Common use of Trip Distribution Clause in Contracts

Trip Distribution. The most common trip distribution model used in statewide modeling are the gravity models (NASEM, 2017), which are based on the mathematical function form of the law of gravity in that travel activities between two TAZs are assumed to be positively proportional to the product of trip production at one TAZ and trip attraction at the other, weighted inversely by a function of travel time between the two TAZs (Ortúzar and Willumsen, 2011). The strength of the gravity models is that they are easy to implement (i.e., only three variables needed in the simplest form) and easy to calibrate. Calibration of a gravity model involves adjusting parameters of the gravity function until the observed average trip length distribution is matched by the model (▇▇▇▇▇▇▇ and ▇▇▇▇▇▇▇▇▇, 2011). However, gravity models cannot effectively model long-distance trips (e.g., trips over 50 miles) because the inverse weight of travel time increases drastically as distance increases such that the curve of the gravity function flattens out to values close to zero after a certain distance (Ortúzar and Willumsen, 2011). This is one of the reasons that statewide models usually incorporate separate long-distance passenger travel models (NASEM, 2017). The logit-based destination choice model is the other commonly used model for trip distribution (NASEM, 2017). A logit destination model hypothesizes that the probability of choosing one particular TAZ depends on the ratio of the TAZ’s utility, which is expressed as a function of land use characteristics of the TAZ (e.g., population, employment, and distance to the TAZ), to the sum of the utilities of all TAZs (Ortúzar and Willumsen, 2011). Some consider logit destination distribution models the best practice for trip distribution (VDOT, 2014), because the logit model form enables consideration of multiple factors that can affect destination choice, while the gravity models theoretically only allow for three variables (i.e., without considering composite variables) in the model form. In addition, the logit models do not have the limitation as gravity models in modeling long-distance trips. In practice, gravity models are far more commonly used for statewide modeling (i.e., 22 gravity models versus 11 logit models according NCHRP synthesis 514). Generally, use of the gravity model for trip distribution is considered acceptable practice in all regions. In small regions, the gravity model for trip distribution also is considered recommended practice. In large regions, the destination choice model formulation is considered recommended practice (VDOT, 2014). It is noted that some states that used logit-based destination choice models also implemented separate long-distance travel model (NASEM, 2017), mostly because long-distance trips are different from short-distance trips in many aspects of trip-making behavior, not just destination choice (Ortúzar and Willumsen, 2011).