Abstract
Travel time reliability in New York City was analyzed with three travel time reliability measures. A classification and regression tree model was used for the analysis. Instead of analysis of conventional peak and off-peak periods, day-of-week (DOW) and time-of-day (TOD) periods were determined on the basis of each travel time reliability measure. DOW and TOD periods were identified on the basis of average travel time and each selected measure. Travel time reliability measures formulated to explain the same phenomenon classified different periods as having similar characteristics. The results agreed with the literature that reliability measures should be based on temporal periods such as DOW and TOD; however, the selection of time periods should be measure specific. The impact of New York City's urban grid network on travel time and speed distributions is also discussed. The travel time distribution patterns reported in the literature for freeways do not exist for the city. Therefore, caution is suggested for transferring reliability measures across different network structures.
Get full access to this article
View all access options for this article.