Uncertainty can be found at every stage of travel demand model, where passed from each stage to another and propagated over the whole model. Thus, studying the uncertainty in the last stage (traffic assignment) is more important because it represents the result of uncertainty in the travel demand model.
The purpose of this paper is to assist transportation modelers and decision makers, to have a fresh look at the uncertainty in traffic assignments of transportation models. By building a new methodology to predict the likelihood of traffic assignment probability distribution and compare predicted values to real values or another prediction methods, the paper shows the uncertainty in traffic volumes, and the amounts of errors and biases in the results as well.
The methodology quantifies the uncertainty in modeling by Monte Carlo simulation. A probability distribution is assigned to all cells of the OD matrix, considering them as stochastic input variables. The distributions of the output values of traffic assignment are classified and into four cases according to errors and bias. Finally, the results are drawn into figures to visual the uncertainty in traffic assignments. The paper constructs three types of probability distributions to the input data. For each type of distributions different parameter assignments, such as different variation values; are analyzed. For each of these parameter assignments, one thousand Monte Carlo samples were made, with the classification and visualization of the results.

Uncertainty quantification, Uncertainty visualization, Monte Carlo simulation, Traffic assignment.