AGADE Traffic 2.0 - A Knowledge-Based Approach for Multi-agent Traffic Simulations

Abstract

AGADE Traffic 2.0 (funded by Karl-Vossloh-Stiftung) is a multi-agent based tool for traffic simulations with focus on modelling dynamic decision-making processes for individual traffic participants. For this purpose, a revised modelling component has been integrated that uses semantic technologies (OWL and SWRL) to express agent behaviour. At the same time, this approach enables new options to make agent behaviour explainable. In this paper, we demonstrate that a rule-based modelling approach combined with a logging mechanism can be applied to provide data that can be used to identify specific emerging behavioural patterns and their consequences for the results of the simulations.