A Tabu Search and Genetic Algorithm for Traffic Light Scheduling*

The actual traffic light system with its prefixed timers suffer from critical traffic jams due to several factors, including the significant amount of vehicles increasing in the cities. This situation requires a more complex signal control, and hence, a more dynamic scheduling. In this paper, we propose two separated metaheuristic: the Tabu Search Algorithm (TSA) and Genetic Algorithm (GA) for the optimization of the traffic signal in order to compare them and seek for the fittest traffic lights strategy (phases and timings) suitable for a road network with multiple intersections. Employing a urban simulator, both approaches are analysed and their results compared in terms of the following key metrics: Number of vehicles passing through the intersections, the global waiting time and trip time.