Traffic assignment with stochastic flows and the estimation of travel time reliability

The adaptation of equilibrium traffic assignment models for the prediction of travel time variability is considered from the standpoint of pre- and post-analysis approaches. A modelling procedure is presented which simultaneously estimates mean performance and reliability in a consistent framework, which relates travel time/cost variability to flow variability, which makes maximum use of network data typically available, and which possesses a computationally-efficient solution algorithm. Two test examples are given of the model which could be used to evaluate hypothetical policies on a risk basis, rather than on their performance under 'average conditions'.