Demand-responsive transit (DRT) has emerged as a flexible mobility solution for addressing service blind spots in areas underserved by conventional public transportation. However, empirical research on DRT demand estimation remains limited, especially in the context of regional heterogeneity and zero-inflated demand. This study proposes a data-driven DRT demand estimation framework using operational records from two contrasting regions in Incheon, South Korea. Yeongjongdo is a tourism- and airport-oriented area with a high floating population and strong temporal variability, while Geomdan New Town is a residential district with relatively stable travel patterns. A grid-based origin–destination (O–D) modeling structure was employed, incorporating spatial variables such as land use, population dynamics, facility distribution, and public transport accessibility. Multiple region-specific machine learning models were developed and evaluated to estimate planning-oriented mean daily demand for each O–D-hour. A comparative analysis of alternative model configurations showed that the appropriate model structure varied according to regional demand conditions. In Yeongjongdo, the proposed two-stage model, which combines demand-occurrence classification and conditional regression, achieved the best overall performance, with a test MAE of 0.0022 and RMSE of 0.0105. These values represented reductions of 37.1% and 42.0%, respectively, relative to the strongest direct regression benchmarks. In contrast, direct regression was more effective in Geomdan New Town, achieving a test MAE of 0.0071 and RMSE of 0.0190. These results indicate that the relative performance of direct regression and two-stage prediction may vary across regional demand contexts.
This paper evaluates how a semi-flexible demand-responsive transport (DRT) design, combined with fixed-route school bus optimization, can improve high-school commute accessibility in low-density and underserved areas of El Paso, Texas, USA.
The analysis integrated Geographic Information Systems, the Google Maps Directions Application Programming Interface, General Transit Feed Specification datasets, school attendance zones, and school bus route information. The analysis first assessed existing public transit and school bus accessibility, then identified low-demand bus stops, formed DRT service zones, tested stop consolidation rules, and restructured affected fixed routes.
The results indicate that a school commute-oriented, semi-flexible DRT concept can reduce travel time for low-demand areas when paired with fixed-route restructuring. Five DRT zones were identified. The largest modeled benefits occurred where low-demand stops were spatially clustered and could be served by smaller vehicles without retaining long fixed-route deviations. The analysis also shows that some benefits arise from route optimization rather than demand responsiveness alone.
The analysis is limited to planning-level modeling for the school district in El Paso and relies on available route, stop, and travel-time data. It does not include primary interviews with students, school staff, or transit operators; observed DRT ridership; or a full cost-benefit assessment. These issues are identified as priorities for pre-implementation testing and future research.
The study contributes to the DRT literature by examining a relatively underexplored application: semi-flexible DRT service design for high-school commutes in an underserved, low-density border-city setting. It evaluates how predefined low-demand zones, smaller vehicles, and fixed-route restructuring can be assessed collectively without treating the intervention as a fully dynamic DRT service.
Abdy Eunice Cervantes, Adeeba A. Raheem, R. Cheu et al.· Journal of Transport and Sus...· 0 citations
Public transport equity is usually assessed after a network has been designed rather than treated as a requirement that shapes the design itself. This study develops an equity-oriented optimization framework for hierarchical multimodal transit networks in which a fixed rail backbone, main-bus routes, and feeder-bus routes are designed jointly. The model minimizes an integrated objective combining total system cost and the modal accessibility gap between public transport and private cars, while spatial equity is imposed as a binding constraint through two alternative Gini-based standards: a demand-proportional (horizontal) index and a need-sensitive (vertical) index. The problem is solved by a genetic algorithm that embeds a strategy-based passenger assignment with crowding effects and tracks the best-so-far solution across generations. Experiments on Mandl’s benchmark network across eight scenarios, combining the two equity standards with four threshold levels, yield average travel times of 11.8–13.9 min. The two standards produce structurally different networks, and stricter equity does not necessarily degrade performance: the strictest need-sensitive scenario attains the lowest average travel time, although it also records the largest modal gap relative to cars. The framework thus supports scenario-based comparison, in which the equity formulation and threshold serve as explicit policy levers rather than fixed technical bounds.
Reliable Origin–Destination (OD) demand information is fundamental to transport planning, fleet scheduling, infrastructure investment, and policy formulation. However, in many developing cities, particularly those dominated by informal paratransit systems, comprehensive passenger-movement data are often unavailable, making conventional demand estimation difficult. This study developed and validated an integrated Gravity–Poisson framework for the synthetic estimation of passenger demand along the Nairobi CBD–Ongata Rongai–Kiserian public transport corridor, a major commuter route characterised by severe data limitations and highly directional travel patterns.
The framework combines a doubly constrained gravity model for OD trip distribution with a Poisson stochastic process for modelling passenger arrivals. Spatial interactions were estimated using demographic and geospatial proxy variables derived from population distributions and network distances, while model calibration was achieved through parameter optimisation and the Iterative Proportional Fitting Procedure (IPFP). Theoretical properties of the framework, including the existence and uniqueness of balanced OD solutions, positivity of passenger flows, trip conservation, convergence of the balancing algorithm, distance elasticity, and parameter identifiability, were formally established. Model performance was evaluated using mean absolute error (MAE), root mean square error (RMSE), sensitivity analysis, and benchmarking against XGBoost and Long Short-Term Memory (LSTM) machine-learning models.
The results revealed a highly concentrated morning commuter flow towards the Nairobi CBD, with approximately 97% of corridor demand converging at the Railways terminal. Calibration yielded a distance-decay parameter of β = 0.1000, indicating relatively weak sensitivity of commuter demand to travel distance. The balanced OD matrix satisfied all theoretical conservation and positivity conditions, while the Poisson arrival model indicated that rainfall, holidays, demonstrations, and network disruptions reduced passenger arrivals. Model calibration substantially improved predictive accuracy, and the sensitivity analysis supported the robustness of the framework under varying operational conditions.
The study demonstrates that reliable and operationally meaningful transport-demand information can be generated even in environments where conventional OD survey data are unavailable. Beyond providing a practical decision-support tool for public transport planning and fleet allocation, the framework contributes to the theoretical foundations of synthetic demand estimation by integrating spatial interaction theory and stochastic arrival modelling within a mathematically rigorous and computationally validated framework. The approach offers a scalable methodology for transport-demand analysis and mobility planning in rapidly urbanising, data-constrained regions.
G. Mocheche· Asian Research Journal of Ma...· 0 citations
This study introduces a systematic redesign of trip-based travel demand models to explicitly incorporate connected automated vehicles (CAVs). The framework models auto ownership of CAVs and human-driven households separately, accounts for zero-occupancy vehicle (ZOV) trips, and modifies trip distribution, mode choice, and temporal patterns to reflect anticipated behavioral and operational changes because of CAV presence. The framework is then applied to the Triangle Region of North Carolina, using Triangle Regional Model Generation 2, and reveals that CAV adoption increases vehicle miles traveled (VMT) and encourages longer trips, with average discretionary and work trips rising by 28%–31% under a 70% CAV adoption rate. Although VMT increases, effective capacity gains reduce total network delay by as much as 60% relative to the 2050 baseline. At the facility level, demand-to-capacity ratios decrease, indicating that some roadway expansion projects could potentially be deferred. Sensitivity analysis reveals that system performance is highly dependent on realized capacity improvements and cautious assumptions, producing nearly 40% more delay than the expected scenario. The proposed redesign framework provides transportation agencies with a scalable and practical approach to incorporating CAVs into long-range planning, project prioritization, and investment decision-making.
Fahim Kafashan, Si Shi, Leta Huntsinger et al.· Transportation Research Reco...· 0 citations
This study proposes a dynamic spatial allocation framework for estimating future non-industrial CO
2
emissions at the municipal level. Previous studies have largely relied on static allocation approaches that apply present-day spatial distributions based on land-cover, population, or nighttime-light data to future periods without accounting for changes in regional activity patterns. To address this limitation, this study developed dynamic spatial allocation factors by linking purpose-based Origin–Destination (OD) trip forecasts with gross floor area (GFA) data. Building-use-specific trip indicators were developed and used to estimate future changes in GFA for residential, commercial, and public sectors. The estimated future GFA was then used to derive dynamic regional allocation factors for spatially distributing national-scale emission projections. The results showed that changes in GFA exhibited strong correlations with changes in building energy consumption (r = 0.577–0.772), substantially outperforming land-cover-based indicators (r = 0.201–0.279). The calibrated GFA estimation model reproduced the observed GFA with mean absolute percentage errors (MAPE) ranging from 8.31 to 13.47% and coefficients of determination (R
2
) ranging from 0.978 to 0.996. Independent framework validation using municipal greenhouse gas inventories demonstrated that GFA-based allocation factors effectively represented the observed spatial distribution of emissions. The proposed framework dynamically updates regional allocation factors using future travel demand forecasts, enabling future changes in socioeconomic conditions and land-use patterns to be reflected in emission spatialization. The framework can be used to transform national-scale emission projections into high-resolution regional emission inventories and provides a practical approach for future climate policy assessment and regional carbon-neutrality planning.
Hyejung Hu, Youjung Jang, Seungyeop Kang et al.· Scientific Reports· 0 citations
Travel demand forecasting is essential for evaluating large-scale infrastructure projects, yet the traditional sequential four-step process can produce inconsistencies across trip distribution, mode choice, and traffic assignment. Although combined models address these inconsistencies, their practical use has been limited by simplified behavioral assumptions, computational burden, and the lack of a unified parameter-estimation framework. We propose a convex programming approach that integrates destination, mode, and route choices within a hierarchical extended logit model. The optimal primal solution characterizes the joint travel-demand equilibrium, while the optimal dual variables recover taste coefficients, alternative-specific constants, and destination- and mode-level scale parameters. The model captures mode correlations through nested logit and route overlap through path-size logit. Its convex structure provides global optimality guarantees and enables efficient solution using off-the-shelf conic solvers. Observed travel patterns are incorporated through moment and conditional-entropy constraints, while the route-level dispersion parameter is calibrated separately using partial link counts. Numerical experiments on eight benchmark networks demonstrate the scalability and computational efficiency of the formulation.
Youngseo Kim, Gioele Zardini, S. Samaranayake et al.· 6 citations· ⚡1
We use cookies to run the site and, with your consent, for analytics and to show ads.
See our Cookie Policy.
