Physica A: Statistical Mechanics and its Applications最新文献

{"title":"Grouping mechanisms of vehicles in heterogeneous traffic with weak lane discipline: A single-site observational study focusing on leader–follower relations","authors":"Akihito Nagahama ,&nbsp;Katsuhiro Nishinari","doi":"10.1016/j.physa.2025.131032","DOIUrl":"10.1016/j.physa.2025.131032","url":null,"abstract":"<div><div>The widespread adoption of automobiles has accelerated global economic growth and improved daily convenience; however, the increase in the number of automobiles has led to severe traffic congestion, especially in developing countries with two-dimensional (2D) mixed traffic. In these mixed traffic conditions, each vehicle type exhibits distinct behaviors, which influence both microscopic and macroscopic traffic characteristics. Previous studies have shown that the composition or sequence of vehicle types in one-dimensional mixed traffic shapes traffic characteristics. Although the local composition and collective dynamics of motorcycles in 2D mixed traffic have been extensively investigated, an analysis that accounts for the dynamics of all vehicle types remains scarce. This study aims to detect “LF-groups (Leader–Follower groups)” in which vehicles tend to maintain leader–follower relationships in various traffic situations and identify the reasons for such group formation. Using video traffic observations on a single road segment in Mumbai, India, leader–follower estimation, graph mining techniques, and statistical comparisons, we enumerated all LF-groups formed in different traffic situations and their tendencies with respect to the compositions of vehicle types. As a hypothesis-generating result, our findings suggest that LF-group formation and nonformation in each traffic situation can be explained by the following three factors: similarity in speed, acceleration, and deceleration (maneuver similarity); surrounding space within the focusing combination of vehicle types (spatial confinement); and surrounding space for vehicles outside the focusing combination of vehicle types (permeability). The interaction among these factors across all vehicle type combinations leads to the formation of LF-groups with multiple vehicle types. Moreover, our findings offer a novel perspective: mixed traffic comprises not only groups but also “collections” of adjacent vehicles that continue to travel closely while their leader–follower relationships keep changing, as well as residual vehicles. Our findings may facilitate the active creation or elimination of LF-groups to improve 2D mixed traffic flow.</div></div>","PeriodicalId":20152,"journal":{"name":"Physica A: Statistical Mechanics and its Applications","volume":"680 ","pages":"Article 131032"},"PeriodicalIF":3.1,"publicationDate":"2025-10-11","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"145334427","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":"Pedestrian trajectory prediction method based on social force – Dynamic risk field coupled graph attention network","authors":"Yuan Gao,&nbsp;Yunfeng Wu","doi":"10.1016/j.physa.2025.131040","DOIUrl":"10.1016/j.physa.2025.131040","url":null,"abstract":"<div><div>Accurate pedestrian trajectory prediction plays a critical role in enhancing traffic safety at unsignalized intersections and advancing the deployment of autonomous driving technologies. To address the limitation of existing models in fully capturing the complex pedestrian-vehicle interactions at such intersections, this paper proposes a pedestrian trajectory prediction method based on a dual-domain coupling graph attention network that integrates social force and dynamic risk field models. The method employs an improved social force model to characterize pedestrian-to-pedestrian interactions and a dynamic risk field model to describe pedestrian-vehicle interactions. These interaction representations are mapped to the edge weights of the graph attention network, enabling adaptive fusion of multi-modal interaction effects. Furthermore, residual connections and a dynamic gating mechanism are incorporated to enhance feature propagation and adaptively balance the contributions of pedestrian and vehicle features. Finally, a LSTM-based encoder-decoder framework is utilized to generate the predicted trajectories. Experimental results on the DUT (Dalian University of Technology Anti-UAV Dataset) and SDD (Stanford Drone Dataset) demonstrate that the proposed method significantly improves the accuracy and reliability of pedestrian trajectory prediction in complex pedestrian-vehicle interaction scenarios.</div></div>","PeriodicalId":20152,"journal":{"name":"Physica A: Statistical Mechanics and its Applications","volume":"680 ","pages":"Article 131040"},"PeriodicalIF":3.1,"publicationDate":"2025-10-10","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"145270568","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":"DDMGPN: A derivative-driven multi-graph propagation network with traffic knowledge graph for traffic flow prediction","authors":"Jiayi Cao,&nbsp;Jianzhong Chen","doi":"10.1016/j.physa.2025.131038","DOIUrl":"10.1016/j.physa.2025.131038","url":null,"abstract":"<div><div>In dynamic urban environments, accurate traffic flow prediction faces three major challenges: intricate spatio-temporal dependencies, integration of heterogeneous data, and abrupt state changes. This paper proposes a novel <strong><u>D</u></strong>erivative-<strong><u>D</u></strong>riven <strong><u>M</u></strong>ulti-<strong><u>G</u></strong>raph <strong><u>P</u></strong>ropagation <strong><u>N</u></strong>etwork (DDMGPN), synergized with a Traffic Knowledge Graph (TKG) to address these challenges. The TKG integrates multi-source data (e.g., road topology, points of interest, overhead view images) into a unified knowledge representation to systematically encode prior knowledge. Building upon this foundation, DDMGPN introduces three innovative components to enhance spatio-temporal modeling. First, a derivative-driven feature modulation mechanism integrates first and second derivatives of traffic flow data, enabling joint modeling of trend evolution and abrupt state changes in traffic flow. Second, a multi-graph synergistic architecture combines a knowledge-guided static prior graph, a flow evolution dynamic graph, and a flow variation dynamic graph, establishing a three-stage knowledge propagation paradigm for spatio-temporal modeling. Finally, a temporal propagation amplifier (TPA) incorporates adaptive attention and derivative amplification, mitigating error accumulation in multi-step predictions. Comprehensive experimental evaluations conducted on two real-world datasets show that DDMGPN achieves state-of-the-art performance, both for short-term predictions and long-term predictions. Moreover, we visualize the learned spatio-temporal adjacency matrix to enhance the interpretability of our proposed model.</div></div>","PeriodicalId":20152,"journal":{"name":"Physica A: Statistical Mechanics and its Applications","volume":"680 ","pages":"Article 131038"},"PeriodicalIF":3.1,"publicationDate":"2025-10-10","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"145334276","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":"A logarithmically deformed entropy functional","authors":"José Weberszpil","doi":"10.1016/j.physa.2025.131029","DOIUrl":"10.1016/j.physa.2025.131029","url":null,"abstract":"<div><div>Stretched exponential distributions appear in disordered systems, glassy dynamics, and anomalous diffusion, yet their thermodynamic origin is often phenomenological. In this work, we propose a deformed entropy functional of the form <span><math><mrow><msub><mrow><mi>S</mi></mrow><mrow><mi>γ</mi></mrow></msub><mrow><mo>[</mo><mi>p</mi><mo>]</mo></mrow><mo>=</mo><mo>−</mo><msub><mrow><mo>∑</mo></mrow><mrow><mi>i</mi></mrow></msub><msub><mrow><mi>p</mi></mrow><mrow><mi>i</mi></mrow></msub><msup><mrow><mfenced><mrow><mo>ln</mo><mfrac><mrow><mn>1</mn></mrow><mrow><msub><mrow><mi>p</mi></mrow><mrow><mi>i</mi></mrow></msub></mrow></mfrac></mrow></mfenced></mrow><mrow><mn>1</mn><mo>/</mo><mi>γ</mi></mrow></msup></mrow></math></span>, which generalizes the Shannon entropy through a logarithmic deformation parameter <span><math><mi>γ</mi></math></span>. We show that, when maximized under standard constraints, this entropy leads asymptotically to stretched exponential (Weibull-type) distributions without requiring nonlinear constraints. The entropy is non-additive for <span><math><mrow><mi>γ</mi><mo>≠</mo><mn>1</mn></mrow></math></span>, tunably extensive, and concave in well-defined regimes. We establish its Lesche stability and derive its asymptotic variational behavior analytically. This framework offers a consistent thermodynamic foundation for modeling systems with memory, heterogeneity, or long-range correlations. The approach extends the Havrda–Charvát–Tsallis paradigm and contributes to the ongoing development of generalized thermodynamics by introducing a stretched-logarithmic entropy consistent with stretched exponential statistics.</div></div>","PeriodicalId":20152,"journal":{"name":"Physica A: Statistical Mechanics and its Applications","volume":"680 ","pages":"Article 131029"},"PeriodicalIF":3.1,"publicationDate":"2025-10-09","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"145271049","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":"Unveiling complex nonlinear dynamics in stock markets through topological data analysis","authors":"Chun-Xiao Nie","doi":"10.1016/j.physa.2025.131025","DOIUrl":"10.1016/j.physa.2025.131025","url":null,"abstract":"<div><div>Testing and characterizing nonlinear serial dependence in financial time series constitutes a critical research focus, extensively applied in examining weak-form market efficiency. This study demonstrates ATCC’s capability to capture nonlinear dependence and employs it to analyze equity market return series. Our findings reveal that rolling-window ATCC can characterize high-resolution dynamics of dependence. For instance, using minute-level data, we document how the Russia–Ukraine conflict information significantly impacted dependence structures in the Chinese market. Furthermore, based on daily index data, the 2025 Trump tariff policies are shown to have substantially influenced dependence patterns in both Chinese and U.S. market indices. Notably, through combined ATCC and linear modeling of SSE 50 constituent returns, we find that while linear models adequately characterize dependence in most daily returns, a minority of stocks exhibit nonlinear serial dependence. This research establishes an ATCC-based analytical framework, providing an effective quantitative tool for investigating nonlinear serial dependence and its high-resolution dynamics.</div></div>","PeriodicalId":20152,"journal":{"name":"Physica A: Statistical Mechanics and its Applications","volume":"680 ","pages":"Article 131025"},"PeriodicalIF":3.1,"publicationDate":"2025-10-09","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"145271045","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":"A modified lattice hydrodynamic traffic flow model incorporating preceding vehicle jerk dynamics under V2V communication","authors":"Md. Zakir Hosen ,&nbsp;Md. Anowar Hossain ,&nbsp;Jun Tanimoto","doi":"10.1016/j.physa.2025.131037","DOIUrl":"10.1016/j.physa.2025.131037","url":null,"abstract":"<div><div>In this letter, we investigate a novel aspect of the jerk sensitivity effect at the preceding site of the lattice hydrodynamics model, which emerges from the motion of motor vehicles equipped with vehicle-to-vehicle (V2V) communication. This study aims to reduce traffic congestion, smooth traffic flow, and optimize flow stability in both rural and urban areas by integrating V2V technologies. The influence of this preceding vehicle jerk sensitivity has been thoroughly analyzed using both linear and nonlinear analytical approaches. The phase diagrams demonstrate that our proposed model contributes significantly to alleviating traffic instabilities by enhancing predictive jerk sensitivity information from the preceding vehicle. A nonlinear stability analysis has been conducted to derive the modified Korteweg-de Vries (mKdV) equation near the critical point and to characterize complex traffic phenomena such as stop-and-go waves, fluctuations, wave propagation, and phase transitions. All theoretical and numerical results have been systematically compared to the Nagatani and Redhu models. The theoretical core findings of this work have been examined through numerical simulations, which demonstrate that incorporating the preceding vehicle's predictive jerk sensitivity substantially reduces traffic jams and enhances flow stability.</div></div>","PeriodicalId":20152,"journal":{"name":"Physica A: Statistical Mechanics and its Applications","volume":"680 ","pages":"Article 131037"},"PeriodicalIF":3.1,"publicationDate":"2025-10-08","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"145270569","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":"A black-box attack method of machine learning algorithms based on quantum autoencoders","authors":"Dong Tan,&nbsp;Lili Yan,&nbsp;Jiayu Zhao,&nbsp;Yan Chang,&nbsp;Shibin Zhang","doi":"10.1016/j.physa.2025.131033","DOIUrl":"10.1016/j.physa.2025.131033","url":null,"abstract":"<div><div>Currently, researchers have conducted extensive studies on adversarial attacks in the field of machine learning. With the development of quantum computing technology, quantum computing has provided new ideas and methods for implementing machine learning algorithms. Meanwhile, the issue of adversarial attacks in quantum machine learning has increasingly become a research hotspot. This paper proposes a new black-box attack method against quantum machine learning models based on a quantum autoencoder (QAE). The method first obtains a basic dataset through a small number of queries to the model, then expands this basic dataset to obtain a training dataset. The training dataset is used to train a surrogate model to generate adversarial examples, and then the transferability of the adversarial examples is utilized to launch attacks, ultimately achieving a black-box attack on the target model. Experiments show that the proposed method only requires 20 queries to the target model. Based on the results of these queries, the quantum autoencoder can be used to expand the basic dataset, and the accuracy of the surrogate model for attacking the target model is improved by 8% on the generated test set. Moreover, compared with the deep convolutional generative adversarial network (DCGAN) model, this method can achieve faster fitting. After training, the effectiveness of transfer based attacks on the surrogate model only decreases by less than 20% under strong perturbation conditions, and under certain conditions, the attack effect on the target model is stronger than that on the surrogate model itself. In addition, using the surrogate model to attack another quantum neural network model also achieves similar effects to those on the target model, thereby further verifying the universality of the proposed attack method.</div></div>","PeriodicalId":20152,"journal":{"name":"Physica A: Statistical Mechanics and its Applications","volume":"680 ","pages":"Article 131033"},"PeriodicalIF":3.1,"publicationDate":"2025-10-08","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"145334428","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":"Impact of headway uncertainty on traffic flow: Asymptotic and local stability analysis of car-following model","authors":"Darshana Yadav ,&nbsp;Vikash Siwach ,&nbsp;Poonam Redhu","doi":"10.1016/j.physa.2025.130998","DOIUrl":"10.1016/j.physa.2025.130998","url":null,"abstract":"<div><div>In car-following behavior, a vehicle’s motion is primarily influenced by the headway and velocity of the preceding vehicle. However, uncertainties in these parameters arising from poor road surface quality, malfunctioning sensors, adverse weather conditions, and driver variability can significantly affect traffic flow dynamics. This study proposes an extended car-following model based on the “Full Velocity Difference” (FVD) model, which incorporates headway and velocity uncertainties along with a cooperative driving mechanism. Control theory is employed to derive both local and asymptotic stability criteria, allowing an investigation into the influence of uncertainties on traffic behavior. Through linear stability analysis and nonlinear analysis, the model’s neutral stability condition is obtained, along with the derivation of the associated Burgers and “modified Korteweg–de Vries” (mKdV) equations. The proposed model outperforms the existing models in the literature with respect to the size of the stability region. Numerical simulations demonstrate that headway and velocity uncertainties notably affect vehicle startup behavior and the stability of traffic flow. Additionally, power spectrum theory is used to analyze spectral entropy, offering deeper insights into the impact of uncertainty on traffic dynamics through simulation.</div></div>","PeriodicalId":20152,"journal":{"name":"Physica A: Statistical Mechanics and its Applications","volume":"680 ","pages":"Article 130998"},"PeriodicalIF":3.1,"publicationDate":"2025-10-08","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"145271046","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":"Efficiency paradox and robustness of elevator-assisted evacuation in a deeply buried subway station","authors":"Shanshan He ,&nbsp;Qiao Wang ,&nbsp;Juan Chen ,&nbsp;Heng Ding ,&nbsp;Jian Ma","doi":"10.1016/j.physa.2025.131035","DOIUrl":"10.1016/j.physa.2025.131035","url":null,"abstract":"<div><div>In deeply buried subway stations, passengers using stairs and escalators need to walk a long vertical distance to reach the ground, which leads to prolonged evacuation times. On the contrast, elevators can vertically transport passengers directly to the ground. This study aims to study the evacuation efficiency after adding an elevator exit during daily commutes under varying buried depths and passenger flows. For this purpose, models for elevator scheduling and the subway station were developed. Three types of simulation scenarios were set up. Scenario A analyzed the baseline condition without elevator exits. Scenario B incorporated varying proportions of passengers opting for the elevator exit. Scenario C examined the combined effects of route-changing behavior alongside variations in elevator capacity and operating speed. The acceptable queue size was considered to reflect the psychology of passengers who change evacuation routes. An evacuation time ratio (ETR) was used to analyze evacuation efficiency quantitatively. It turns out that adding an evacuation elevator without proper allocation of passengers to evacuation routes can result in the evacuation efficiency paradox (EEP). To avoid EEP, the robustness of evacuation efficiency was analyzed. Results show that directing passengers to optimize their evacuation routes is more effective for improving overall efficiency than merely increasing elevator capacity and speed. When an acceptable queue size is under 30 passengers, evacuation efficiency improves by 15 % across all depths, with greater gains at shallower depths. Taking construction and operating costs into consideration, it is better to maintain a proper queue size and select an elevator with around 15-passenger capacity and 1–2 m/s speed. When a subway station is deeply buried or higher evacuation efficiency is required, increasing the elevator’s rated capacity and speed becomes necessary. This paper can provide novel ideas for the control strategy of large passenger flow and offer guidance on the design of elevator exits in deeply buried stations.</div></div>","PeriodicalId":20152,"journal":{"name":"Physica A: Statistical Mechanics and its Applications","volume":"680 ","pages":"Article 131035"},"PeriodicalIF":3.1,"publicationDate":"2025-10-08","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"145334275","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":"Dynamic passenger flow assignment in urban rail transit networks based on AFC data","authors":"Haoyu Wu,&nbsp;Junwei Zeng,&nbsp;Yongsheng Qian,&nbsp;Xu Wei","doi":"10.1016/j.physa.2025.131034","DOIUrl":"10.1016/j.physa.2025.131034","url":null,"abstract":"<div><div>In order to study the optimization of urban rail transit train schedules, this research aims to derive dynamic passenger flow data for network sections over time based on time-varying OD (origin-destination) passenger flow data. AFC (Automatic Fare Collection) data is divided into several groups of OD passenger flow data based on a set time granularity. By utilizing passenger travel time parameters derived from AFC data, as well as the actual inter-station train running times, and considering the impact of network passenger flow on travel time, a congestion coefficient is applied to the path cost in order to describe the travel time cost of passengers. The Method of Successive Algorithm (MSA) is employed to dynamically assign passenger flow for each time segment of the 4-hour calculation period, using a 5-minute granularity. The results of the multi-dimensional analysis of dynamic passenger flow assignment show that: (1) In large-scale networks, the efficiency of passenger flow assignment for high passenger volumes is at least 58.34 % higher than before the improvement, with each iteration’s convergence progress error not exceeding 3 % for different time periods. (2) While ensuring the output path results and quantities are consistent, the efficiency is significantly improved compared to traditional path search algorithms. (3) The introduction of the transfer count in the objective function improved the optimal objective function value by 5.49 % compared to the generalized path cost function alone.</div></div>","PeriodicalId":20152,"journal":{"name":"Physica A: Statistical Mechanics and its Applications","volume":"680 ","pages":"Article 131034"},"PeriodicalIF":3.1,"publicationDate":"2025-10-07","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"145270565","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}