2015•Transportation Planning and TechnologyRequires access
A comparative analysis of traditional four-step and activity-based travel demand modeling: a case study of Tampa, Florida
Ming Shou Zhong, Rong Shan, Donglei Du, Chunyu Lu
Open publisher page 23 citations
Abstract
Activity-based travel demand modeling (ABTDM) has often been viewed as an advanced approach, due to its higher fidelity and better policy sensitivity. However, a review of the literature indicates that no study has been undertaken to investigate quantitatively the differences and accuracy between an ABTDM approach and a traditional four-step travel demand model. In this paper we provide a comparative analysis against each step – trip generation, trip distribution, mode split, and network assignment – between an ABTDM developed using travel diary data from the Tampa Bay Region in Florida and the Tampa Bay Regional Planning Model (TBRPM), an existing traditional four-step model for the same area. Results show salient differences between the TBRPM and the ABTDM, in terms of modeling performance and accuracy, in each of the four steps. For example, trip production rates calculated from the travel diary data are found to be either double or a quarter less than those used in the TBRPM. On the other hand, trip attraction rates computed from activity-based travel simulations are found to be either more than double or one tenth less than those used in the TBRPM. The trip distribution curves from the two models are similar, but both average and peak trip lengths of the two are significantly different. Mode split analyses show that the TBRPM may underestimate driving trips and fail to capture any usage of alternative modes, such as taxi and nonmotorized (e.g., walking and bicycling) modes. In addition, the ABTDMs are found to be less capable of reproducing observed traffic counts when compared to the TBRPM, most likely due to not considering external and through trips. The comparative results presented can help transportation engineers and planners better understand the strengths and weaknesses of the two types of model and this should assist decision-makers in choosing a better modeling tool for their planning initiatives.