A discrete shipment size choice model with latent classes of shippers’ attributes.

Kurzfassung

Modelling the choice of shipment size is an important aspect in developing a freight transport model which also considers logistic choices. We developed a discrete shipment size choice model for road transports as a first step enabling a consistent conjunction between logistic choices and the discrete mode choice. Based on the categorization of the shipment sizes into three classes describing piece goods, partial loads and (multiple) full loads, the common Economic-Order-Quantity-model (EOQ-model) was validated for the discrete case. As there exist a huge variety of shippers the induced behavioral heterogeneity should be taken into account. Due to the insufficient explanatory power of the underlying industrial sectors and an inflation of the model in the statistical sense by adding the logistical attributes an attribute-based latent-class-analysis approach has been applied. Including latent classes improved the discrete choice model in comparison to the pure EOQ-Model as well as in comparison to the EOQ-model with shippers’ attributes.