The San Francisco County Transportation Authority (the Authority) has traditionally depended primarily on SF-CHAMP, a regional Activity-Based Model, to evaluate transportation system performance for geographies ranging from neighborhood to regional scale. The travel frequency-, destination choice- and mode choice- models in SF-CHAMP do not represent capacity limitations in any input variables. Vehicle and transit passenger assignment in SF-CHAMP depend on capacity-informed static user equilibrium, and iterative transit assignment (see Zorn et al, 2012), respectfully, that are both sensitive to capacity, but are not capacity constrained. As such, some system performance metrics, such as V/C, can exceed realistic levels. An unrealistic flow indicates latent demand and the likelihood of poor system performance; the magnitude of the difference between capacity and flow indicates the scale of the problem.

Recently the Authority has begun working with dynamic traffic assignment models (DTA) and person-based transit assignment models (i.e. FAST-TrIPs) that have hard capacity limitations. In these models demand that cannot be accommodated on a desired facility is held at origin centroids or at network boundary entry points. The surplus demand is then accommodated in subsequent time periods when capacity becomes available. Rather than showing a level of demand that exceeds supply, dynamic capacity-constrained models represent over-saturated networks by extending the duration of capacity-level activity until unserved demand is accommodated. The transfer of demand to later travel periods produces flow-based metrics that may not indicate the full extent of the disutility caused by congestion. It is apparent that the system performance and user utility metrics historically used in conjunction with a static model, which show “total demand, considering the network conditions”, are not appropriate for application in a hard-capacity model, which shows “physically possible demand”.

This presentation will discuss the efforts undertaken by the Authority to develop a transportation system performance indicator framework for use with capacity constrained traffic and transit simulation models that respects how these models differ from static models without hard capacity limits.