Joint and sequential models for freight vehicle type and shipment size choice

Abstract

The choice of freight vehicle type and shipment size are among the most important logistics decisions made by firms. An important aspect is the nature of the choice process i.e., whether the two choices are sequential or joint in nature. In this study, we investigate the factors that influence the two choices and develop sequential and nested logit models with both possible sequence or nesting structures i.e., vehicle type first or at upper level and shipment size second or at lower level and vice-versa. A commercial travel survey for the Greater Toronto and Hamilton Area is used to estimate the models. Characteristics of firms including industry type and employment, and characteristics of shipments including commodity type, destination location, and density value are tested. Shipment size is categorized into four categories and four vehicle types are considered. The results show that both sequences and nesting structures are possible. The nested logit model results show a potential correlation among unobserved components of utility for vehicle types (80%) and shipment sizes (38%) which should be considered. Model performance is assessed using rho-squared and BIC value. The results show that the sequential logit model with shipment size first and vehicle type second sequence has the best model fit. However, based on the strong correlation indicated by the nested logit model for vehicle type nested within shipment size choice (second best model), a model reflecting the joint nature of the choice process might be suitable.