The issue of congestion on road networks is becoming more and more apparent. The need for effective and cost-oriented solutions to improve roadways has reached a state of high importance. Traffic simulation models permit engineers to analyse possible solutions to traffic problems before the construction phase, allowing an optimal solution to be chosen.
This document aimed to validate the performance of Paramics in portraying the Brisbane western commuting corridor road network. Critical parameters, such as headway and reaction time were identified and the model’s performance, when these parameters were varied, was investigated.
When considering the total network travel times for Coronation Drive and Milton Road, two combinations of reaction time and headway were found, which best represented real field data, collected by Abacus Surveys. The optimal combinations were a reaction time of 2.0 seconds with headway of 1.5 seconds and a reaction time of 1.0 second with a corresponding headway of 2.5 seconds.
During the calibration process, the network was discovered to be extremely sensitive to changes in reaction time and headway. Choosing a slightly different arrangement of reaction time and headway greatly varied results. A reaction time of 2.5 seconds and headway of 1.5 seconds would produce a 22% aberrance from the field data.
In order to choose one final set of parameters a statistical analysis was undertaken. The statistical analysis showed the former combinations (i.e. a reaction time of 2 seconds and a headway of 1.5 seconds), result in the best replication of field conditions. The resulting p-value for the best combination was 0.506. The results of the t-test indicated a disparity between the simulated data and field data collected for Coronation Drive. This was attributed to the ongoing road works along the inbound stretch of Coronation Drive.
To remedy these results, field data for Coronation Drive would be regathered during a period free of road-works in order to provide more valid data for the validation of this model.