改进BP神经网络的城市交通流预测研究

由于交通流预测具有高度的非线性特点,这与BP神经网络能够处理非线性问题的特征相符合。但BP神经网络算法易使解陷入局部极小,而遗传算法的全局优化能力则恰恰可以克服这一缺点。文中将遗传算法应用于对BP神经网络模型的改进来对交通流进行预测。通过对预测数据与实测数据的比较分析,证实了改进后的方法更为有效。

Study of Improved BP Neural Network on Forecasting City Traffic Flow

As the traffic flow forecasts are highly non-linear characteristics,which is in line with the BP neural network capable of dealing with the characteristics of non-linear problems.But neural network algorithm is easy to make solution into a local minimum,nevertheless the genetic algorithm are just able to overcome this shortcoming by its global optimization capability,so a traffic flow prediction method based on BP neural network model improved by Genetic algorithm is put forward in this paper.Case study proves the validity of the method by comparing and analyzing the prediction data and the real data,which makes the study have a more extensive application value.