Activity-Based Model Sensitivity Testing – Showing the Model’s Sensitive Side
Corresponding Author: Yijing Lu, Baltimore Metropolitan Council
Presented By: David Kurth, Cambridge Systematics, Inc.
Abstract
The Baltimore Metropolitan Council (BMC) has developed an advanced Activity Based Model (ABM) known locally as the Initiative to Simulate Individual Travel Events (InSITE). InSITE simulates long term choices, household person tour rosters for mandatory and non-mandatory travel, and intra-household person travel for fully joint non-mandatory and half-tour school escorting. Scenario tests reflecting anticipated MPO studies were designed to evaluate InSITE’s behavior model component choice sensitivity to the following scenarios:
• Household person age distributions – Regional demographers have forecasted older regional median age for some time. The synthetic population inputs were revised to reflect an older population.
• Re-development at an urban brownfield – The Baltimore region is highly urbanized with slow growth typical of the Northeast. A significant amount of forecasted growth is targeted in areas for re-development, which is consistent with State of Maryland Smart Growth policies. A segment of synthetic population and employment was reassigned to a selected urban brownfield TAZ.
• Freeway segment capacity – Envisioned vehicle-to-vehicle and vehicle-to-infrastructure technology promises more efficient highway travel as a result of less congestion and higher vehicle travel speeds. Using the region’s performance monitoring reports, an unreliable segment of the Baltimore Beltway was identified. The capacities on selected highway network links for this segment were increased.
The scenario tests were designed to gain better understanding of InSITE’s sensitivity to transportation system changes, policy testing, and traveler characteristics. During InSITE development, BMC began looking for ways to connect the simulation work to MPO planning activities. This presentation will discuss a number of issues important to the MPO and member governments such as changing demographics, land use , and highway capacity, and re-development from single family to attached housing. The example scenario tests have demonstrated some of the issues that are encountered in constructing the input datasets required for future scenario testing as well as some of the reasonableness tests required when analyzing model outputs. Sensitivity to demographic changes, in particular, is an advantage that ABMs have over traditional trip based models (TBMs). Comparisons of InSITE results to past TBM results will be shown to highlight the differences in the models’ sensitivity to key input variables.