A new model for disaggregate traffic assignment making explicit the spatial distribution of trip extremities
Résumé
Traffic demand modelling relies on the partitioning of the region studied into smaller Transport Analysis Zones (TAZ) inside which data are aggregated. The spatial distriubtion of trip extremities inside these zones influence greatly traffic assignment. To account for this effect, we model network access and egress times as random variables, instead of deterministic quantities as would be the case if centroids were used. Logit and probit models are compared. The probit model gives the most satisfactory results because of the wide range of covariance structure allowed. The new model is then tested on a portion of the Paris region, yielding results that vary significantly from those obtained with the standard procedure.