A Case of Multiagent Decision Support: Using Autonomous Agents for Urban Traffic Control

This paper presents the experimental TRYSA2 distributed decision support system, that has been designed for the management of the urban motorway network around Barcelona. It shows how different technologies in the area of intelligent agents can be combined to solve this real-world decision support problem: on the one hand, a preestablished distribution of the loci of decision-making and the nature of local traffic management tasks, suggested to apply “deliberate” problem-solving agents; on the other, the complexity of the co-ordination task called for an “emergent” approach, in order that the overall decision support functionality be the result of non-benevolent agent interactions. The paper sets out from a description of our particular traffic management problem. Subsequently, the architecture of TRYSA2 is outlined, pointing out the design strategy followed and describing how the different design steps have been realised. Finally, we discuss the lessons learnt from building this multiagent application.