摘要
采用遗传算法优化BP神经网络来建立一个道路交通事故宏观预测的模型.该模型结合遗传算法和神经网络两者的优点,具有更好的运算性能、更快的收敛速度和更高的精度.模型以交通事故死亡人数为输出变量,以机动车保有量、公路里程、人均GDP为输入变量,利用1978年至2006年的道路交通事故数据进行训练及检验.实例计算表明,建立的基于遗传算法的BP神经网络模型可以很好的适用于道路交通事故宏观预测,为制定交通安全对策提供理论依据.
This paper uses genetic algorithms to train BP neural network, and form a neural network predictive model in macroscopic forecast of road traffic accident. The model combines the advantage of genetic algorithms and neural network. It is of better performance, faster speed and high accurate. In the model, deaths of traffic accidents are the output variables, and vehicles quantity, highway mileage, average GDP of per person are input variables. The road traffic accidents data from 1978 to 2006 are used to train model and test forecast result. The experimental results indicate that the model is effective and can offer theoretical basis for establishing traffic safety strategy.
出处
《浙江工业大学学报》
CAS
2008年第3期338-342,共5页
Journal of Zhejiang University of Technology
关键词
遗传算法
BP神经网络
遗传神经网络
道路交通事故
宏观预测
genetic algorithm
BP neural network
genetic neural network
road traffic accident
macroscopic prediction
