Traffic light optimization using Non-Dominated Sorting Genetic Algorithm (NSGA2)

The number of vehicles in the world has increased significantly. The optimization of traffic light control plans in large cities is an extremely important area of research in traffic engineering, as poorly coordinated traffic lights can lead to increased waiting times for vehicles at traffic lights and thus to long delays at intersections. In addition, suboptimal control plans could contribute to traffic congestion, which has become a major concern for public administrations, as it is one of the factors that reduce productivity and the quality of life of people. In this context, this paper presents a method for Active Control of Traffic Signals (ACTS) associated with the use of NSGA2 multi-objective algorithm, aimed at optimizing the traffic signal control plan in response to the traffic demand of a region of the city of Belo Horizonte (Brazil), at each cycle time interval. The objectives are to minimize the average delay time and the number of vehicles stops per cycle. Actual data collected by the Company of Transport and Traffic of Belo Horizonte (BHTrans) were used with the AIMSUN simulator and the experiments were compared with the current signal control plan used by BHTrans to determine if NSGA2 is efficient in optimizing the traffic control plan of the region under study.