R. Wang

Pages: 3-12

Abstract
In order to solve the problems of high average delay and high green signal ratio in traditional traffic signal timing optimization methods. An intelligent traffic signal timing optimization method based on genetic neural network is proposed. The basic principle of genetic neural network is analyzed, and the neural network model optimized by genetic algorithm is constructed. The optimal solution of traffic flow prediction model is transformed into the problem of calculating the best individual and its representative optimal solution. The coding genetic algorithm is used to code the random weight threshold of neural network to get chromosome coding string, calculate the error function value of neural network, and obtain all of them According to the value of fitness function, the population is operated by genetic algorithm to obtain the optimal solution of traffic flow prediction model; based on the optimal prediction results of traffic flow, the effective green light time of each phase is obtained according to the optimal time allocation period of traffic lights to complete the optimization of intelligent traffic signal light timing. The experimental results show that the minimum average delay time is about 46.8 s, and the green signal ratio of this method can reach 95%, which can effectively reduce vehicle congestion.
Keywords: genetic neural network; smart transportation; signal light timing; real number encoding