Route Decision Behaviour in a Commuting Scenario:
Simple Heuristics Adaptation and Effect of Traffic Forecast
One challenge to researchers dealing with traffic management
is to find efficient ways to model and predict traffic flow. Due to the social
nature of traffic, most of the decisions are not independent. Thus, traffic
systems can be seen as organisations where the inter-dependence of actions
leads to a high frequency of implicit co-ordination decisions. Although there
are already systems designed to assist drivers in these tasks (broadcast,
Internet, etc.), such systems do not consider or even have a model of the
way drivers decide. Our research goal is the study of commuting scenarios,
drivers' decision-making, and its influence on the system as a whole. The
present paper addresses three key issues: simulation of various forms of driver
decisions, the role of a traffic forecast component, and how these can be
used to understand complex traffic systems and how they can be used to explain
the way drivers actually make decisions. The former is realised by a naïve
model for the route choice adaptation, where commuters behaviour is based
on heuristics they evolve. The second issue is realised via a traffic control
system which perceives drivers' decisions and returns a forecast, thus allowing
drivers to decide the actual route selection. For validation, we first compare
our simulation results with empirical data from real experiments and show
that the heuristics drivers evolve lead to a situation similar to that obtained
in the real experiments. As for the forecast scenario, our results confirm
that a traffic system where a large share of drivers reacts to the forecast
will not develop to equilibrium. However, we also can show that introducing
some
individual tolerance for sub-optimal forecasts turns the system more stable.
Ana Lúcia C. Bazzan
University of Wuerzburg
Department of Computer Science
UFRGS and Artificial Intelligence
www.inf.ufrgs.br/~bazzan
bazzan@inf.ufrgs.br
Franziska Kluegl
University of Wuerzburg
Department of Computer Science
UFRGS and Artificial Intelligence
kluegl@informatik.uni-wuerzburg.de