Regresar

Traffic signal optimization and coordination using neighborhood mutation

Abstract:

Urban planners face increasing challenges to design and optimize sustainable cities. Evolutionary algorithms are an important tool for design optimization and can help urban planners finding alternative optimal designs to increase the sustainability of cities. Mobility and transportation are two important components of modern cities that are amenable to simulation and their design can be improved by evolutionary means. However, traffic simulation is computationally expensive and puts a serious constraint on the number of generations allowed to artificial evolution. In addition, to grasp the implication of traffic policies for sustainability usually a significant part of the traffic in the city must be simulated. This implies that we must design our evolutionary algorithms for an effective short-term evolution on large-scale problems. This paper investigates neighborhood mutation operators to explore efficiently in few generations a large space of cycle lengths, offsets and green time settings of traffic lights. Our aim is to find settings that allow a better coordination of signals. In addition, we analyze clusters of signal settings to gain knowledge about geographical coordination patterns to provide valuable information to city planners for micro-zonification.