Investigation of attributes determining trip chaining behavior in hybrid microsimulation urban freight models

A hybrid microsimulation modeling framework is proposed to construct goods-related vehicle tours that satisfy a known commodity flow origindestination (O-D) matrix in an urban freight network. One distinct issue of urban freight systems - trip chaining behavior - is addressed in the modeling procedure. To shed light on the variables that play significant roles in affecting trip chaining behavior, two types of discrete choice models are estimated by using a data set from the integrative freight market simulation. These models generate probabilities that help choose destination locations and make the decision about whether to return to the base or not for each tour until the known commodity O-D matrix is satisfied. The proposed modeling framework was applied to an 84-node test network. Results show that the estimated trip length distribution is consistent with the underlying data set. Meanwhile, several attributes are found to have significant effects on the trip chaining behavior. The choice of the next destination is negatively affected by the distance from the current location to the potential destination and positively affected by the amount of cargo available for pickup and delivery. For the tour termination decision, the perceived utility of returning decreases with the increase of the return distance and increases with the accumulation of cargoes delivered.