ValidationTargets

Two sources are considered for validation standards for this project.

The recently updated Travel Model Validation and Reasonableness Checking Manual# provides guidance on validating all aspects of travel models, but is clear to recommend a balanced approach between matching observed data and staying true to the theoretical foundations of the model. This manual provides a range of example validation guidelines (but no specific standards). The example guideline selected for this project is the Ohio RMSE Curve, which offers a target percent root mean squared error by volume group.

The second guideline is from the Caltrans Travel Forecasting Guidelines#. These guidelines include a standard that is based on the maximum desirable deviation for individual link counts. They recommend that at least 75% of count locations on freeways and principal arterials fall within the maximum desirable deviation.

In addition, given the importance of convergence to a DTA model, a self-imposed standard is adopted for the level of convergence. The target convergence is less than a mean of a 5% critical gap across all time periods, and more importantly a gap that appears stable upon visual inspection.

Travel times are evaluated and additional visual inspections are completed as part of the validation, but no specific standards are adopted for those evaluations.