Transportation Research Part C-Emerging Technologies最新文献

{"title":"Real-time traffic simulation and management for large-scale urban air mobility: Integrating route guidance and collision avoidance","authors":"Canqiang Weng ,&nbsp;Can Chen ,&nbsp;Jingjun Tan ,&nbsp;Tianlu Pan ,&nbsp;Renxin Zhong","doi":"10.1016/j.trc.2025.105477","DOIUrl":"10.1016/j.trc.2025.105477","url":null,"abstract":"<div><div>With a vision to expand transportation supply using low-altitude airspace, urban air mobility (UAM) has emerged as a promising alternative to provide point-to-point travel in congested areas. The rapid development of electric vertical take-off and landing vehicles is expected to make UAM a viable and sustainable transportation mode. Given the spatial heterogeneity of land use patterns in most cities, large-scale UAM deployments will likely focus on specific areas, such as intertransfer traffic between suburbs and city centers. However, large-scale UAM operations connecting multiple origin-destination pairs raise concerns about air traffic safety and efficiency due to potential conflict movements, particularly at major conflict points analogous to roadway junctions. To meet the safety and efficiency requirements of future UAM operations, this work proposes an air traffic management framework that integrates route guidance and collision avoidance. The route guidance mechanism optimizes aircraft distribution across both spatial and temporal dimensions by regulating their paths (composed of waypoints). Given the optimized paths, the collision avoidance algorithm generates collision-free aircraft trajectories between waypoints in the 3D space. To enable large-scale applications, we develop fast approximation methods for centralized path planning and adopt the velocity obstacle model for distributed collision avoidance. To our knowledge, this work is one of the first to integrate route guidance and collision avoidance for UAM. Simulation results demonstrate that the proposed framework enables efficient and flexible UAM operations, including air traffic assignment, local congestion mitigation, and dynamic no-fly zone management. Compared with a collision-free baseline strategy, the proposed framework achieves considerable improvements in traffic safety and efficiency, with increases in the average minimum separation (+98.2 %), the average travel speed (+70.2 %), and the trip completion rate (+130 %), along with a reduction in the energy consumption (-23.0 %). The proposed framework demonstrates its potential for real-time traffic simulation and management in large-scale UAM systems.</div></div>","PeriodicalId":54417,"journal":{"name":"Transportation Research Part C-Emerging Technologies","volume":"183 ","pages":"Article 105477"},"PeriodicalIF":7.6,"publicationDate":"2026-02-01","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"145689766","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":1,"RegionCategory":"工程技术","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}

{"title":"Which type of backpressure is more stable? – Comparative analysis based on two-movement intersections","authors":"Dianchao Lin ,&nbsp;Li Li","doi":"10.1016/j.trc.2025.105478","DOIUrl":"10.1016/j.trc.2025.105478","url":null,"abstract":"<div><div>Stability, which indicates that queues do not grow infinitely over time, is a key concept in control policies such as BackPressure (BP). However, its abstract nature and diverse definitions make its comparative analysis difficult both theoretically and experimentally. As a result, simulations in existing studies often use alternative metrics, such as average delay, to evaluate the performance of different control policies. Little research directly compares different stabilities through theory or experiments. In this paper, we compare seven common stability definitions and theoretically demonstrates that they are equivalent in simulations and applications. Furthermore, we propose a <em>t</em>-test method for identifying whether a queue is stable based on the sequence of queueing differences. This method allows us to classify any sampled demand as stable or unstable based on simulated queues for a given control policy. Therefore, if the network’s dimension, i.e., the number of movements, does not exceed three, we can directly draw the stability region (SR) for all policies and compare their sizes. To accurately reproduce various BP theories, ensure fair comparisons, and facilitate the visualization of SRs, we use simulation codes to simulate a two-movement intersection scenario and discuss its extension to networks. Six distinct types of BP policies are compared, along with analysis for fixed-time and actuated controls. We obtain many insights that are difficult to achieve through purely theoretical analysis and delay-based simulations, including: 1) variability in BP’s SR: the SR typically varies when the BP changes its queue status weight or efficiency weight; 2) size hierarchy of SR: BPs generally outperform actuated controls in terms of SR, and actuated controls tend to outperform fixed-time controls; 3) non-cyclic vs. cyclic BP: non-cyclic BP usually has a larger SR than cyclic BP; 4) effect of real-time supply information: using real-time supply increases the SR of BP, even under the assumption of fixed saturation headway; and 5) SR degradation phenomenon: longer cycle lengths in cyclic BP may cause its SR to degenerate into a rectangular shape typical of fixed-time control.</div></div>","PeriodicalId":54417,"journal":{"name":"Transportation Research Part C-Emerging Technologies","volume":"183 ","pages":"Article 105478"},"PeriodicalIF":7.6,"publicationDate":"2026-02-01","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"145657758","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":1,"RegionCategory":"工程技术","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}

{"title":"Learning to retrieve containers: A scale-diverse deep reinforcement learning approach for the container retrieval problem","authors":"Woo-Jin Shin,&nbsp;Inguk Choi,&nbsp;Sang-Hyun Cho,&nbsp;Hyun-Jung Kim","doi":"10.1016/j.trc.2025.105496","DOIUrl":"10.1016/j.trc.2025.105496","url":null,"abstract":"<div><div>This study addresses the container retrieval problem (CRP), a key challenge in the storage yards of automated container terminals where operational efficiency directly affects vessel turnaround time and yard congestion. In storage yards, containers are stacked vertically to maximize space utilization; however, accessing one located below others requires relocating the blocking containers, leading to additional crane movements and delays. The CRP involves retrieving containers from multiple bays in a specified order while minimizing the total working time of the yard crane, with relocation position decisions being critical. The CRP poses several practical challenges: despite being <span><math><mi>NP</mi></math></span>-hard, real-world instances often involve hundreds of containers, requiring high-quality solutions in real time; yard configurations also vary widely and change frequently, demanding methods that adapt effectively to arbitrary layouts. We propose a novel deep reinforcement learning approach incorporating (1) a size-agnostic network architecture, enabling a single trained network to handle diverse yard configurations, and (2) a scale-diverse learning framework, which trains on a various yard scales using a normalized loss to improve generalization and scalability. Experiments on well-known benchmarks with several hundred containers show that the proposed method substantially outperforms existing baselines across a wide range of yard sizes. It also scales to instances with thousands of containers and maintains strong performance in dynamic settings where retrieval orders are revealed online. Solutions are produced within a second for realistic instances, confirming its effectiveness and practical applicability in real-world automated container terminals. The implementation and datasets used in this study are publicly available in the GitHub repository: <span><span>https://github.com/operagang/CRP_RL</span><svg><path></path></svg></span>.</div></div>","PeriodicalId":54417,"journal":{"name":"Transportation Research Part C-Emerging Technologies","volume":"183 ","pages":"Article 105496"},"PeriodicalIF":7.6,"publicationDate":"2026-02-01","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"145785149","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":1,"RegionCategory":"工程技术","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}

{"title":"Statistical inference of boarding and alighting counts in transit systems with incomplete data","authors":"Xiaoxu Chen ,&nbsp;Marc-Olivier Thibault ,&nbsp;Martin Trépanier ,&nbsp;Lijun Sun","doi":"10.1016/j.trc.2025.105484","DOIUrl":"10.1016/j.trc.2025.105484","url":null,"abstract":"<div><div>Automatic passenger counting (APC) systems have been widely used in public transit systems to collect boarding and alighting counts, which are essential for understanding travel demand, optimizing transit operations, and improving transit service quality. However, missing boarding and alighting counts remain a pervasive problem due to APC deployment, hardware malfunctions, or operational disruptions. The reconstruction of these missing data is particularly challenging because boarding and alighting counts must satisfy real-world constraints, such as balance conditions and onboard passenger limits. To address this issue, we propose a probabilistic framework that integrates passenger travel behavior and operational constraints to estimate missing boarding and alighting counts. The framework builds a time-varying Poisson model to estimate boarding demand and employs a method to infer time-varying alighting probabilities. Further, the alighting counts are derived by assigning estimated boarding counts to downstream stops with time-varying alighting probabilities, ensuring that the reconstructed data meet operational constraints. We validate the proposed framework using real-world transit data. The results demonstrate the method’s accuracy and robustness in estimating missing APC data, while also providing valuable insights into time-varying passenger travel behaviors, including arrival rates and alighting probabilities. This framework offers a practical and interpretable solution for reconstructing incomplete boarding and alighting data, with significant implications for improving transit planning and operational decision-making.</div></div>","PeriodicalId":54417,"journal":{"name":"Transportation Research Part C-Emerging Technologies","volume":"183 ","pages":"Article 105484"},"PeriodicalIF":7.6,"publicationDate":"2026-02-01","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"145732480","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":1,"RegionCategory":"工程技术","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}

{"title":"Iterative physics-enhanced residual learning for context-aware traffic assignment under biased inputs","authors":"Yuxin Shi ,&nbsp;Keke Long ,&nbsp;William H.K. Lam ,&nbsp;Xiaopeng Li ,&nbsp;Wei Ma","doi":"10.1016/j.trc.2025.105498","DOIUrl":"10.1016/j.trc.2025.105498","url":null,"abstract":"<div><div>Hybrid approaches combining physics-based models and data-driven methods have shown promise in traffic modeling by leveraging physical structure while enhancing learning flexibility. One representative example is Physics-Enhanced Residual Learning (PERL), which augments physics-based predictions with learned residuals to correct modeling errors. However, the effectiveness of physics-based model can degrade under biased input features. To address this limitation, we propose iterative Physics-Enhanced Residual Learning (iPERL), an end-to-end framework designed to improve the robustness of physics-guided models under biased inputs. We apply iPERL to context-aware traffic assignment, in which explanatory inputs such as OD demand, link and node characteristics (e.g., capacity, free-flow speed), and performance function parameters may be biased due to indirect observations and calibration errors, while traffic conditions simultaneously vary with contextual factors like time and weather. iPERL extends standard PERL by incorporating a residual-based input correction mechanism that iteratively calibrates these biased inputs using feedback from residuals between predicted and observed flows. By integrating contextual features, iPERL further enables adaptive correction strategies under diverse traffic scenarios. We evaluate the framework on both synthetic and real-world networks. Results show that iPERL consistently outperforms baseline methods, including standard PERL, particularly when input bias or data scarcity is present. The proposed framework offers a robust, interpretable, and data-efficient solution for traffic flow estimation, with potential for generalization across networks and practical applications.</div></div>","PeriodicalId":54417,"journal":{"name":"Transportation Research Part C-Emerging Technologies","volume":"183 ","pages":"Article 105498"},"PeriodicalIF":7.6,"publicationDate":"2026-02-01","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"145823196","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":1,"RegionCategory":"工程技术","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}

{"title":"Knowledge-data fusion oriented traffic state estimation: A stochastic physics-informed deep learning approach","authors":"Ting Wang ,&nbsp;Ye Li ,&nbsp;Rongjun Cheng ,&nbsp;Guojian Zou ,&nbsp;Takao Dantsuji ,&nbsp;Dong Ngoduy","doi":"10.1016/j.trc.2025.105422","DOIUrl":"10.1016/j.trc.2025.105422","url":null,"abstract":"<div><div>Physics-informed deep learning (PIDL)-based models have recently garnered remarkable success in Traffic State Estimation (TSE). However, the prior knowledge used to guide regularization training in current mainstream architectures is based on deterministic physical models. The drawback is that a solely deterministic model fails to capture the universally observed traffic flow dynamic scattering effect. Considering the existence of more realistic stochastic physical models that can reproduce the relationship between speed and flow, they can provide better bounds for neural network models with uncertainty. Therefore, this study, for the first time, incorporates stochastic physics information to improve the PIDL architecture and propose stochastic physics-informed deep learning (SPIDL) for traffic state estimation. The idea behind such SPIDL is simple and is based on the fact that a stochastic fundamental diagram provides the entire range of possible speeds for any given density with associated probabilities. Specifically, we select percentile-based fundamental diagram and distribution-based fundamental diagram as stochastic physics knowledge and design corresponding physics-uninformed neural networks for effective fusion, thereby realizing two specific SPIDL models, namely <span><math><mi>α</mi></math></span>-SPIDL and <span><math><mi>B</mi></math></span>-SPIDL. The main contribution of SPIDL lies in addressing the “overly centralized guidance” caused by the one-to-one speed-density relationship in deterministic models during neural network training, enabling the network to digest more reliable knowledge-based constraints. Experiments on real-world datasets indicate that proposed SPIDL models achieve accurate traffic state estimation in sparse data scenarios. More importantly, as expected, SPIDL models reproduce well the scattering effect of field observations, demonstrating the effectiveness of fusing stochastic physics model knowledge with deep learning frameworks.</div></div>","PeriodicalId":54417,"journal":{"name":"Transportation Research Part C-Emerging Technologies","volume":"182 ","pages":"Article 105422"},"PeriodicalIF":7.6,"publicationDate":"2026-01-01","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"145461632","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":1,"RegionCategory":"工程技术","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}

{"title":"The emergence of rideshare buddy-pooling behavior: Dynamic user choices and mobility platform operation decisions","authors":"Zhuoye Zhang,&nbsp;Jie Lin,&nbsp;Zhenwei Gong,&nbsp;Fangni Zhang","doi":"10.1016/j.trc.2025.105427","DOIUrl":"10.1016/j.trc.2025.105427","url":null,"abstract":"<div><div>Ride-pooling services have become globally popular due to their cost-effectiveness, offering discounted fares for passengers with similar origins and destinations. However, these fares are typically higher than if the cost of a single ride was evenly split among passengers. As a result, many travelers are now utilizing social media or group chats to find appropriate rideshare partners themselves, and then request a single ride as a group. This practice, henceforth referred to as ‘buddy-pooling behavior’, is particularly popular among certain communities such as residents and college students, who seek suitable partners within their community for rideshare to city centers, airports, and train stations. This paper presents a comprehensive doubly dynamical framework (considering within-day and day-to-day dynamics) to model, optimize, and evaluate the decision-making processes of travelers, drivers and ride-sourcing platforms with the buddy-pooling behavior. The proposed model considers several typical travel options, including ride-pooling (facilitated by the ride-sourcing platform), buddy-pooling (organized by customers themselves through social media or third-party matching service), multihoming for both ride-pooling and buddy-pooling, non-pooling services, and public transit. Our study begins by examining the time-dependent decision-making of travelers and drivers within a day, subsequently characterizing the evolution of the system on a day-to-day basis. Furthermore, we introduce a bi-level framework to optimize the pricing strategies of ride-sourcing platforms with the aim of enhancing system efficiency and platform profitability. Our results demonstrate that the emerging buddy-pooling behavior will benefit the drivers, travelers and the overall system. Moreover, the results suggest that multihoming behaviors combined with buddy-pooling behavior have the potential to generate more profit for the ride-sourcing platform. The proposed doubly dynamical model offers a framework to model and simulate the operational strategies of ride-sourcing platforms across different time periods, effectively capturing the complex choice behaviors of travelers.</div></div>","PeriodicalId":54417,"journal":{"name":"Transportation Research Part C-Emerging Technologies","volume":"182 ","pages":"Article 105427"},"PeriodicalIF":7.6,"publicationDate":"2026-01-01","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"145473238","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":1,"RegionCategory":"工程技术","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}

{"title":"Hybrid machine learning and physics-based modeling of pedestrian pushing behaviors","authors":"Qiancheng Xu ,&nbsp;Ezel Üsten ,&nbsp;Ahmed Alia ,&nbsp;Biao He ,&nbsp;Renzhong Guo ,&nbsp;Mohcine Chraibi","doi":"10.1016/j.trc.2025.105395","DOIUrl":"10.1016/j.trc.2025.105395","url":null,"abstract":"<div><div>In high-density crowds, close proximity between pedestrians makes the steady state highly vulnerable to disruption by pushing behaviors, potentially leading to serious accidents. However, the scarcity of experimental data on pushing behaviors has hindered systematic investigations into the underlying mechanisms and the development of accurate models. Using behavioral data from bottleneck experiments, we analyze the heterogeneity of pedestrians’ internal pushing tendencies, revealing that pedestrians tend to push under high-motivation conditions and in wider corridors. In addition, we introduce a spatial discretization method to encode the state of pedestrian neighbors into feature vectors, serving together with pedestrian internal pushing tendency as the input of random forest classifiers to predict whether a pedestrian would engage in pushing behaviors. By analyzing speed-headway relationships, we reveal that pushing behaviors correspond to an aggressive space-utilization movement strategy. Consequently, we propose a hybrid machine learning and physics-based model integrating the heterogeneity of internal pushing tendencies, the random forest-based prediction of pushing behaviors, and multiple movement strategies associated with pushing and non-pushing behaviors. The proposed model is calibrated using experimental data, and parameter sensitivity analysis is conducted. Validation results demonstrate that the hybrid model effectively reproduces experimental crowd dynamics, particularly in high-motivation scenarios. Moreover, the hybrid structure of the proposed model is suitable for incorporating additional behaviors, providing a solid foundation for advancing the understanding and simulation of complex pedestrian dynamics.</div></div>","PeriodicalId":54417,"journal":{"name":"Transportation Research Part C-Emerging Technologies","volume":"182 ","pages":"Article 105395"},"PeriodicalIF":7.6,"publicationDate":"2026-01-01","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"145365689","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":1,"RegionCategory":"工程技术","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}

{"title":"Scalable analysis of stop-and-go waves: Representation, measurements and insights","authors":"Junyi Ji ,&nbsp;Derek Gloudemans ,&nbsp;Yanbing Wang ,&nbsp;Gergely Zachár ,&nbsp;William Barbour ,&nbsp;Jonathan Sprinkle ,&nbsp;Benedetto Piccoli ,&nbsp;Daniel B. Work","doi":"10.1016/j.trc.2025.105385","DOIUrl":"10.1016/j.trc.2025.105385","url":null,"abstract":"<div><div>Analyzing stop-and-go waves at the scale of miles and hours of data is an emerging challenge in traffic research. The past 5 years have seen an explosion in the availability of large-scale traffic data containing traffic waves and complex congestion patterns, making existing approaches unsuitable for repeatable and scalable analysis of traffic waves in these data. This paper makes a first step towards addressing this challenge by introducing an automatic and scalable stop-and-go wave identification method capable of capturing wave generation, propagation, dissipation, as well as bifurcation and merging, which have previously been observed only very rarely. Using a concise and simple critical-speed based definition of a stop-and-go wave, the proposed method identifies all wave boundaries that encompass spatio-temporal points where vehicle speed is below a chosen critical speed. The method is built upon a graph representation of the spatio-temporal points associated with stop-and-go waves, specifically wave front (start) points and wave tail (end) points, and approaches the solution as a graph component identification problem. It enables the measurement of wave properties at scale. The method is implemented in Python and demonstrated on a large-scale dataset, I-24 MOTION INCEPTION. Our results show insights on the complexity of traffic waves. Traffic waves can bifurcate and merge at a scale that has never been observed or described before. The clustering analysis of all the identified wave components reveals the different topological structures of traffic waves. We explored that the wave merge or bifurcation points can be explained by spatial features. The gallery of all the identified wave topologies is demonstrated at <span><span>https://trafficwaves.github.io/</span><svg><path></path></svg></span>.</div></div>","PeriodicalId":54417,"journal":{"name":"Transportation Research Part C-Emerging Technologies","volume":"182 ","pages":"Article 105385"},"PeriodicalIF":7.6,"publicationDate":"2026-01-01","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"145554543","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":1,"RegionCategory":"工程技术","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}

{"title":"Integrating graph and neural relational inference for network-wide train delay prediction: An equilibrium between accuracy and interpretability","authors":"Zhongcan Li ,&nbsp;Wei Dong ,&nbsp;Yindong Ji ,&nbsp;Jie Luo ,&nbsp;Ping Huang","doi":"10.1016/j.trc.2025.105418","DOIUrl":"10.1016/j.trc.2025.105418","url":null,"abstract":"<div><div>Accurate train delay predictions contribute to real-time decision-making, and comprehending the intricate interactions among diverse elements of delay evolution holds paramount significance for tactical timetabling. However, existing research struggles to strike an equilibrium between the accuracy and interpretability of delay prediction models. This paper introduces NRI-GraphSAGE, a predictive model for railway network delay evolution, successfully harmonizing interpretability and accuracy by integrating neural relational inference (NRI) and Graph Neural Networks (GNNs). The proposed model follows a standard encoder-decoder structure. The model’s encoder module employs a variational autoencoder structure to learn train-train interactions. In the model’s decoder module, heterogeneous GNNs are used to process the acquired train-train interactions and other information guided by domain knowledge. Case studies on two local networks of the Chinese high-speed railway affirm the rationality of each module within NRI-GraphSAGE and showcase its outstanding predictive accuracy. Through experiments, we affirm the significance of interactions between elements (station-train, disturbance-train, station-station) in the railway network, alongside the sensitivity of influencing features. Furthermore, an analysis of the learned train-train interactions reveals that multiple adjacent trains can interact, and the strength of interactions increases with the decrease of headways or growth of train delays. Compared with existing approaches that rely on predefined relationships, our model automatically infers these interactions from historical data, more accurately capturing critical train interactions. Consequently, the high predictive accuracy of NRI-GraphSAGE furnishes dispatchers with a foundation for crafting rescheduling decisions, while explaining the interactions of different elements during the delay evolution lends support to the allocation of recovery time in timetable planning.</div></div>","PeriodicalId":54417,"journal":{"name":"Transportation Research Part C-Emerging Technologies","volume":"182 ","pages":"Article 105418"},"PeriodicalIF":7.6,"publicationDate":"2026-01-01","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"145473239","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":1,"RegionCategory":"工程技术","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}