Traffic Light Control through Agent-based Coordination

J.H. Domínguez Sánchez and J.L. Aguirre (Mexico)


Multiagent traffic simulation, Negotiation, Auction, Coor dination.


A common mechanism that allows control of vehicular flow in big cities is the use of traffic lights. Generally, each traffic light on an intersection is assigned a constant green time. In last years several agent-based traffic simu lators have been proposed to show how better results can be achieved with respect to traffic flow variables, like vehi cles’ route time and vehicles’ waiting time. It is possible to propose agent-based coordination schemes where the green time of the traffic lights is assigned based on the present conditions of traffic. Due to those intelligent assignations on the traffic lights’ green time, it’s reasonable to think that vehicles’ waiting time could be reduced. In this paper we compare two traffic light coordina tion mechanisms of an intersection that assign traffic lights’ green time based on approximate information about traffic conditions. It is compared the vehicles’ waiting time on the traditional mechanism with respect to an auction mech anism, in which the traffic lights are “selfish”, and to a negotiation mechanism where the traffic lights work with a “non-selfish” approach in the intersection. Our results show that the negotiation mechanism, under certain con ditions, gets better results over the traditional and auction mechanisms.

Important Links:

Go Back