ABSTRACT. Women and gender non-conforming individuals (GNC) often feel less safe at and around public transport (PT) stations, compared to men. While the significant influence of built environment (BE) on perceived safety is agreed upon, its interaction with gender remains underexplored. Using a tailor-made survey data (N=3,101) from East Denmark, we investigate BE features effective at addressing women & GNC’s safety concerns at the home and activity travel environments. Linear regression models reveal that lighting, cleanliness, trees and greenery benefit all travellers, while activating isolated areas around stations would especially benefit women & GNC, and opening up facades would enhance safety at the activity end. These findings are relevant for public transport agencies and local authorities seeking to address women & GNC's safety concerns while contributing to a more inclusive and attractive public transport system.

ABSTRACT. Public transport companies (PTCs) are facing significant challenge in phasing out high-polluting diesel vehicles and selecting alternative propulsion solution for their city bus fleet renewal. To support PTCs in this complex multi-criteria decision making process, this paper presents an integrated method, that includes the improved Analytical Hierarchy Process (AHP) to define the weights of criteria, and the Technique for Order Preference by Similarity to Ideal Solution (TOPSIS) to assess and rank considered solutions. The method is applied in a case study of Arriva, one of the Dutch major PTCs, with battery-electric system with opportunity charging identified as the optimal configuration.

ABSTRACT. We classified and evaluated main railway stations in Chukyo Metropolitan Area with advanced node-place model based on potential factors considering feeder networks. Using five factors of rail-based node, car assisted node, and three places, stations are classified into six clusters. The results showed that firstly, node and place could be divided based on feeder modes and activity types, suggesting the necessity of advancement in model. Secondly, stations could be evaluated by various isochrones. Finally, stations were in monocentric structure while different patterns were identified with respect to access/egress potential between transit and car, suggesting delicate approaches for multimodal policies.

ABSTRACT. Urban congestion remains a critical challenge for sustainable urban mobility. This paper examines how road-rationing policies, such as the odd-even vehicle regulation, can act as a trigger to increase public transit (PT) usage when integrated with enhanced last-mile solutions. Using real-world data from Delhi, India, we analyze the impact of road rationing on public transit speeds, demonstrating improvements of up to 50%, alongside significant reductions in travel costs. We propose a policy framework leveraging these short-term benefits to encourage long-term shifts towards public transit adoption. Our findings highlight the potential of combining road-rationing with strategic multimodal transit planning to mitigate congestion and promote sustainable urban mobility.

ABSTRACT. This study assesses the interrelationships between public transportation demand, shared micro-mobility usage, and land use characteristics in a car-centric urban environment using a two-stage least squares modeling approach. ZIP code-level analysis reveals that proximity to urban centers, road infrastructure, and land use types significantly influence population density and mobility patterns. Public transportation ridership benefits from high-density areas and frequent service, while shared micro- mobility is less utilized in densely populated zones, complementing public transit in lower-density areas. The findings emphasize the importance of integrating land use and transportation planning to mitigate urban sprawl and promote sustainable urban mobility.

ABSTRACT. Equitable access to opportunities is essential for building connected and inclusive communities. Yet, many Canadian cities continue to show disparities in how different population groups reach key destinations. This research presents the Mobilizing Justice’s Transportation Equity Dashboard, a new interactive platform designed to support data-driven transportation infrastructure planning across Canada.

The Dashboard combines open data on transportation networks, land use, and census demographics to measure access to seven essential destinations: employment, primary schools, groceries, healthcare, recreation, cultural venues, and post-secondary institutions. It does so across three travel modes —walking, cycling, and public transit (both peak and off-peak periods)— and disaggregates results by 20 equity-deserving population groups, including low-income, seniors, recent immigrants, and visible minorities.

Covering all 41 Census Metropolitan Areas in Canada, the Dashboard allows users to identify spatial inequalities in access, compare across cities and groups, and prioritize neighbourhoods where low accessibility levels and high concentration of equity-seeking groups intersect. By offering a national, standardized tool for accessibility and equity analysis, this tool is designed for planners, policymakers, researchers, and advocates seeking to integrate equity into transportation decisions.

ABSTRACT. Urban transit hubs face increasing congestion due to growing transit demand and complex trip patterns. Effective management is essential to improve connectivity and reduce travel time. This study analyzes transfer hubs by integrating user trip patterns and station area environments. Utilizing correlation analysis, PCA, and clustering methods, diverse hub types are identified, providing foundational data for systematic management strategies. The findings offer significant insights for future facility expansion, congestion mitigation, and demand forecasting, serving as critical baseline data for efficient transit operations in metropolitan areas.

ABSTRACT. In this paper, we propose an LLM-based knowledge system that transforms historical incident reports into institutional knowledge for incident management in public transit. By converting unstructured data into a structured knowledge graph, the system captures seasoned personnel’s practical experience and merges it with official guidelines, providing decision support. A retrieval augmented generation approach integrates these data sources to yield context-specific recommendations for route supervisors. This framework underscores the potential of advanced language models for consistent, transparent, and explainable incident control.

ABSTRACT. In Japan, the transportation gap between urban and rural areas is widening, emphasizing the need for flexible mobility options for rural communities. Demand Responsive Transport (DRT) offers a potential solution with reservation-based flexibility. However, it may compete with existing public transportation services, which could undermine the efficiency of the transportation network. This study proposes a method for identifying areas suitable for introducing transfer-based DRT that complements existing bus services. The method uses generalized cost to evaluate travel options from the user’s perspective. The method was applied to Kameyama City, Japan, and considers three transportation mode combinations: 1. Traveling directly from the origin to the destination using a taxi, 2. Walking to the nearest bus stop, taking buses to the stop closest to the destination, and then walking to the final destination, 3. Using DRT to reach a specific bus stop, taking buses to the stop nearest to the destination, and then walking to the final destination. The results indicated that generalized costs are minimized in regions with few bus routes situated far from the city center, as well as in areas surrounding the city center that have better access to hub bus stops situated within the center.

ABSTRACT. The COVID-19 pandemic has led to enduring changes in urban mobility and real estate sectors, with transit ridership stabilizing at 73% of 2019 levels by 2024 and remote work patterns remaining consistent since 2022. Disaggregated analysis of ridership trajectories in Washington DC, however, reveals significant heterogeneity which can be partially explained by real estate characteristics. This study jointly analyzes transit ridership and corresponding characteristics and changes to the real estate markets around stations, employing time-series clustering station ridership trajectories to identify distinct recovery patterns. These findings underscore the significant potential of strategic TOD planning to play a pivotal role in revitalizing urban transit systems in the post-pandemic era.

ABSTRACT. If aging individuals who rely heavily on private vehicles lose the ability to drive, they must either use alternative transportation or forgo certain activities. With the expansion of online services, some daily activities can potentially be conducted virtually, serving as substitutes for physical mobility. Although many accessibility indicators have been proposed in the literature, none has accounted for the complementarity between online services and physical transportation.

This study develops a new accessibility index that incorporates the acceptability of online alternatives and quantifies the impact of losing access to private vehicles. Based on a questionnaire survey, the study also evaluates the extent to which online services can compensate for reduced physical mobility.

The results reveal that 10 to 15% of individuals are likely to stop traveling when they can no longer use a private car, and about half of them do not accept online alternatives. This means that approximately 5% of people may be at risk of being unable to continue their daily activities in the future. This tendency is particularly pronounced in less densely populated areas—especially those far from train stations and with a high proportion of elderly residents—suggesting the need for both enhanced physical mobility support and improved access to digital services in such regions.

ABSTRACT. This paper presents a scalable methodology for real-time transfer synchronization in urban bus networks, using online stochastic optimization (OSO). The approach integrates three key components. First, an offline arc-flow model captures all control tactics—hold, speedup, and skip-stop—for a main line and its feeder connections, using a graph-based representation over a fixed control horizon. Second, the REGRET (R) algorithm operates in real time within an OSO framework, leveraging the offline model to evaluate multiple stochastic scenarios and select robust control tactics. Third, a network-wide simulator (NWS) integrates the full OSO framework and re-optimizes decisions dynamically at each bus departure from any stop, allowing the coordination of multiple interconnected lines. The NWS is applied to the transit network of Laval, Canada, using historical vehicle positions and smart-card validations to replicate real-time stochastic conditions. Results show significant improvements in both passenger travel and transfer times across a variety of network structures, highlighting the scalability and applicability of real-time transfer synchronization for urban multi-line transit networks.

ABSTRACT. Automatic Passenger Count (APC) data are used to estimate city-wide origin destination (OD) passenger flows for the purpose of investigating changes in travel patterns resulting from exogenous events. Quantified changes from such “before and after” analyses can assess impacts of events and be used to inform service improvements. APC data collected on Central Ohio Transit Authority buses over 33 months are used to estimate route-direction-bus-trip stop-to-stop passenger OD flows that are aggregated into city-wide zone to zone flows by time-of-day. Measures derived from these flows are analyzed over time to quantify the impacts of the covid-19 pandemic on travel patterns.

ABSTRACT. We developed an approach to estimate OD flows, including activities performed at destinations, using crowdsourced Google Popular Times (GPT) data and mobile spatial statistics on population presence. Our method is suggested to be of particular relevance to transit operators to understand the activities that public transport users engage in after their journey, enabling insights into demand sensitivities to urban activities. The study uses data from Kyoto, Japan with a focus on the Kyoto Station area. Results show that GPT data effectively estimate the time-varying population in the station vicinity. Further analysis illustrates the origin of station users and the activities they engage in.

ABSTRACT. This research introduces a comprehensive toolbox designed to evaluate the equitable distribution of spatial accessibility across a given territory. The toolbox is demonstrated through a case study comparing accessibility levels to public transit and urban opportunities for potentially vulnerable and non-vulnerable populations within two municipal sectors in the Montreal region. The tools leverage primarily open data, offer straightforward result interpretation, support multi-scale analyses, and enable scenario-based comparisons. These features position the toolbox as a valuable resource for territorial planning and public transit decision-making.

ABSTRACT. This paper proposes the Livability Index by Public Transport (LIPT), a novel accessibility indicator designed to assess quantitively the feasibility of completing trip chains using public transport. Alongside, we present LIPT-sim, a tool enabling users to compute several accessibily indicators using GTFS public transport data. A case study in the Nagoya region reveals that traditional coverage-based methods may overestimate accessibility in some rural regions, highlighting the need for trip chain–oriented assessments. Through these proposal and prototyping, this paper propose our direction to improve public transport assessment and designing to enable it surely support people’s daily activities.

ABSTRACT. This study examines intra-household ride-giving through the lens of Mobility of Care, using person trip survey data from Nagoya City. Findings reveal pronounced gender differences in mobility patterns: women make more trips, travel shorter distances, and bear a greater share of caregiving-related travel, especially ride-giving for children. These patterns reflect structural inequalities and invisible care labor embedded in daily transport. The study also highlights growing burdens associated with children’s extracurricular mobility in urban Japan. Addressing caregiving mobility is essential for ensuring equitable access to opportunities. Policy implications include the need for care-sensitive transport planning and inclusive mobility systems.

ABSTRACT. This study aims to construct and verify a quantitative evaluation method for providing opportunities for residents of the same town to meet, based on public transport services. We defined situations where multiple travellers are in the same space at the same time as 'Co-exist', and situations where specific pairs of residents repeatedly encounter each other in the same place as 'RCM' (repeated chance meetings). We developed a method to quantify these indicators using Multi-Agent Simulation (MAS). Based on individual data from a person-trip survey conducted in the target area, agents were generated to reproduce the intra-district travel patterns of around 18,000 residents. These agents were then simulated, enabling us to analyse the frequency and trends of Co-exits and RCMs. The results revealed that high-density Co-exits are distributed around community facilities and stations, with approximately 6,000 RCMs occurring daily across the entire target area. Additionally, the generation intensity of Co-exits and RCMs varies significantly depending on the travel mode; RCMs associated with bus and shared taxi travel were found to be significantly more common than those associated with walking or car travel.

ABSTRACT. Recent research on Bus Rapid Transit (BRT) systems has mostly focused on ridership forecasting and scheduled travel time gains, with little empirical evidence on potential operational improvements. This study examines the short-term impacts of implementing a new BRT corridor in Montreal, Canada, on key bus performance indicators: running time, running time deviation, and headway deviation. Using Automatic Vehicle Location (AVL) and Automated Passenger Count (APC) data from 2022 to 2023, we compare the performance of the BRT to a parallel local bus route operating along the same corridor, before and after the BRT implementation. Our findings indicate that the BRT significantly reduced trip durations (over five minutes on average) primarily due to infrastructure features such as dedicated lanes and all-door boarding policy. The local route experienced modest running time improvements post-BRT, suggesting potential corridor-wide benefits. However, both run time deviation and headway deviation were significantly higher for the BRT, particularly during peak periods. These findings highlight the importance of integrating infrastructure investments with dynamic operational strategies such as real-time dispatching and headway control. It emphasizes the need for schedule calibration following implementation to ensure that planned service aligns with actual performance. These findings offer practical insights for transit agencies planning or managing BRT systems.

ABSTRACT. Train headway and passenger flow are key factors affecting MRT station transportation efficiency and commercial benefits. This study proposes an integrated operational model that involves signalling system adjustments and pedestrian flow simulations that can enhance overall MRT services. Using the Taiwan Taoyuan International Airport MRT as a case study, an OpenTrack simulation was run to demonstrate how signalling system modifications can be applied to reduce headways along the entire line. In addition, BIM and pedestrian flow simulation tool PTV Viswalk were used to assess the impact of reduced headways at A19 station of the Airport MRT, and the results informed recommendations for allocating underutilized retail space within the station. Simulation outcomes show that the proposed model can improve both transportation efficiency and commercial benefits.

ABSTRACT. Autonomous Rail Rapid Transit (ART) has been adopted in several cities around the world. Moreover, with the ongoing expansion of China's high-speed rail (HSR) network, the number of newly constructed HSR stations is also increasing. To address the last-mile connection between existing ART lines and new HSR stations, this study proposes the establishment of demand-responsive HSR-ART branch lines. With the addition of the HSR branch, the ART network forms a Y-type structure. By adding a turnaround track at the node station and utilizing virtual coupling technology, the system can accommodate passenger travel demands in all directions—between trunk lines as well as between trunk and branch lines. To simultaneously meet passenger travel demand and minimize the operator's operating costs, this study proposes a MINLP model, which is then transformed into a MILP model to facilitate solving with Gurobi. Finally, the proposed model is applied to a real-world case in Yibin, China. The solution proposed in this study reduced the total cost by 9.0% and 33.7% compared to Scenario 2 and Scenario 3, respectively.

ABSTRACT. Bridging the last-mile service of the backbone/trunk lines such as mass transit system is indeed a challenging problem. This study proposes a flexible last-mile transit service using modular vehicles, which can can dynamically adjust their formation size by coupling or decoupling modules in response to real-time, spatially heterogeneous passenger demand along different routes and stops. We propose a model that explicitly considers the dynamic passenger arrivals and the optimal modular vehicle formations to accommodate the time and spatial demand variation. Specifically, passenger-to-route assignments, limited number of modules, and early return of modules at intermediate stops are all considered in the model formulation. To solve the dynamic operations of the modular-vehicle last-mile service, we propose a two-stage route-stop reinforcement learning-based approach, in which the Advantage Actor-Critic (A2C) algorithm and the optimization model are integrated into the framework. Specifically, the A2C based reinforcement learning approach is proposed to predict the initial dispatching decisions before departures on each service route. Then an optimization model is adopted to derive more detailed operation decisions for both the modules and the passengers. Numerical experiments are conducted to demonstrate that the proposed model can provide a high level of service and significantly reduce total costs.

ABSTRACT. Accurately estimating travel times is essential for developing reliable rail schedules and ensuring operational efficiency. Run time allowances—also known as time buffers, recovery margins, time supplements, or schedule padding—are added to minimum run times to absorb potential delays and support punctuality. This paper examines the wide range of methodologies used globally to integrate these allowances into train schedules.

Through a literature review, expert consultations, and interviews with U.S.-based rail agencies, the study identifies both simple and complex approaches, from fixed percentage-based allowances to geographically distributed models. Key findings include the absence of universal guidelines, inconsistent terminology, reliance on institutional knowledge, and significant impacts on both capital costs and operational behavior.

This paper is intended to assist rail practitioners by presenting not only academic research but also practical recommendations. It summarizes various methods with associated advantages and limitations, and provides a glossary to clarify terms and their roles in scheduling and capacity planning.

Given the locally dependent nature of current practices, the paper also aims to encourage further research and collaboration. Its broader goal is to promote the development of structured, adaptable guidelines for implementing run time allowances, leading to more reliable, cost-effective, and passenger-focused rail operations.

ABSTRACT. Public transport (PT) interventions such as metro openings can vastly enhance the accessibility within a city. However the impact is more nuanced to the connectivity and the resulting demand flows at the origin-destination (OD) level. We have developed a multilevel regression model that uses the longitudinal PT schedule and OD-level demand data for 11 years from Santiago de Chile to evaluate the impact of OD flows after the opening of 2 metros lines. The result shows a large variation of positive and negative effects to OD flows along and beyond the alignment of the new lines.

ABSTRACT. Light rail can flexibly share road infrastructure, making its capacity in mixed right-of-way conditions critical for system planning. Existing studies focus on dedicated lanes, overlooking shared lanes with highway traffic, where bottlenecks often occur at stations. This study develops models to calculate station capacity under three scenarios: no nearby signals, signals behind, and signals ahead. Using VISSIM, the models are validated with capacity prediction errors mostly within 10%. Combined with turnback operations, these models aid in designing efficient, reliable light rail systems by addressing capacity variations in mixed-traffic environments.

ABSTRACT. In response to the issue of mismatched travel demand and accessibility at urban transportation hubs, this study proposes a comprehensive analytical framework that takes into account travel demand, accessibility, and spatial patterns. Firstly, identification rules are introduced to detect grid cells imbalanced between travel demand and accessibility to transportation hubs. Subsequently, interpretable machine learning models are developed to analyze the relationships between travel demand with accessibility, built environment, and transportation facilities. Taking Shanghai's Hongqiao Transportation Hub as a case study, policy implications are provided based on the results of the analysis.

ABSTRACT. This study proposes a new method for the estimation of the supply level of bus services for urban areas. Italy’s framework for subsidy allocation includes provisions for: a) ensuring minimum service levels in all areas, as mandated by national mobility policies; b) funding allocation based on the computation of the standard costs depending on operational and urban characteristics. In the last years, the revision of funding allocation is started but subsidies are still distributed according historical data. While a standard costs computation model was provided by Ministry of Transport, the first process is still not developed and, according to regulations, should be defined for ensuring, in quality and quantity, the mobility needs of the population. For overcoming the existing limits, the proposed method is based on the identification of the effective transport needs as output of a Cobb-Douglas function. Such production function is widely used to model relationship between inputs and output. Input factors are the reference length of the public transport network and the average number of runs made by each line. The new method derives from the idea of reproducing the usual process of design of public transport services where the two factors computed can refer to the phase of route identification and frequency setting.

ABSTRACT. This study investigates train timetable scheduling for a metro line, utilizing virtual coupling and stop-skipping to address spatially and temporally imbalanced passenger demand. The problem is formulated as a mixed integer linear programming model, aiming to minimize train operational costs and total passenger travel time. Multiple operational strategies are integrated within a time-space network to accommodate time-dependent passenger demand patterns. A tailored branch-and-price algorithm is developed to solve the model for large-scale networks efficiently.

ABSTRACT. China is developing the train group operation technology to enhance the railway capacity. Under this technology, trains are tracked closely to present a flow state. In this research, the fundamental diagram model under train group operation is established based on the analysis of train operation. The group train flow characteristics under different influencing factors are studied and validated. The results show that the train type, line gradient, train group size and train order all have different effects on train flow. These findings can provide references for the transportation organization under train group operation.

ABSTRACT. Post-pandemic, the rapid and drastic changes in people’s mobility styles have left transit agencies grappling with financial stress. California’s transit ridership has generally tracked alongside national ridership trends in the US with a substantial dip in ridership and then slow recovery. However, post-pandemic mode shares in Northern California appear to be remaining lower than pre-pandemic shares, most noticeably on commuter rail services. To reverse some of the declines in commuter rail ridership, we are using survey research to better understand how rail markets have evolved since the pandemic and how the potential user base views rail services.

ABSTRACT. The rapid growth of transportation network companies (TNCs) over the past decades has accompanied several challenges, including competition with transit and increased congestion. To address these issues, some US cities started taxing TNC trips. It has successfully collected tax but is less efficient in tackling congestion and loss in transit ridership. Hence, there is a need to seek a pricing scheme that can position TNCs as a complementary mode rather than as a competitor to transit while minimizing their impact on congestion. This research explores zero-net pricing schemes that encourage TNC trips where transit is insufficient and discourage others. We use tax and subsidy that sum up to zero to nudge a modal shift with Chicago's peak hour data. We found that with a fixed criterion for pricing, there is a tax level that affects demand the most, and a generalized-cost-based pricing scheme has more impact on downtown trips. This indicates that a zero-net pricing scheme can effectively target TNC trips in congested areas with competitive transit for tax and subsidize those without good transit connections. Further, it diminishes the competition between TNC and transit and mitigates congestion.

ABSTRACT. Dockless bike-sharing systems have emerged as a promising transportation mode that is low-carbon, environment-friendly, and sustainable. The nature of one-way usage in bike-sharing systems inherently poses the issue of demand-supply incongruence over time and space. The spatial flexibility of parking and the uncertainty of travel demand in dockless systems can accelerate the incongruence, resulting in a significant decrease in user satisfaction with bike-sharing services. This study deals with the stochastic version of a user-based rebalancing problem in dockless bike-sharing systems that integrates strategic bike fleet sizing, operational dynamic pricing, and parking recommendations. The problem is formulated as a multi-stage stochastic convex programming model, which determines bike fleet size planning at first and then optimizes bike fares and parking recommendations sequentially and recursively while observing the spatial distribution of idle bikes and realizations of stochastic trip requests. The objective is to minimize the investment cost of bike fleets and maximize the revenue from dynamic pricing. An iterative algorithm based on a multi-stage Benders decomposition provides the optimal solution to the proposed model with guaranteed convergence. We conduct numerical experiments to demonstrate the Pareto efficiency and robustness of user-based rebalancing with parking recommendations.

ABSTRACT. The integration of rail and air services has been attracting increasing attention with the growing emphasis on multimodal transport systems. In this paper, we propose an air-rail timetable synchronization model to improve the passenger transfer experience in integrated air-rail transport networks. The model applies the time shift and rerouting strategy to existing rail and air timetables to provide more connections and smoother transfers for multimodal travelers. It also captures the passenger itinerary shifts resulting from timetable adjustments. The problem is formulated as a mixed-integer linear program. The effectiveness of the proposed method is demonstrated through a real-world case study of the Spanish rail and air network. The results show that the rerouting strategy can significantly reduce passenger transfer times.

ABSTRACT. This study investigates the relationship between ride-hailing services and public transit (PT) in a saturated market environment, using data from 35.94 million trips in Shanghai, September 2022. Our findings indicate nearly comparable ratios of complementary trips (9.22%) and substitute trips (9.06%), contrasting sharply with the findings of prior studies. The results show significant nonlinear effects in some variables, including the distance to the nearest metro station and the density of bus stops. These findings offer valuable insights for policymakers to promote urban multimodal mobility systems integrating ride-hailing and public transit.

ABSTRACT. In this paper, we explore a two-stage model with Γ-robustness for line planning in public transport systems, focusing on link blockages and re-routing of passengers in the second stage. We keep changes to the line concept small, running the line before and after closures with the same frequency if unblocked. Our network includes bypass links, allowing alternative modes and ensuring feasible solutions. We analyse the benefits of Γ-robustness and the addition of bypass links, highlighting their impact on the reliability of public transport systems.

ABSTRACT. This study tackles the Train Timetabling Adjustment Problem with integrated track assignments in response to planned maintenance. Building upon existing models that primarily operate at a macroscopic level, we propose a mesoscopic approach and extend the standard event-activity network, allowing for assigning tracks, modelling train short-turning and track capacity directly, and accounting for partial open track possessions which are often overlooked in previous studies. The problem is formulated as a mixed-integer linear programming model and tested on real-world instances with multiple possession scenarios. Results indicate that the model can find optimal alternative timetables and feasible train routings within reasonable times.

ABSTRACT. There are significant fluctuations in passenger railway demand throughout the day. Despite these fluctuations, many European countries use a fixed line plan and cyclic timetable that is the same throughout the day. Conversely, the multi-period railway timetable is designed to address fluctuating demand patterns throughout the day, while maintaining the memorability of cyclic schedules. This study evaluates how the railway demand would be impacted by implementing such a timetable. A case study conducted on a segment of the Dutch railway network, compares the passengers' Generalised Journey Time (GJT) in the multi-period timetable with their GJT in the cyclic reference timetable. Based on the change in GJT, passenger demand is then altered using incremental elasticity analysis with time elasticities. Our analysis of the case study shows improved average journey times and a slight increase in passenger demand, particularly during the off-peak period. However, during the morning peak the loss of direct connections and resulting increased journey times cause significant decreases in demand. The findings underscore the importance of determining an optimal line plan for each period and improving waiting times during transitions between cyclic schedules.

ABSTRACT. This study proposes a new approach to assess Demand-Responsive Transport. We developed the DRT Equilibrium (DRT EQ) within an agent-based simulation environment, an iterative procedure in which DRT routes are defined at each procedure iteration and treated as conventional bus lines, with the routes refined at each iteration based on equilibrium performance. To assess the performance of this methodology, we tested it against MATSim’s dynamic DRT module using a case study from Luxembourg. We show that under a stress test scenario, the DRT EQ performs better and more consistently than the DRT Module from both the customers' and provider's perspective.

ABSTRACT. Many rail operators are shifting from a traditional commuter rail model towards all-day, frequent bidirectional service. However, rail terminals configured for infrequent, peak-direction travel may not provide sufficient capacity to meet these new demands, leading to delays. This study investigates operational, tactical, and strategic solutions to reduce delays at terminals while increasing service frequencies, using Boston’s South Station as a case study and SUMO as a simulation engine. The analysis shows that minimization of train movements to and from the yard, reallocation of platforms to meet each line’s capacity demands, and a simplified interlocking yield a higher-performance terminal.

ABSTRACT. This study proposes an AI-driven generative framework to produce high-quality synthetic personal trip activity data for transportation behaviour analysis. Using Tokyo personal trip data, we apply Conditional Tabular GAN (CTGAN) combined with Long Short-Term Memory (LSTM) and a Triplet-based Variational Autoencoder (TVAE) with Autoencoder architecture to capture complex with-in-day activity patterns. To improve sequence modelling, we introduce a novel LSTM-GAN that integrates LSTM layers into generative models, enhancing the realism of synthetic Activity sequences. Generated datasets are evaluated against real-world data using statistical metrics such as KL divergence and likelihood, as well as predictive model performance. Results indicate that TVAE excels in data quality and generation speed, while CTGAN performs better in likelihood. LSTM-GAN demonstrates the best structural fidelity, accurately preserving temporal and categorical distributions of travel time, stay time, activity patterns, and transportation modes. Visualizations confirm the ability of LSTM-GAN to reflect real-world behaviour, highlighting its applicability in mobility studies. This research enhances data generation methodologies for transportation modelling, offering a cost-effective alternative to extensive survey-based collection.

ABSTRACT. This study develops a simulation model specifically tailored to the characteristics and precisely analyzes the capacity of the intermediate turn-back stations within high-speed rail systems. Additionally, this research conducts in-depth case studies to explore the most suitable track utilization in various scenarios. The study also examines how variations in train turn-back times and the ratio of train types affect capacity. The results not only enhance understanding of the operational efficiency and capacity management of intermediate turn-back stations but also provide effective evaluation tools for planning and capacity analysis of high-speed rail systems.

ABSTRACT. Municipality-led evacuation planning often fails to account for the significant impact of commuters and mass transport, even though both are central to urban functioning and everyday mobility. This study addresses that gap by integrating both commuter and local transport systems into an agent-based model (ABM) to simulate flood-related evacuation in Okazaki, Japan. The model captures the dynamics of both resident and commuter populations, reflecting statistically grounded travel patterns and public transport dependencies. By comparing scenarios with and without commuting populations within the target area, the results reveal significant differences in evacuation outcomes, particularly in the number of people trapped by flood inundation. Commuters, often excluded from local evacuation planning, are disproportionately affected, highlighting critical vulnerabilities. The simulation demonstrates how rail station closures due to flooding increase congestion and leave thousands unable to evacuate or return home. This case study underscores the need to account for transient populations and commuter transport systems in evacuation models, especially in workplace cluster regions. The findings suggest that integrating mass transport systems into ABMs provides valuable insights for municipality-led evacuation planning. This approach offers a scalable framework for enhancing urban evacuation strategies worldwide.

ABSTRACT. Abstract: This study explores the fare management challenges in Ulaanbaatar’s public transportation system, focusing on revenue losses caused by cash payments and fare evasion. In October 2024, we surveyed over 6,945 passengers across 38 routes and found that only 56% used smart cards, while others paid cash or evaded fares. We validated these findings with smart card data and simulated alternative fare policies to explore the implications for revenue. Our results show that promoting smart card adoption and aligning fares with travel distance can boost revenue. This work highlights practical strategies to enhance Ulaanbaatar’s transit efficiency and financial sustainability.

ABSTRACT. This research develops a novel machine-learning based transportation prediction model, integrating environmental variables as adjustable parameters to enable flexible scenario analysis in response to environmental changes. The model considers the network effect of bike docks, unmet demand hidden in ridership, and long-term environmental changes, using Temporal Fusion Transformers farmwork, addressing the limitations of traditional transportation prediction methods. Applied to the public bike system in Seoul, South Korea, the model showcases its ability to dynamically predict the potential ridership under various environmental scenarios. We also measured the difference in extrapolation ability between the transformer and linear models. Simulation results demonstrate how adjustments like new docks or e-bikes can improve user experience and bike usage, informing strategic level and operational level of decision making on public bike systems.

ABSTRACT. In this study, we combine a microscopic traffic simulation with the eFLIPS scheduling approach to analyze bus electrification in part of Kyoto's bus network. We compare two scenarios -- solely depot charging and terminus fast charging -- to estimate vehicle requirements, charger allocation, and peak power loads. The microscopic simulation provides detailed trip-level energy consumption, while eFLIPS optimally assigns trips and simulates charging events. Results show terminus charging demands extra vehicles but uses smaller onboard batteries, whereas depot charging lowers peak electric loads. The method yields rapid insight into operational and economic trade-offs and can adapt to larger networks or additional constraints.

ABSTRACT. Cooperative Connected Automated Mobility services are envisioned to increase the quality of passenger mobility. The design of such services and the evaluation of their impacts constitute an active field of research, where quantitative methodologies such as agent-based simulations are extensively deployed. Several contributions regarding various aspects of CCAM services were made using open-source frameworks, and today, the methodologies and tools have reached a level of maturity that allows them to be deployed at scale. In this paper, a decision-support tool that builds upon well-established methods is proposed to allow non-experts to use simulation models for the evaluation of CCAM solutions.

ABSTRACT. This study examines energy consumption differences in simulated electric bus operations using microsimulation networks and open data. Eight baseline scenarios incorporating varying vehicle and environmental parameters were extended with advanced traffic control strategies, including actuated signals and cycle adaptation based on Webster (1958). Results demonstrate that traffic conditions and driver behavior significantly impact energy use, with some scenarios showing reduced consumption due to enhanced driver abilities. The study highlights the value of integrating traffic calibration and trajectory validation for practitioners utilizing open data and cost-effective tracking experiments, providing insights for energy-efficient e-bus operations under diverse traffic conditions.

ABSTRACT. We propose the introduction of bus incentives to reduce long bus dwell times at stops near demand hotspots. We designed a microscopic traffic scenario for the Kyoto City Bus line 201 in the SUMO simulation. An intermodal routing of a novel agent was introduced to extract boarding and alighting times. Our results show that an appropriate bus incentive is effective in reducing bus dwell time. Additionally, this suggests that in-vehicle crowding is a determining factor for bus dwell time.

ABSTRACT. The adoption of electric buses is often limited by battery capacity and slow recharging. Accurate energy consumption estimation is therefore crucial. This research simulates the daily operations of several bus routes of Kyoto city bus using actual GPS data and the SUMO simulation, combined with the vehicle energy model (VEM). We analyze energy consumption under different scenarios, considering temperature and passenger load variations, and assess the impact of traffic stochasticity. The findings offer insights into bus fleet electrification and the optimization of charging strategies.

ABSTRACT. Building on advances in intelligent transportation, this study proposes a novel real-time control framework that integrates Graph Neural Networks (GNNs) with multi-agent reinforcement learning for distributed control of traffic signals and Connected Autonomous Vehicles (CAVs). The Heterogeneous Graph Reinforcement Learning (HGRL) framework facilitates dynamic and mutual interactions among agents, optimizing traffic efficiency, safety, and environmental sustainability. Comparative evaluations show that HGRL outperforms state-of-the-art approaches, reducing travel time by 21.9%, delays by 57.3%, time-to-collision by 85%, and both fuel consumption and CO2 emissions by 15.5%. Sensitivity analysis across different CAV penetration rates and GNN configurations further indicates HGRL’s robust performance.

ABSTRACT. In the context of carbon-neutral, the carbon generalized system of preferences (CGSP) policy is newly implemented to encourage passengers to shift from automobiles to public transport, in which the government obtains the excess income by trading travelers’carbon emission reduction (CER) and offers a corresponding subsidy to public transport enterprises, while enterprises provide fare discounts to attract more passengers. In this study, an evolutionary game model is established to simulate the behaviors of the government, public transport enterprises and heterogenous traveler groups, which is distinguished by the value of time. Evolutionary stability strategies in different scenarios are investigated with both numerical analysis and real-world based simulations. Results are expected to provide references to pre-evaluate the policy implementation.

ABSTRACT. Demand Responsive Transport (DRT) systems can struggle with user retention due to variability in ride-time experiences, even within guaranteed service levels. Under constant aggregate demand, the stochasticity of users’ daily travel decisions can alter the order of pick-ups and hence ride-time. Simulation results indicate that users can experience very high levels of ride-time volatility, regardless of their origin. This work investigates how the detour factor, as an individual level-of-service constraint, can limit experienced ride-time fluctuations and hence mitigate against users leaving the system.

ABSTRACT. To address the challenge of predicting individual-level behavior evolution in multimodal transit systems, we propose MBSE-Net, a novel end-to-end deep learning framework incorporating a Multi-view Attributed Graph Model (MAGM) to encode multimodal trip chains and estimate spatiotemporal similarities among riders. MBSE-Net synergistically performs behavior status identification and evolution prediction by integrating MAGMs with advanced deep learning techniques. This data-driven approach eliminates the need for manual feature selection, advancing the analysis of multimodal transit rider behavior. Our framework provides valuable insights for personalized interventions in intelligent public transit systems.

ABSTRACT. This study investigates the spatiotemporal relationship between land use patterns and transit ridership along Bangkok’s MRT Blue Line using stepwise and geographically weighted regression (GWR) analysis. Drawing on ridership data and analysing land use composition within 500-meter buffers, the study identifies residential, commercial-office, and transport utility land uses as the primary predictors of both weekday and weekend ridership. Additionally, malls and markets emerge as significant predictors during late evening weekday and afternoon weekend periods, reflecting non-commuting travel behaviour. The GWR results highlight substantial spatial heterogeneity, demonstrating how the influence of land use on ridership varies across time and location.

ABSTRACT. Mobility as a Service (MaaS) integrates multiple transport options into one platform to reduce car ownership and congestion. In July 2021, the ODIN PASS MaaS trial at the University of Queensland (UQ) gathered data on UQ staff and students' trip characteristics, reasons for using MaaS, subscriptions, and feedback. This study aims to identify user travel and subscription patterns via latent class analysis, and explore attitudes and improvement suggestions through thematic analysis. Findings show non-car users favour public transport bundles, while users with fewer trips prefer bundles with micro-mobility; feedback varies by user type, emphasizing fare policy and app functionality.

ABSTRACT. As megacities grow, urban transit must address increasing travel times by offering both efficiency and comfort. The few emerging premium options in intracity transit lack scientific discussion on the optimal fare level and structure. This study develops a pricing model for premium-class seating, integrating passenger segmentation and willingness-to-pay using a discrete choice and fare optimization framework. The model obtains profit-maximizing fares given crowding levels, value-of-time distributions, and in-vehicle productivity provided by the premium class. Results provide insights into fare structures to balance enhanced passenger experiences and operational profitability. This research offers guidance for implementing differentiated services in public transport systems.

ABSTRACT. With the popularity of emerging transportation such as e-hailing, the traditional public transit (PT) faces an intensive competition. Thus, how to regulate the competition between PT and e-hailing mode, and enhance the share rate of PT has become an important issue in urban transportation management. This study develops a bi-level programming model to explore the collaborative pricing problem between e-hailing and PT in a multimodal urban transportation network. The upper-level model maximizes e-hailing service revenue through the optimal flag-down fare. Constraints are targeted at the PT mode share rate and the initial cost balance. The initial travel cost refers to the difference between the flag-down fare of e-hailing and the PT fare within the same OD pair. The lower-level model characterizes the mode choice behavior of choice travellers and loyal travellers by a dogit-nested logit (DNL) model. Results show that setting an initial cost balance increases the PT mode share rate and facilitates low-carbon mobility. Also, a suitable collaborative pricing strategy can balance PT share rate and e-hailing profitability under different levels of PT service.

ABSTRACT. Previous work proposed “premium fare tickets” that offer passengers free access to alternative modes if public transport services are delayed beyond a set threshold, enhancing travel time reliability. A problem with the concept is the need to guarantee the availability of sufficient alternative transport options in case of a delay. In this work we address this by providing alternative mode companies compensations for deploying more vehicles to a station then the normal expected demand might justify. The model is applied to a railway line network considering multiple origin stations. In the resulting objective function we consider a social cost perspective. The results show that even considering incentives, the premium fare can reduce the expected travel cost and still be beneficial for the PT operator.

ABSTRACT. In order to lessen passengers’ burden to select and prepay the correct zonal fare, to reduce the reliance of fare inspection and to open up other fare options, the Montreal Transit Authority (ARTM) is examining the implementation of a partial “tap-in tap-out” scheme to replace the current “tap-in” only practice. This paper describes the datasets and the methodology used to evaluate the potential impacts on flow, user experience and revenue. Passive smart card fare validation data are the main inputs in the modelling process. The research aims 1) to estimate the passenger flow at the exit faregates, 2) to characterize passenger flow according to time of the day, fare types, previous and subsequent journey leg, and other relevant variables, 3) to perform sensitivity analyses on contributing factors, such as the number of exit faregates and the impact of concurrent entry-exit dynamics, and 4) to model the impacts on fare revenue when the exit function is deactivated. This paper incorporates several past advancements of smart card data research, namely in data processing, fusion and enrichment. The results, though specific to Montréal, demonstrate a practice-ready application of passive data in a real-world environment.

ABSTRACT. In developing cities, rail transit systems are planned based on transportation masterplans, often lacking detailed analyses of accessibility provided by public transportation. This study assessed the impact of railway network expansion using cumulative accessibility index. The results revealed 10% increase in citywide accessibility, with 50% of facilities within 20 km of the city center reachable within 60 minutes due to the expanded rail transit network. While rail development significantly enhances accessibility, some areas may continue to experience low service levels. These findings underscore the need for detailed evaluations to guide future rail transit planning and ensure equitable regional access.

ABSTRACT. This study aims to apply XGBoost and Random Forest models to explore the relationship between safety-related road environment factors and Advanced driver assistance systems (ADAS) multi-warning events based on data from Taipei City buses. It is shown that XGBoost model has comparatively better performance in terms of revealing key factors that significantly affect forward collision warnings (FCW), lane departures, and speeding incidents. Key findings indicate that FCW events are strongly associated with large vehicle ratios, lane numbers, and occupancy rates, while lane departure warnings are linked to traffic conditions and geometric designs. Speeding risks, however, are primarily influenced by arcade-no-motorcycle road segments and bus lanes. These insights will help of formulating strategies for road safety improvements. It is also recommended that optimizing road designs, re-assessing heavy vehicle management schemes, and revising bus lane speed limits or bus lane layouts are worth further studies.

ABSTRACT. Many cities have introduced free or low-cost public transit programs to increase ridership and promote equity. However, few studies have examined how these programs operate in Asia. This study examines Kaohsiung’s Mobility-as-a-Service (MAAS) program, MeNGo, focusing on its impact on bus ridership and travel behavior. First, we applied clustering methods to bus data from November 2023, identifying five distinct user groups. Most program users are students; their trips occur primarily on weekdays; and travel is largely concentrated in the city center. Although the MeNGo program improves transit access, challenges such as overcrowding and limited coverage of intercity routes remain. In a second analysis, a flow clustering approach revealed especially high demand in city-center areas and at major transit hubs, highlighting the need for targeted service improvements. These findings offer insights for designing more effective and inclusive transit systems.

ABSTRACT. This study investigates individuals’ alternative acceptance in tour-based travel behavior in integrated passive and active public transit data. We build the data by integrating National Household Travel Survey data in Korea and four types of passive data (navigation, smartcard, bike-sharing, and GIS data). By developing a tour-based truncated choice model, we assume that individuals do not consider unattractive alternatives (those exceeding a certain threshold) in their choice set. We empirically demonstrate that assessing alternative acceptance is significant for describing tour-based travel behavior and increases the value of travel time estimates for subway, bus, and taxi.

ABSTRACT. It is important for bus operators to understand the number of passengers on their buses in order to create operation plans, such as the number of buses and routes. This study uses Wi-Fi packet sensors on small buses to estimate passenger volume by detecting Wi-Fi signals from smartphones. It distinguishes passenger signals from non-passenger signals, treating the latter as background noise. The background noise is pre-estimated through regression analysis, using route characteristics like road width, land use, and the number of buildings as dependent variables. Two estimation methods, linear and non-linear regression were tried and compared. The regression equation method estimating the amount of background noise did not have a high coefficient of determination. It will be necessary to increase the number of independent variables for estimating background noise to improve the goodness of fit.

ABSTRACT. Despite its practical potential, the current adoption of Digital Twins (DT) in the transportation domain and especially in Public Transport (PT) is relatively slow. A significant barrier is the substantial effort and investment in resources required during the development phase, especially for informative visualizations, thus limiting its accessibility to PT agencies. The study proposes an automated development pipeline of DT for PT, which uses Open Source software and data that make it easy to access and extend. We demonstrate the functionality of the pipeline using a scenario in Kista, Stockholm and discuss the potentials and limitations of DT for practical use cases from a conceptual perspective.

ABSTRACT. This paper examines the impact of real-time information apps on passenger satisfaction and their interaction with actual public transport (PT) service quality in 13 Chilean cities. We introduce two methodological innovations. First, we combine a discrete choice model with sentiment analysis using a Large Language Model (ChatGPT-3.5-turbo) to classify responses to an open-ended satisfaction question, integrating qualitative insights into a quantitative framework. Second, we incorporate an objective service quality measure—a headway reliability index based on bus GPS data—improving upon traditional models that rely solely on perceived service attributes. The model captures how satisfaction is shaped by service attributes, app-induced behavior, and perception shifts. Findings reveal a synergistic effect: real-time information enhances satisfaction, particularly under high-quality service conditions. Sentiment analysis shows that 70% of satisfied users value wait time information and time management enabled by the app, while 12% of dissatisfied users criticize low bus frequency and data inaccuracies. These results underscore the importance of real-time information in improving user experience and support the case for user-centered policies. Expanding access to reliable real-time data is especially critical in developing countries, where it can significantly boost satisfaction with PT systems.

ABSTRACT. As extreme weather events become more frequent, public transport users face growing challenges due to unreliable services and changing environmental conditions. This study examines the impact of weather conditions on bus ridership in the West Midlands, United Kingdom, using six years (2016-2022) of transit smart card data and hourly weather observations. Through a matched-pair analysis, we examine the relationships between key weather variables (e.g. precipitation and temperature) and bus ridership across different passenger groups, including commuters, students, children, elderly and disabled individuals. To better capture outdoor thermal conditions, we also incorporate thermal comfort indices that combine temperature with other meteorological factors. The findings reveal that the elderly and disabled passengers are the most sensitive to weather variations, primarily due to their greater flexibility in travel scheduling and higher vulnerability to extreme conditions. Precipitation exerted a strong negative impact on ridership. A warm environment also influences travel behaviour, particularly among the elderly and disabled, with thermal comfort indices providing a better measure of these impacts than temperature alone. These findings contribute to a more comprehensive understanding of weather-related impacts on urban bus usage and provide valuable insights for developing more inclusive and weather-resilient transport systems.

ABSTRACT. This study proposes a methodology to generate GTFS (General Transit Feed Specification) data reflecting actual driving patterns using smart card data and bus operation information. GTFS is a data standard that provides public transportation service information in a standardized format, but existing GTFS reflects only the planned schedule and differs from the actual driving situation. In this study, a realistic GTFS data generation process based on smart card data was developed and applied to bus routes in Seoul.

ABSTRACT. Mobile applications have become cost-effective tools to obtain both long-term crowd sourced trajectory data for travel behaviour and app interaction logs for transit information inquiry patterns. This study used long-term data collected from 445 Puget Sound (Washington State, US) residents through the “OneBusAway” transit application. The strength of the data is the ability to obtain travel records continuously and that participating users provide data over long-time. The focus of this study is on the captured inquiry information regarding bus arrivals at stops. The study discusses users’ behaviour in inquiring real-time bus information and then shows the relationship between app usage and actual travel behaviour. Regular transit users appear to inquire as much as infrequent ones.

ABSTRACT. We present a novel approach to clustering run time anomalies using the Density-Based Spatial Clustering of Applications with Noise (DBSCAN) algorithm. Our methodology applies the DBSCAN machine-learning clustering algorithm to detect, group, and classify anomalies based on spatial and temporal characteristics. Our study focuses on real-time vehicle location data collected by the transit operator in Geneva, Switzerland. The objective is to enhance the efficiency of anomaly investigation by categorizing anomalies into three types: isolated, periodic, and continuous. Isolated anomalies occur sporadically over very short time intervals. Continuous anomalies, on the other hand, are persistent and ongoing, often indicating a systemic issue that requires immediate attention. Periodic anomalies are characterized by their recurring nature at regular intervals. The results of our study demonstrated the model’s efficacy, successfully identifying all clusters of anomalies previously recognized by planners in Geneva. By forming clusters based on similar spatial and temporal characteristics, our approach enables targeted investigations. For instance, on-site verifications of recurring disruptions can be conducted more efficiently, thereby improving the overall reliability of public transit systems.

ABSTRACT. Abstract: Crowding in public transit impacts system efficiency and reliability. This study introduces a novel Crowding Contribution Index (CCI) to quantify how destination stations contribute to link-level congestion using Automated Fare Collection (AFC) data collected from Delhi Metro Rail Corporation, India. A linear mixed-effects model was developed to assess the influence of built environment factors, including Point of Interest (POI) entropy, POI density, and relative wealth variables. Results indicate that higher land-use diversity reduces a station’s contribution to link-level crowding, while higher POI density increases it. Findings emphasize the need for decentralized development strategies and transit-oriented policies to alleviate the lopsided contribution of a few stations to link-level crowding in metro systems.

ABSTRACT. While graphical train schedules are commonly used in the public transport research community, there seems to be little choice for open and free software to draw these diagrams from timetable data. This poster aims to discuss an effort to address this. By employing a modular and extensive design, open source software could be used to experiment with novel ways to visualize public transport data easily and conveniently.