To search, Click below search items.


All Published Papers Search Service


CITL: A Multi-Agent System to Control the Intersection Traffic Lights


Fatemeh Khezerlou, Farnaz Derakhshan


Vol. 17  No. 3  pp. 93-102


In this paper, we used an intelligent method to control traffic lights at intersections based on multi-agent system called CITL. Our proposed method uses clustering and evolutionary computing algorithms to schedule the traffic lights, in addition to use exchanging of messages between different agents to control possible fluctuations. The main capability of our proposal is that our method performed distributed processing centers which avoid excessive computational load processing in each cycle and it reduces delay time compare with other methods. For the design and implementation of our method, we used Tropos methodology and JADE library. For evaluation, this method tested with 2000 laboratory data in both low and heavy traffic areas. The average run-time in the sequential cycles in our combinatorial method is less than using the other single methods compared to the fixed-time methods, using our method leads to the improvement of the delay time of vehicles 24.5% and 35.5% in heavy-traffic and low-traffic areas, respectively. Also, CITL get improved the delay time 9.58% and 9.79% in heavy-traffic and low-traffic areas, compared to the similarmulti-agent systems.


Intelligent traffic control Multi-agent systems Urban traffic control.