Making Way for Emergency Vehicles

One of the major problems large metropolitan areas have been facing is traffic congestion (saturation of the transportation network infrastructure). The problem’s cause is well-known: as demand gets higher, the infrastructure (roads, rails and so on) is unable to fulfil it, because of its limited capacity. The alternative to the permanent enlarging of the infrastructure is to make use of more efficient control systems. An even more recent trend is to expand these control systems with capabilities to influence the behaviour of the network users. This paper introduces some experimental work being done in pursuit of a computational solution for traffic congestion. We will describe a very basic Multi-agent System (MAS) for simulation of traffic control and present the results from two distinct points of view: traffic simulation and simulator performance. This work represents some initial experiments of a larger work aimed at three objectives: (a) the development of a decision support system for micro-simulation of wide urban areas of traffic, (b) the characterization of control strategies in heterogeneous MAS and (c) the development of new agent architectures with emphasis on adaptable, “mentalistic” and emotional capabilities, as well as group learning.