On July 5th, 2012 TrafficQuest organised, together with TRAIL and Delft University of Technology, a masterclass on human behaviour in traffic models. Over the past years a lot of knowledge and data on traffic behaviour have become available, but it does not yet find its way into commonly used traffic models. From dynamic macroscopic planning tools to microscopic simulation models: all models contain assumptions about human behaviour. But how realistic are these assumptions? And what are the consequences? During the masterclass 4 experts shared their opinion about this topic with about 40 participants.
Gerard Tertoolen (XTNT) shared his doubts about the assumption of the decreasing status of the car. He presented some arguments why the status value of the car will rise. In the second part of his presentation he showed the audience the results of behavioural research and the link with traffic and traffic management.
Marieke Martens (TNO and University Twente) focussed on the gap between behavioural experiments and modelling. Among other things she concluded that having predictive and valid models for driver behaviour remains a challenge.
An attempt to include behavioural aspects in macroscopic traffic flow models was demonstrated by Chris Tampère (KU Leuven). He showed that an extra parameter in the model improved the simulation of certain traffic phenomena, such as the capacity drop.
The final speaker of the masterclass was Hani Mahmassani (Northwestern University, Evanston, USA). He talked about expanding the boundaries of the modelling of behaviour and the implications for managing the transportation system. He concluded that the main focus of the research and simulation community to date was on HOW TO SIMULATE, but a greater focus is needed on WHAT WE SIMULATE.
All four speakers gave inspiring lectures, which are worth to reread and to rethink.