Research on time-of-day (TOD) choices has predominantly focused on weekday trips or activities while there was few studies covering behavior of TOD choices on weekends. This paper examines TOD choices on weekends using a tour-based approach. The effects of household and individual sociodemographics and tour related travel characteristics are examined, and a multinomial discrete choice model was calibrated to forecast the TOD choices of various social-economic and demographic groups. In light of unique travel patterns on weekends, the authors have combined both Saturday and Sunday as a single travel period, which is then divided into six TOD segments. The methodology is based on the observations that, unlike weekdays, what people do on Saturday is less likely to be repeated on Sundays. This manuscript demonstrates the application of a tour-based TOD weekend forecast model using the 2001 Atlanta Household Survey data. The 2001 Atlanta Household Survey contains detailed travel information on weekends. After presenting the exploratory analysis of weekend travel pattern in Atlanta, this paper also compares the results of a tour-based and a trip-based models with the same data set to further examine the differences between the two approaches. The study validates that tour-based model does improve the overall goodness of fit of the model and produces better forecast.
Abstract