A multi-agent approach towards modeling urban traffic scenarios

Abstract

This paper investigates the use of the multi-agent paradigm in modeling urban traffic scenarios. It demonstrates that vehicles, pedestrians, traffic-lights, car-parks and even streets can be considered agents in a heterogeneous multi-agent system. Different types of agents in such scenarios will be identified, characterized and constructed by virtue of a general agent model. The various kinds of relationships and interactions, generally called cooperations between these agents will be modeled; some examples sketch the major issues developed in this paper.