TRB 2016 Blue Ribbon Committee
16th National Transportation Planning Applications Conference

Stepping closer to ABM: Hybrid 4-Step Models

Corresponding Author: Gaurav Vyas, WSP

Presented By: Gaurav Vyas, WSP


Activity-based travel demand models (ABMs) are significantly better in capturing individual travel patterns and are better in testing sensitivity to different travel demand management policies compared to 4-step models. The advantages of using ABMs are well known in research and practice. However, there is a cost associated with the advancements in behavioral realism and policy sensitivity of the travel model. ABMs are generally characterized by a long development cycle, substantial budget, and certain additional data requirements. Moreover, there is a learning curve for the model users to switch from a 4-step model to an ABM in terms of model understanding, data preparation, as well as processing and interpretation of model outputs. Because of this, many MPOs (especially smaller ones) are still using 4-step travel demand models and show reluctance in adoption of the ABM paradigm.

With the commonly adopted population synthesis technique which generates a population of residents of the modeled region at the fully disaggregate level it is possible to incorporate some of the benefits of ABMs in a 4-step framework, with the possibility of eventually migrating to a full ABM. This can be done by adopting a subset of the sub-models typically used in an ABM framework in a 4-step model, resulting in hybrid 4-step structure. There is a wide spectrum of options in terms of the number of ABM sub-models incorporated in the 4-step model system; with the inclusion of synthetic population and individual trip generation as a practical minimum. The synthetic population allows the model to incorporate a rich set of person and household characteristics in trip generation, and potentially also in some other models (such as car ownership or usual workplace/school choice), while a traditional 4-step model can only use a limited number of variables. Depending on an MPO’s policy requirement and resources, a hybrid 4-step model can be developed which can provide many advantages of ABMs while preserving the simplicity of the existing 4-step model system. This paper discusses an example of such a hybrid 4-step model developed for the city of Jerusalem, Israel.


Discuss This Abstract