Choice Modeling in Urban Mobility Systems

Choice modelling seeks to represent how urban travellers select between alternatives—such as walking, cycling, private car use, public transport or shared services—by quantifying the factors that influence these decisions. These models combine socio-demographic attributes, level-of-service measures (for example travel time, cost, waiting time) and psychological or contextual variables (such as habit, risk perception and service variability) to estimate the utility associated with each option. Approaches range from discrete choice frameworks—rooted in random utility theory—to memory-based and machine-learning variants that capture individual learning and habit formation over time. Advances in data collection, such as smartphone-based activity tracking and digital demand data, have expanded the empirical foundation for choice modelling, enabling richer representations of heterogeneity and dynamics. Practical applications include demand forecasting under new mobility services, testing congestion-pricing regimes, designing incentives for mode shift and evaluating the equity impacts of emerging transport innovations. Through integration with traffic simulation and land-use models, choice modelling in urban mobility systems supports both strategic planning and real-time operational optimisation, fostering more sustainable and responsive urban transport networks.

Research from Nature Portfolio

No recent Nature Portfolio content available.

Choice Modeling in Urban Mobility Systems publication trend

The graph below shows the total number of articles in choice modeling in urban mobility systems across all publications each year (not limited to Nature Index journals).

Technical terms

Discrete choice model: A statistical framework that represents decision making by assigning a utility to each alternative and predicting choice probabilities based on comparative utility. Mixed logit model: A flexible form of discrete choice model that allows taste parameters to vary randomly across individuals, capturing unobserved heterogeneity. Instance-based learning: A cognitive modelling approach where past observed instances are stored in memory and retrieved to evaluate and predict future choices, with memory decay and similarity measures. Panel effects: The influence of repeated observations from the same individual over time, which can reveal intrapersonal variation in behaviour. Inertia: The behavioural tendency to repeat previous choices independent of current attributes, reflecting habit or status quo bias.

References

1. An instance-based learning approach for evaluating the perception of ride-hailing waiting time variability. Travel Behaviour and Society (2023).
2. The influence of panel effects and inertia on travel cost elasticities for car use and public transport. Transportation (2021).
3. Modeling Departure Time Choice of Car Commuters in Dhaka, Bangladesh. Transportation Research Record Journal of the Transportation Research Board (2021).