Virtual coupling is considered a key technology for increasing line capacity and enhancing recovery capabilities in emergencies. Real-time generation and optimization of train speed trajectory are fundamental to ensuring safe and efficient train operation under virtual coupling. The autonomous train enables autonomous request line resources and makes decisions, allowing for more flexible coupling and reliable and efficient operation. This paper constructs a speed trajectory optimization model for the autonomous train under virtual coupling, which meticulously considers the effects of line resources, such as switches and routes, on virtual coupling. A Twin Delayed Deep Deterministic Policy Gradient (TD3) is utilized to train the agent to optimize the train speed trajectory in real-time. By integrating the Long Short-Term Memory (LSTM), the agent has a longer history memory and learns a better policy. Moreover, two protection mechanisms involving safe following and switch protection are designed to ensure absolute operation safety of the autonomous train and improve training efficiency. Three numerical experiments based on real data from the Beijing-Shanghai High-Speed Railway are conducted. The proposed method can generate a higher quality train speed trajectory within seconds, achieving an average reduction of over 10% in the objective function compared to the commonly used driving strategy and commercial solver. The protection mechanisms always ensure the safety of the trains, even in unknown operating scenarios. Furthermore, the effect of memory and its length have been analyzed by comparing the proposed method with other deep reinforcement learning methods.

Traditional cognitive travel modeling typically employs a unified cognitive model to simulate representative travel behaviors, which may usually result in a weak characterization of user heterogeneity in paths, modes, and other factors. Large language model (LLM), by contrast, has significantly enhanced the anthropomorphic and personalized features of intelligent systems. To integrate their advantages, this article proposes LLM-driven cognitive modeling to generate more diverse and personalized travel demands. The new method sufficiently exploits LLM such as the llama as a basis and provides personalized travel plans so that more heterogenous travel demands could be generated. Additionally, introducing LLM into cognitive modeling can significantly reduce the time of model development, thus accelerating the research or engineering deployment. By calibrating and testing with one month’s data from public transportation (buses and subways) in Beijing, our method, compared to traditional cognitive models, not only achieves better accuracy in reproducing typical travel patterns, but also generates more diverse ones, providing a more comprehensive input for computational experiments on traffic management and control strategies.

With the continuous growth in passenger of high-speed railway, the existing line passing capacity (LPC) is unable to meet the increasing transportation demands. The train tracking interval (TTI) serves as an important parameter for evaluating LPC. The traditional train control system (TCS) considers the transmission latency as a fixed constant measured in the worst environment, which restricts the effectiveness of TTI optimization. In this article, an optimization strategy of TTI for train autonomous operation is proposed based on stochastic network calculus (SNC), aiming at enhancing LPC. Different with the existing TCS, SNC can calculate the transmission latency as a dynamic value, thereby more accurately reflecting the complexity and variability of the train operating environment and speed. This strategy not only guarantees the safety of train operations but also achieves smaller and more appropriate transmission latency. In addition, this article employs a moving block system for train autonomous operation and further decreases the TTI which is realized by train-to-train communication. Simulation studies were conducted to explore the relationship between transmission latency and the dynamic operational speed and environmental changes. The results demonstrate that the proposed method can enhance LPC by 2% to 9%. Meantime, the developed control algorithm can prove the effectiveness and availability of the TTI adjustment method.

The safe, punctual, and reliable operation of high-speed railway (HSR) networks is crucial for ensuring system efficiency and enhancing passenger experience. However, due to the complexity of HSR systems and long operation routes, failures in system components can lead to unexpected incidents, causing deviations from the scheduled operations and, subsequently, delays. These delays, particularly large delays, significantly impact the overall performance and passenger experience of the network. Thus, accurate prediction of train delays, especially in large-delay scenarios, is essential for optimizing train scheduling and restoring normal operations in a timely manner. To address this challenge, this study proposes a deep learning architecture based on spatio-temporal feature extraction for accurate train delay prediction. A graph attention network-long short-term memory block is introduced to capture the spatio-temporal evolution features of different trains. In addition, a sequence forecasting approach and a mixture of experts module are integrated to model the complex relationships between the target train’s delay and its previous states. The delay evolution features are incorporated into the loss function, allowing for more accurate predictions. Experimental results show that the proposed model outperforms the baseline models, achieving at least an improvement of 50.62% and 27.89% in root-mean-squared error and mean absolute error, respectively. When the error tolerance is set within 3 min, the prediction accuracy reaches 96.68%. Experimental results demonstrate the superior performance of the proposed model.

Escaping the station during the golden time is the main reference for managers to develop optimization strategies under floods. However, regular evacuation drills often fail to incorporate the impact of flood evolution and involve ethical considerations. The evacuation time obtained from microscopic simulations makes it difficult to meet real-time requirements, which is time-consuming and labor-intensive. To address this issue, this paper proposes a flood-induced evacuation time prediction model for metro stations, with high accuracy and strong real-time performance. The hydrodynamic model and passenger dynamics model for the metro station are built based on the Fluent and the PathFinder software, to obtain the evacuation time corresponding to different scene parameters and evacuation scales. The flood invasion speed and gate availability parameters in the scene parameters, as well as the number of passengers and the proportion of passengers with different attributes, are used as inputs to the prediction model, and the evacuation time is used as the output. The accuracy of the constructed time prediction model is mainly concentrated between 0.85 and 1.00, with a high fit between the predicted values and the actual values. An evacuation path optimization strategy is further provided to reduce evacuation time and improve evacuation safety. Taking evacuation time, evacuation path length, and human stability risk as optimization objectives, the optimal path decision under a multi-objective game is obtained. Results show that adopting an optimization strategy significantly reduces evacuation time and increases the maximum passenger capacity threshold allowed by the platform during the golden evacuation time. The proposed method provides feasible suggestions for the rapid evaluation of station evacuation efficiency and improvement of escape strategies under flood conditions.

High-speed trains are susceptible to unexpected events such as strong winds and equipment failures, which can result in deviations from the scheduled timetable. As the density of traffic increases, these delays can quickly spread to other trains, eventually leading to conflicts in the timetable. To ensure the efficiency of high-speed railways, quickly resolving potential conflicts and generating appropriate rescheduling schemes are essential. The existing hierarchical structure of train control and online rescheduling tends to be inefficient in terms of information communication and can even lead to unfeasible rescheduled timetables and trajectories. To address these issues, an integrated structure of timetable rescheduling and train trajectory optimization is proposed by introducing the train minimum running time into the process of timetable rescheduling and using the adjusted running time as the objective of trajectory optimization. The integration model is formulated by considering the constraints of timetable rescheduling such as the maximum number of trains overtaking trains, platforms at stations, and the priority of the train, as well as the constraints of trajectory optimization. A deep reinforcement learning (DRL)-based approach is proposed to solve the problem. Numerical experiments are conducted on a segment of the Beijing-Shanghai high-speed railway line, using adapted data to demonstrate the effectiveness of the proposed method in rescheduling timetables and optimizing train trajectories. The results show that the integrated rescheduled timetable and the optimized train trajectory can be generated simultaneously and the computation time exhibits a linear increase with respect to the size of the problem.

China’s high-speed railway (HSR) has entered the era of networked operation. Any internal disturbance or eternal disturbance may result in delays of some trains and even cascading delays, which will not only reduce the traffic efficiency of HSR, but also break passengers’ travel and lower their satisfaction. Studying the delay propagation mechanism could assist the dispatcher in suppressing the negative effect of disturbances. However, current studies seldom consider the withholding strategy’s impact on delay propagation. Inspired by this, this article proposes a novel bi-directional delay propagation model combined with the trains’ operation trajectory and stations’ withholding strategy. Moreover, the operation constraint, station capacity constraint, and interlocking constraint are also considered. Then, the primary delay under section disruption (SD) and section temporary speed limit (STSL) are derived based on the location of the disturbance, duration time of the disturbance, and the operation strategy. Then, a max-plus algebra-based delay propagation model is established to compute the corresponding secondary delays. Also, the All Pair Critical Path algorithm is modified to incorporate the station capacity constraint in the searching process. Simulations based on the real China HSR subnetwork are implemented to verify the proposed model. Compared with the current study, the proposed model could accurately unfold the delay propagation in the opposite train heading direction. Besides, the relationship among disturbance duration, primary delay, and accumulative delay for the SD scenario and the relationship among temporarily limited velocity, primary delay, and accumulative delay for the STSL scenario are revealed.

As the demand for coal and other ore resources increases, the hauling capacity of heavy-haul railways is severely challenged. Virtual coupling technology has gained attention for its ability to improve operational efficiency in bottleneck sections and reduce the time it takes for trains operating on the line to resume normal operation during emergencies. In this article, virtual coupling-based timetable rescheduling method is proposed to reduce the delays under disruptions and improve the line capacity. A mixed-integer linear program (MILP) model that allows trains to be coupled either at departure or by sharing the same arrival and departure line is formulated to reduce the delay time and its propagation range. The strategies of retiming, rearranging tracks, and virtual coupling are adopted to collaboratively optimize the deviation in train schedules and track utilization under disruptions, aiming to enhance the occupancy capacity of arrival and departure lines while simultaneously reducing train delays. A heuristic algorithm utilizing simulated annealing (SA)-particle swarm optimization (PSO) algorithm is developed to generate optimal train coupling and stopping schemes. Numerical experiments are conducted to verify the effectiveness of the proposed model and heuristic algorithm on a real heavy-haul railway configuration. The results demonstrate that our method effectively reduces train delays and minimizes the impact of track utilization on adjacent stations, as well as the repercussions of train delays on subsequent stations.

On a passenger-freight collinear railway, the freight train operation level is comparatively low, frequently resulting in significant deviations from the original timetable and crew plan in the presence of various interferences. This article focuses on the problem of coordinated rescheduling of train timetable and crew scheme in the presence of disruptions on a double-track passenger-freight collinear railway. We develop a mixed-integer linear program (MILP) model considering the distinct priorities of passenger and freight trains, as well as crew operations, thereby surpassing the current practice of independently adjusting train timetable and crew plan to achieve a collaborative solution. The objective is to minimize delays for passenger trains and deviations in crew schedule, while maximizing the delivery rate of freight trains at railway Bureau boundary stations prior to the settlement time. Furthermore, for large-scale delays, we design a solution algorithm based on the rolling horizon approach to enhance computational efficiency. To validate the effectiveness of the proposed model, simulation experiments are conducted using actual running data from the Beijing–Shanghai railway. The experimental results illustrate that our coordinated model enhances the feasibility of adjustment outcomes during emergencies, in contrast to the model that neglects crew connections. Additionally, our proposed algorithm guarantees a solving error of under 5% and reduces solving time by over 60% compared with the results obtained by CPLEX. Moreover, three additional comparison experiments are conducted to further demonstrate the impact of crew activities on train operation adjustments, which also indicate that our approach can provide dispatchers with more feasible train operation adjustment schemes in terms of crew utilization.

The independent hierarchy of train dispatching command and train operation control in the existing urban rail transit systems restricts the improvement of operational efficiency and emergency handling capability. This article focuses on integrating train regulation and speed profile optimization by utilizing a feature learning and hybrid search algorithm. Specifically, a genetic algorithm (GA) is used to optimize the train speed profile for a fixed interval running time, and then, the generated labeled sample data are used to train a convolutional neural network (CNN) to learn and extract the features of the optimal speed profile. The nonlinear mapping relationship between input and output variables in trajectory optimization is characterized by a well-trained CNN to reduce the computation time of the optimal speed profile during train regulation. The input variables comprise line conditions and interval running times, while the output variables include the corresponding energy consumption and operating condition switching points of the optimal speed profile. An integrated model of train regulation and operation control is developed with the objective of minimizing total train delay time and energy consumption. To ensure convergence and global search capability, we design a hybrid search algorithm-based train regulation algorithm. Simulation experiments are conducted using data from the Beijing Yizhuang line to validate the effectiveness of the proposed model and algorithms. The experimental results demonstrate that the proposed method can provide an optimal scheme for train regulation and speed profiles.

The efficient and safe evacuation of passengers is an important foundation for ensuring the service level of subway stations and an important part of promoting the development of smart urban rail. Due to the complex structure of stations, strong heterogeneity of passenger flow, and high uncertainty of road interruption in emergencies, passenger evacuation dynamics have received widespread attention and become an inevitable research trend. In order to deeply understand the current research focus and development trend related to emergency evacuation of passengers in subway stations, 196 relevant published literature from 2000 to 2023 are analyzed, and a systematic review is provided. The publication trend of literature, the distribution of countries and institutions, the co-occurrence of keywords, and the types of evacuation are comprehensively reviewed from the perspective of bibliometrics. It can be observed that the number of literature related to passenger evacuation in subway stations has shown explosive growth, especially in the past three years. Special attentions are paid to the current main research hot spots, where passenger evacuation behavior is studied from multiple dimensions, including passenger evacuation dynamics modeling, evacuation simulation, evacuation behavior analysis, and evacuation optimization. Challenges restricting the improvement of passenger emergency evacuation capability are identified, and corresponding possible research directions are proposed and discussed.

Setting up guidance equipment and leaders is widely used as an effective measure to improve the operation and evacuation efficiency of subway stations for the safety of passengers. Cooperation among different kinds of guidance modals can benefit the passenger evacuation process and reduce the operating and management costs as well as the risk of injury to people in subway stations. This article proposes a framework of multimodal cooperative guidance (MMCG) systems for commanding the crowd evacuation in case of emergency, where three types of guidance modes are considered. The bi-level MMCG optimization models are constructed to determine the optimal quantities and initial locations of multimodal guidance. The MMCG schemes are designed by minimizing cost functions taking into account the constraints of the number of guidance, valid coverage, and guiding expectation. An extended social force (SF) model is proposed to study crowd evacuation dynamics with multimodal guidance. The computational experiments are conducted to evaluate the performance of the proposed cooperative guidance schemes at the subway platform scenario. Three unimodal guidance schemes and a contrasted scheme without guidance are also proposed as comparison schemes. The results show that the crowd evacuation efficiency and the utilization ratio of exits are improved by taking into account the cooperation among different guidance modals.

China’s High-Speed Railway (HSR) has entered the era of networked operation. Unpredictable emergencies may disrupt the normal operation of the railway, resulting in train delays. The delays may even cause cascading delays due to the track constraint. This paper proposed a delay propagation model by max-plus algebra, considering the constraint of the HSR network, to predict the spatial-temporal range of cascading delays under emergencies. Moreover, the operation strategy re-management in sections and the re-timing strategy due to the station capacity constraint are studied under emergencies, forming a chain model of emergencies, primary delays, and secondary delays. Simulation results show that the proposed model could accurately predict the secondary delay under different emergencies.

The high-speed train control system is essential to the safety and efficiency of train operation. With the rapid increase of high-speed railway (HSR) operating mileage and development of information technology, the disposal flow and methods in emergency response are still based on dispatchers and drivers’ experience within the “layered” architecture of current system. There is a certain gap between current processing methods and effective resolution, which may even cause the spread of delay along with the railway networks. Therefore, we propose an integration system of operation control and online rescheduling to improve the recovery ability of HSR carrying capacity. We first describe the framework, information flow, and disposal process of the current system and analyze the shortcomings in handling emergencies. Then, the basic concept, system structure, and framework of the integration system are introduced. Finally, taking temporary speed restriction caused by strong wind as an example, we also analyze the principle of why and how the integration system can promote the recovery ability of HSR carrying capacity.

Under the condition of network operation of high-speed railways (HSRs), the influence of disruptions on the train control and dispatching at the current line and related lines is more and more significant. This article focuses on the timetable cooperative rescheduling problem with multidispatching sections of HSRs from a macroscopic perspective in the case of large disruptions. The problem is formulated as a mixed-integer linear program (MILP) model on the objectives of minimizing the weighted sum of the arrival delay time of trains, the delay time of depart trains at the handover station, and the number of delays of trains at all stations. The strategies of retiming and reordering are adopted to generate the rescheduling scheme and reduce delay propagation by making full use of three kinds of buffer time reserved in the timetable, i.e., buffer times of train operation in the station, train running in the section, and electric multiple unit (EMU) connection. A case study of the timetable rescheduling at the two adjacent dispatching sections of the Beijing–Shanghai HSRs line is conducted to evaluate the performance of the proposed integrated rescheduling approach. The relationship between computing time and the quality of rescheduling schemes is also investigated. The computational results show that the proposed approach can generate a conflict-free timetable with the minimum arrival delay time and the number of delays of all trains at all stations compared to the nonintegrated rescheduling approach and the benchmark solution of the first-come-first-serve (FCFS) approach. The propagation of delay between dispatching sections is also greatly suppressed. The results can provide support for dispatchers to making reasonable rescheduling decisions in the case of large disruptions.

This paper proposes a novel train trajectory optimization approach for high-speed railways. We restrict our attention to single train operation scenarios with different scheduled/rescheduled running times aiming at generating optimal train recommended trajectories in real time, which can ensure punctuality and energy efficiency of train operation. A learning-based approach deep deterministic policy gradient (DDPG) is designed to generate optimal train trajectories based on the offline training from the interaction between the agent and the trajectory simulation environment. An allocating running time and selecting operation modes (ARTSOM) algorithm is proposed to improve train punctuality and give a series of discrete operation modes (full traction, cruising, coasting, full braking), and thus to produce a feasible training set for DDPG, which can speed up the training process. Numerical experiments show that an optimized speed profile can be generated by DDPG within seconds on a realistic railway line. In addition, the results demonstrate the generalization ability of trained DDPG in solving TTO problems with different running times and line conditions.

Emergency signage systems provide effective route guidance and evacuation information for pedestrians in case of emergencies such as fire, blackout, and explosion. This paper proposes a modified social force (SF) model to investigate crowd evacuation dynamics taking into account the influence of emergency signs. The perceiving probability model is formulated for the quantitative description of the probability that pedestrians can successfully notice the sign and clearly perceive the guidance information. Simulation experiments and controlled experiments are designed to calibrate the parameters of the proposed model. The effectiveness of the modified SF model is preliminarily verified by comparing the simulation results with experimental data and/or empirical results such as fundamental diagrams and self-organization phenomena. A case study of crowd evacuation simulations at a typical Beijing subway station is conducted to evaluate evacuation performance of three signage distribution schemes, i.e., Maximal Covering (MaxCover), Uniform, and Random, which are proposed by the maximal covering location and empirical approaches, as well as contrasted scheme without emergency signs. The effects of the quantity and distribution of emergency signs on crowd evacuation efficiency are studied quantitatively by simulations. The results show that installing emergency signs can improve evacuation efficiency no matter what distribution scheme is adopted. By choosing an appropriate distribution scheme i.e., MaxCover, the evacuation performance can be further improved and the evacuation time can be significantly reduced.

Considering the distances to exits, exits' capacities, the sizes of queues at exits, distances to the nuclear power plant, as well as individual characteristics, the exit choice model for pedestrians in the plume planning area is established based on a random forest model. This model is trained and verified with the survey data of residents around the Daya Bay Nuclear Power Plant collected from a serious game-based questionnaire system. Combining the pedestrian choice with the agent-based pedestrian behavior simulation model, the evacuation process of a nuclear accident is simulated. Based on the detailed evacuation simulation model, a comparative experiment is performed to evaluate the staged evacuation strategy in such scenarios. Simulation results indicate that staged evacuation may not be the best strategy all the time, and the number of groups highly impacts its performance.

This paper provides a critical review on the state-of-the-art pedestrian and evacuation dynamics so as to comprehensively comprehend the motion behaviors of pedestrians from observations to simulation aspects. Types of typical data collection methods, namely the field survey, the controlled experiment, and the animal experiment, are classified, and the connections and differences of these three observation methods are explored. Pedestrians' complex behaviors characterized by the self-organization phenomena and movement data characterized by the fundamental diagram are then studied after the data collections, which can be used to calibrate and validate the pedestrian models. The mathematical models for pedestrian dynamics from both tactical level and operational level are also highlighted. The simulation data produced by the mathematical models could further reproduce pedestrian behaviors during the observations and contribute to decision makings for improving the evacuation efficiency. The applications of pedestrian models for behavior analysis, evacuation simulation, and layout design are also presented. Some challenges and future directions in the pedestrian and evacuation dynamics are also put forward. Findings presented in this study are helpful for researchers who want to understand the pedestrian and evacuation dynamics and to perform further research in this field.

In this paper, pedestrian evacuation in the view-limited condition is investigated by using an extended social force model that considers both visibility distance and herding behavior. At first, the relations between visibility distance and mean evacuation time, density evolution, respectively, are explored. Then, the effects of herding behavior on the evacuation features, i.e., mean value and frequency distribution of evacuation times, and the evacuation process are investigated. The results show that in a certain range, the larger the visibility distance is, the faster is the evacuation process, and different visibilities lead to different tendencies of density fluctuations. It is found that herding behavior plays a beneficial role in evacuation and group formation appears when herding behavior dominates the selection of directions. In addition, the dual effects of pedestrian density on mean evacuation time are discovered. Our research can provide the theoretical guidance for the formulation of evacuation strategies in the view-limited condition.

The adoption of passenger leaders could make crowd evacuation in urban railway transit (URT) stations more efficient. The number, location, and the actions of the leaders are the most important elements in an evacuation strategy and have a great impact on evacuation efficiency. This paper proposes a hybrid bi-level model to optimize the number and initial locations of leaders as well as the routes of leaders during the evacuation, which explicitly incorporates the passengers' guidance demand and multi-leader coordination mechanism. The leaders' initial locations are generated by solving the maximal covering location problem (upper level model) and their evacuation routes are determined by a co-simulation heuristic approach (lower level model). The social force model and its modifications are used to model the dynamics of common evacuees, leaders, and followers in simulation models. The convergence performance of the proposed co-simulation heuristic approach and the effectiveness of the optimal evacuation strategy have been investigated and demonstrated using a case study of a typical island platform of Beijing's URT station. Three other evacuation strategies are considered for comparison purposes in order to show the influence of the number and initial locations of leaders as well as the multi-leader coordination mechanism during the evacuation process. Our analysis supported by simulations shows the following: 1) the optimal number of leaders exists for a given human cost and guidance demand constraints; 2) the distribution of leaders for maximal covering makes the evacuation of the followers more efficient; and 3) the proposed optimal evacuation strategy has better performance in terms of shorter evacuation time and higher utilization of exits compared with other considered strategies.

This paper focuses on the bibliographic analyses and collaboration patterns in the field of pedestrian and evacuation dynamics (PED) research covering the period of 1991-2017. The statistic analyses of most productive authors, institutions, countries/regions, and cited papers, as well as keywords and their trends, are conducted based on the data set collected from the Web of Science. The most productive and high-impact authors, institutions, and countries are identified. The results of bibliographic analyses show that Europe researchers dominate and guide the research of PED in that it not only has the most papers but also has six out of the ten most-cited papers. Helbing Dirk is an influential author in the research of PED field in that he has five out of the ten most-cited papers. Chinese institutions account for 70% of the top 10 most productive institution and three among the top 5 ranks. Meanwhile, researchers from China and USA have published nearly half of the papers in this field. In addition, we generate four networks, i.e., coauthorship network, coinstitution network, cocountry/region network, and document cocitation network to analyze collaboration patterns and evolution of PED research at different levels. The software of Citespace is adopted to visualize the topological interactions among authors, institutions, and countries/regions as well as document cocitations. The degree, betweenness, burst, and PageRank are selected and measured as indicators to identify the key nodes. Finally, some future directions are put forward. The results of this paper provide a better understanding of patterns, trends, and other important factors as a basis for directing research activities, sharing knowledge, and collaborating in the field of PED research.

Modeling and analyzing the behaviors and characteristics of crowds in emergency is a challenging task with significant practical meanings. In this paper, a fuzzy logic approach is proposed to describe crowd evacuation behaviors, taking into account the effect of assailants. First, the microscopic pedestrian model and the assailant model are developed according to their different intentions in evacuation scenarios. Pedestrians are further divided into three categories depending upon whether they are affected by assailants. The individual's behaviors are determined by the integration of recommendations of local obstacle-avoiding behavior, regional path-searching behavior, and global goal-seeking behavior with adjustable weighting factors, which are automatically adjusted based on the perceptual information obtained from the complex interaction with surrounding environments. Then, the proposed pedestrian model is validated by comparing the simulated fundamental diagram with a large variety of empirical and experimental data. Finally, simulations in a hall with a single exit are implemented. It is shown that the model can truly reappear typical collective phenomena such as “arching and clogging” and “faster-is-slower effect.” The variations of the model and scenario parameters, such as pedestrian's desired speed, exit width, assailant's desired speed, and duration of attack, greatly influence the evacuation efficiency. In addition, a novel “circuity phenomenon,” i.e., pedestrians will give up the direction of goal when they encounter assailants or they see assailants and, at the same time, perceive a very crowded exit, is observed in crowd evacuation simulations.