Page 4 of 5
While originally conceived as a purely traffic assignment package with a
fixed user-defined trip matrix SATURN now also provides a number of standard
modelling steps for matrix estimation, e.g. modal split and distribution. These
models have been integrated with the assignment in a self-consistent mode to
provide a mathematically convergent form of “Variable Demand Modelling”.
A further long-standing example of matrix estimation within SATURN is the ME2
Model (Matrix Estimation from Maximum Entropy) developed at Leeds and University
College London by Dr. L.G. Willumsen, enabling O-D trip matrices to be estimated
directly from traffic counts, thus reducing significantly the survey costs
required to run an assignment model such as SATURN as well as increasing its
accuracy. In effect ME2 generates the ‘most likely’ trip matrix which is
consistent with all available traffic information, including where possible any
prior estimates of the trip matrix.
SATURN has over 300 users worldwide in some 30 countries over all six continents
and has been tested and applied by Local Authority planners, consultants and
universities for the evaluation of road schemes. It is also widely used for
research projects, e.g. for the evaluation of road charging and route guidance
systems.