摘要
针对城市交通“智能运输系统” ,提出了基于改进BP神经网络理论模型的路面交通流动态时序的预测算法。在BP算法的自适应学习率 ,在动量法优化网络收敛性等方面 ,进行了深入研究 ,并改进了基本BP算法中的收敛速度慢和易陷入局部最小点等问题。文章给出了基于改进BP算法的交通流动态时序的预测算法仿真实验 ,结果验证了该算法的可行性和先进性。
Error Back Propagation (EBP) neural network is one of the popular and perfect neural network models since it was put forward by Rumelhart and others. Aiming at “Intelligent Transportation System”, this paper presents a traffic flow prognostication algorithm based on improved BP neural network. Adaptive study rate of BP neural network is improved in this BP neural network and optimization of astringency based on momentum is realized, which solves problems of slow astringency and local minimum. At the end of this paper, simulation experiments of this algorithm are illustrated,the result of which proves the feasibility and efficiency of the algorithm and high application value in traffic flow prognostication.
出处
《交通与计算机》
2001年第3期11-14,共4页
Computer and Communications
关键词
动态时序预测
交通流量
预测算法
BP神经网络算法
dynamic sequence prognostication
flow prognostication
improved BP neural network
