There are two main modeling approaches in travel demand forecasting today: one is the traditional four-step travel demand model (FSM) that is being used by the majority of transportation planning agencies, and the other is activity-based model, which simulates individual and household activities at much more detailed levels. The activity-based approach has been viewed as a more advanced method than the traditional four-step model. This paper reports the partial results – trip generations only from a larger research project investigating the differences between the two modeling approaches. In order to do so, a micro-simulation model is developed and used to simulate 24-hour individual daily travel behaviors of the entire population in the Citrus County from Florida, based on survey data that are obtained from Tampa Bay Regional Transportation travel diary records. Corresponding to each step in the traditional fourstep travel demand model, trip generation rates, trip distribution, mode choice percentages, and trip assignment results are either directly calculated from the observed or simulated individual daily travel records and are compared with the corresponding results from the Tampa four-step travel demand model (developed elsewhere). Study results show salient differences in modeling performance and accuracy in each of four steps above between four-step and activity-based travel demand approaches. In particular, it is found based on the collected travel diary data from Tampa Bay Region that traditional four-step model tends to overestimate trip production for the low-density area, but tend to underestimate trip production for the high-density area, when compared to the activity-based model. Moreover, analyses to simulated results show that production rates are higher in the four-step model for home-based shopping, home-based social and home-based work trips, but lower for home-based school and home-based other trips. On the other hand, the four-step model shows higher attraction rates for home-based work and homebased sociality trips, but lower rates for all other trips when compared to the activity-based model. It is believed that such differences most likely result from the fundamental modeling philosophy of the two modeling approaches.
Menu
