Increasing demands on the road networks have resulted in the requirement to seek alternative cost-effective solutions, apart from increasing road capacity, in order to improve network performance. Traffic simulation models enable the Engineer to analyse a number of proposed alternatives and choose the best one.
This thesis is concerned with validating the performance of the Brisbane western commuting corridor using Paramics. Reaction time and headway were the driver-related parameters that were identified as having the greatest effect on network performance. Traffic flows were simulated and an analysis of travel times showed that the model best replicated reality for combinations of reaction time of 1.0 second with a headway of 2.5 seconds and reaction time of 2.0 seconds with a headway of 1.5 seconds.
A statistical analysis showed that the modelled traffic flow rates were not similar to actual flow rates for a reaction time of 1.0 second and headway of 2.5 seconds. A resultant p-value of 0.0 meant that the mean flow rates are different and the distributions of the data are different. The analysis also showed that the simulated traffic flow rate correlated with actual flow rate for a reaction time of 2.0 seconds with a headway of 1.5 seconds. A two-sample t-test on flows showed that the model poorly replicated flows down Coronation Drive. Roadworks on the inbound lanes of Coronation Drive is believed to be the source of the discrepancy between the simulated flows and actual flows. Simulated traffic flows agreed with actual flows down Milton Road for a 2.0 second reaction time and 1.5 second headway. The resultant p-value of 0.51 proved that the means and data spreads are from the same distribution.
A sensitivity analysis on reaction time and headway showed that a ±1.0 second difference to the calibrated reaction time results in approximately a ±30 percent difference in the travel time. A 1.0 second increase in headway results in an increase to the travel time of 40 percent and a 1.0 second decrease yields a 16 percent decrease in the travel time.