From the perspective of rail transit enterprise management, this study addresses the problem of coordinating and balancing rolling stock circulation efficiency, train load factor, and operational economics in cross-line operation, and constructs a multi-dimensional collaborative optimization model for cross-line train timetables with the objective of minimizing operational train-kilometers. The study focuses on the economic optimization of the spatiotemporal allocation of transport resources. By integrating core elements such as timetable scheduling, turnaround operations, depot operation plans, and cross-line connections, a linkage constraint mechanism between maximum headway and train load factor is established, enabling coordinated control of operating costs and service reliability. To improve solution efficiency, the FSAE-CLO (fast solution algorithm for engineering-cross line operation) algorithm is proposed, which transforms cross-line timetable optimization into a combinatorial search problem for cross-line decisions, gradually screens feasible schemes according to the sequence of trains available for cross-line operation, and adjusts cross-line decisions based on an iterative probability update mechanism, thereby compressing the solution space while reducing ineffective searches. Taking the morning peak hours of Beijing Metro Fangshan Line and Line 9 as an example, the empirical results show that, compared with the single-line operation mode, the proposed scheme reduces train-kilometers by 6.83%; compared with the genetic algorithm, it reduces operational mileage by 4% and shortens solution time by 24%. This study provides a new theoretical method and practical support from the perspective of cost reduction and efficiency improvement for transport enterprises in making decisions on adopting cross-line transportation organization methods, and its resource coupling optimization mechanism has reference value for improving the benefits of networked operation.

To achieve accurate identification and intelligent monitoring of metro rail corrugation, this study proposes a novel method combining WPD and 1DCNN for rail corrugation identification. Based on the vertical axle-box acceleration signals collected from in-service metro trains, sixteen sub-band energy features were extracted through four-level WPD and normalized to construct energy vectors. A lightweight 1DCNN architecture was designed to perform multiscale feature learning using the energy vectors as input. Three corrugation types—no corrugation, long-wavelength corrugation, and short-wavelength corrugation—were selected as classification labels for model training and testing. Based on the differences in energy distribution among different corrugation types in low- and high-frequency sub-bands, the feature extraction results were analyzed. The results demonstrated that the proposed method achieved an average recognition accuracy of 97.60%, with a recall rate of 100% for long-wavelength corrugation. Compared with SVM, LSTM, and 2DCNN models, the 1DCNN model exhibited higher accuracy and stronger generalization capability. It enables intelligent rail corrugation identification under in-service operating conditions without reliance on specialized inspection systems. The consistency between the wavelength-domain evaluation framework of conventional rail corrugation testing and the identification results of this study was analyzed, and a complementary engineering application workflow integrating online identification with on-site precision measurement was proposed. This study verified the effectiveness of the WPD-1DCNN fusion method in metro rail corrugation recognition and provided a new approach for online monitoring and intelligent maintenance of track conditions.

To address the issues of high reliance on manual experience, poor dynamic adaptability of emergency decision-making, and low efficiency of decision generation in urban rail transit emergency response, a decision-making model for emergency response is proposed, which integrates a "predicate-centered event representation framework" and a "causal graph-retrieval augmented generation (RAG) integrated architecture". Firstly, a non-static trigger word is adopted to construct a process-oriented knowledge structure for emergency response that is superior to the static knowledge graph. Then, by combining information grouping, situation assessment, and an "event-situation" two-stage query rewriting strategy, the proposed model enhances traditional RAG with a stronger ability to understand the logic of event chains, thereby generating emergency response decisions that are more actionable and better adapted to specific scenarios. Finally, evaluation indicators such as the decision compliance rate, decision generation time, and decision efficiency improvement are defined and applied to comparative experiments to evaluate the model's performance. The research results show that compared with the traditional RAG method, the proposed model improves key-node accuracy by 29.84% in the decision compliance evaluation and achieves a logical-relation correctness rate of 84.74%. Meanwhile, compared with traditional manual decision-making, the proposed model improves generation efficiency by 57.14% indicating that the model can shorten the response time while meeting compliance requirements, providing a reference for constructing an intelligent dispatch decision generation system for urban rail transit.

With the rapid development of large AI models, the demand for their application in urban rail transit is increasing rapidly. However, there is still a lack of evaluation standards and methods for large models specific to the urban rail transit industry. To address this, this paper establishes an evaluation system for large models tailored to the urban rail transit industry. This evaluation system covers the construction of evaluation datasets, the design of evaluation dimensions, the selection of evaluation methods, and the formulation of evaluation metrics. Through this system, the performance of urban rail transit large models for urban rail transit in tasks such as language understanding, knowledge reasoning, image-text alignment, and spatiotemporal consistency can be evaluated comprehensively and objectively. Meanwhile, a dedicated evaluation system for large models for urban rail transit is designed and developed, and the effectiveness of the evaluation system is verified through practical application. The verification results show that the established evaluation system can provide an evaluation basis for the optimization and application of large models in urban rail transit, and promote their implementation and application in the urban rail transit industry.

To promote equity and expand the spatial and social reach of multimodal rail systems, this paper examines the intersection of multimodal transit and social equity. It analyzes key research approaches, critical challenges, and influencing factors, and proposes future research directions and policy strategies. Findings suggest that two fundamental issues underpin the inequity of multimodal rail systems, namely spatial disparities in network coverage and insufficient attention to the diverse needs of users. Rail station and line densities tend to be lower in low-density urban neighborhoods and peri-urban or rural areas, making access more difficult. Vulnerable groups, including the elderly, people with disabilities, low-income populations, and individuals with lower educational attainment, are particularly affected by economic constraints, digital exclusion, and time-space coordination challenges in their daily mobility. Future research on social equity in multimodal transit systems should adopt a holistic perspective that integrates built environment characteristics, transport infrastructure provision, temporal factors, individual attributes, economic policies, intermodal connectivity, and digital literacy. Advancing equity in this context requires innovation in methodological frameworks, model development, as well as the identification and quantification of influencing mechanisms. As a capital-intensive foundational infrastructure, urban rail transit plays a critical role in shaping inclusive and equitable transport systems. Addressing social equity in multimodal transit is vital for enhancing population mobility, increasing employment opportunities, and promoting balanced and sustainable regional development.

With the increasingly prominent role of bike-sharing in the "last mile" of urban transportation, the issue of service equity has attracted extensive attention. Based on tshared bike usage data and metro station passenger flow data in Shenzhen in August 2021, this research evaluates the equity characteristics of bike-sharing services in Shenzhen metro catchments within 15 min cycling catchments and analyzes their spatial patterns and temporal differences. By constructing two types of indicators, namely the Gini coefficient of per capita bicycle use, and the Gini coefficient of supply-demand matching, this study quantifies the imbalance in bike-sharing use in space and crowd. A higher Gini coefficient indicates a greater degree of inequity across spatial and demographic dimensions. The results show that there is a certain inequity in the use of shared bicycles in Shenzhen, with significant differences across spatial and temporal contexts. The distribution of bike-sharing use within metro catchments in central urban areas is relatively balanced, and the use of bicycles in the suburban new city is highly concentrated. Some metro lines (such as those connecting the central urban area and the outer new towns) have significantly lower levels of bike-sharing use equity than those in the central area. Bike-sharing usage is closer to the distribution of commuting demand during the morning and evening rush hours on weekdays, while there is a mismatch between supply and demand during weekends. Finally, based on the results, this paper proposes targeted recommendations for optimizing the layout and scheduling of bike-sharing services in metro catchments to improve the equity of bike-sharing feeder services for metro stations.

To address the limitations of existing research on land value premiums of urban rail transit, which are characterized by short time spans and insufficient exploration of the long-term dynamic impacts of economic cycles, this study, based on data from the Tokyo Metropolitan Area spanning 1984-2023, for the first time, constructs a long-time-series framework covering the bubble period, recession period, and recovery period. It employs the Geographical and Temporal Weighted Regression (GTWR) model to quantitatively capture the spatiotemporal heterogeneity of land value premiums. The results indicate that during the bubble period, the rail transit-induced land value premium exhibits a core-periphery gradient attenuation, with commercial land demonstrating the highest sensitivity; during the recession period, the premium effect reverses direction, leading to strip-shaped depreciation along rail transit corridors in urban areas while exurban land maintains stronger stability in its premium levels; during the recovery period, the overall premium elasticity weakens, yet peri-urban commercial land shows prominent growth potential, with elasticity approximately 1.5 times that of urban areas. The study reveals that there exists a dynamic interaction between macroeconomics and transit accessibility, necessitating the establishment of a cross-cycle response mechanism. This research uncovers the nonlinear dynamic interaction mechanism between macroeconomic cycles and rail transit accessibility, fills the research gap in the quantitative study of cross-cycle spatiotemporal heterogeneity of rail transit- induced land value premiums, provides data support for quantifying such premiums in metropolitan areas under economic fluctuations, and offers theoretical support for mitigating systemic risks and enhancing planning resilience.

Based on a case study of Beijing's Chaoyang District, this research employs a multiscale geographically weighted regression (MGWR) model to systematically examine the spatial heterogeneity of housing prices around subway stations and the mechanisms through which multiple factors affect them. Using second-hand housing listing data from Lianjia in 2023, the study incorporates both internal estate characteristics (e.g., average housing age, property fees, greening rate, and plot ratio) and external location-based attributes (e.g., proximity to subway stations, bus stops, commercial centers, schools, hospitals, and parks) to construct the MGWR model, which is compared against ordinary least squares (OLS) model and the geographically weighted regression (GWR) model. The results show that the MGWR model achieves the highest goodness-of-fit (R²=0.691) and effectively captures the scale-varying effects of different factors. Distance to subway stations exhibits a significant negative correlation with housing prices, with a stronger effect in peripheral areas than in the urban core. Proximity to top-tier schools shows a heterogeneous pattern, positive in clustered areas but negative in peripheral zones, reflecting a localized saturation effect of educational resources. The influences of commercial, park, and hospital accessibility also demonstrate notable spatial variations, highlighting the complex interplay and contextual dominance of transportation, education, commerce, ecology, and medical services in the capitalization of housing prices. This study provides a multiscale analytical framework for understanding the formation mechanisms of housing prices in the context of urban rail transit, offering valuable insights for spatial governance and public service allocation.

This paper focuses on the operational organization issues of metro logistics systems, drawing on domestic and international practices and taking literature from 2016 to 2025 in databases like Web of Science and CNKI as samples. By integrating bibliometric and content analysis methods, it examines research progress in this field. Subsequently, it provides an in-depth review of studies using quantitative analysis methods from perspectives such as research objects, stakeholders, transportation organization modes, train scheduling, capacity allocation, freight flow distribution, modeling strategies, and solution algorithms. The findings show that international practice and research started earlier but have not been continuously promoted, while domestic research, despite considerable attention, remains in a small-scale exploratory phase. Current research lacks systematic connections when analyzing key issues like transportation organization modes, train scheduling, capacity allocation, and freight flow distribution, highlighting the urgent need to construct an integrated optimization framework for multi-element collaboration. The research objects need to be further extended from single-line scenarios to complex networks to enhance the practicality and applicability of the theoretical findings. Under the green and low-carbon objective, incorporating energy consumption indicators into scheduling optimization and integrating low-carbon strategies such as regenerative braking energy utilization is of great significance. Meanwhile, moving beyond fixed demand assumptions and strengthening the ability to respond to uncertainty and elastic demand can improve the alignment between research and reality. Additionally, developing optimization algorithms suitable for large-scale networks remains a key direction for future research.

To address the significant limitations of using urban rail transit for freight transport and the lack of large-scale operation despite the gradual launch of pilot urban rail freight projects in some cities, this paper presents a systematic statistical analysis of the current status of urban rail freight in China based on a synthesis of existing practices and research. First, three typical operation modes are summarized: mixed passenger-freight transport, shared-line trailer operation, and shared-line separated operation. Second, using statistical data from 18 pilot cities, a quantitative analysis is conducted from three perspectives: spatial distribution, temporal evolution, and comparison of pilot projects in representative cities. Finally, the main challenges in the development of urban rail freight are examined, and corresponding development strategies are summarized. The results indicate that urban rail freight projects in China exhibit a spatial pattern radiating from the eastern coastal regions to hub cities in central and western China, with high concentrations in the Beijing-Tianjin-Hebei region and the Yangtze River Delta; urban rail freight is transitioning from the initial exploration stage to large-scale pilot implementation, with the number of pilot projects increasing year by year; most pilot cases adopt the mixed passenger-freight mode, a few use the shared trailer mode, and operations are mostly scheduled during off-peak passenger hours. Urban rail freight can effectively utilize idle capacity, alleviate urban logistics pressure, and promote green development, but it still faces challenges such as incompatibility of existing infrastructure, operational scheduling complexity, poor connectivity, inadequate safety control, data barriers, and profitability difficulties. The development trend of urban rail freight points toward the coordinated advancement of line retrofitting, market-oriented mechanisms, the construction of an integrated ground-air distribution system, and technological innovation.