Calibration and validation of micro-simulation models of medium-size networks

Traffic micro-simulation models are widely used to quantify the benefits and limitations of traffic control options. The usefulness of these models in making design and traffic control decisions will mainly depend on their accuracy and reliability. Accordingly, thorough and careful efforts are required in the development, calibration, and validation of these models. This paper describes a procedure for the calibration and validation of medium-sized simulation network models. Most calibration efforts of small scale networks in the existing literature focus only on driver behaviour and lane changing parameters. On the other hand, calibration of medium and large scale networks was directed to OD matrix estimation with little attention given to driver behaviour parameters calibration. The procedure used in this paper provides a comprehensive calibration of the micro-simulation model in a sequential manner. A case study was applied to a VISSIM model of downtown Vancouver, British Columbia, which is an urban condensed grid network of more than 100 signalized intersections. The calibration procedure included OD matrix estimation using recent traffic volumes, route choice calibration by manipulating link surcharges, and driver behaviour parameters calibration using an experimental design and multiple runs. Real-life travel time data were collected for model calibration and validation. The observed travel times were found to reasonably match simulation travel times of the calibrated model. Therefore, it was concluded that the model was calibrated and validated with an acceptable level of accuracy.