Physica A: Statistical Mechanics and its Applications最新文献

{"title":"6-point tripled Ashkin–Teller global phase diagrams in two and three dimensions","authors":"Deniz Ipek Zeynioğlu ,&nbsp;A. Nihat Berker","doi":"10.1016/j.physa.2025.131049","DOIUrl":"10.1016/j.physa.2025.131049","url":null,"abstract":"<div><div>The tripled Ashkin–Teller model including 6-point interactions is solved in <span><math><mrow><mi>d</mi><mo>=</mo><mn>2</mn></mrow></math></span> and 3 by renormalization-group theory that is exact on the hierarchical lattice and approximate on the recently first/second-order-transition improved Migdal–Kadanoff procedure. Five different ordered phases occur in the dimensionally distinct global phase diagrams. 16 different phase diagram cross-sections in the 2-point and 4-point interaction space are obtained, with first- and second-order phase transitions, multiple tricritical points and critical endpoints.</div></div>","PeriodicalId":20152,"journal":{"name":"Physica A: Statistical Mechanics and its Applications","volume":"680 ","pages":"Article 131049"},"PeriodicalIF":3.1,"publicationDate":"2025-10-15","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"145334423","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":3,"RegionCategory":"物理与天体物理","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}

{"title":"Bayesian Kolmogorov-Arnold networks: Uncertainty-aware interpretable modeling through probabilistic spline decomposition","authors":"Masoud Muhammed Hassan","doi":"10.1016/j.physa.2025.131041","DOIUrl":"10.1016/j.physa.2025.131041","url":null,"abstract":"<div><div>Deep learning has emerged as an essential tool in many industries, including healthcare. Traditional deep learning models lack interpretability and omit to take prediction uncertainty into account, two crucial components of clinical decision-making. In order to produce explainable and uncertainty-aware predictions, we present Bayesian Kolmogorov-Arnold Networks (Bayesian-KANs), a novel neural architecture that rigorously implements the Kolmogorov-Arnold representation theorem while providing quantified uncertainty estimates. Unlike existing KAN implementations, our method formally connects each network component to the theorem's mathematical structure through probabilistic splines. It introduces Bayesian uncertainty quantification in both inner (<span><math><mi>ψ</mi></math></span>) and outer (<span><math><mi>φ</mi></math></span>) functions, and hence enables interpretable feature interaction analysis via uncertainty-aware decomposition. We evaluated Bayesian-KANs on four benchmark medical datasets: Pima Indians Diabetes, Cleveland Heart Disease, Breast Cancer Wisconsin (Diagnostic), and Hepatitis, and observed consistently superior performance over baseline models. Bayesian-KAN achieves accuracies of 80.1 %, 85.7 %, 96.2 %, and 88.5 %, respectively, with significantly tighter confidence intervals and higher AUC-ROC and F1 scores. Comparative analyses showed that Bayesian-KAN not only outperforms logistic regression, SVMs, traditional neural networks, and deterministic KANs in predictive accuracy, but also provides more calibrated and trustworthy uncertainty estimates. Additionally, Bayesian-KAN successfully highlights clinically relevant features in all datasets, enhancing transparency in decision-making. Moreover, Bayesian-KANs' capacity to represent aleatoric and epistemic uncertainty guarantees doctors receive more solid and trustworthy decision support. Our Bayesian strategy improves the interpretability of the model and considerably minimises overfitting, which is important for tiny and imbalanced medical datasets, according to experimental results. We present possible expansions to further use Bayesian-KANs in more complicated multimodal datasets and address the significance of these discoveries for future research in building reliable AI systems for healthcare. This work paves the way for a new paradigm in deep learning model deployment in vital sectors where transparency and reliability are crucial.</div></div>","PeriodicalId":20152,"journal":{"name":"Physica A: Statistical Mechanics and its Applications","volume":"680 ","pages":"Article 131041"},"PeriodicalIF":3.1,"publicationDate":"2025-10-14","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"145333577","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":3,"RegionCategory":"物理与天体物理","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}

{"title":"Spontaneous order in cryptocurrency markets: Crash dynamics and complexity–entropy causality plane","authors":"Huy Quoc Bui ,&nbsp;Christophe Schinckus ,&nbsp;Hamdan Al-Jaifi ,&nbsp;Boon-Keong Lim","doi":"10.1016/j.physa.2025.131050","DOIUrl":"10.1016/j.physa.2025.131050","url":null,"abstract":"<div><div>This study examines the dynamic evolution of spontaneous order in cryptocurrency markets using the Complexity–Entropy Causality Plane (CECP) methodology. By analyzing closing price data from seven major cryptocurrencies between 2017 and 2025, we find that although these assets generally exhibit behavior consistent with a random walk, a clear bifurcation emerges. Newer and less established cryptocurrencies—such as DOGE, ADA, and BNB—display higher complexity and shift toward chaotic regimes, while BTC, XRP, ETH, and TRX maintain lower complexity and greater randomness, suggesting higher informational efficiency. Additionally, the analysis reveals distinct disorder-to-order transitions preceding several major market downturns, indicating that spontaneous order may function as an early warning signal of systemic stress.</div></div>","PeriodicalId":20152,"journal":{"name":"Physica A: Statistical Mechanics and its Applications","volume":"680 ","pages":"Article 131050"},"PeriodicalIF":3.1,"publicationDate":"2025-10-14","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"145334280","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":3,"RegionCategory":"物理与天体物理","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}

{"title":"MADDPG-GST for coordinated variable speed limit and ramp metering: A hybrid action deep reinforcement learning approach to bottleneck congestion mitigation","authors":"Tun Qiu,&nbsp;Pan Liu,&nbsp;Zhibin Li,&nbsp;Chengcheng Xu,&nbsp;Kailai Qiu,&nbsp;Shunchao Wang","doi":"10.1016/j.physa.2025.131045","DOIUrl":"10.1016/j.physa.2025.131045","url":null,"abstract":"<div><div>Expressway merging bottlenecks are major sources of traffic congestion, where insufficient coordination among multiple traffic streams leads to severe flow disruptions. Although variable speed limits (VSL) and ramp metering (RM) are commonly used to mitigate congestion, their independent operation and mismatched control scopes often result in suboptimal outcomes. To address this, this study proposes a coordinated VSL–RM strategy based on multi-agent deep reinforcement learning. The control task is modeled as a Markov Decision Process (MDP), allowing joint policy learning between decentralized VSL and RM agents. A customized Multi-Agent Deep Deterministic Policy Gradient (MADDPG) algorithm is employed to dynamically optimize coordination policies. To bridge the gap between discrete VSL and continuous RM action spaces, the Gumbel-Softmax Trick (GST) is integrated into the learning process for differentiable hybrid action optimization. Additionally, a transfer learning mechanism is incorporated to ensure efficient policy adaptation across diverse traffic scenarios. Simulation results under varying demand levels show that the proposed strategy achieves 7.3 %–34.1 % improvements in traffic efficiency and stability compared to traditional methods. It also demonstrates strong transferability, reducing retraining time by up to 63.7 % and traffic delays by up to 62.7 %, while maintaining robust control under overspeed disturbances and control lag conditions.</div></div>","PeriodicalId":20152,"journal":{"name":"Physica A: Statistical Mechanics and its Applications","volume":"680 ","pages":"Article 131045"},"PeriodicalIF":3.1,"publicationDate":"2025-10-14","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"145334277","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":3,"RegionCategory":"物理与天体物理","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}

{"title":"Adaptive hybrid spatial hypergraph convolution module with data embedding optimization for stock ranking prediction","authors":"Yicheng Qian,&nbsp;Pufan Pang","doi":"10.1016/j.physa.2025.131046","DOIUrl":"10.1016/j.physa.2025.131046","url":null,"abstract":"<div><div>Spatio-temporal data mining have various applications in the domains of finance, transportation, and sociology. Predicting stock rankings is a typical case that presents certain challenges. These challenges include: (1) The inability of existing graph learning methods, Regardless of how comprehensive their prior knowledge used for constructing graph relationships are, to generate significant improvements due to their lack of adaptive capturing of highdimensional data structures. (2) In time series forecasting. Stationarizing the data can more effectively capture data trends, but this operation may lead to the loss of important non-stationary factor information. (3) Spatio-temporal data mining models typically integrate time, space, and graph learning modules. Different modules with different functionalities often require different training environments. Many models rely solely on end-to-end optimization through the loss function. This leads to insufficient driving force for downstream modules, gradient environment disruptions, training blockages, and other issues. To address these challenges, we propose the Adaptive Hybrid Spatial Hypergraph Convolution Network (<em>AHS-HGCN</em>). Specifically, multiple multi-functional attention mechanisms are introduced to model the main task and provide suitable training environments for downstream modules. Among them, the HSCA module is capable of outputting hybrid spatial hyperedge weights and rewriting upstream outputs, maximizing the efficiency of model training. We evaluate the framework on three large-scale real stock datasets (NASDAQ, NYSE, and TSE). Compared to the baseline models, it achieved a minimum improvement of 27.22 % and a maximum improvement of 30.89 %.</div></div>","PeriodicalId":20152,"journal":{"name":"Physica A: Statistical Mechanics and its Applications","volume":"680 ","pages":"Article 131046"},"PeriodicalIF":3.1,"publicationDate":"2025-10-14","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"145334274","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":3,"RegionCategory":"物理与天体物理","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}

{"title":"Binary option market manipulation by influencing belief dynamics","authors":"Henry Waldhausen ,&nbsp;Christopher Griffin","doi":"10.1016/j.physa.2025.131036","DOIUrl":"10.1016/j.physa.2025.131036","url":null,"abstract":"<div><div>Using techniques from information geometry, we construct a semi-Hamiltonian system modelling trader beliefs in a binary asset market and study the impact of inequality or asymmetry in beliefs, information, and power on price dynamics. We show that in a market with no inequality and <span><math><mi>N</mi></math></span> completely symmetric traders, the resulting dynamics evolve on a <span><math><mrow><mn>2</mn><mi>N</mi><mo>+</mo><mn>1</mn></mrow></math></span> dimensional manifold consisting of a <span><math><mrow><mn>2</mn><mi>N</mi><mo>−</mo><mn>2</mn></mrow></math></span> dimensional centre manifold, a 2 dimensional stable manifold and a 1 dimensional slow manifold. Introducing asymmetry into the traders has the potential to decrease the dimension of the centre manifold, which we prove using a parameter analysis. Using the belief model, we also study the impact of inter-agent communication, exogenous information and asymmetric purchasing power on price dynamics, showing that market bubbles can emerge when powerful traders produce outsize influence in the market, thus impacting other traders’ beliefs as well as the price. This process is exacerbated when back-channel communication is permitted. The impact of areas of high curvature in belief space is also discussed.</div></div>","PeriodicalId":20152,"journal":{"name":"Physica A: Statistical Mechanics and its Applications","volume":"680 ","pages":"Article 131036"},"PeriodicalIF":3.1,"publicationDate":"2025-10-13","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"145334420","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":3,"RegionCategory":"物理与天体物理","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}

{"title":"Turning dynamics of pedestrian social groups on a stair landing: A field study","authors":"Tuantuan Lu ,&nbsp;Pengfei Zhu","doi":"10.1016/j.physa.2025.131043","DOIUrl":"10.1016/j.physa.2025.131043","url":null,"abstract":"<div><div>Pedestrian social groups play a critical role in crowd dynamics. However, how group members interact during turning on stair landings remains unclear. In this study, a field observation was conducted on a stair landing to investigate the turning dynamics of social groups. A total of 122 descending social groups and 40 individuals were observed, and their trajectories were extracted. The results highlight that the position of group members contributes significantly to the movement characteristics. Specifically, within the same group, outer members (with larger turning radii) have higher walking speed but lower angular speed, whereas inner members exhibit the opposite pattern. As the outermost member tends to arrive at the landing earlier and adjust movement direction sooner than others, a left-right asymmetry in the relative positions is observed within social groups. During turning, group members dynamically regulate their movement by reducing speed and shortening interpersonal distances, enabling them to re-establish a spatiotemporally cohesive structure. Moreover, a stepping strategy for maintaining group cohesion is found, whereby social groups exhibit shorter step lengths and step durations compared to non-group pedestrians when the speed exceeds 0.8 m/s. Finally, the relation between movement time and walking radius is established to estimate the travel paths of social groups on stair landings. These findings have implications for group walking modeling and the design of public spaces.</div></div>","PeriodicalId":20152,"journal":{"name":"Physica A: Statistical Mechanics and its Applications","volume":"680 ","pages":"Article 131043"},"PeriodicalIF":3.1,"publicationDate":"2025-10-13","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"145334424","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":3,"RegionCategory":"物理与天体物理","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}

{"title":"Review of the percolation threshold for spherocylinder-based systems in a continuum model","authors":"Meysam Khodaei ,&nbsp;Mohsen Jafaraghaei ,&nbsp;Ashkan Ajrian ,&nbsp;Sina Giahkar","doi":"10.1016/j.physa.2025.131044","DOIUrl":"10.1016/j.physa.2025.131044","url":null,"abstract":"<div><div>This study presents a comprehensive review and comparative analysis of various methods for determining the percolation threshold in systems of spherocylinders—a critical parameter in the design of advanced composite materials. We evaluated a range of approaches, including analytical models based on excluded volume theory (soft- and hard-core), computational Monte Carlo simulations, and established experimental techniques. A central focus was reconciling the discrepancies between theoretical models, which often assume infinite aspect ratios, and experimental results from fillers with finite aspect ratios. Our analysis reveals that while traditional analytical bounds and hard-core models exhibit limited predictive accuracy, computational soft-core simulations for finite-sized fillers provide robust predictions that align well with experimental data. Moreover, empirical approximations fitted to numerical results demonstrate strong agreement across all aspect ratio regimes. The primary contribution of this work is a novel interpolation formula that unifies the distinct asymptotic behaviours observed at very low and very high aspect ratios. This formula shows excellent agreement with extensive simulation data and serves as a highly accurate, unified predictive tool. By clarifying the strengths and weaknesses of existing methods, this investigation provides a reliable framework for accurately predicting the percolation threshold in spherocylinder-based systems.</div></div>","PeriodicalId":20152,"journal":{"name":"Physica A: Statistical Mechanics and its Applications","volume":"680 ","pages":"Article 131044"},"PeriodicalIF":3.1,"publicationDate":"2025-10-13","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"145334414","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":3,"RegionCategory":"物理与天体物理","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}

{"title":"Experimental study on the impact of visibility on bidirectional pedestrian flow under high and low urgency conditions","authors":"Qing Deng ,&nbsp;Zheng Zhou ,&nbsp;Ziyue Xu ,&nbsp;Yanchao Ye ,&nbsp;Ye Xu ,&nbsp;Quanyi Liu ,&nbsp;Huiling Jiang ,&nbsp;Xiaole Zhang ,&nbsp;Lida Huang ,&nbsp;Guoray Cai","doi":"10.1016/j.physa.2025.131042","DOIUrl":"10.1016/j.physa.2025.131042","url":null,"abstract":"<div><div>Bidirectional pedestrian flow often leads to significant casualties and property damage when out of control, especially in emergencies. This study aims to investigate the characteristics of bidirectional pedestrian flow evacuation under different visibility and urgency conditions. Bidirectional pedestrian evacuation experiments are conducted under three different visibility conditions and two levels of urgency. The evacuation behavioral patterns are analyzed, including typical evacuation behaviors, evacuation speeds, and evacuation time of bidirectional pedestrian flows. Overtaking behavior, following behavior and boundary fast effect are observed during the evacuation process. As visibility decreased, pedestrian evacuation speeds also declined. The reduction in pedestrian speed due to diminished visibility was more pronounced under high-urgency conditions compared to low-urgency scenarios. Compared with non-emergency situations, in emergency situations, the moving speeds of pedestrians moving in the opposite direction are 61.90 %, 34.62 %, and 26.25 % faster when they are not wearing eye-patches, wearing an eye patch with a light transmittance of 27 %, and wearing eye-patches with a light transmittance of 16 %, respectively. The results highlight the critical impact of visibility on evacuation efficiency in emergencies. These findings provide empirical data on the influence of visibility and urgency level on emergency evacuation performance and offer valuable insights for crowd management.</div></div>","PeriodicalId":20152,"journal":{"name":"Physica A: Statistical Mechanics and its Applications","volume":"680 ","pages":"Article 131042"},"PeriodicalIF":3.1,"publicationDate":"2025-10-11","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"145334417","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":3,"RegionCategory":"物理与天体物理","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}

{"title":"Modeling the self-organizing dynamics of pedestrian flow in subway stations under the constraints of guide barriers","authors":"Tao Yu ,&nbsp;Mengxuan Jie ,&nbsp;Liqiang Zhao ,&nbsp;Shuixiong Tang ,&nbsp;Jinjin Tang","doi":"10.1016/j.physa.2025.131039","DOIUrl":"10.1016/j.physa.2025.131039","url":null,"abstract":"<div><div>With the rapid development of dense urban rail transit networks, subway stations are experiencing increasingly frequent large-scale passenger flow events, which has led to the adoption of pedestrian guide barriers as a standard practice for ensuring orderly movement in crowd management. However, the underlying mechanisms through which these barriers influence pedestrian flow dynamics remain insufficiently understood. Traditional social force models fail to capture pedestrians' adaptive path selection under such constraints, limiting their ability to simulate pedestrian trajectory distribution in varying crowd densities. To address this, we propose a Dynamic-Adaptive Social Force Model (DSFM) that integrates time-urgent behaviors, collision avoidance, and an adaptive path selection strategy based on spatial occupancy. Two typical scenarios, a serpentine barrier and a linear platform barrier, were simulated using real-world station data. The DSFM demonstrates significant superiority over the traditional Social Force Model (SFM). In parametric analyses of the serpentine barrier, the DSFM reduced average travel times by 11.7–36.4 % and traffic conflicts by 21.8–57.9 %. Furthermore, it accelerated spatial utilization, with the cumulative growth rate of channel occupancy peaking at 85.95 % relative to the SFM. In the high-density linear barrier scenario, the DSFM improved passage efficiency by 12.92 %. These results, validated across various crowd densities and geometries, confirm the DSFM's robustness and accuracy. This research provides a novel, validated simulation tool for optimizing guide barriers, enhancing both passenger flow management and station service quality.</div></div>","PeriodicalId":20152,"journal":{"name":"Physica A: Statistical Mechanics and its Applications","volume":"680 ","pages":"Article 131039"},"PeriodicalIF":3.1,"publicationDate":"2025-10-11","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"145333579","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":3,"RegionCategory":"物理与天体物理","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}