Modeling interregional freight flow by distribution systems

Distribution Centers with a warehousing function have an important influence on the flow of
goods from production to consumption, generating substantial goods flow and vehicle
movements. This paper extends the classical 4-step freight modeling framework with a
logistics chain model, explicitly modeling freight flows via distribution centers and
warehouses. Modeling logistics at the regional level establishes a link between trade flow
and transport flow, allows determination of the warehouse and distribution centers locations
and corresponding freight flows, permits more detailed and accurate policy decision support
systems. This paper describes a two-stage logistics model that estimates interregional goods
flow via logistics chains. The first stage estimates interregional trade flows by the means of a
gravity model application starting from regional production and consumption volumes. The
second stage, the logistics chain model, creates logistics chains, which split the productionconsumption
flow between direct shipments and shipments via warehousing facilities. We
use an aggregate multinomial nested logit discrete choice model to determine flow volumes
for each of the possible logistics chains. We achieve consistency between the gravity and
logistics chain models by a joint estimation of unknown parameters. The proposed two stage
model estimates interregional OD freight flows that match the observed real world volumes.
We use a new transport flow survey dataset produced by Statistics Netherlands with
information on loading and unloading location types. This dataset allows the required model
calibration with respect to different types of flow. We present estimated and observed
interregional freight flow volumes, estimated parameter values and their interpretation.