Comprehensive Exploratory Analysis of Truck Route Choice Diversity in Florida

This study presents a comprehensive exploratory analysis of truck route choice diversity in the state of Florida, for both long-haul and short-haul truck travel segments. It employs six metrics to measure three different dimensions of diversity in truck route choice between any given origin–destination (OD) pair. These dimensions are: (a) number of distinct routes used to travel between the OD pair, (b) the extent of overlap (or lack thereof) among the routes, and (c) the evenness (or dominance) in the usage of different unique routes. The diversity metrics were applied to a large database of 73,000 truck routes derived from 200 million GPS records. Descriptive analysis and statistical modeling of the diversity metrics offered insights into the determinants of various dimensions of truck route choice diversity between any OD pair. The results are useful for improving choice set generation algorithms for truck route choice modeling and in truck route policies and investments. An essential step toward enhancing highway freight mobility is to improve our understanding of freight-truck route choice behavior. Specifically, analysis of the routes that trucks use to travel between different origins and destinations can support the design of truck routing policies aimed at mitigating congestion, improving travel time reliability, and facilitating truck movement during network disruptions. However, research on truck route choice has been limited due to insufficient data on truck movements. The recent availability of global positioning systems (GPS) data has encouraged research on truck route choice modeling (1–3) and highway freight performance measures (4, 5). Yet little attention has been paid to exploring the diversity or variability of truck route choice between travel origins and destinations. An improved understanding of truck route diversity has significance in both freight modeling and planning. For modeling applications, understanding truck route choice diversity can help determine the number and structure (i.e., extent of overlap) of route alternatives to be used in route choice models and traffic assignment procedures (6). For planning applications, analyzing the diversity of truck route choices observed in the field can help inform truck routing decisions during regular and emergency situations. For example, identifying origin–destination (OD) pairs with high travel demand but low diversity in the routes used (e.g., a single route used) can help identify critical segments of the network and inform routine infrastructure maintenance scheduling as well as re-routing efforts during emergency recovery. Also, one can apply route diversity measures to evaluate the redundancy of (or lack of) truck routes in existing transportation infrastructure to justify long-term investment in truck corridors to increase network redundancy. This research presents a comprehensive exploratory analysis of truck route choice diversity in the State of Florida for both long-haul and short-haul travel segments. Specifically, the paper addresses two broad questions: How to measure the degrees of diversity in the routes trucks use to travel between an OD pair? and, What factors influence the diversity of truck route choice between an OD pair? To this end, six metrics are utilized to measure the following three different dimensions of diversity in route choice between a given OD pair: (a) number of different routes used between the OD pair, (b) extent of overlap (or lack thereof) among the routes, and (c) evenness (or dominance) of the use of different unique routes between that OD pair. These metrics are applied to quantify truck route choice diversity using a database of about 73,000 routes derived from more than 200 million truck-GPS records. Next, statistical models are estimated to explore the influence of various determinants on the three dimensions of route choice diversity between hundreds of OD pairs. The models provide insights into the influence of the characteristics of truck travel demand, OD location, and network structure on the diversity of truck route choice between an OD pair. Potentially, these insights can help travel modelers in improving choice set generation algorithms for modeling truck route choice and help planners in devising truck routing policies. The next section reviews past studies on variability in route choice behavior. The following section summarizes the truck-GPS data used for this study. The metrics used to quantify diversity in truck route choice are then elaborated. The next section presents the statistical modeling methodologies used in this study. Subsequently, empirical results are presented and discussed. The last section concludes the paper