A general framework for classification of road traffic models is presented below. This two-dimensional framework considers the nature and level of detail offered by a traffic model in handling road geometry and traffic elements. The focus is on the movement of vehicle traffic (see the note below about different traffic modes).
Software developers and users are invited to submit comments on this framework, and request inclusion of their models in appropriate categories.
Contrasting models as macroscopic vs microsimulation, deterministic vs microsimulation, empirical vs theoretical, empirical vs analytical, etc are not valid ways of qualifying models. Models never fall into clear-cut categories, but there is a spectrum (continuum) of models. The framework presented here may be helpful to understand that analytical models or simulation models can be microscopic or macroscopic (and in between).
The US Highway Capacity Manual (HCM) defines an analytical model as "A model that relates system components using theoretical considerations that are tempered, validated, and calibrated by field data.", whereas it defines a simulation model as "A computer program that uses mathematical models to conduct experiments with traffic events on a transportation facility or system over extended periods of time." HCM Glossary (Chapter 5) gives various other definitions related to models and HCM Chapter 31 presents a very useful discussion of different types of models.
For the purpose of the framework presented below, the analytical models are defined as those that use direct mathematical computations to determine system states, and simulation models as those that use various rules (mostly in the form of mathematical equations) for movement of vehicles in a system (individually or in platoons). The HCM models are analytical. HCM Chapter 31 states that "The HCM methods represent traffic flows with variables that reflect flow dynamics. These methods stop short of representing the movements of individual vehicles. The intent is to employ calculations that can be done by hand, using a set of worksheets, or by computer ...".
In the framework presented here,
(i) a simulation model can be microscopic, macroscopic or mesoscopic;
(ii) an analytical model can be microscopic, macroscopic or mesoscopic;
(iii) a simulation model can be deterministic or stochastic.
Analytical traffic models usually incorporate stochastic elements (e.g. overflow queue models for traffic at intersections) although each application of the model may produce the same outcome (deterministic). The distinction "Stochastic vs Deterministic" does not necessarily imply model quality since parameters of traffic elements at every level of detail (individual vehicle. platoon, traffic flow, etc) can be randomised.
Contrasting models as "Empirical vs Theoretical" represents a simplistic view since most models have basis in traffic behaviour theory and are empirical at the same time. However, the term "empirical model" is usually used to mean "based on statistical analysis of field data without any direct basis in traffic theory".
The framework presented here is limited to vehicle traffic. The issues of different vehicle types and driver types, and the size of the area modelled (single intersection, arterial, network, etc) are further considerations in this context.
Different modes of traffic (pedestrians, cyclists, public transport) could be added as a third dimension to this framework, each with its own special considerations. For example, for pedestrians, drive cycles are not applicable, and pathways rather than lanes would be relevant.
A general framework for road traffic models
|ROAD GEOMETRY ELEMENTS|
(all Lanes Aggregated)
|Lane group |
(or lane segments)
|Speed - flow models||Macro-analytical||NA|
(1) Drive cycle may be defined as the vehicle speed-time trace consisting of acceleration, deceleration, cruise and idling elements.
(2) Lane Group is a set of lanes with one or two shared lanes (e.g. Lane 1: Left-Turn and Through, Lane 2: Through) or a set of exclusive turn lanes (e.g. a single Right-Turn lane).
Traffic modeling software packages according to the above framework
|Level of traffic and road geometry detail|
|*e.g. EMME/2, QRS II, TRIPS, TRAFFIX|
Traffic modeling software packages in each group are listed in alphabetical order. Although it is simplistic to place software packages into separate boxes (since different aspects of a package may employ different levels of analysis), the above table is intended to indicate the dominant model characteristic in each package. Links are provided to software supported by Akcelik & Associates.
First published: 30 August 2004
Last Revised: 15 July 2007