A Cellular Automata Approach for Simulation-Based Evolutionary Optimization of Self-Organizing Traffic Signal Control

Self-organizing traffic signals are controlled autonomously by control rules that rely on adaptation to local variations in traffic state and enable effective coordination of the vehicular traffic at a network level. In this study a cellular automata model of self-organizing traffic signal system is proposed, which enables evolutionary optimization of the control rules. Fitness function, which guides the evolution of control rules, is evaluated via traffic simulation by using a microscopic cellular automata model. According to the proposed approach, relevant control rules are initially determined by using a clustering algorithm. Subsequently, an evolutionary strategy is applied to optimize the control rules. Decisions about switching the traffic signals are made by using control rules that are applicable for current traffic state. The k-nearest neighbours algorithm is employed for selection and fusion of the applicable control rules. Results of simulation experiments clearly show that the proposed approach can achieve a significantly reduced vehicular delay when compared with state-of-the-art algorithms for the self-organizing traffic signals.