Train Stop Research Articles

A framework for measuring the training efficiency of a neural architecture

Measuring Efficiency in neural network system development is an open research problem. This paper presents an experimental framework to measure the training efficiency of a neural architecture. To demonstrate our approach, we analyze the training efficiency of Convolutional Neural Networks and Bayesian equivalents on the MNIST and CIFAR-10 tasks. Our results show that training efficiency decays as training progresses and varies across different stopping criteria for a given neural model and learning task. We also find a non-linear relationship between training stopping criteria, training Efficiency, model size, and training Efficiency. Furthermore, we illustrate the potential confounding effects of overtraining on measuring the training efficiency of a neural architecture. Regarding relative training efficiency across different architectures, our results indicate that CNNs are more efficient than BCNNs on both datasets. More generally, as a learning task becomes more complex, the relative difference in training efficiency between different architectures becomes more pronounced.

Effective passenger flow congestion propagation model for multi-mode rail transit networks

ABSTRACT During peak passenger flow periods, congestion propagation directly affects the operational safety and efficiency of multi-mode rail transit interconnections. By analyzing the key factors affecting congestion propagation, such as the train stop schedule, and considering parameters such as the basic reproduction number and propagation threshold, this study proposes a multi-mode rail transit susceptible-infected-recovered-susceptible (MRT-SIRS) epidemic model to analyze passenger flow congestion propagation. Simulation experiments and sensitivity analyses using data from the multi-mode rail transit in Beijing, China, were conducted to examine the influence mechanism of key factors on congestion propagation. The degree of influence of each factor was investigated using Gray correlation analysis. Each key factor, including the propagation and recovery rates, influences congestion propagation differently. The results of this study may provide theoretical support for the efficient operation and management of multi-mode rail transit systems.

Traces of Post-Soviet revolution poetic cinema in the films of Sergei Loznitsa

ABSTRACT This study analyzes the post-war documentaries ‘The Train Stop’, ‘Landscape’, and ‘Portrait’, which reflect the aftermath of the Post-Soviet revolution, through the lens of Brecht’s Aesthetics of Alienation. Combining elements of poetic cinema and Brechtian aesthetics, these films engage the viewer intellectually while depicting the societal and human impact of post-Soviet destruction. Loznitsa’s documentaries invite critical reflection, highlighting the legacies of the Soviet era and prompting a reconsideration of historical and contemporary issues. Through visual storytelling, they offer a critical perspective on resilience and societal transformation following political upheaval.

Research on prediction of human oral bioavailability of drugs based on improved deep forest

Human oral bioavailability is a crucial factor in drug discovery. In recent years, researchers have constructed a variety of different prediction models. However, given the limited size of human oral bioavailability data sets, the challenge of making accurate predictions with small sample sizes has become a critical issue in the field. The deep forest model, with its adaptively determinable number of cascade levels, can perform exceptionally well even on small-scale data. However, the original deep forest suffers unbalanced multi-grained scanning process and premature stopping of cascade forest training. In this paper, we propose a human oral bioavailability predict method based on an improved deep forest, called balanced multi-grained scanning mapping cascade forest (bgmc-forest). Firstly, the mordred descriptor method is selected to feature extraction, then enhanced features are obtained by the improved balanced multi-grained scanning, which solves the problem of missing features at both ends. And finally, the prediction results are obtained by feature mapping cascaded forests, which is based on principal component analysis and cascade forests, ensures the effectiveness of the cascade forest. The superiority of the model constructed in this paper is demonstrated through comparative experiments, while the effectiveness of the improved module is verified through ablation experiments. Finally the decision-making process of the model is explained by the shapley additive explanations interpretation algorithm.

Integrated optimization of train stop planning and seat allocation for high-speed railways based on active management for dynamic demand

ABSTRACT Proactive demand management methods can achieve a trade-off between supply and demand for high-speed railways. Focusing on multi-type demand configurations, four active demand management methods were designed, two single-scenario and two multi-scenario methods. An optimization model of stop planning and seat allocation was constructed to minimize the difference between demand and supply, number of trains, and train-stop costs. The number of train stops, transport capacity, and train occupancy rates were considered. Particle swarm optimization and CPLEX were combined to handle this significant linear programming problem. A real case study of different demand scenarios based on the Hohhot-Beijing high-speed railway in China verified the feasibility of active demand management methods. The results indicate that integrated optimization can be justified with mixed load patterns. Compared with the original scheme, the optimized scheme reduced the number of trains by 23.4% and increased train occupancy rate by more than 36.7%.

Mechanical Properties of Adjacent Pile Bases in Collapsible Loess under Metro Depot

Metro transit construction has begun to develop rapidly in northwest China because of the acceleration of urbanization. Accordingly, metro depots are also regarded as an essential auxiliary facility for stopping, operation, and maintenance of trains. Meanwhile, many commercial buildings are constructed over metro depots to improve the utilization rate of land due to the increasingly scarce urban land resources, known as transit-oriented development (TOD). These buildings have a large covered area and transfer concentrated loads to the bases. Therefore, pile bases under metro depots have the bearing characteristics of undertaking large concentrated loads, while lesser loads are placed on the soil between the adjacent pile bases. Additionally, the main ground in northwest China is collapsible loess, so the collapsibility should also be considered. Based on the above background, this research performed static loading tests with and without immersion in a reduced scale of adjacent pile bases under a metro depot in Xi’an. The remolding process of natural loess could destroy its structure and the anisotropy of natural loess could also affect the test results. Therefore, four kinds of artificial collapsible loess with different mass ratios of barite powder, kaolin, river sand, cement, industrial salt, and calcium oxide were made by the free-drop method. This method could make the artificial loess simulate the structure of natural loess reasonably. Then, the artificial loess with the most similar properties to intact loess was selected by comparison. Finally, static loading tests with this artificial loess were implemented. The results showed that the ultimate bearing capacity was 4.5 kN. At the same time, the axial force decreased along depth, since the pile shaft friction was positive, and the load sharing ratio of pile tip force increased to 0.58 when the load exceeded 4.5 kN in the situation without immersion; the settlement of pile bases increased significantly after immersion, while the negative shaft friction occurred at the depth of −8 cm~−35 cm, and the load sharing ratio of pile tip force reached 0.92.

Metro Train Stopping Scheme Decision Based on Multisource Data in Express-Local Train Mode

Metro Train Stopping Scheme Decision Based on Multisource Data in Express-Local Train Mode

Xception transfer learning with early stopping for facial age estimation

The rapid development of deep learning attracts more attention to the analysis of person's face images. Deep learning methodsof facial age estimation are more effective compared to methods based on anthropometric models, models of active appearance, texture models, subspace of aging patterns. However, deep learning networks require more computing power to process images. Pre-trained models do not need a large training set and their training time is less. However, the parameters obtained as a result of transfer learning of the pre-training network significantly affect its efficiency. It is also necessary to take into account the properties of the processed images, in particular, the conditions under which they were obtained.Recently, the facial age estimation is implemented in applications in devices with limited resources of computing, for example, in smartphones. The memory size and power consumption of such applications are limited by the computing power of mobile devices. In addition, when photographing a person's face with a smartphone camera, it is very difficult to ensure the uniform lighting. The aim of the research is reducing the error of facial age estimation from uneven illuminated images by applying an early stopping of transfer learning of the Xception network. The proposed technique of transfer learning includes an early stopping of training, if the improvement of the results is not observed within a certain number of epochs. Then the network weights from the epoch with the lowest validation loss are saved. As a result of the proposed technique applying, the average absolute error of age estimation was about five years from unevenly illuminated test images. A number of parameters of the used in this case Xceptionnetwork is less than that of other deep learning neural networks which solved the age estimation problem. Then applying of the Xception network reduces the resource consumption of devices with limited computing power. Prospects for further research are reducing the unevenness of facial image lighting to decrease the error of age estimation. Also, to reduce the computing resources, it is promising to use fast transforms in the Xception convolutional layers.

Energy dissipation and seismic response reduction system for high-speed railway bridges based on multiple performance requirements

The emergence of high-speed railway bridges in regions with high seismic intensity has brought attention to the growing seismic challenges associated with these structures. Although the prevalent use of friction pendulum bearings for seismic isolation in China has proven effective in mitigating the seismic response of bridges, it falls short of meeting the safety standards required for high-speed trains. To address the limitations of existing isolation systems, this study employs a combined approach of numerical simulation and experimental research, leading to the invention of performance-controllable isolation bearing and multi-directional energy dissipation dampers. Mechanical models for both components were introduced, and design methods for their key parameters were established based on multi-level seismic defense requirements. By integrating these advancements, a novel cooperative seismic isolation system for high-speed railway bridges was created. Under the influence of moderate earthquakes, the system relies on the shear key to meet the stiffness requirements for normal service. During large earthquakes, the shear key is sheared off, and the damper provides momentary stiffness to control track deformation, enabling the safe stopping of trains. This combination forms a seismic isolation system during large earthquakes, ensuring the safety of railway bridges. The numerical analysis of a high-speed railway bridge employing a cooperative seismic isolation system demonstrates substantial mitigation in the response of both the pier bottom and pier top. This aligns with seismic requirements for the bridge structure. Additionally, the reduction in train track deformation meets safety standards, confirming the seismic performance of bridges and ensuring the reliability of train operations.

A comparative study on deep learning models for condition monitoring of advanced reactor piping systems

Advanced nuclear reactors offer innovative applications due to their portability, reliability, resiliency, and high capacity factors. To operate them on a wider scale, reducing maintenance life-cycle costs while ensuring their integrity is essential. Autonomous operations in advanced nuclear reactors using augmented Digital Twin (DT) technology can serve as a cost-effective solution by increasing awareness about the system’s health. A key component of nuclear DT frameworks is the condition monitoring of safety systems, such as piping-equipment systems, which involves acquiring and monitoring the plant’s sensor data. This research proposes a condition monitoring methodology utilizing deep learning algorithms, such as multilayer perceptions (MLP) and convolutional neural networks (CNNs), to detect degradation and its severity in nuclear piping-equipment systems. Sensor signals are processed to obtain the power spectral density and the Short-Time Fourier transform, and feature extraction methodologies are proposed to develop degradation-sensitive data repositories. The performance of MLP, one-dimensional (1D) CNN, and 2D CNN within the proposed condition monitoring framework is compared using a finite element model of a 3D piping system subjected to seismic loads as the application case study. Various approaches, such as dropout, k-Fold validation, regularization, and early stopping of training the network, are investigated to avoid overfitting the models to the input sensor data. The predictive capability and computational capacity of the deep learning algorithms are also compared to detect degradation in the Z-pipe system of the Experimental Breeder Reactor II (EBRII). The Z-pipe system is subjected to harmonic excitations that represent normal operating loads, such as pump-induced vibrations. The findings of the study indicate that the proposed artificial intelligence (AI)-driven condition monitoring framework demonstrates superior prediction accuracies with a 2D CNN, whereas the MLP exhibits higher computational efficiency.

A flexible train composition strategy with extra-long trains for high-speed railway corridors with time-varying demand

To accommodate the uneven spatio-temporal distribution of passenger demand and improve the maximum transportation capacity for high-speed railway (HSR) corridors, this study proposes a demand-oriented flexible train composition strategy. The proposed strategy allows flexibility in selecting the number of train composition units (e.g., train carriages) for each train to accommodate the demand variations, where extra-long trains (longer than the length of the train station platform) might be utilized. Under such a strategy, a seat allocation method coupled with a train stopping position control strategy is proposed to avoid the need for cross-carriage walking during passenger boarding or alighting (for those extra-long trains). The train stopping position control strategy specifies the carriages that dock at or extend beyond the station platform, while the seat allocation method allocates seats within each carriage to specific origin-destination pairs. The studied problem can be formulated as an integer nonlinear programming model, where the weighted sum of the HSR operating cost and passenger travel time cost is minimized. The proposed model is then reformulated into an integer linear programming model via a series of linearization techniques. A tailored heuristic algorithm based on the variable neighborhood search (VNS) and GUROBI solver is designed to produce high-quality solutions for large-scale problems. Two numerical examples, i.e., a small example and a real-world Shanghai-Hangzhou HSR line example, are presented to illustrate the effectiveness of the proposed approach. The computation results show that, in comparison to traditional fixed and flexible train composition strategies without extra-long trains, the proposed strategy can significantly reduce the total operating cost and total travel time cost.

철도 역명 제정 및 개정에 관한 사회적 갈등 양상과 합리적 해결 방안 연구

In modern society, railways are Korea's core means of transportation that provide equitable convenience to all citizens, making a profound contribution to the welfare of the people. The railway station is not just a point where trains stop, but also a place of history and culture that informs an area, a place where various cultural events and social exchanges take place, and a space for various commercial and economic activities. Therefore, the station name, the name of the railway station, is very symbolic and has a complex influence in history, culture, society, and economy. The enactment or revision of the station name is a key means of transportation that contributes to the welfare of the people, and the task is to be reasonably and objectively improved in consideration of the symbolic and complex influence and ripple effect of the railway station name involved in regional identity. This paper aims to improve the standard of route name, station name, and station name bookkeeping that meets public satisfaction and expectations by improving the standard of reasonable railway line name, station name, and station name bookkeeping. The public's interest in railway station names is growing more than ever, and measures to improve objective standards were proposed not only by preventing unnecessary conflicts between different local governments and social organizations, but also by reviewing reasonable station names and standards for station bookkeeping to meet public interest and expectations. In this analysis, a plan was also suggested to consider the historical meaning of the railway station name together.

A Distributed Multiparticle Precise Stopping Control Model Based on the Distributed Model Predictive Control Algorithm for High-Speed Trains

Abstract The fixed-point stopping of high-speed trains in stations is generally accomplished through manual operation in China. This situation often leads to a failure to stop at fixed-point signs and causes a fluctuation in the longitudinal acceleration due to the lack of experience of the drivers. To achieve precise, stable, and automatic stopping, a distributed multiparticle precise stopping control model based on the distributed model predictive control (MPC) algorithm is developed in this paper. A two-level hierarchical control structure for the subcontroller of each vehicle is adopted to bring itself to a controlled stop. In the upper control of subcontroller, the MPC algorithm is designed in turn based on the multiparticle mechanism model of the train. In the lower control of subcontroller, the target input from the upper control is converted and distributed. The controlled object, a comprehensive numerical computing model including the spatial dynamic model of the train and its electropneumatic blending braking model, is established and controlled by the corresponding subcontroller and employed to verify the performance of the controller. The influence of the model control parameters on the stopping performance is discussed, and the optimal combination of control parameters is selected. The proposed control model using the optimization parameters is tested and verified through the comprehensive numerical computing model. The results indicate that the stopping error is 0.0075 m, which is much less than the accuracy requirements for fixed-point stopping. The computing time of each subcontroller in real-time is stable at 0.09 s. The coupler impact force between two adjacent vehicles can also be effectively inhibited and eliminated. Its control performance outperforms a proportional–integral–derivative (PID) algorithm. The proposed precise stopping model and comprehensive numerical computing model provide references for the application and algorithm optimization of automatic train operation technology in high-speed trains.

On-time and energy-saving train operation strategy based on improved AGA multi-objective optimization

On-time and energy-saving train operation is important for the sustainable development of rail transit. As for the problems of traction energy consumption and on-time arrival at stations faced by trains in rail transit, an optimization strategy of energy-saving speed curves of trains based on an improved adaptive genetic algorithm (AGA) was proposed in this paper. First, weight coefficients of operation time and energy consumption were designed through an analytic hierarchy process, and an optimization model that targets train operation time and energy consumption was established according to a basic train operation model with constraints such as speed limits and precise train stopping. Then, on-time and energy-saving speed curves of trains were generated based on the improved AGA. Finally, a simulation was carried out with actual rail transit lines. The results show that the proposed method has strong efficiency for energy conservation and better optimization performance than the simple genetic algorithm in solving train trajectory optimization problem.

Integrated optimization of train stopping plan and seat allocation scheme for railway systems under equilibrium travel choice and elastic demand

This paper examines the integrated optimization of train stopping plan and seat allocation scheme in railway systems, where equilibrium passenger travel choice under elastic demand is considered. The integrated optimization problem is formulated as a non-concave and non-linear mixed-integer mathematical model, where the objective is to maximize the system net benefit considering ticket revenue, consumer surplus, and cost associated with train stoppings. The integrated optimization model can be reformulated into a mixed-integer linear programming (MILP) model based on a series of linearization, relaxation, and outer-approximation techniques, which can then be solved by commercial MILP solvers (e.g., GUROBI). We also compare the integrated optimization approach with that when the train stopping plan and seat allocation are optimized separately and identify the potential benefits. Numerical studies have been conducted on a small-scale example, the Zhengzhou-Xi’an and Shanghai-Beijing high-speed railway corridors to illustrate the proposed model and solution approach.

Train stop scheduling problem: An exact approach using valid inequalities and polar duality

Consider the problem of minimizing the number of train stops on a particular rail line. The objective is to assign passengers of each origin-destination pair to different trains in such a way that train capacity and passenger demand constraints are satisfied with minimal stoppages. The literature refers to this problem as the train stop scheduling problem. The problem has been extensively studied for decades, yet an exact solution approach has not been proposed. This paper proposes several valid inequalities to strengthen the mixed-integer programming formulation and solve the problem exactly in a reasonable amount of CPU time. The concept of polar duality has been utilized to find more complex valid inequalities which may be hard to find otherwise. Despite the problem’s high practical relevance, valid inequalities for the problem have not yet been studied in the literature. An aggregation procedure has been proposed to solve large size problem instances exactly. Computational study demonstrate the efficacy of the proposed valid inequalities.

Latent feature reconstruction for unsupervised anomaly detection

AbstractAnomalies (or outliers) indicate a minority of data items that are quite different from the majority (inliers) of a dataset in a certain aspect. Unsupervised anomaly detection (UAD) is an important but not yet extensively studied research topic. Recent deep learning based methods exploit the reconstruction gap between inliers and outliers to discriminate them. However, it is observed that the reconstruction gap often decreases rapidly as the training process goes. And there is no reasonable way to set the training stop point. To support effective UAD, we propose a new UAD framework by introducing a Latent Feature Reconstruction (LFR) layer that can be applied to recent UAD methods. The LFR layer acts as a regularizer to constrain the latent features in a low-rank subspace from which inliers can be reconstructed well while outliers cannot. We develop two new UAD methods by implementing the proposed framework with autoencoder architecture and geometric transformation scheme. Experiments on five benchmarks show that our proposed methods can achieve state-of-the-art performance in most cases.

Analysis of Construction Networks and Structural Characteristics of Pearl River Delta and Surrounding Cities Based on Multiple Connections

The Pearl River Delta (PRD) is one of three world-class city clusters in China, which is important for the strategical deployment of the national “Belt and Road”. Based on nighttime lighting data, Baidu index, and train stopping times, this study analyzed the network of spatial patterns and structural evolution of the PRD and surrounding cities via social network analysis and dynamic network visualization, providing new perspectives and ideas for the study of intercity linkages and urban networks. The results provide decision support to the government for urban cluster planning. From 2014 to 2020, the economic network evolved from a uniaxial structure to an “inverted V” structure. The transportation network evolved from a uniaxial structure to a “△” structure. The information network did not show any obvious structural changes during its development, except for a star-shaped radial structure. The PRD city cluster and its surrounding cities exhibited a spatially non-uniform distribution in terms of spatial connections. The total connections between Guangzhou and Foshan and the surrounding cities in terms of economic, transportation, and information functions account for 30%, 28%, and 10% of the total urban connections, respectively. The graph entropy growth rates of the PRD city cluster and surrounding cities in economic, transportation, and information networks from 2014 to 2020 were 39.9%, 115.4%, and 5.1%, respectively. The network structures of economic and transportation networks stabilized eventually. The information network structures are stable.

Intra-individual variability adaptively increases following inhibition training during middle childhood

There is ongoing debate on the relationship between intra-individual variability (IIV) of cognitive processes and task performance. While psychological research has traditionally assumed that lower intra-individual variability (IIV) aids consistent task performance, some studies suggest that greater IIV can also be adaptive, especially when flexible responding is required. Here we selectively manipulate inhibitory control (Stopping) and response speed (Going) by means of a training paradigm to 1) assess how this manipulation impacts Stopping IIV and its relationship to task performance, and 2) replicate previous findings showing that reductions in Going IIV are adaptive. A group of 208 6–13-year-old children were randomly allocated to an 8-week training targeting Stopping (experimental group) or Going (control group). The stop signal task was administered before and after training. Training Stopping led to adaptive increases in Stopping IIV, where greater flexibility in cognitive processing may be required to meet higher task demands. In line with previous studies, training Going led to adaptive reductions in Going IIV, which allows more consistent and efficient Going performance. These findings provide systematic and causal evidence of the process-dependent relationship of IIV and task performance in the context of Stopping and Going, suggesting a more nuanced perspective on IIV with implications for developmental, ageing and intervention studies.

Network periodic train timetabling with integrated stop planning and passenger routing: A periodic time–space network construct and ADMM algorithm

Designing a passenger-centric railway service is of great importance for railway companies to attract more passengers and increase their revenue. Focusing on the passenger-centric operational plans, this paper considers the problem of integrated periodic train timetabling and train stop planning in high speed railway networks. In view of the immediate dependency of passengers’ routes on the train timetable and stop plan, simultaneous passenger routing is included to directly shape the train stops and timetable in a passenger-friendly way. Besides travel time, ticket price is also considered in passenger utility to better depict passengers’ route choices in the railway network. A railway network-based time–space network with train stop-skip technique, periodic properties and detailed consideration of different train headways is designed, based on which a multi-commodity network flow model is developed. With a novel linearization technique of train capacity penalty term, the alternating direction method of multipliers (ADMM) is used to relax and decompose the track and train capacity constraints. Further, time deviation of regular train services is achieved by a new shortest path algorithm to design more flexible timetables, however its application is still limited to small instances due to computational efforts required. The algorithm is finally tested on the Shandong high-speed railway network in China, and demonstrates its capability of providing tighter lower bounds than Lagrangian relaxation method and generating good feasible solutions. Besides, the impacts of different cycle times on service quality and train cost are evaluated.