Short-term traffic flow forecasting plays a very important role in urban traffic management and control. With the shortening of the forecasting term, the uncertainty of traffic flow becomes more and more seriously. Former forecasting method can not reveal the uncertainty and chaos or reduce stochastic disturb factors. So with the shortening of the forecasting term, the forecasting precision becomes lower and lower.This paper, according to the property of urban traffic flow, applies the analysis method of chaotic theory, and study on the traffic flow forecasting. It helps to grasp the regularity of traffic flow system. We compute the parameters of phrase space reconstruction for traffic flow system. Based on chaotic theory, it can portray traffic flow system inherent randomness.This paper uses the intellectual method to forecast the short-term traffic flow. According to the parameters of phrase space reconstruction for traffic flow system, we choose the best introduction mode of the neural network. Aiming at the characteristic of BP neural network operating, this paper optimizes the neural network with genetic algorithm based on isolation niche technique. And it chooses the best structure of latent layer of neural network through that method. The new NN modeling method has the superiority of computation complicacy, model performance evaluation and whole search efficiency. Use this model to forecast the short-term traffic flow ofhighway section of urban road, have made the comparatively satisfactory result.
|