摘要
针对神经网络交通事件检测算法的缺陷,提出遗传算法与神经网络相结合的事件检测算法。应用遗传算法优化交通事件检测的神经网络模型参数,得到事件发生与交通参数间的映射关系。最后,用实测数据对模型进行校验。结果表明该算法有很好的鲁棒性,能提高事件检测的效率。
To solve the problem in the neural network prediction algorithm, a hybrid algerithm which combines genetic algorithm and neural network was proposed. The genetic algorithm was used to optimize the parameters of neural network model. The function between incident and traffic parameters was determined. Finally,the testification of algorithm validation was undertaken by measured data of a real freeway. The results indicated that the genetic algorithm has a better robust and can greatly improve the efficiency of traffic incident detection. 2 figs, 9 refs.
出处
《长安大学学报(自然科学版)》
EI
CAS
CSCD
北大核心
2005年第4期70-72,共3页
Journal of Chang’an University(Natural Science Edition)
关键词
交通工程
交通事件检测
神经网络
遗传算法
实数编码
traffic engineering
traffic incident detection
neural network
genetic algorithm
real code