TRCLC 15-11

Transit-oriented development (TOD) was coined by Peter Calthorpe as “moderate and high density housing, along with complementary public uses, jobs, retail and services, concentrated in mixed-use developments at strategic points along transit systems”. It emphasizes the physical integration and linkage of public transportation investments and urban land development at or near a transit station. Although the theoretical definitions for a TOD vary in the literature, typical TODs share some core features: moderate- to high-density development, mixed land use, and high-quality transit services. TOD has emerged in recent years as a promising paradigm to promote public transportation, increase active transportation usage, mitigate congestion, and alleviate air pollution.

 

Studies have shown that household auto ownership level (i.e., car availability) has a huge impact on travelers’ mode choice behavior in TODs. Travelers who have no access to personal transportation or unable to drive because of age, disability, income or family circumferences are often defined as transit captive users, or transit-dependent users. In contrast, transit choice users are those who feel the transit option is superior to other travel options in terms of time, cost, convenience and comfort. Failing to consider differences in feasible travel options available to different groups of travelers may lead to inaccurate representations of travelers’ mode choice behavior. In addition, route overlapping is one of the major concerns in the route choice models used in the traffic assignment problem for predicting traffic pattern in the transportation network. Therefore, it is critically important to explicitly consider captive travel behavior in mode choice and route overlapping in route choice to better predict the modes and routes that trips will take, resulting in traffic forecasts for the highway system and ridership forecast for the transit system. With a better behavioral model that explicitly considers captive travel behavior in mode choice and route overlapping in route choice, evaluating TOD strategies is expected to be more accurate and effective.

 

In this study, we provided a new mathematical programming formulation for the CMSTA problem that explicitly considers captive travel behavior in mode choice and route overlapping in route choice. Specifically, the dogit model was adopted to permit flexibility in modeling captive travel behavior by allowing certain choice to be IIA as in the MNL model and other choice to be non-IIA, and the PSL model selected to account for route overlapping problem in a transportation network. The dogit-PSL CMSTA model simultaneously determines both mode choice and route choice with consideration of captive travelers and route overlapping problems.

 

Numerical examples were performed to demonstrate the captive travel behavior in mode choice and route overlapping in route choice. Through the numerical results, we found that ignoring mode captivity can lead to biased results in evaluating the impacts of TOD strategies. When transit captivity is in presence, the MNL-PSL model tends to generate higher transit mode share and larger percentage of increase in transit VMTs compared with the dogit-PSL model, and thus overestimate the benefits of TOD strategies. However, we are cautious about generalizing this conclusion because the results could be sensitive to parameter settings and network topology. In future research, we plan to conduct the evaluation of TOD strategies in more realistic network settings using the proposed dogit-PSL model.