NC truck network model development research

              NC Truck Network Model
7
Truck Traffic Modeling Methods
Various truck traffic modeling methodologies have been developed and applied. The simplest and least
data intensive methods are statistical models based on past historical counts of truck traffic, and
sometimes direct and indirect causal variables like indicators for business activity. Such trend models
focus on highway segments that are candidates for near- term construction and repair. But statistical
models for estimating truck traffic on individual highway segments do not permit evaluation of network
effects like congestion, alternative network improvements, and changing land use and economy. For more
information on statistical models, see NC Truck Traffic Forecasting ( Stone, 2006).
More complicated and geographically broad truck network models share planning procedures – trip
generation, distribution, mode choice and assignment – that are characteristic of traditional urban and
regional traffic models. Freight network models have three categories: 1) urban truck network models, 2)
statewide or regional commodity flow models, and 3) statewide truck network models. The different
approaches are the result of differing priorities at each level. Planners in metropolitan areas are typically
concerned with alleviating roadway congestion, while regional planners tend to focus on issues of
economic competitiveness and efficiencies. The planners focus upon how they view freight movements.
Urban planners tend to deal with the externalities associated with trucks such as urban traffic congestion,
while regional planners tend to focus on the economic interchanges between various zones within the
region that accompany freight movements. In general, network models assume that shippers and carriers
use minimum cost paths on a network where the cost is a combination of price and time. The networks are
modeled with an array of traffic analysis zones and origins and destinations that produce and attract
vehicle trips and freight in response to system demand. Network models for freight logistics hold promise
for modeling urban and intercity truck flows. They are more complex to implement than other methods
and have more intensive data requirements. Yet, a number of cities and states have truck and freight
network models. Thus, the implementation of network models for truck traffic is a viable strategy for
statewide truck traffic demand forecasting. At the interregional and national levels the FAF, GeoFreight,
and LATTS models have been used by federal and state agencies. They are discussed subsequently.
There are two modeling techniques employed in forecasting truck network flows – commodity flow
models and trip based models ( Raothanachonkun, 2007). Both approaches are typically employed in a
“ four- step” sequential process that uses a gravity- model distribution, a mode- split step and trip
assignment. The only significant difference is that the trip generation step is based on freight flow data
( usually classified by industry groups) in commodity flow models, instead of regression equations for
employment and population, as with trip based models.
A good example of the commodity approach is the Indiana Freight Model ( Bernardin, Lochmueller &
Associates, 2004). The Indiana model predicts both truck and rail traffic volumes. For each of 21
commodity groups, trip generation equations are developed based on a regression of data available from
1993 Commodity Flow Survey ( CFS). Following trip generation, freight shipments are distributed by a
gravity model, which is also calibrated using the CFS data. The mode split step also utilizes the 1993
CFS, projecting the 1993 national shares into the future. Next, the model divides the freight tonnages into
an equivalent number of vehicles, with tons- per- vehicle payload factors determined separately for each
commodity group. Finally, the traffic is assigned to the network. This approach builds the relationship
between commodity flow and truck traffic. It takes economic activities into account using the familiar
four- step travel demand methodology to forecast future truck volumes. However, such a model is
complicated and requires much survey and quantitative data.
The commodity flow- based technique is also used in the Wisconsin Intermodal Freight Model ( Wilbur
Smith Associates, 2004), Kentucky Freight Model ( Wilbur Smith Associates, 2005), and the Southern
California Freight Planning Model ( Fischer, 2003). Based on commodity flow forecasts and economic