong resonances (1 : 2, 1 : 3, and 1 : 4)—occurring at the positive fixed point to uncover their ecological significance. Bifurcation theory is applied to study these dynamics, and the theoretical findings are validated through numerical simulations performed with the MATLAB tool MatContM. In addition, a control mechanism is introduced to mitigate severe instabilities within the system. The results show that predation rate is key to ecological balance, while prey refuge has limited impact on stability. The study offers important insights for conserving biodiversity and managing ecosystems.

The scientific community is aware that the great scientific revolution of this century will be the formulation of a theory of complex systems and formalize it in mathematical terms. In this article, we formulate a unified theory of living systems as complex systems called the general biological relativity theory, which states that at every level of organization of a living system, there is no privileged or absolute scale, which would determine the dynamics of the living system, only interactions between the biological space–time scales of a level of organization of the structurally organized living system form and the biological size–time scales of a microlevel–macrolevel class of levels of organization of the functionally organized living system form. To date, such a theory has found little content because there has been very little that has been established that is common about the multilevel and multiscale organization of living systems. Drawing on a structurally organized living system form of lymphatic filariasis disease system as an example, we illustrate how this theory can be applied to extend the conceptual and multiscale modeling framework of living systems as complex systems.

Multiple datasets enable a deep learning model to achieve a wide range of classifications, while the diversity of datasets reduces classification accuracy. To solve this problem, a method based on multidimensional information is proposed. The first dimension is the outputs of multiple models on different datasets. Through this information, we can predict the dataset that may contain the testing samples. The second one is the outputs of multiple models on the same dataset, through which the labels of testing samples can be classified. The third one is the distribution of labels on the same testing sample, which further increases accuracy. Experimental results show that our method achieves the best performance compared to the existing methods while ensuring good scalability.

The aim of this work is to study the alumina–tantalum/motor oil hybrid nanoliquid flow in a porous cavity subjected to a uniform magnetic field. We have used the Darcy–Bénard convection model for the momentum equation and a new local thermal nonequilibrium formulation for heat transport. Linear stability theory and nonlinear stability theory based on the double Fourier series representation are used to study the onset of stationary and chaotic convection in the hybrid nanofluid. The analytical expression of the stationary thermal Rayleigh–Darcy number has been found as a function of dimensionless parameters and physicochemical properties of the hybrid nanoliquid. Tools such as bifurcation diagrams, Lyapunov exponent, phase spaces, and time histories were used to analyze the chaotic and regular behaviors of the resulting five-dimensional system of nonlinear equations. The added value of this work lies in the stabilization and control of chaotic convection in motor oil flow using hybrid alumina and tantalum nanoparticles and a uniform magnetic field.

Apple maturity detection algorithms based on deep learning typically involve a large number of parameters, resulting in high computational costs, long processing times, and dependence on high computational power graphics processing units (GPUs). This paper proposes an improved YOLOv8 model to address the issues related to the maturity detection of Fuji apples grown in China using image-based methods. The model was optimized in several ways according to the characteristics of apple targets and scenes. First, a lightweight MobileNetV3 is used as the backbone network, replacing the original CSPDarknet-53 backbone network, which reduces the model parameters and computational complexity and increases the inference speed. Second, by introducing the efficient multiscale attention (EMA) module and using the bidirectional feature pyramid network (BiFPN) in the neck part, the model enhances the extraction capability of important features and suppresses redundant features, thus improving the model’s generalization ability. Experimental results show that the size of the model is 2.6 megabytes. On the apple dataset, its precision, recall, F1 score, and mean average precision reach 90.2%, 88.5%, 89.3%, and 91.3%, respectively, with improvements of 4.3%, 3.2%, 3.7%, and 2.6% compared to the original model. Based on this model, an Android application has been developed for real-time apple maturity detection. The improved model proposed in this paper achieves real-time apple target recognition and maturity detection, providing quick and accurate target recognition guidance for the mechanical automatic harvesting of apples.

The rapid proliferation of Internet of Things (IoT) infrastructures has introduced significant security challenges due to device heterogeneity, dynamic interactions, and resource limitations. Traditional intrusion detection systems (IDSs) often struggle to capture temporal dependencies and emergent behaviors inherent in modern IoT cyber threats. This study presents a novel hybrid framework that combines deep recurrent neural networks (RNNs), specifically long short-term memory (LSTM) architectures, with complex network modeling to enhance the detection and classification of sophisticated attacks. The proposed system leverages normalized and labeled IoT traffic data, encompassing multiple attack classes (e.g., DoS, DDoS, Brute Force, MITM, and Replay) to train an LSTM-based IDS capable of multiclass temporal analysis. Simultaneously, an IoT network environment is simulated using graph-theoretic principles, where each node represents a device characterized by parameters such as latency, energy usage, and communication protocols. Cyberattack scenarios are emulated within this network to facilitate real-time detection of anomalous behaviors. Experimental results demonstrate the effectiveness of the proposed model in capturing sequential patterns and improving detection accuracy in complex IoT environments.

Influence maximization (IM) is a concept in social network analysis and data science that focuses on finding the most influential nodes (people, users, etc.) in a network to maximize the spread of information, behavior, or influence. IM studies have become more crucial due to the quick uptake of social media and networking technologies, which have revolutionized communication and information sharing. Using information from the Scopus database, this study conducts a thorough bibliometric analysis of the literature on instant messaging from 2006 to 2024 to investigate publishing trends, significant contributors, and developing themes. The three primary issues the study attempts to answer are finding the most productive journals, nations, and scholars in IM research; assessing the growth and influence of publications; and predicting future research trends. The results show that IM research is dominated by China and the US, with significant contributions from organizations like the Department of Computer Science and Microsoft Research Asia. The development of the field toward scalable algorithms and practical applications is highlighted by highly cited articles, such as Chen’s (2009) work on successful instant messaging. The investigation also shows the possibility of incorporating AI into future advancements and points out shortcomings in behaviorally informed techniques. This study offers a valuable summary of information management research for academics and professionals trying to understand this ever-evolving topic.

The emergence of hubs in scale-free networks plays a critical role in understanding dynamic complex networks such as social interactions, transportation networks, and biological processes. Given that real-world scale-free networks are dynamic and time based, a temporal-scale-free network (TSF network) is proposed in this paper. To predict the emergence of hubs, proposed a temporal graph convolutional neural network (T-GCN) that integrates graph convolutional networks (GCNs) for spatial feature extraction and long short-term memory (LSTM) networks for modeling temporal dynamics. Our framework effectively learns both the structural evolution and dynamic node interactions in scale-free networks, allowing accurate prediction of hub emergence. The proposed model is trained on synthetic and real-world datasets, demonstrating superior predictive accuracy compared to traditional methods. Our findings provide valuable insights into the mechanisms governing hub formation and offer a robust framework for forecasting influential nodes in evolving networks.

In this paper, we develop a periodic impulsive switching stage-structured model to investigate the population dynamics of species exhibiting hibernation behavior. The model incorporates stage structure (larvae and adults), birth pulses occurring exclusively in the active season, and impulsive harvesting events taking place immediately after hibernation. By combining switched dynamical systems with impulsive differential equations, we accurately capture the seasonal alternation between active and dormant states along with discrete reproductive and harvesting pulses. Using the Jury criterion, we establish sufficient conditions for the local asymptotic stability of both the population extinction periodic solution and the positive periodic solution. Furthermore, we identify an explicit extinction-survival threshold Γ and analyze how key parameters such as hibernation duration, harvesting rate, and birth pulse intensity govern population persistence. Numerical simulations not only validate the analytical results but also uncover complex nonlinear dynamics, including period-doubling bifurcations and chaotic oscillations, as the birth coefficient increases. These findings provide theoretical insights for wildlife conservation and sustainable harvesting strategies concerning hibernating species.