Integrated transport hubs require reliable, fine-grained forecasts of crowd distribution to safeguard operations and sustainable urban mobility. We present Group Evolution Mechanism Embedded Network (GEME-Net), a passenger flow distribution forecasting architecture that fuses multimodal data, including video-derived counts, digital twin-based mobility chains, and railway/metro operations information, via multi-graph spatial representations and event-aware temporal modules, with a distilled lightweight student model for deployment. In a real-world case at Shanghai Hongqiao, GEME-Net consistently outperforms statistical, convolutional, recurrent, graph-based and Transformer baselines across MAE, RMSE and WMAPE, while retaining inference latency compatible with near-real-time use. Ablations indicate that schedule encoding and event-driven frequency enhancement, together with learned long-range and community graphs, are principal contributors to accuracy. By coupling operational signals with spatial semantics, our approach improves hub-scale situation awareness and short-horizon decision support, offering a practical route to resilient crowd management without asserting broader societal or policy impacts.

Efficient utilization of rolling stock stands as an important goal for railway enterprises. Achieving it requires assigning appropriate rolling stock to train paths that cover a given set of routes in the operational plan, while accounting for practical constraints such as maintenance schedules and depot capacity. Although numerous studies have focused on minimizing operating costs by reducing the number of rolling stocks to develop operational plan, they often ignore practical requirements or neglect rolling stock assignment, rendering plans infeasible in practice. To tackle this problem, this paper develops an integrated approach for train timetabling, stop planning, rolling stock maintenance and assignment, while considering circulation rules and operational constraints. The model incorporates constraints such as time- and mileage-based maintenance rules, rolling stock assignments, the initial number of available rolling stocks, the number of stored rolling stocks in the depot. A rolling stock selection variable is introduced to represent the state of available rolling stock and the train formation, and a nonlinear mixed-integer programming model is developed to minimize train operating costs and the passenger travel costs. To solve large-scale real-world problem, the adaptive large neighborhood search (ALNS) algorithm is employed. The effectiveness of the formulated method is verified through numerical experiments and a real-world case study. Results demonstrate that the proposed approach not only guarantees operational feasibility but also reduces the required rolling stock by 5.4-8.6 % and operational costs by approximately 4.9-8.4 % compared to sequential optimization strategies, without significantly compromising passenger service quality. Furthermore, the method offers valuable guidance for decision-makers with diverse preferences.

Efficient weekly scheduling, including task assignments, maintenance, washing, and parking, for urban rail transit vehicles is essential for reducing operational costs and ensuring smooth depot operations. However, optimizing these components separately may cause inefficiencies, conflicts, and suboptimal vehicle utilization, ultimately compromising system performance and reliability. This study presents an integrated optimization model to minimize mileage variance, driving distances, and shunting operations. To address the multiobjective nature of the problem while satisfying operational constraints, a novel utility-balancing simulated annealing (U-SA) algorithm is developed. In this algorithm, the utility-balancing mechanism is embedded in the SA framework and guided by fuzzy programming to coordinate multiple objectives effectively. The algorithm utilizes three distinct operators—mileage-balancing operator, distance-minimization operator, and shunting-optimization operator—to steer the optimization toward an efficient solution. The model and algorithm are validated through a real-world case study of Beijing Subway’s Line 12. Computational results demonstrate that the approach effectively balances objectives, reduces operational costs, and enhances resource utilization. Sensitivity analysis reveals that adding parking tracks reduces shunting operations, while earlier washing windows helps shorten driving distances. This method offers a practical decision-support tool to improve the efficiency of urban rail transit operations.

The COVID-19 pandemic posed a great challenge in railway industry, as it changed passengers' behavior towards travelling, which affected their mobility choice in turn. This paper focuses on factors influencing railway passengers' behavior in the new normal, based on 3,318 valid responses collected through an online survey. Variance analysis and exploratory factor analysis were conducted to identify key determinants of passengers' mode choice. Building on these results, a structural equation model (SEM) was developed to describe the interrelationships among passengers' personal attributes, mode characteristics, and travel intentions. The whole modelling process involved selecting latent variables, designing the initial theoretical framework, developing the questionnaire, and estimating the model using maximum likelihood in AMOS, followed by calibration and modification until acceptable parameter significance and model fit were achieved. Considering the trends and uncertainties related to railway industry, 4 development scenarios were constructed based on political, economic, and social factors. The sample data from 4 scenarios were input into the modified SEM model separately. In the end, we could obtain 4 similar SEMs adapting to different scenarios by adjustment. This will provide a scientific basis for formulating railway development strategies in the future for the new normal of post-COVID-19 era.

Inferring travel mode distribution is key to understanding individual travel behaviour and reconstructing travel chains, which is crucial for personalized travel services. An end-to-end model architecture is proposed to enhance discriminative feature representation through context-aware geographic embeddings that integrate bus stop and intersection information. An innovative feature encoding framework augmented with spectral and attention mechanisms is developed: periodic movement patterns are extracted through spectral analysis, while spatiotemporal dependencies are jointly modeled via combined self-attention and positional attention. Furthermore, a hybrid Transformer-CRF decoding layer is constructed to ensure precise point-wise classification, with plausible mode transitions being modeled through state transition probabilities, significantly enhancing trajectory coherence. Experiments on real GPS data show our model achieves an 88.9% accuracy, outperforming existing methods. Ablation and comparison tests validate the effectiveness of our approach, confirming its superiority in travel mode identification.

Balancing passenger service efficiency and train energy consumption is a key focus of rail transit operation optimisation in recent years. These aspects are influenced by train stop plans, timetables and speed profiles. While some progress has been made in integrating three aspects in previous studies, achieving comprehensive optimisation of the train stop plan, timetable and speed profile remains a challenge. This study proposes a collaborative approach to integrate train stop plans, timetables and speed profiles within a unified framework while considering passenger demand variations over time. This approach treats train quantities as a variable and explores the correlation among train energy consumption, passenger waiting time and in-vehicle time by considering variable timetables, stop plans and speed profiles. A large-scale comprehensive optimisation model is introduced and effectively solved using a simulation-embedded adaptive large-neighbourhood search (S-ALNS) algorithm. A numerical case is indicated to validate its effectiveness, and a real-world case is investigated through the S-ALNS algorithm. The findings demonstrate significant reductions in passenger waiting time up to 2.16%, in-vehicle time up to 25.17%, and train energy consumption up to 16.31%, while maintaining an acceptable convergence speed.

This study proposes a deep reinforcement learning-based optimization framework for integrated train scheduling and rolling stock circulation planning under dynamic passenger demand. The problem is formulated as a Markov decision process (MDP) with a hybrid action space that simultaneously captures continuous timetable decisions and discrete rolling stock allocations. The objective is to minimize passenger waiting time and operator costs while adhering to complex operational constraints. To address the challenge of simultaneously coordinating continuous and discrete decision variables in a high-dimensional operational context, we adopt a Hybrid Proximal Policy Optimization (HPPO) algorithm, incorporating separate actor networks for discrete and continuous actions, and employing constraint-handling techniques such as action masking and action space embedding. Furthermore, a potential-based reward shaping function is introduced to enhance learning efficiency by addressing issues of sparse and delayed rewards. The proposed approach is validated on the Beijing Metro Changping Line. Experimental results demonstrate that the HPPO algorithm effectively improves system efficiency and policy robustness.

Mechanical coupling operations require specific equipment and locations, leading to longer operation times and inadequate alignments between transportation supply and passenger demand. Virtual coupling can replace mechanical connections with train-to-train wireless communication, reducing operation times and expanding the operation locations. Thus, train timetables and rolling stock circulation are closely linked to the effective implementation of virtual coupling. This study investigates the integrated optimization of train timetables and rolling stock circulation based on virtual coupling to increase service efficiency and operational efficiency. The key challenge is determining train operation times, rolling stock connections, and virtual coupling operations under relaxed operation times and locations. The problem is formulated as a multi-objective mixed-integer nonlinear programming (MINLP) model to reduce passenger costs, including the total passengers’ waiting time and the total passengers’ in-vehicle time, as well as operating cost, including the number of rolling stocks and the number of trains. To address the multi-objective nature of the optimization model, fuzzy programming with a utility function is integrated into an adaptive large neighbourhood search (ALNS) algorithm, resulting in a utility-based ALNS (U-ALNS) algorithm. Six series of operators are designed according to specific operation mechanisms and are subject to certain constraints to accelerate the optimization process. The model and algorithm are tested with a numerical case study and a real-world case study of the Jinshan Line in Shanghai. The results demonstrate that the U-ALNS efficiently generates superior solutions within an acceptable time. Compared with the two-stage approach, the integrated approach outperforms in terms of operating costs. Furthermore, compared with mechanical coupling, virtual coupling enhances the alignment between the transportation supply and passenger demand and effectively reduces both passenger and operating costs.

The dynamic evolution and heterogeneity of passenger wayfinding decisions in integrated transport hubs have a significant impact on both operational efficiency and user experience. However, existing static models fall short in capturing the temporal variability of passengers' cognitive states in response to environmental and situational changes. This study develops a virtual reality scenario of an integrated transport hub and conducts non-immersive behavioral experiments to support the construction of a novel dynamic modeling framework-Dynamic Hidden Markov Model-Logit (DHMM-Logit), which integrates a multi-state hidden Markov model with a Logit model. For the first time, decision cascading analysis is introduced into this framework, utilizing mutual information theory to uncover the temporal dependency and decay mechanism of historical decisions on current choices. These insights guide both the hyperparameter setting and discretization of decision sequences in the DHMM-Logit model. The framework comprehensively incorporates spatial syntax metrics, the use of 2D navigation tools and travel purposes to account for spatial and individual heterogeneity. In addition, a graph embedding-based high-order semantic encoding of nodes is introduced as explanatory variables, enhancing the model's ability to fit and generalize sequential pedestrian decision-making processes. Empirical validation in the Shanghai Hongqiao Integrated Transport Hub demonstrates that the proposed DHMM-Logit model significantly outperforms baseline methods. The findings reveal pronounced latent cognitive state transitions during pedestrian wayfinding, with travel mode and navigation usage exerting significant influence on passengers' spatial sensitivity and cognitive processes. This research provides a solid theoretical and empirical foundation for the optimization of hub spatial design and the implementation of personalized information guidance strategies.

This study investigates a rolling stock circulation planning problem under enhanced operational conditions, including the consideration of multi-day maintenance periods, the feasibility of coupling and decoupling operations, and the arrangement of deadhead movements. A tailored branch-and-price algorithm and a decomposition structure with observable upper and lower bounds are proposed to solve this problem. First, the problem is formulated into a path-based generation model for single-day routes and an assignment model to obtain multi-day circulation plans satisfying maintenance constraints. Then, regarding the feasibility of coupling and decoupling operations, circulation type variables with set covering constraints are introduced, and tailored type-and-state-based branching rules are designed to explore the solution space within the branch-and-price framework effectively. Next, to improve computational efficiency, additional relaxation techniques based on the deadhead movements in the network are designed. Lastly, extensive experiments are conducted on large-scale instances with up to 785 trips and 66 stations, which demonstrate the effectiveness of the proposed decomposition structure, additional relaxation techniques, and branching rules in both solution quality and performance. Comparisons under different maintenance rules and coupling/decoupling modes further validate the practical significance of the problem under enhanced operational conditions considered in this study.

Urban rail transit is operated with network scenario mostly. Train routing determines spatial service range on the network, which will affect whether passengers transfer during the trip. Timetable determines arrival and departure time of each train at each station. To achieve balance between demand and supply and enhance service quality, simultaneous optimization of train routing and timetable is of significant importance for rail companies with the network. This will reduce the waiting time of passengers effectively, and the stations where train serve will be convenient for most passengers. However, existing studies in train operation planning with time-varying passenger demand haven’t addressed this aspect adequately. To bridge this gap, we develop an integrated optimization model for the Train Routing and Timetable Problem (TRTP) within a rail transit network, considering the dynamic of passenger demand and passenger path choice. Our objective is to minimize the operating costs for companies and to reduce the total waiting time experienced by passengers. The constrains of TRTP include time constraint, path constraint and train constraint. What’s more, the compatibility of headways on different routings is the focus when modeling. To verify the effectiveness of our proposed model, we conduct a numerical experiment with CPLEX. After that, the results of integrated optimization and staged optimization are compared to show the beneficial of integration with train routing and timetable. Moreover, we devise an algorithm tailored to this problem to solve the model, and a network case study is presented. As a case study, we apply the model to Guangzhou Metro Line 3 to validate its performance further, which is a Y-type line, the most common example of operation with network of urban rail transit in China. The results of our study demonstrate that the proposed integrated optimization model can achieve a reduction in operating costs and total waiting time for passengers effectively, providing convenience for passenger travel.

Balancing fairness and efficiency is a challenge for urban rail transit operators. While skip-stop operations can enhance efficiency, their impact on fairness remains unclear. This paper explores the trade-off between efficiency and fairness using two allocation methods: min-max fairness and alpha-fairness. Passenger fairness is measured by differences in travel efficiency across stations. We develop a MILP model with min-max fairness, a MINLP model with alpha-fairness, and a MILP model focused solely on efficiency for comparison. An extended Adaptive Large Neighborhood Search (ALNS) algorithm is proposed and tested on the Jinshan railway in Shanghai. Results show that our approach optimizes timetables by maintaining efficiency while improving fairness under skip-stop mode. Moreover, min-max fairness proves to be a more practical method for real-world operations, balancing efficiency for the majority of passengers.

Demand-responsive transportation has been introduced in many cities around the world. However, whether it is applicable in the railway is still questionable, an exploration of passenger choice behavior between demandresponsive trains and pre-scheduled trains is pivotal in addressing this issue. To delve into passengers' choice preferences when facing demand-responsive trains and to dissect the feasibility of implementing demandresponsive service in high-speed railways, the stated preference survey method is employed to investigate travel intention of passengers. Based on the survey data obtained in China, the heterogeneity of passengers is analyzed from three aspects: personal socio-economic characteristics, travel characteristics, and travel mode choice.Considering the situation that demand-responsive train cannot operate, the risk attributes are considered. To bolster the appeal of demand-responsive trains, personalized service product attributes are added. Mixed Logit mode, which takes into account the heterogeneous travel choice behavior of passengers, is developed, and Maximum Likelihood Estimation and the Monte Carlo method are used to calibrate model parameters. The willingness to pay in terms of different factors of passengers is determined. The results indicate that early arrival deviation time, late arrival deviation time, demand response time, and success rate of ticket purchase remarkable influence passengers' decision regarding demand-responsive train, with only the success rate of ticket purchase positively impacting train choice. Moreover, the significant difference in train ticket price is observed solely in the self-funded long distance scenario, while demand-responsive trains are found to be particularly appealing in self-funded short distance scenario. Through the Willingness To Pay(WTP) analysis, it is discovered that by shortening demand response time, enhancing the success rate of ticket purchase, and minimizing the deviation times of early arrival and late arrival of trains, the attractiveness of the demand-responsive train to passengers under three travel scenarios can be augmented. This study provides profound insights into the possibility of railway enterprises operating demand-responsive trains.

Railways and aviation are major modes of international origin–destination (OD) travel, and the intermodal model is influenced by external factors such as the layout of airports and train stations, as well as passenger demands and preferences. At the same time, internal factors like infrastructure, service protocols, and coordination mechanisms also impose constraints. This study focuses on the passenger transfer process, luggage handling procedures, and ticketing mechanisms in air-rail intermodality, systematically identifying the factors that affect air-rail intermodal services from the perspective of passenger preferences throughout the “end-to-end” travel process. Using international air-rail intermodality in cities such as Guangzhou, Hangzhou, and Tianjin as case studies, the research analyzes the key service metrics that passengers prioritize in air-rail intermodal travel and establishes a correlation mechanism between service components and these metrics. A SWOT analysis is employed to summarize the strengths and weaknesses of various air-rail intermodal models, identify the bottlenecks and challenges in air-rail intermodal cooperation, and propose development strategies for air-rail services through the application of new technologies, product optimization, and infrastructure improvements. The study finds that business travelers prioritize convenience, while leisure travelers focus more on affordability. Despite regional differences in international air-rail intermodal services, common challenges persist, including insufficient cooperation among operators, a lack of service refinement, and increasing infrastructure demands, often accompanied by high construction costs. In air-rail services with spatial separation between airports and stations, particular emphasis should be placed on transfer connectivity, especially in developing luggage and check-in facilities before transfers, which plays a crucial role in managing seasonal surges in passenger volume. For air-rail services where airports and stations are in close proximity, optimizing the coordination of air and rail timetables holds significant potential, requiring personalized ticketing solutions tailored to different passenger groups. Integrated air-rail intermodal services should prioritize enhancing “soft power,” not only by offering flexible, personalized services but also by improving the overall service level through specialized third-party intermodal platforms and collaborative mechanisms.

The Yangtze block is a crucial component of the Rodinia and Columbia supercontinents, providing insights into their evolutionary history. The newly identified Caiziyuan-Tongan accretionary complex (CAC), situated on the southwestern margin of the Yangtze block, serves as an ideal window for understanding its Precambrian evolution. The exposure of the CAC suggests the presence of an ancient ocean basin that divides the Yangtze block into northern and southern micro-blocks. During the convergence of the Rodinia supercontinent, this ocean basin underwent subduction and eventual closure; however, its associated subduction polarity remains ambiguous. To address this issue, this study employed magnetotelluric in the Caiziyuan-Tongan area. Through threedimensional inversion, three distinct features were delineated: an upper crustal conductor in the southern CAC, a middle-upper crustal resistor trending stepwise north in the northern CAC, and a middle-lower crustal conductor encompassing the entire area. In conjunction with previously published geological and seismological observations, it is proposed that the first two features may be remnants associated with the northward subduction of the south Yangtze micro-block, while the third feature may indicate the existence of crustal flow related to the collision of the Indo-Asian plate. This finding enhances our understanding of the Precambrian evolution within the Yangtze block and contributes to the reconstruction of paleogeographic frameworks associated with the Rodinia supercontinent.

This paper presents an improved gravitational search algorithm (GSA) for the one-dimensional magnetotelluric (MT) inversion. The inversion of MT data is inherently ill-posed and nonlinear. Traditional deterministic inversion methods, which rely on approximate linearized solutions, are highly sensitive to the initial model and are prone to converging to local optimal solutions. The GSA is a novel metaheuristic algorithm inspired by the law of gravity and mass interactions to deal with nonlinear optimization problems. It is improved through hybrid methods involving population initialization and calculation of gravitational force, and it is employed for one-dimensional MT inversion to search for the global optimum solution. Using synthetic data derived from two geoelectric models, the effectiveness of main parameters in GSA for MT inversion was investigated. The inversion results from the improved GSA were comparable to the true models and demonstrated robustness, as evidenced by repeatability and noise tests. Additionally, it also shows good agreement with previous results obtained using other methods when dealing with the field MT dataset COPROD2. Therefore, both synthetic and field data tests demonstrate that the improved GSA has good performance for MT inversion. In particular, the improved GSA is suitable for inverting MT data and does not require an initial model.

This paper proposed a travel plan recommendation system that can provide multi-modal, personalized, and door-to-door travel plans to solve travelers' difficulty in choosing when facing vast and complex travel information. First, we established a dynamic Travel Choice Behavior Graph (TCBG) model, which considers the travel plan candidate set and the temporal characteristics (time-decay and periodicity) of travelers' behavioral preferences. Next, to effectively learn from TCBG, we constructed a Unified Candidate Set Representation Module (UCSRM) and a new graph neural network called Continuous Dynamic Heterogeneous Graph Attention Networks (CDHAN). UCSRM can employ a multi-head self-attention mechanism for a unified representation of travel plan candidate sets with inconsistent lengths. CDHAN can capture the temporal characteristics of travelers' preferences by combining the improved Hawkes process. Finally, we validated the effectiveness of the model and framework on multi-modal travel datasets and achieved 0.8172, 0.7994, 0.7859, and 0.9345 on the evaluation metrics of Pre, Rec, F1, and NDCG, respectively. These results show that our model/framework outperforms six existing stateof-the-art models/frameworks in these four evaluation metrics. This study provided a new model and learning framework for travel plan recommendation systems, essential for improving the efficiency of urban transportation and travelers' travel experience.

Accurate understanding of passenger flow distribution is crucial for effective station crowd management. However, due to the complexity and randomness of passenger flow and the unclear spatial-temporal correlation between functional areas within the station, predicting the spatiotemporal distribution dynamics of inflow and future short-term distribution trends is challenging. Emerging deep learning models offer valuable insights for accurately predicting passenger flow distribution. Thus, we propose a deep learning architecture, named “ST-Bi-LSTM,” which combines a bidirectional long short-term memory network with a spatial-temporal attention mechanism. Initially, we outline the methodologies of Bi-LSTM, the DeepWalk-based spatial attention mechanism, and the temporal attention mechanism. The spatial attention mechanism is employed to extract station spatial network topology information and enhance the representation of passenger flow characteristics in highly correlated areas during the forecasting process. Simultaneously, the temporal attention Bi-LSTM is utilized for capturing temporal correlations. The architecture comprises four branches dedicated to station real-time video monitoring data, spatial network topology, function area attributes, and train timetables. Subsequently, leveraging in-station CCTV data, passenger travel behavior data, and train timetables, we apply the architecture to the Tianjin West High-Speed Railway Station. We conduct a comparative analysis of the prediction performance and time complexity of the proposed architecture against existing baseline models, demonstrating superior performance and robustness exhibited by the ST-Bi-LSTM model (achieving a reduction in RMSE of over 10%). This study facilitates the transition of station management from passive response to active prediction of station passenger flow dynamics.

ObjectiveThe resilience of transportation networks is a prominent research area in transportation safety. However, current studies on transportation network resilience often inadequately measure the changes in spatiotemporal travel costs for passengers, primarily focusing on the recovery phase rather than the resistance phase in two-stage resilience. There is also insufficient identification and analysis of critical segments, and a lack of suitable resilience simulation and evaluation methods for urban agglomeration railway passenger transport networks. This paper proposes a resistance resilience assessment model and a resistance resilience simulation evaluation process for urban agglomeration railway passenger transport networks centered on spatiotemporal accessibility for passengers. The aim is to evaluate the resistance resilience of these networks and identify critical segments.MethodsThis paper explores the concepts of resistance resilience and recovery resilience within transportation networks. Utilizing the complex network Space L modeling method, this paper develops a spatiotemporal weighted urban agglomeration railway passenger transport network model that considers actual railway passenger stations as network nodes. Segment interruption scenarios were simulated using attack modes involving single segment deletion and multiple segment continuous deletion. A dynamic resistance resilience evaluation index termed the network performance retention rate, was introduced based on the performance response function and spatiotemporal accessibility of passengers. This paper devises a resistance resilience assessment model and simulation evaluation process to evaluate the substitutability of segments and the overall network resistance resilience. The Chengdu—Chongqing urban agglomeration was selected as a case study to identify and compare critical segments and resistance resilience across unweighted, spatially weighted, and temporally weighted railway networks.ResultsThe results of this paper were as follows: (1) The interruption of critical segments near railway hub cities could lead to a maximum network performance loss of 12.23%. It was necessary to identify critical segments through predisaster simulations. (2) Significant differences were found in the critical segments identified through resistance resilience simulations across unweighted, spatially weighted, and temporally weighted railway networks. The Spearman correlation coefficient indicated a relatively poor correlation between the critical segment rankings of unweighted and weighted railway networks. (3) The resistance resilience indices of the three railway networks highlighted that single segment interruptions significantly affected travel time. (4) Continuous interruption of identified critical segments severely affected network performance, with temporally weighted railway networks experiencing a stronger impact than spatially weighted and unweighted railway networks. Predisaster simulations solely based on topological structure or spatial distance might underestimate the consequences of risk interference.ConclusionsThe methods proposed in this paper address the gap in targeted research on the resistance resilience of railway passenger transport networks in urban agglomerations. Simulations of single segment interruption and multiple segment continuous interruption enable the identification and verification of key network segments. Additionally, analyzing the network resistance to interruptions provides a scientific foundation for transportation network planning and decision-making. Furthermore, analyzing the network's resilience evaluation index of the network performance retention rate proposed in this paper offsets the impact of disturbance time uncertainty, providing a scientific foundation for transportation network resilience research.