This explainer series is split into three parts. It is intended as a brief overview of how agent based transport modelling works and would be useful for anyone who wants to know a little more how the analysis is produced.
A great advantage of ABM modelling is that the basic concepts are relatively straightforward. Therefore by reading this overview, you should quickly (in under 15 minutes) gain an understanding of how ABM modelling works and how it can be used for transport planning.
This first two parts will focus on how the MATSim model works and the data inputs that are needed.
The third part will focus more on the model outputs and how these can be used for transport planning.
Agents are fundamental to an Agent Based Model. In MATSim an agent is representative of a person within the modelled area.
For the WofE model there are xNumber agents, each with a unique set of attributes. The attributes are derived from census and other data sources such as the National Travel Survey.
Each agent has a home location (shown below), and the overall distribution of home locations corresponds with an adjusted version of the 2011 Census population distribution.
The agents have a number of attributes associated with them. Again these are based upon the 2011 Census. So Agents with home locations in a rural location will have a different attribute profile to agents living in an urban area.
As an example of this, the chart below shows agent population densities for two different attributes.
An agent’s attributes are important for determining their travel options (more on this in part 2), and can also be used in the final analysis (e.g. for determining trends amongst agents with similar attributes).