Given limited room for capacity expansion, transportation network investments need to be evaluated thoroughly. Traditional planning models, while indispensable for demand forecasting purposes, cannot capture the necessary level of traffic dynamics detail for effective supply side analysis. This can be overcome by using microscopic traffic simulation; however, utilizing microsimulation for planning purposes is quite a challenging problem. In this paper, we present the California SR41 Corridor Simulation as a case study of how to utilize microscopic traffic simulation for planning purposes. This paper will emphasize two of the most important components of preparing simulation models for planning purposes; namely, origin-destination matrix calibration and peak spreading for long-term testing (e.g., 20-year horizons) to overcome unrealistic network grid-lock. By streamlining the OD calibration process, we showed that our model reproduced count and travel time information collected from the field. Incorporating peak spreading as a result of congestion for long-term scenarios also yielded performance improvements in the models and overcame network gridlock issues common to such applications.
Abstract