Modeling Reductions in Freeway Capacity due to Incidents in Microscopic Simulation Models

Modeling the impacts of incidents on traffic flow operations is essential for the use of simulation programs to evaluate incident management alternatives. In particular, these programs must be able to model correctly the reductions in highway capacity that are due to incidents and the lane-changing behaviors of drivers ahead of incident locations. Although simulation models have been used for evaluating incident impacts, the abilities of these programs to model incidents have not been adequately assessed. Here the modeling of incidents in microscopic simulation models and the effects of calibration parameters on the simulated reductions in capacity due to incidents are examined. A basic freeway segment was simulated with three widely used microscopic simulation models: CorSim, VisSim, and AIMSUN. The three investigated simulation models allow the analyst to simulate incident blockages either explicitly or by using other events that have similar impacts on traffic operations at the incident locations. Calibration parameters of the three models were varied to determine if it is possible to calibrate the models to achieve target link capacity values for conditions both with and without incidents. The target capacity values used were those presented in the Highway Capacity Manual 2000, although any other target capacities could have been used. For all three models, there is a need to calibrate model parameters to produce acceptable reductions in capacity due to incidents. It is shown that this objective is possible with all investigated models. However, there is a need to introduce incident-specific time-variant calibration parameters in AIMSUN and VisSim. Recommendations are given regarding required improvements to the simulation programs to enhance their ability to model the reduction in capacity and the lane-changing behaviors at incident locations.